ML20134A101
| ML20134A101 | |
| Person / Time | |
|---|---|
| Site: | Peach Bottom, Surry, 05000000 |
| Issue date: | 02/20/1985 |
| From: | Cybulskis P Battelle Memorial Institute, COLUMBUS LABORATORIES |
| To: | Jeffrey Mitchell NRC |
| Shared Package | |
| ML20133Q389 | List: |
| References | |
| FOIA-85-447 NUDOCS 8508150189 | |
| Download: ML20134A101 (3) | |
Text
{{#Wiki_filter:- gm . , - -
;,7
, c QBallelle Columbus Lalmratories soning Asenue Columbus. ()hio 41201 269 l Telephone H,14) 424-6424 Tries 14 5454 February 20, 1985 Ms. Jocelyn Mitchell U.S. Nuclear Regulatory Commission Willste Building - P-822 7915 Eastern Avenue Silver Spring, Maryland 20910
Dear Jocelyn:
As discussed during last month's meeting in Columbus, we have reviewed the MARCH results for the Surry and Peach Bottom sequences in order
'to quantify the noble gas releases that would be expected in each of these cases. As you are well aware, MARCH is not designed to address questions of fission product transport and deposition; however, for purposes of fission productThedecay heat distribution to be us MARCH model is quite appro-of the various fission product groups. The noble gas release re-priate for the behavior of the noble gases.
sults obtained are discussed below. Surry Sequences AB-BETA The BMI-2104 analyses included two variations on the treatment of the AB-BETA sequences. The first, labeled 2-volume, considered the primary as well as the secondary containments, with each modeled as a single separate volume. The se'ond, labeled 4-volume, considered only the The primary containment but subdivided it into four compartments. These values analyses for the two cases are 0.89 and 0.82, respectively. are at the end of the calculations; it is reasonable to infer that essen-tially all the noble gases would be released eventually for these cases. AB-GAMMA The results for the AB-GAMMA sequence indicate release of 0.83 of the noble gases at the end of the calculation; complete release would be exp'ected eventually. l I , AB-EPSILON ( For this sequence, the fractional noble gas release at the end of the
, This value was relatively low because the analysi- .
calculation was 0.15.
' took into account the hydrostatic pressure outside the containment basemat; 8508150189 850710 l j PDR FOIA SHOLLYB5-447 PDR ,
[
. .._ _~_.
m__ _ 4 2 February 20, 1985 , Ms. Jocelyn Mitchell USNRC 4 since the containment pressure at the time of predicted meltthrough is relatively low, only a fraction of the containment atmosphere is Since the generation of noncondensables is predicted to be released. quite low, further leakage to the environment would be limited. S20-GAMMA The fractional noble gas release for this sequence at the end of the calculation is 0.46. The operation of the containment sprays in this sequence keeps the containment pressure low, except for the large hy-drogen burn that is postulated to lead to failure of containm the total release. S20-EPSILON The MARCH calculation for this sequence was not carried out to the point of containment meltthrough, thus leakage to the environment must be inferred. The operation of the containment sprays during this sequence At the time of expected meltthrough keeps the containment pressure low.the containment pressure was calcula little leakage to the environment would be indicated even at the time
! of basemat penetration. The principal leakage to the environment would take place relatively early in the sequence when the containment pressure is somewhat elevated. This leakage is estimated to release on the order of 0.01 of the noble gases to the environment.
TMLB-BETA Complete release of the noble gases was calculated for this sequence; , this particular MARCH calculation was carried out for a longer time than a number of the other sequences. TMLB-DELTA The fractional noble gas release at the end of the calculation for this sequence was found to be 0.85; complete release would be expected eventually. TMLB-EPSILON The fractional release of the noble gases for this sequence at the end The large release in this case compared of the calculation was 0.82. to the other basemat meltthrough cases is due to the combination of higher containment pressure and the assumption of depressurization down to atmospheric pressure. l V SEQUENCE The fractional gas release for this sequence at the end of the calculation was 0.95; complete release would be expected eventually. e "i. t J .
- t 6 I
. . - - . . . .. ~ . _ . - ,
i
',~' . . ...
February 20, 1985 3 Hs. Jocelyn Mitchell USNRC 7 Peach Botton Sequences ( Complete release of the noble gases was The indicated for all relatively small three of the free t Peach Bottom sequences analyzed in BitI-2104. volume of this containment together with higher noncondensable generation from concrete decomposition in comparison with the Surry results leads to little or no retention of the noble gases in any of these sequences. I trust that the above will be of benefit in your preparation of NUREG-0956 Sinceret j' / l/ 'n 00lsh eter Cybulskis Nuclear Systems Section cc: R. Meyer, NRC PC:dem
! i I
0 A i n
.4
- e 1 l
p ..a . - - s Sandia National Lab 0ratories Albuquerque. New Mexico 87185 i February 1, 1985 Mr. O. E. Bassett, Direct'or Division of Accident Evaluation i U.S. Nuclear Regulatory Commission Nashington, DC 20555
Dear Mr. Bassett:
Submitted herewith is a summary report of technical & program-natic activities and results for FY84. The report condenses details described in bi-r.onthly progress reports submitted throughout the year into summaries that should be of interest to those dealing with current severe accident issues. It should be noted that some observations are, of course, preliminary and subject to revision as more work is completed. The report is prefaced with a list of highlights, each cross referenced to more detailed discussion in the main text. . Our major FY84 focus has been on matters important to the defini-tion of severe accident radiological source terms. Specifically, the QUEST study provided insights into the nature and magnitude of source term uncertainties and defined the important severe accident phenomena whose uncertainties dominate source term un-certainties. Other programs provided data and computational support to the NRC source term reevaluation effort (BMI 2104 and CLWG). Of most importance were chemistry data describing the interaction of the important fission product species with struc-tural materials which were used in the TRAP-MELT code calculation of fission product transport and retention in the RCS. Also, detailed mechanistic computations of ex-vessel release for each i sequence analyzed in BMI 2104, using our newly developed VANESA model, were provided to BCL. An important perspective from QUEST was that a necessary condi-tion for technically defending lower severe accident source terms in safety deliberations requires that early containment failure probabilities be shown to be low. To this end a concentrated effort was wade to develop models necessary to quantify contain-ment threats from hydrogen burning and detonations, from direct containment heating due to melt ejected from the reactor vessel in.high pressure meltdown scenarios and from in-vessel steam explosions. This model development was accompanied by experimen-tal programs to define and quantify the important phenomena. Considerable progress has been made in quantifying the first two phenomena this year and the focus on dealing with the latter has been considerably sharpened.
- d, y o
- t .
\
l
as - ! , Mr. O. E. Banestt February 1, 1985 f* t This has been a productive year for the program developing the , mechanistic in-vessel melt progression code MELPROG with comple-tion and demonstration of the first useful code version. Code development has emphasized the phenomena shown by QUEST to be important to accurately describe fission product release and transport. The HECTR hydrogen containment response code and the advanced core melt concrete interaction code CORCON MOD 2 were also completed and released to the safety analysis community. Finally, nine ACRR tests were conducted in FY84: (1) DF-1 LWR fuel melt progression, (2) DCC-2 LWR degraded core coolability, (3) STAR 1 and 2 fuel and clad motion, (4) TRAN B-3, B-4, B-5 molten fuel hydrodynamics, (5) D-10 bottom cooled LMFBR debris coolability, and (6) DC-2 UO 2 melt progression. All tests were successfully completed and provided considerable data which are currently being used to develop and verify accident analysis code models. Questions on topics discussed in the report should be directed to me or to the individual authors designated in the text. Very truly yours, V0! ~
. V. Walker, Manager Reactor Safety Research Department 6420 Enc: As stated i
O 9
.\ . .
February 1, 1985 Mr. O. E. Baccott .. R. W. Barber, USDOE Z. R. Rosztoczy, NRC/NRR R. B. Minogue, NRC/RES L. C. Shao, NRC/D/DET D. F. Ross, NRC/RES K. G. Steyer, NRC/CEBR
- 0. E. Bassett, NRC/RES T. M. Su, NRC/GIB R. M. Bernero, NRC/NRR R. H. Vollmer, NRC/D/DE C. N. Kelber, NRC/RES P. Baybutt, BCL W. M. Morrison, NRC/RES R. Denning, BCL G. P. Marino, NRC/RES W. Kato, BNL M. Silberberg, NRC/RES M. Stevenson, LANL R. T. Curtis, NRC/RES T. Kress, ORNL R. W. Wright, NRC/RES M. L. Corradini, University T. J. Walker, NRC/RES of Wisconsin P. M. Wood, NRC/RES J. H. S. Lee, McGill S. B. Burson, NRC/RES University J. T. Han, NRC/RES T. G. Theofanous, Purdue L. K. Chan, NRC/RES University J. T. Larkins, NRC/RES R. Strehlow, University T. M. Lee, NRC/RES of Illinois C. L. Allen, NRC/NRR R. C. Vogel, EPRI F. H. Rowsome, NRC/NRR L. Baker, Jr., ANL J. F. Meyer, NRC/0CM R. Anderson, ANL J. Austin, NRC/0CM E. A."Warman, Stone &
R. O. Meyer, NRC/ASTP0 Webster J. A. Mitchell, NRC/ASTP0 M. W. Fontana, TEC C. Ryder, NRC/ASTP0 M. W. Jankowski, IAEA W. C. Lyon, NRC/NRR L. G. Hulman, NRC/NRR R. Van Houten, NRC/RES P. Worthington, NRC/RES C. G. Tinkler, NRC/RES C. W. Nilsen, NRC/RES T. P. Speis, NRC/NRR J. E. Rosenthal, NRC/NRR G. R. Burdick, NRC/RES J. A. Martin, Jr., NRC/RES P. M. Williams, NRC/NRR J. Reed, NRC/NRR J. L. Telford, NRC/RES M. A. Cunningham, NRC/RES F. P. Gillespie, NRC/RES V. Benaroya, NRC/CHEB W. R. Butler, NRC/CSB R. L. Palla, NRC/CSB K. I. Parczewski, NRC/CHEB J. Austin, NRC/0CM G. Quittschreiber, NRC/ACRS R. P. Savio, NRC/ACRS O s
\
c Reactor Safety Research
~
Summary of Activities and Results - FY84 ! for The Nuclesr Regulatory Commission (NRC)
' Editor, Jack V. Walker, Manager Reactor Safety Research October 1984 i
f i Printed January 1985 h Sandia National Laboratories c o'6)n o: g, e< .n 1 . l \ 1 . 8
, w v o n .~ ,. - - - - , . -, .
r
- 2. FISSIOM-PRODUCT SOURCE TERM 2.1 HIGH-TEMPERATURE FISSION-PRODUCT CHEMISTRY r
(D. A. Powers, 6422; R. M. Elrick, 6422; R. A. Sallach, '
.ho l 1846)
Q3 '
~
Or Release of radionuclides from the reactor fuel during a go ' severe a'ccident is but the first phenomenon that must occur 3, to create an inventory of radionuclides available for release
;w from the plant. Once released from the fuel, the radionu- i clides must travel to a point where they can escape the plant before these radionuclides have any safety consequences. Two I Processes can limit the efficiency of radionuclide transport within the reactor coolant system. Fission-products can 3 . react and bind to cool surfaces of structural materials
- h within the reactor coolant system. Aerosol processes can cause radionuclides to deposit or sediment on structures within the coolant system and eliminate these radionuclides from the inventory of releasable radioactivity.
p, The objective of the High-Temperature Fission-Product Chemistry program is to ex'perimentally characterize the chemical interactions of radionuclides that should affect 3 the efficiency of radionuclide transport'in reactor coolant 3 systems. Most of the work in this program has been focused on the reactions of volatile fission products such as Cs. I, and Te with structural materials. Results obtained in the g quantitative characterization of these chemical processes
> are used .in reactor accident source term analyses such as
. that sponsored by the NRC's Accident Source Term Project L Office. The tellurium retention on structures predicted in these analyses (the analyses are commonly called BMI-2104) is based on deposition velocities obtained in the High-r Temperature Fission-Product Chemistry program. Data obtained in this program on CsOH vaporization have also had an impor-tant use in the source term analyses sponsored by the IDCOR group.
- Recently, it has been recognized that radionuclides deposited on structures will heat these structures because of radionuclide decay. Heating of structur.es may be suffi-cient to cause deposited radionuclides to revaporize and again be available for release from the plant. Results obtained in the High-Temperature Fission-Product Chemistry program have a significant bearing on such predictions.
Experimental studies in this program have shown that depos-ited radionuclides undergo significant chemical transforma-tions. CsOH. i.s converted to cesium silicate, cesium phosphate, and the like. Tellurium is converted to metal
. tellurides. These chemical transformations alter the volatility of the radionuclides-frequently reducing the volatility. Analyses of revaporization phenomena in severe accidents have been, to date, scoping in nature and have not
. '\ * .
__e___________ _ _ _
e%' l l .' . V l
, at taken into account chemical trantformations of deposited ze 4
i radionuclides. Che'mical transformations may be more impor- at tant than gas flow for revaporization. r e-yses of revaporization will require More realistic anal- f.i data such as that o generated by the High-Temperature Fission-Product Chemistry program. p1 38, wB The activities in the High-Temperature Fission-Product tG) Chemistry loving areas: program during the year were focused in the fol- tg tt'
- 1. Definition of the chemical form of cesium deposited on stainless steel; si
! 2. Interactions of cesium hydroxide and cesium iodide tt Tl with boron carbide in steam; and p,
- 3. Effects of simultaneous steel oxidation on f.
i the g deposition of tellurium,
. c 2.1.1 Definition of Steel the Chemical Form of Cesium Denosited ) on Stainless a v'
i Very detailed studies of the microstructure and chemical
- composition of stainless steel exposed to cesium hydroxide tt vapors in steam and hydrogen were undertaken this year. The i
objective of these studies was to ascertain the chemical g nature CsOH. of the so-called " water insoluble form" of deposited Cesium hydroxide has been shown in this program to i deposit on stainless steel in two forms. One form is read- . I ily removed by (1). temperatures in excess of about 800*C. and (2) water leaching the specimen. The other form of i deposited cesium hydroxide is not water soluble. This water insoluble form persists on sta.inless steel coupons at tem-peratures up to at least 1000*C. This is the form of cesium i found on Island Unit 2. lead screws taken from the reactor at Three Mile
\
Quantitative analyses of the elemental distributions 1 within oxide coating formed on stainless steel in steam show that cesium is present as cesium silicate of 81 490Si). ! Silica is formed in the oxide by steam oxidation(Cs2 impurity in stainless steel. the The presence of cesium sili-i cate rather than cesium chromate may be surprising to some t i familiar fast reactors. with the literature of fuel / clad interactions for , Cesium chromate is not stable in the steam-hydrogen atmospheres of interest for radionuclide transport j in the reactor coolant system during LWR accidents. - f i Demonstration of the chemical form of deposited cesium is of major significance in reactor accident source term analyses. Quite clearly, cesium bound in the steel as cesium silicate is not susceptible to mechanical resuspen-l I sion as a result ,. of vessel depressurization or ~in-vessel r F
.,..p
_-.--._. ___..._..-._,.2.___,__._______.________,,___....______.__,.___._.,._._._ __
6- .
, l.
~
f f steam explosions. The susceptibility of cesium silicate to jo ' revaporization as a result of temperature excursions of the
; steel may also be limited. Analyses to date of cesium revaporization have presumed that the cesium remains in the 3
form of cesium hydroxide, which is quite volatile. Results . 7 e of the wo,rk in the High-Temperature Fission-Product Chemistry l Program show deposited cesium is transformed chemically to a ; v much less volatile form. At worst, much higher temperatures will be required to revaporize cesium from cesium silicate fP than f rom cesium hydroxide. The timing and even the magni-tude of cesium revaporization then should be quite different than has been supposed up to now. -1 Identification of cesium silicate also calls into ques-i tion the mathematical description of fission-product deposi-g tion found in accident analysis codes such as TRAPMELT. These codes hypothesize that the rate of deposition is pro-i portional to the vapor phase concentration. Other limiting factors exist. For one, the availability of silica to react
; to form cesium silicate imposes a limit on the extent of cesium deposition. Further, the rate of deposition must approach zero when the vapor concentration of cesium ~
approaches the cesium vapor concentration in equilibrium with cesium silicate. The importance of having proper chemical characteriza-
' tion of deposited radionuclides for revaporization analyses was graphically demonstrated during the year. Industry analyses of revaporization using literature data for the
. vapor pressure of cesium hydroxide and more recent measure-ments obtained in this program have been done. Results of the analyses show that the lower vapor pressures measured in the High-Temperature Fission-Product Chemistry program reduce the importance of revaporization.
Quite clearly, the detailed studies of the chemical form of deposited cesium conducted in the High-Temperature Fission-Product Chemistry program have had remarkable effects on the quantitative analysis of severe accidents. radionuclide behavior in ! 2.1.2 Interactions of CsOH and CsI with BqC in Steam A " conventional wisdom" is growing that iodine released from degrading fuel will quickly become an iodide and be ! l subject to aerosol physical processes. Such aerosol pro-cesses will mitigate the consequences of iodine' release from the fuel. The conviction that iodine will be in the form of Cs! after release from the fuel is based, to a large extent. on results of simplistic thermochemical analyses of the Cs-I-0-H system. Such simplistic calculations fail to take into account the wide diversity of material encountered when Cs and I are released into the atmosphere surrounding a degrading core.
* . ., h
.h
.i
. . *(.
Work in the High-Temperature Fission-Product Chemistry program has demonstrated several instances where the pre-ferred form of iodine and cesium is not CsI, as held by the
" conventional wisdom". One of these instances is particu-larly important for Boiling Water Reactors (BWRs). Cs! will react with the products of steam oxidation of BgC control blades according to a formal stoichiometry that is approxi.
mately: CsI + HB02 CsB02 + HI . The gaseous species HI is susceptible to the equilibrium 2HI - H2 + 2I . Detailed thermochemical analyses of the B-Cs-I-0-H system were reported this year by investigators in the High-Temperature Fission-Product Chemistry program. These analyses show that the reaction requires some dilution of the activity of either CsBO2 or HI to proceed signifi-cantly in the direction of Cs! dissociation. One easy way to reduce the activity of CsB02 is to simply dissolve it in another oxide such as condensed phase B023 that forms on B 4 C exposed to steam. Results of the thermochemical calculations show that cesium activity., especially if it reacts with boron-bearing l species, will be reduced relative to that of CsOH. This is yet another rea' son revaporization analyses hypothesizing the chemical form of cesium to be CsOH would be in error. The behavior of I released by reactions of Cs1 is not yet well understood. Clearly, it passed through the reaction system L, , in experiments to date. It might not be so nobile in the
!. complex chemical mixture flowing out of the reactor in a severe accident. .
o ,t Another known cause of CsI dissociation is the photoly-sis of Cs1 in a radiation field. Cesium is readily ionized 4 I h d by gamma rays. Cesium iodide, when ionized, dissociates to yield Cs+ and atomic iodine. In the environment of a reactor accident, an intense flux of, gamma irradiation would
'b e present and a quasi-steady state concentration of Cs+
and consequently I would be expected. Thermochemical cal-cul.ations to date.have failed to take into account the pho'- tolysis processes. Because the formation and destruction of ions is a dynamic process, it canhot be treated within a ~ strictly thermodynamic formalism. Rather, a set of unknown I kinetic parameters are required. s.
--- -: n.. . -. - . ._, .- , , _ d. . - . . , . - . . . - ~ .
l r .-
1 l ! l
. To evaluate the importance of gamma photolysis on the chemical form of iodine in a severe react'or accident, a series 'of tests are being planned. The test series will "
l consist of identical tests in laboratory environments and in an 60Co. intense gamma irradiation field created by the decay of i I on bothThe tests will evaluate the photolysis t'ie effects ofofirradiation and the revaporization CsI. The i first lab ~ oratory runs for the test program have been con-ducted. Results of tests in the irradiation field could alter dramatically the conventional wisdom on the chemical l form of iodine. 2.1.3 Effects of Deposhtion Simultaneous Steel Oxidation on Tellurium f Tests of the simultaneous oxidation and tellurium deposi-tion on stainless steel were conducted. Tellurium was found to be bound to nickel nodules in the oxide on the stainless steel as hadrather been than as a mixture observed previously. of iron and nickel tellurides This observation again ( shows that fission-product deposition on surfaces in a reac-tor coolant system may not be strictly a function of vapor concentration. The availability of reactive surfaces, in this case nickel, also may be rate-controlling. 2.1.4 Acen Act d ene. Lvit:,es Source Term Proiect Office (ASTPO) Related An important aspect of the NRC's Source Tern Reassess-ment effort has been the peer review and review by the American Physical Society of the accident analyses reported in the documents commonly known as BMI-2104. Investigators for the High-Temperature Fission-Product Chemistry program were diverted to preparing voluminous responses to hosts of technical inquiries made by members of the APS review com-mittee. Detailed responses were made to the inquiries con-carning the following topics: 1. Chemical form of radionuclides in fuel and in the fuel / clad gap: 2. Reactions of CsOH and CsI with Bec : 3. Solubility steam of fission products in high-temperature
- 4. Chemical state of Ba and Sr during fuel / clad inter-actions: '
- 5. Basis for CORSOR rate equations 0
/
I.k a , m .s
=
. . i i
)
- 6. Effects of oxygen poten'tial and pressure on fission- l product release; ;
- 7. Revaporization of deposited rad'ionuclides: ,
- 8. Effects of radiation and radiation-induced nitric acid formation on the chemical form of iodine; and I
- 9. Resuspension of deposited aerosols coolant system.
in the reactor 2.1.5 Other Activities Investigators in the High-Temperature Fission-Product Chemistry program are obligated to assict investigators involved with the PBF severe fuel damage experiments in interpretations of their chemical data. To this and:
- 1. Thermochemical analyses of the Cs-I-O-H system were l conducted to determine the chemical form of radio- l nuclide species during transit through the flow )
network at PBF.
- 2. Modifications were proposed to the CORSOR code to incorporate the effects of pressure and flow on the release of radionuclides as a possible explanation 1 of differences between release results obtained '
out-of-pile at ORNL and in the PBF 1-1 tests. .
- 3. A model of reverse reaction and flow stagnation was formulated to interpret results of the test PBF 1-3.
The models of fission-product release have a larger sig-nificance whether or not they are indeed applicable to the interpretation of complex, integral tests such as those done at PBF. .These models incorporate physical and chemical processes not recognized in the calculations done for the NRC's source term reassessment, but expected to be operative in accidents. These processes will' be most important for I those accidents in which core degradation takes place in a pressurized t vessel such as the TMLB' and TC sequences. j Unfortunately, these accidents are often the risk-dominant accidents for many plants. Quite accurate source term f estimates are most desirable for these accidents. The effects described by the models depress release rates. Thus, real releases may be much lower than anticipated in I the BMI-2104 calculations. Aerosol processes may then be less officient and retention of radionuclides
- within the' reactor' coolant system may be less important. . .
f . Also during the y'ar e investigators in the High- i
* ; Temperature Fission-Product Chemistry program were called upon to review documents produced by IDCOR and to partici-meetings
~
pate in with IDCOR dealing with source term
concerns. Reviewing the documentation produced by IDCOR I was, of course, an unprogrammed task that required a remark- s
, able amount of. time. Critiques of the IDCOR work were pre- J sented at meetings with IDCOR. Among the comments concerning the IDCOR work as it relates to research in the High-Temperature Fission-Product Chemistry program were:
i S
- 1. Chemical transformation of Cs1 by reaction with l borates ought to be considered in sequences for BNRs.
- 2. Chemical transformation of deposited species ought -
to be recognized in revaporization analyses.
- 3. Revaporization of Te ought to be considered..
Finally, investigators working on the High-Temperature Fission-Product' Chemistry program participated in the , preparation of ' a document commonly known as NUREG-1053. i i This document is intended to summarize the state of knowl-
) edge concerning source term phenomena, the data needs, and the modeling capabilities. Specific ~ responsibilities were assumed by investigators in the program to summarize knowl-edge about fission-product release during core degradation.
1 A draft of the chapter on this topic was p,roduced. It is a
; rather thorough review containing in excess of 200 refer-ences dealing with both experiments and modeling. The i review concludes with an importance ranking of information needs and an indication of where these needs will be met.
2.1.6 Documentation R. Sa11ach, Vapor Pressure of Lieulf CsOH, SAND 84-1693 Sandia National Laboratories, Albuquerqua, NM, 1984. R. A. Sa11a cht, R. M. Elrick, A. L. Ouellette, and S. C. Douglas, Reactions Between Some Cesium-Iodine ConDounds and the Reactor Materials 304 Stainless Steel. Inconel 600 and Silver. Volume I. Cesium Hydroxide Reactions, NUREG/CR-3197, SAND 83-0395, Sandia National Laboratories, Albuquerque, NM, 1984.* R. A. Sa11ach, Calculated Vapor Compositions for the H+O+I+Cs+B System, SAND 84-0662, Sandia National Laborato-ries, Albuquerque, NN, 1984. R. A. Sallach, Chemical Assocts of Cs! Interaction in , Steam with 304 Stainless Steel and Inconel 600, SAND 84-0749, 1 Sandia National Laboratories, Albuquerque, NM, 1984. l l
~ 'o 5
- r
,s-
" g.
R. A. Sn11ach and R. M. Elrick, " Chemical CsOH, Cs1 and Proceedinos Te Vapors with Oxidizing of- the Reactions og Topical Meetino Reactor Materials.* Behavior and on Fission-P g 1984. Source Term Research, Snowbird, UT,
- July 15-19*
2. A. R. Taig, LWR Cores," " Release and Retention Phenomena in Degraded Roduct Proceedinos of the Topical Meetino on Fission-Behavior Juif 15-17, 1984. and Source Term Research, Snowbird, W
- A. R.
the Taig, " Interpretation Interactions of of Experimental Results on Fission-Product Vapors with
- Materials," Proceedinos of the Product Behavior and Source Term Research, Topical Meetino Reactor d
July 15-17, 1984. on Fission _- Snowbird. UT, tl Et R. A. Gallach, Jyteraction of C. J. Greenholt, and A. R. Taig, Chemical tl Tellurium NUREG/CR-2921, SAND 82-ll45, Albuquerque, NM, March 1984. Vapors with Sandia Weactor Material _s, National ( Laboratories, r' g D. A. Powers, d Deoradation, Chapter 3Release 'of Radionuclides NUREG-lO53. Durino Core ~ 2.1.7 Meetinos Attended t t Bethesda,SevereMD.Fuel Damage Meeting. 'I October 29-November 2, 1983, ' I lith LWR Safety Research 1 ber 22-26, 1983, Gaithersburg, MD.Information Meeting, Octo- ' American Physical 1984, Boston, MA. Society Review Meeting, January 15-18, American Physical 1984. Albuquerque, NM. Society Review Meeting, March 22-24, Review of IDCOR Source Term 1984, Hunt Valley, MD. Analyses, . February 6-10,
\
Rockville, Review MD.of IDCOR Source Term Analyses, May 13-15, 1984, i querque,Severe NM.Fuel Damage Meeting, April 10-11,,1984, Albu-Review of ! ations. August 27-29,IDCOR Analyses including Source Tara consider-1984, Rockville, MD. Peer Review'BMI-2104, January 25-27, 1984. Peer Review BMI-2104, October 11-14, 1983. ! i i " ~ , .
Planning for NUREG-1053, June 4-5, 1984. g Planning for NUREG-1053, June 27-28, 198k. 2.2 QUANTITATIVE (QUEST) UNCERTAINTY ESTIMATE OF THE SOURCE TERM to (D.,A. Powers, 6422; P. K. Mast, 6425) ' h 2.2.1 Obiective of the Work A key element of NRC research over the last two years in has been the development of a new, mechanistic method for ir g estimating during radioactivity a severe accident.release from a nuclear power plant
, the use of mechanistic The essence of this new method is release and subsequent behavior of codes computer to estimate the radionuclides in ways y thataccidents and explicitlyinrecognize question. the peculiar features of plants '
i The objective of the new method 3 is to provide more accurate and more realistic estimates of radioactivity release during severe reactor accidents than is possible using the prescriptive source term tabulations developed for the Reactor Safety Study. j L As part of the development of the new method for source 1 term estimation, a number of accident analyses were done and 1 documented The computerincodes a series used of reports commonly known as BMI-2104. for these analyses were known, how-ever, knowntotobe imperfect great accuracy. and input data for the codes were not A question then arises about the magnitude of the uncertainty that ought to be attached to source term estimates made using the new methods. The objective of the Quantitative Uncertainty Estimate of the Source Ters (QUEST) program was to develop an esti-mate of uncertainties that ought to be ascribed to mech-anistic source term estimates using the currently available state-of-the-art computer models. Both the general magni-tude of the uncertainties and the disposition of the uncer- l sought.range about a point estimate of the source term was tainty study. Two Thetypesfirst of uncertainty type were recognized in the QUEST : of uncertainty--the so-called uncertainty--relates to the suite of computer codes utilized &c for the analyses of the accidents' considered for the BMI-2104 documents. These computer codes represent the current state of the art. accepted, these As the new method of source term estimation is for implementing the computer method.codes are what will be available variety of The codes rely, however, on a - precision. input data--much of which is not known with justification, differentDifferent analysts may make, with good technical of codes. choices for the inputs to the suite The 4e' uncertainty then provides an estimate 4 1 s - rhb _ _ _ -. Y-- - - - - ~
c.2% .
, .A c er n jur. h L T ' '< ^ 'Y I
)
g ggg cats June 28, 1984 apu nusa to Q-7-84-318
/
( Los Alamos Nationallaboratorg Los Alamos.New Mexico 8754
***57o' K556 innes (505) 667-2023 or FTS 843-2023 f
Energy Devision Dr. James T. Han Reactor Safety Research Branch Mail Stop 1130SS US Nuclear Regulatory Commission Washington, DC 20555
Dear Jim,
Enclosed is a draft letter report that presents the results of the TRAC-PF1 upper-vessel-circulation calculation for the Surry TMLB' accident sequence. The AB sequence calculation is completed and a discussion of the results will be included in the final version of the report. Please call me if you have any questions or require futher information. Sincerely yours, Rudolph . Henninger Safety Analysis RJH:bn Enc. as cited Distribution: CRM-4 (2), MS-A150 C. Kelber, NRC R. Denning, BCL File (RJH)' ( f 5-() h 'I? O An towel opportunity Employer 10perstod by Unweety of ceMornia
n '
,s e e, w 1 TDRAFT h\ /
t VAPOR CIRCULATION IN THE UPPER VESSEL OF THE SURRY
~'
.'s- I. PWR FOR THE TMLB' ACCIDENT SEQUENCE t.
L R. J. Henninger
- 1. INTRODUCTION The above-core structures can provide a significant heat sink during a degraded core acciden,t. In order to determine the extent to which the structures affect an accident, TRAC-PFi l calculations for the Surry Pressurized Water Reactor were performed. The sequence chosen was 'a total loss of feedwater with failure of the energency core cooling (ECC) system (TMLB"). Core outflow conditions, that consisted of time-dependent steam and hydrogen mass flows and vapor temperatures, were used as boundary conditions for the TRAC-PF1
( calculations. These core outflow conditions were calculated by means of the MARCH code and were provided to us by Battelle Columbus Laboratories.2
- 11. MODEL DESCRIPTION A. TRAC Model The TRAC model for the upper part of the Surry vessel is shown in Fig. 1.
The model consists of 7 axial levels, 3 radial regions and 2 azimuthal sectors. The inner two radial nodes model the region inside of the core barrel. The first five axial levels correspond to the region between the core support plate and the upper support plate. The upper two axial levels model the upper head. In each of the four nodes inside of the core barrel there is a pipe that provides a connection between the bottom of the upper plenum and the upper head. These four pipes represent the 53 control rod guide tubes (CRGT) in the Surry vessel. Flow through the CRGTs is restricted by a small total flow area of 0.1294 m2 near their tops. Small-area flow paths between the downconer and r upper head and the downconer and the inner radial regions were yso modeled. The hot leg with the- pressuriser is connected to one of the tho azimuthal [ sectors.
C3r __ . -. _ _ . . _. _ In TRAC-PF1, only one heat slab is allowed per node. The vessel noding was therefore chosen so that " thin" structures within the core barrel could be separated from " thick" structures such as the upper support structure and vessel ! wal(s. Thin' structures were typically 0.006 to 0.008 m thick. The heat slabs in ghe fif th axial region, which includes the upper support are 0.022 m thick and those in the the outer radial node which models the vessel range from 0.15 to 0.30 m thick, dipending upon the presence or absence of nozzles and flanges. The mass of the CRGTs was divided equally between the vessel component and the pipes used to represent the guide tubes. All of the heat-slab masses and surface areas were obtained from Westinghouse via Battelle Columbus Laboratories.3 B. Boundary Conditions One of the boundary conditions for these calculations is the pressure in the hot leg. The other boundary condition, as indicated in Fig. 1, is the flow at the core outlet. The conditions for the TMLB' sequence are given in Figs. 2-4. As shown in Fig. 2, the mass flows decrease from the time of core (E uncovery at 5730 s until approximately 8760 s. At 8760 s, the core slumps into water below the core region producing an increase in core outflow. The vapor temperature shown in Fig. 3 increases with the center (higher power) node leading the outer node. The calculation was stopped when the temperature returned to the saturation temperature and the flow from the core re61on was steam. Figure 4 gives the total pressure and the hydrogen partial pressure at the core outlet central and outer radial regions. The total pressure remains near the relief valve setpoint (16.3 MPa) throughout the transient, and the fraction of flow that is hydrogen increases as the accident proceeds. C. Initial Conditions The TRAC-PF1 calculation of the 'IMLB' sequence was begun when the core was uncovered and the vapor temperature at the core outlet began to increase above the saturation temperature. The initial conditions for the 'IMLB' sequence were J that the vessel and hot leg were at the saturation temperature corresponding to This seems a reasonable assumption especially for the thin 16.3lMPa(622K).
- structures. The vessel temperature is not very important in thesescalculations t
because, as we shall see, there is not much flow to either che uppei head or the ( downconer.
. _ ._...._a. . _ . __. _. +.. .
( t III. RESULTS FOR THE TMLB SEQUENCE j The TMLB' calculation was run at the initial conditions foy 400 s (from 5360, to 5760 s). As the temperature of the vapor flowing from the core inc/ eased alter 5760 s, a flow pattern similar to that depicted in Fig. 5 is t- j devgloped. This pattern consists of two major convection cells. Most of the upflow is in the central radial cell that is on the side of the vessel next to the hot leg. The returning downflow is in the outer radial cells. The axial mass flow at the top of cell 2 in the vessel is given in Fig. 6. This pattern persists until the rapid flow increase that occurs when the core slumps at 8760 s. The flows within the upper plenum can be compared to the inflow from the core and the outflow through the hot leg which are given in Fig. 7. Comparison of Figs. 6 and 7 shows that the flow within the vessel remains large compared to the inflow and outflow. The vapor from the core is therefore well mixed with vapor in the upper plenum before it exits through the hot leg. This is seen in Fig. 8, which gives the core outlet vapor temperature and the vapor temperature in the hot leg. The decrease in temperature is a result of the ( mixing of the vapor from the core with the vapor already in the vessel. Until the flow increase at 8760 s, the temperature in the hot leg remains relatively low. The energy flows in the vessel are given in Fig. 9. This figure indicates that the major energy removal mechanism up to 8760 s is flow out.through the hot leg. The heat slabs participate very little. The flow is therefore driven by density differences between the vapor in the vessel and the vapor leaving the core region. The vapor entering the upper plenum from the core is less dense for two reasons. The first, of course, is that its temperature is higher. The second reason is that the fraction of the flow that is hydrogen is increasing. The importance of these mechanisms in driving flows will be examined by J. Dearing with his two-dimensional MELPROG flow module. Flow through the CRGTs was 5 kg/s or less and not important for energy transport. The flow directions with the CRGTs, as is indicated on Fig. 5, were similar to the flow pattern in the upper plenfs,namelyupthecenterradialpipeonthehot-legsideanddowntheother pipes. FlowsintheCRGTsonthehot-legsidearegiveninFigs.lgand11. The increase in mass flow associated with core slumping results in an f' altered flow pattern. The flow pattern at 8800 s is shown in Fig.12. The convection cells persist but some of the vapor flows more directly to the hot leg. As the flow increases, so does the importance of the heat slabs. Figure 6
p:7., - . . .. - . . . . _ . . . . . - - . - . .-.. .- . - . . . - . - . . . . . - . . 1 i t shows that the energy flow to the heat slabs becomes significant after 8760 s. The temperature of the heat slabs in Figs.13 and 14 increases pignificantly following increased core outflow. .. t- .. E. IV.. CONCLUSIONS AND RECOMMENDATIONS s The important conclusions that result from these calculations are:
- 1. For the flows provided by Battelle Columbus, the vessel structures were not an important heat sink from the time of core uncovery until the time of core slumping;
- 2. Flow driven by differences in density between the vapor exiting the core and vapor present in the vessel resulted in mixing and lower temperature vapor exiting the vessel;
- 3. Flow areas of the connections between the upper plenum, upper head, and downcomer are too small to be of any importance to the energy t
flows; and
- 4. The vessel structures become more important as heat sinks when the
~
([' core outlet flow increases following slumping. I believe that a coupled multi-dimensional analysis of this accident that included the core region would produce higher flow from the core region thereby increasing the importance of above-core structures. I therefore recommend that this calculation be re-run when such a capability exists. This accident sequence will provide an excellent test for the multi-dimensional version of TRAC /MELPROG. REFERENCES
- 1. Safety Code Development Group, " TRAC-PF1, An Advanced Best-Estimate Computer Program for Pressurized Water Reactor Analysis," Los Alamos National Laboratory report LA-9944-MS (NUREG/CR-3567), February 1984. ;
- 2. Roger O. Wooton, Battelle Columbus Laboratories, private communication 1 March 27, 1984. '
r
- 3. Peter Cybulskis, Battelle Columbus Laboratories, private gommunication October 1983. t i
4 l
- -~ -- - --- '
m . . k SURRT UPPER PLINUM AND HEAD I trygt t y .. E 7 k 6 ,
. CONTROL ROD GUIDE TUBE ma see -namnmen mmmmmmme - ma S h COOLING N0ZZ11 4
3 HOT LIG 2 -.- 1 i t t t ' CORI OUTFLOW PROVIDED AS BOUNDART CONDITION Fig. 1.
. TRAC noding diagram for Surry upper vessel.
20 , , , , , , , CENTER CELL } l' e- OUTER RADIAL CELL . 7 N
,o D- .
v a S i w M 3- . 8 s n_
. _a .
~
r -s
$000 5503 6$00 6NO 7dOO 7b3 $$D 85'00 9000 TIME (s) {
Fig. 2. , 1 Outlet mass flow for the two radial regions in the core. The i flows are azimuthally symmetric. j
Gb .. ...:--...-. . . . . . .w . - - . t ,,. ... 2600 . . . , i i . k' 2 00 .......... CENTER CELL - [ ' 22w - OUTER RADIAL CELL 8 7. v 2000- " i
? ? mW l
- t
.s .!
( 8 1600-
,/
~
W00- *** " p ,,...****...***** 6 1200- .*** ~ g
> 1000 ........ -
B00- ,'
..*.**. ~
600- ~ 400 . . . . . S000 1500 6000 6500 7000 7500 8000 8500 9000 TIME (s) Fig. 3. Outlet vapor temperatures for the two radial regions in the core. ( 18000000 , , , , , , , 160000s0- ; , I
- l l u000000- .
12000000-10 tat **tsut l/l ll/ 9 g10000000 -
//
.... ..... mvoctw past.n6 espug or (g ,ge negicy. .
48 pooctu past g Il pagug or Atte pacis ass + g800'000- f n // I6000000- f[ - O. + r
,5 i ,
4000000- - (( 2000000- l/
,l
... /
l ;.
}-
t
, 4_ , .....................- ,,
-2000000 . . , , , .
. SOOO S$00 60'0 s 6500 7000 7LOJ 2000 8500 90'0 r TIME (s)
Fig. 4. Total pressure and partial pressures of hydrogen at the c(ore g outlet. Pressure drops through the system are small, so the total pressure throughout the primary remain near this pressure.
m . . . . - . 8 e, **. i r , I SURRY UPPER PLENUM AND HEAD LEVEL . t t- - 7 t g 6 . . [ .
- m. - ime- mm menm- - amm -men me 5 <
4 - : h U 3 V < ; HOT LIG _ > n m , . 3 '
,_ 3, 1 - -
t t t t CORI OUTFLOW PROVIDED AS BOUNDART CONDITION Fig. 5. Flow pattern in the upper plenum for TMLB' sequence from core uncovery to core slumping. This flow is driven mainly by (' density differences between vapor exiting the core and vapor already present in the vessel. The vessel heat slabs are of little importance because of the limited core outflow in this time regime. Flow in the CRGTs was limited and of little importance. Y l 1 ( 1 l 1 1
..n . + --. .. -.- --..--- .
Q,, , , , , ,, . . . - - . - - - . - - - - - . . - - . - - . . - - . - . - - .-.. . t ,,. ... 100 , , , , , , , g h 44tta 3 utelite it0e l gg. seede ea0**6 St&#W (MHaft El 186
=. e, * ** *-... easd e eaceas eg gsk es'A stG S M
.,'*. - (Wtte SaDias atbah (**($sti east LIG gg. **.,, -- Ehrte eneia6 etteh as,f af& Sati . I n
(e **...,'..,,- i a- 40- , .
&_ D
; s .,... !
. .o 20- % . . . .,'
h v . y o_ bd N . 2 !
. .., ,#i '
o i ~... . Wf, - h ,/.. \ ,
.<0 .
.p./.J i -
. -6 0.-
-80 . . . . . . .
5000 SS:,0 0000 6500 70",0 75e3 2000 21:0 5000 TIME (s) Fig. 6. Vapor mass flows for the four segments within the core barrel at the top of axial level 2. 'lhis figure shows that the vapor flows up the center cells and down the outside cells as {'c depicted in Fig. 5. 40 , , , , . . . 35- -- -- IN FROM CORE ,. 30-i, OUT HOT LEG .
\,
g 25- g , 6 os gg-s :* 0 , C 15 - 1 e i 0 ' s e- '.. 5 '.,
., n.
r
-5 . , . . s 5000 5500 6000 6500 7000 7500 8000 8500 9000 t "
TIME (s) ( Fig. 7. Mass flows entering the upper plenum from the core and exiting through the hot leg for the TMLB' transient.
..~ -
2600 , , , , , . . (- 2400- ....... - CORE OUTLET -
. k 2200- HOT LEG 1 ~
5 g, 20w- k 1
- l'. g
?
n 3 t woo- l i-
;. I L ? ux- ! -
1400- - H ,,...****,,,........**
- s. 1200- -
o g 1000- .**
. . . . . ' . . . * * [- .
l soc- :-
.w. -- .*. J 600- -
400 , , . . . . 5000 5500 6000 6500 7000 7500 8000 8500 5*00 TIME (s) Fig. 8. Core exit and hot-leg vapor temperatures for the TMLB' transient. The vapor temperature in the hot leg does not ( .- increase significantly until the core slumps at 8760 s.
'b 7 35 , , , , . , ,
TO HE AT SL ABS 30- ~
.......... FROM CORE g
--- OUT HOT LEC !
25- 3 - [ 20- :- s : o : C 15 - i-x i o 6 e
,E 10 -
......g {-
.s .
s- ' s,
' ' ~.~....... _.e \_
r -s , , . . 5000 5500 6000 6500 7000 7500 8000 8500 9000 , TIME (s) i. Fig. 9. ( Energy flows from the core, out the hot leg and to the vessel heat slabs for the 'DfLB' transient. The heat slabs are not important until the core slumps and the mass flow from the core region increases.
D_. . . - _ _ _ _ . _ . .. -___ ._ __
. , . . e..
, 1 . . . . . i i O~ ~
CRG1 fio. Outer Rodiol Node
~ '
t y --
- c. .
?-
L ? g -0 s - \ - 6 8 -i- { a m g -ss- - 2
-2 -
-2.s - -
-3. . .
7 5000 5103 5000 c500 7003 75C3 8500 25*J 5;,L0 TIME (s) Fig. 10. Flow in the CRGT in the center on the hot-leg side. ( 6 , , , , , , , s- CRGT Vtow t we r Radiol Node - 4 . 7 N c' 6 3- - a 2 2 . E O 2 1- -
- c. .
7 -1 . , , . . . . 5003 5503 E003 (533 7050 7100 E;03 E!D3 9:03 TIVE (s) Fig. 11. ( I Flow in the CRGT in the outer radial node on the hot-leg side.
~
I I
S'L .. . . . . - . . _ , . _.
. , . e SURRY UPPER PLENUM
.. ANDHEAD LIVEL i i -
7 l 4 e. b 6 i 'na uma mm- mum === - ma 5 < < ,] 4 - : , A " 3 V r
- MHOT LEG
_4__. .. 2 -
-m i M -
, ; ; ;<: 4y t t t t '
CORI OUTFLOW PROVIDED AS BOUNDARY CONDITION L Fig. 12. Flow pattern in the vessel at 8800 s. Some of the vapor from the core exits directly to the hot leg. ( 000 . . LittL 1 leae RADI AL MC10% OPPOStig Hot LIG Uoo-
--- - - Iso p RADIAL K GioN M01 LIG ssK *t ~
- outta BADI AL REGION. o*POSITE Mof LIG
- F3 ,,oo . -- -- ourts sAois mCion Not tro six _
t looo- - I a. E soo- - e I g soo- - o 3 A 700- - _ sn - soo- - M - r 5000 sioo sdoo sioo 7doo 7soo adoo sioo 'sode TlWE (s) Fig. 13. . Surface temperatures of the heat slabs in level 1 for the two ( radial regions within the core barrel. t O F
. _ . ..~.~....-..-_ - - i +- ..... .. l . 1 . l l
MO , , , , , , , I 3
.. WWL &
leste RADie atsiON. OPP 0Sitt M0f Ltc , i !
... ... lastt 94044 atG60N. NOT LIG SIDE '
- Outta ea084 atCION, OsPostTL M01 LEC p g p
7-
-- evita saDiu assiO=. not uD siot t- e i S00- ~
' 3' 0
6 750- - k 3 100- . H xsn s50- - 300. . . . . . . . . . 5000 S500 4000 6500 7000 7S00 8000 8500 0000 TlWE (s) Fig. 14. Surface temperatures of the heat slabs in level 5 for the two radial (_.-g regions within the core barrel. a 6 v (
. ,J .
, *s f0
~
C4Ba11elle Columlius Lalx>ratones c 505 King Awnue Columinus Ohso 41201-2691 Telephone m l 47 414 6414 April 11, 1985 ! I Mr. Ralph Meyer Fuel Behavior Branch Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Commission Washington, D.C. 20555
Dear Ralph:
Contract No. NRC-04-84-127 Task 11 Enclosed is the long promised letter report on our TRAP-MELT / MERGE combination effort. It highlights those aspects of this effort that should be of general concern, as well as a limited selection of results from a demonstration calcu-lation. Details of the novelties of the new code (TRAP-MELT 3) relative to the TRAP-MELT / MERGE codes as used for BMI-2104 will be contained in the documenta-tion for the code package. There is one caveat. The figures show release fractions of core inventory up to vessel failure but do not include the source to containment resulting from suspended nuclides at the time of failure (puff release). The numbers can therefore not be readily compared to tabulated BMI-2104 values. I will send you revised figures that include this release in the near future. Since the puff release will dominate the overall release, it will reduce the difference in calculated with and without decay heat release fractions. Submission of this report completes all efforts and satisfies the reporting requirements for Subtask 5 for Task 11, NUREG-0956 Support Calculations. If you have any questions or comments, please give me a call. Si cerely, James A. Gieseke Physico-Chemical Systems Section JAG:drr Enc. cc: Sharon Wollett Division of Contracts 9 5~ O 09 O HO L
w -
. . s
- j 1
+
.i l REPORT 1 i on TRAP-ELT/ MERGE MERGING to U.S. NUCLEAR REGULATORY COMMISSION i , April 11, 1985 4 by Hans Jordan BATTELLE Columbus Laboratories 505 King Avenue Columbus, Ohio 43201 4 4
TRAP-MELT / MERGE MERGING by Hans Jordan April 11, 1985 ) At the beginning of 1984, the U.S. NRC initiated a task at Battelle intended to investigate the effect of distributed decay heating on radionuclide transport in the primary coolant systems of LWR's during meltdown accidents. Such effects manifest themselves through a backcoupling of radionuclide trans-port (TRAP-MELT code) to PCS thermal hydraulics (MERGE code) and require an intimate coupling of these two (or similar) codes for their proper treatment. This task therefore consisted of a code modification and writing effort to combine the TRAP-MELT and MERGE codes and a debugging and code running effort to verify the viability of the resultant code and to illustrate, by suitable example, the consequences of including decay heating effects in the treatment of radionuclide transport. This letter gives an overview of the modifications undertaken in TRAP-MELT and MERGE in order to combine these two codes in one code, the TRAP-MELT 3 code. Included are also selected results of runs for the Surry TMLB sequence that illustrate what are expected to be the maximum consequences of the effect of decay heating on radionuclide transport during the pre-meltthrough period. In concurrence with our previous observations, based on an iterative procedure, and with the observations of others, decay heating effects appear to be of minor consequence to radionuclide transport during this period. The potentially more interesting post-meltthrough period has not yet, however, been explored, as will be explained in the following. This letter report constitutes the conclusion of this TRAP-MELT / MERGE ! combination effort for the NRC. l l i Modifications of TRAP-MELT and MERGE and the Architecture of TRAP-MELT 3 l TRAP-MELT 3 is a new driver routine that calls on the modified MERGE ! (as a subroutine) to calculate system thermal hydraulics including gas flows, gas and structure temperatures, gas pressure, gas composition, gas transport l l
n . 2 properties, and heat transfer coefficients as averages over a MARCH time step and on the modified TRAP (as a subroutine) to calculate radionuclide transport based on these system conditions. Previously TBAP-MELT required precalculated thermal hydraulic conditions to be read as input. These are now transmitted In addition, previous concurrently from the MERGE to the TRAP subroutine. subroutines of TRAP-MELT that were used to calculate gas properties and heat transfer coefficients could be, and have been, eliminated since these quanti-ties are needed, and therefore calculated, in MERGE. This treatment is not In only expedient, it ensures a consistency that was not achieved before. particular, full account is taken of the steam / hydrogen mixture calculated for and control volume at each moment in time. To allow such a consistent tr ment, TRAP was modified to allow for internal structures in each control volume. The primary system is therefore now nodalized in a consistent manner for both the thermal hydraulic and the nuclide transport treatment. TRAP now tracks the fraction of core inventory of each radionuclide Deposition is group that is suspended or deposited in each control volume. further resolved by structure. This is used to calculate decay heating of gas and structures and is the only TRAP generated data used by the MERGE subroutine. All decay energy associated with gamma rays is assumed to he absorbed The distribution by the control volume structures, i.e., no gas attenuation. of the decay heat among the structures is based on the following assumptions: (1) Half of the energy associated with the decay heat from deposited material is assumed to be absorbed by the structure on which the material is deposited. (2) The other half of this energy is distributed . among the other structures in the control volume based upon their relative surf ace areas (3) The energy associated with the decay heat from airborne material is distributed among the control volume structures, based upon their relative surface areas. Beta decay energy is treated similarly except that allowance is taken for energy lost to the gas traversed by the beta rays. A simple model is included that approximates the amount of energy lost to the gas before impacting a structure by examining the mean range of the beta particle in the gas as a function of the gas density and particle energy.
3 The fraction of the beta decay energy absorbed by a control volume gas is assumed to be the ratio of the mean distance a beta particle must before reaching a structure (d) to the mean range of the beta particle in the gas (R). These quantities are approximated by: d = h for particles emitted from deposited radionuclides
= f for particles emitted from airborne radionuclides
.09541 in Eg )*
R = .412 1p Eo(1.265 where V = volume of control volume A = total surf ace area of control volume structures R = mean beta particle group (cm) o
= gas density (g/cc)
Eo
= initial beta energy (Mev).
This step is chosen MERGE and TRAP communicate at system time step intervals. to be the MARCH code time step. Over this time step MERGE iterates an internal time step that is chosen such that fractional changes in mass and energy of a control volume are limited. This internal time step also drives TRAP. Results of a Demonstration Calculation To illustrate the operability of TRAP-ELT3 and the effects of the superposition of decay heating on the thermal hydraulic analysis, TRAP-ELT3 Data from the MARCH was run for the Surry TMLB sequence treated in BMI-2104. code csiculations for that document were used to drive the calculations. Nodalization of the primary system was however reorganized to reflect the more
*Katz, L. and Penfold, A. S., " Range-Energy Relations for Electrons and the Determination of Beta-Ray End Point Energies by Absorption", Rev. Mod. Phys.,
B , 28 (1952).
4 rational treatment now possible in TRAP due to the inclusion of internal struc-tures (parallel treatment of multiple structures of different thermal inertia but in the same convectively mixed gas space is now possible, for example). Thus the core region, the upper plenum, the relevant hot leg, the surge line, and the pressurizer were taken as separate control volumes. In the upper plenum, the core plate, the guide tubes (and support columns), the top plate and the annulus walls were differentiated as separate structures. The appended figures show selected output for the comparison runs: oecay heat (DH) considered; decay heat not considered (NO). The legends include
- the mnemonics MT (mass transfer) and FB (f all back) to indicate the use of two TRAP modifications not present for the BMI-2104 runs. These are the considera-tion of gas space mass transfer in the calculation of chemisorption of radio-nuclide vapors on surf aces and the transfer of aerosol particles from a down-stream to an upstream control volume by counterflow settling (as through the core plate). The latter does not appear to have a strong influence on release to containment. The former has a pronounced effect on Te adsorption, which is now totally controlled by gas phase mass transfer.
In addition to the plot of core (grid) plate temperature as a func-tion of time from start of melt to vessel failure (Figure 1), the cumulative fractions of core inventory of Csi, Cs0H, and Te that are released to contain-ment as a function of time are exhibited. For each species the total fractional mass as well as that attributable to the vapor and particle phases are shown. Sunrnary and Conclusions Inclusion of decay heat effects in PCS modeling raises the core plate temperature by about 100 K for the Surry TR.B sequence. The core plate _is the site of lowest thermal inertia and highest concentration of deposits. For other sequences the expected deposit and therefore change in temperature is expected to be less. A rise in structural surface temperature raises the vapor pressure of radionuclides deposited there and hence reduces deposition by con-densation. A rise in structural temperature does not affect chemisorption rates in TRAP-MELT 3 since these are taken as independent of temperature, though in general one might expect an increase in these rates. In TRAP-ELT3 only Cs0H and Te are assigned a (constant) sorption rate. Under TR.B conditions 9
~
s . 5 (high pressure, low diffusivity) these are sufficiently high for this mechanism to be gas phase mass transfer limited. Thus more precisely determined chemi-l sorption rates as functions of temperature would not alter the picture signifi-cantly. The remaining figures show a marginally higher release to containment when decay heating is considered. This is uniformly true for all three species The and for all times. Note that the bulk release is in particulate form. In distribution of particle sizes released to containment is not affected. Such a every case the release is increased by less than a factor of two. limited variation is obviously r. asked by larger uncertainties in the model. Several limitations of the present approach should be mentioned. Foremost is the lack of a flow treatment past vessel f ailure. The problem here is that while HERGE does treat expansive flow, the direction of flow can-not now be determined mechanistically since flow resistances in the various sections of the PCS are not calculated. Ways for removing this limitation are being investigated. The question of interest here is of course the possibility In this vein, also, the of long-term reevolution of deposited radionucildes. possibility of natural convection loops through the intact legs was not consi-dered and for the same reason. This phenomenon is of interest in that it holds promise for either additional retention or a new accident sequence! If one calls the above limitations first order perturbations on the results, the class of second order perturbations should include aspects dealing with the extreme density of radionuclide material above the melt, such as the contribution of vapor flow to the thermal hydraulics and gas properties and the heat associated with phase changes of radionuclides. In addition, consi-deration must be given to the adequacy of the transport models in TRAP-MELT under these conditions.
)
o . w 0 - 4 .
~ -
=
,5 3
F O N H O- O
~ d +
nB D N
+
B
,3 0 I T
C N
\ eF g T+
F T+ U F S
\ L eM M
,5 2
A E R 4 U T A R
)n i
ET PL ME m EM
/ 0
,2( T F
E EO M I T A. TR T t PA T
)S
,5 D 1
IM RO GR (F EE RM OI CT
,0 1
1 E R U G I F
,5 0
- 0 0 0 0 0 0 0 0 0 0 0 6 4 2 0 8 1
MJm L 4mEyN
, f' 4
m -a sa - G O
+
7
/
O i i e
- O W t
- 5 a
b 8 l2 e N O
.E 5,-
N, Ov E dd
==
LLI $5 2 me
~~
p & 5* p5
\ $E
'M l EC
~3 $8 I ol wz
\ $
c$ 0 e e
==
SE Gd & a 3ss w 8 O g g....... g....... g....... g....... C O O O f f C O O O O ! 6 d O O 03SV3138 NOl10VB.3
f t l l 8 , o 4 k _W m l E
~o
' M g N W N 5
;5 N _e 5 5 N vH J
ow HE Q O u. E 3 _oN vE wea <c W F-La.! am 3 -E to o Od w LA. W owr z-o F-e-o
>- k.
uo a Nz
- u. o o" >- e-
)
20 i o= EE W z oi Q Z
+
EN t+ I-c
~
m c2 O 2 wo m a. o ct O v> m
, , , , , , , , , , , , , o w
.u,, ,,y....,,,g.,,,,,y...,,,, o o a [
C o o o o o o S l f o o o o o 6 f C o o o o o o o
- u. i i
f f C o O o l 6 o o 6 o 03SY3138 NOl10V8J J
0.1; 0.01-O i;j :
~
Legend WT+FB+DH h _ r -- " 0.001: / e 5 i / 5<
/
E g p
/ .
0.0001: p 0.00001 , , , , , , 30 35 40 O 5 10 15 20 25 TIME (min) FIGURE 4. CORE INVENTORY FRACTION OF Csl RELEASED TO CONTAINMENT AS PARTICLES AS FUNCTION OF TIME FROM START OF MELT
o.1. O 0.01-y m i Legend
< ~
d w MT+FB+DH
/ " + S /
o.ool:
/
- / o i-- '
y _ E o.0001-o.00001 4 fo io 40 TIME (min) FIGURE 5. CORE INVENTORY FRACTION OF Cs0H RELEASED TO CONTAINMENT AS FUNCTION OF TIME FROM START OF MELT 4
--- M--4 m <t --- - --
} U 0 4 S A T N E M S N _
'3 I A
T - N O C T OL TE 0 T ,3 D EF M
/ SO A
/ ET LR EA
/ RT
[ ,5 2 H 0M S .
' )n sO CR F
OE F
-_ ,0 (m i
NI OT I M _ 2 TF E CO A M I RN FO T I YT RC ON
,5 TU 1
NF E VS g-NA I R EO RP OA
,10 CV 6
d+ nB eF H D g T+ O-N
+
B F T+
- ,5 E
R U G I F L eW W _
.' : - E : - : . - 3 : . 0
- : . . - 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 _ 8 E zo{<E _
e 0 12
,l o
T l n _M m l r
!2 W W u.
l 5o _o L " i$ z< b
\ b oz
\ _n a u.
w
\ N k$
\ ?
'lp E u.
\ _o 4 ssn O h N y Uw LU 2
P 5 r !E '
- te $E
( 5t3
\ 5d WU
\ _o 8f
\ 4 I of
\ u, o z \ a:
D+ + S e k @ .. O
~J e
o a; . g' ' ' ' ' ' ' 8 o d E o 8 o o o d 03SYT3d NOIDY83
0.1: 0.01= Q E W i Legend 4 -
~-
W MT+FB+DH
._j w
0.001_ -
/
/
m i MT+FB+NO z :
/
O 0.0001-
/
/
O ~
/
4 E
/
/
0.00001; /p
- /
- /
0.000001 3 I g 0 10 20 30 40 TIME (min) FIGURE 8. CORE INVENTORY OF Te RELEASED TO CONTAINMENT AS FUNCTION OF TIME FROM START OF MELT _ _ _ _ _ __ ._ _ - _ _ - - _ ~ _ _ . - - _ _ . . - _ _ _
n -
.A *.
14 O M _n m a E Wb _O
- WW z
N u. N $ N Ee v<
% H N
h -Q 8"r co "E T w es "H _ON o u. W Ho 2 u. z F S ab _n
~
RE z u. M z< e
\i wo m a.
\ -e 86 .
Z Q Of Z
\ $
w e T+ + D Cg g -n S u-e+ + 2 2
. . . . . . . . g...... .
g...... . g...... . _O O O O O O. O O O O O. O O O o. o o
- 6 O O
03SY3138 N0110V83
P. ( 0 4
,5 3
T
'~
SL AE T M
~ NF EO
~ 0 M
,3 NT f IR AA TT
/ NS O
CM
/ ,5 OR TF O
[ 2 DE [ EM SI AT
/ )n i
E LF EO [ ,0 (m 2 R eO N E TI T M FC I T ON U YF R OS
~ ,5 1 TA N
ES VE NL IC . I ET RR
< ,o t OA CP f 0 H
D O N
+
[ 1 E R d+ ,5 U nB B G F I eF g T+ F T+ eM M L
- . . 0 _
1
;i 1
;i
'1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 _
0 0 aw y${$
% V y ,'
=
.t ;
3
!1 i; M__
_ g.3 , y~ r ire KfK 3554 e a j
- i. ff August 1983 I i ! d WP '1 sdhl 2
I t en W i !.Y N '{
~ n M,#m.. ,
NAUA Mod 4 I Es!- ms A Code for Calculating Aerosol gg
?i Behaviour in LWR Core Melt Accidents NRe Code Description and Users Manual I
?'
s MW N$i m
! 3
- t. tp::n
.~r j 33% H. Bunz, M. Koyro, W. Schuck l
[jh Laboratorium fur Aerosolphysik und Filtertechnik t ' was nu Projekt Nukleare Sicherheit 1 vu9tM^ gl . p k *-'. V f biUN
'lI- ll iMM, met;;
'b ? V,1 g 1 .~.
1 -
', ip#n
] . n. =3 -
- [d [,6 w
l k ti :! sas.V l
) '
1- ,j i'-lnn l w
; y , a.: '
;r t AQ. ' .. .
I, ': -
- I .MA v' ejg$l na -
f -4
!- L
#2Y w- ji '?*Q, 1 t q
> i 'i. . .n$
r ! . g# 4- :
- - i .DEC). ;
l-F l m :. .-w f'I'!j,0.' u
% h )$
1if %f m: Kernforschungszentrum Karlsruhe Yy }; m ;
- SW
- .,:sn n it
.n j .,fd.*,y $
L .:.,GR d;; / 3-hll ONM.f G',,
- d. O G U,, ,'- i i f sid '
k . J, ^ $N%
E5" ,
- 't KERNFORSCHUNGSZENTRUM KARLSRUHE ,
Laboratorium fur Aerosolphysik und Filtertechnik Projekt Nukleare Sicherheit l
! KfK 3554 I
i i NAUA Mod 4 A Code for Calculating Aerosol Behaviour in LWR Core Melt Accidents Code Description and Users Manual H. Bunz, M. Koyro, W. Sch6ck Kernforschungszentrum Karlsruhe GmbH, Karlsruhe t
1 - Abstract This report describes the computer program NAUA Mod 4. Its purpose is to calculate the behaviour of a polydisperse aerosol system in [ a closed vessel containing a condensing atmosphere as a function of the time. The main object is to explain the physical background and to describe the structure of the code and the input and output in detail. NAUA Mod 4 Ein Computerprogramm zur Berechnung des Aeroso1verhaltens bei einem LWR-Kernschmelzenunfall Programmbeschreibung und Benutzeranleitung Zusammenfassung Dieser Bericht beinhaltet eine Beschreibung des Computerprogramms NAUA Mod 4, dessen Aufgabe die Berechnung des zeitlichen Verhaltens eines polydispersen Aerosolsystems in einem geschlossenen Beh31ter unter dem EinfluS von Wasserdampfkondensation ist. Der Schwerpunkt liegt dabei auf der Er1Huterung der physikalischen ZusammenhHnge und der Struktur des Programms sowie der genauen Beschreibung der Ein- und Ausgabe. l I
>1 l
l / - III - Contents Page J
~
Abstract I Introduction 1
- 1. The Physical Model. 2 1.1 Removal Processes 3 1.2 Interaction Processes 5 1.3 Transport Processes 7 1.4 Shape Factor Consideration 8 1.5 Combination of Precesses ,
9
- 2. The- Numerical Model 9 2.1 The Model Equation 9 2.2 The Solution Technique 11 2.3 Program Units 11 2.4 Input 14 2.5 Output 15
- 3. References 17
- 4. Figures and Tables 18
- 5. Appendix 26
,1 . . _g_
INTRODUCTION J NAUA-Mod 4 is a computer code for calculating the aerosol behaviour in closed vessels containing a condensing steam atmosphere. The code was designed and - developed for applications in the field of fission product transpc- and depletion during core melt accidents in light water reactors, especially pressurized water reactors. Although the equations for aerosol behaviour do not depend on the container itself and also not on the environmental boundary conditions, the code works with assumptions which are directly related to a melt dowa accident. There-fore, the code should not be used for completely different applications with-out careful consideration of its validity. The authors will be happy to assist in such cases. NAUA-Mod 4 describes the aerosol behaviour only. The containment the rmodynamics the source terms for aerosols and steam, and the leakages are external input functions for this code. This implies of course no feedback mechanism between the aerosol and these functions. This is true for a core melt scenario. Finally, NAUA-Mod 4 calculates physical processes only. Chemical ef fects. in fission product behaviour, as e.g. iodine chemistry, are not included. They have to be accounted for by proper use of the very versatile input capabili-ties (separate tracking of differene species).
~
NAUA-Mod 4 replaces the older version NAUA-Mod 3 f 1_7. The main physical im-provement was the substitution of the condensation model by an experimentally verified new one f 2_7. This having been done, all the model equations in the code are verified by single effect experiments. The authors, therefore, con-sider the code ready for release to interested users. In order to f acilitate the application, some changes in the code structure were made to provide input and process options for a wide variety of conceivable cases to be calculated. The most important input capabilities are: 1
- Time dependent aerosol source structuring. '
Up to 30 arbitrary time intervals may be specified in which instantaneous or constant rate sources occur. ! l Bimodal aerosol source size distribution. Only for convenience the size distribution is input as a bimodal lognormal 1 i 9
2 - a ! l -) function. The code itself is not restricted to any type of size distribution, and tables of arbitrary size distributions could be used as input if they l were given. Source rates for the two modes can be specified independently in every source time interval as well as the size distribution parameters and particle density. Multi-species source composition. Every source time interval and every mode may contain independent fraction values for up to 50 'nuclides', which are separately tracked throughout the calculation. i Multi-compartment option. Sequential multi-compartment calculations can be done in one job. ! Restart option. .- The restart option can be used when running new cases. A short run can be made and after inspection of the intermediate results the computation can be continued. Also long running cases can be split into more manageable . sections. Large scale demonstration experiments are planned for the emmediate future to show the validity of the calculated results. This appears to be necessary in j spite of the fact that the physical models used in the code have been validated experimentally. The authors, however, are confident that no changes in the modeling will become necessary. Additional numerical improvements which are on going concern mainly the im-provement of the problem time / computing time ratio and the interface to con- { l tainment codes. I
- 1. THE PHYSICAL MODEL The basic physical model in NAUA-Mod 4 is the same as in mode of the advanced aerosolbehaviourmodelsforclosedcontainers[3).Thefollowingassumptions are used in order to keep computing time at an adequate level:
Particles are homogeneously distributed in a control volume except for the boundary layers at the walls. Within one particle size class no difference in particle composition is allowed.
- i a.. <
1 1
. . , i l
- 3 - l
(
- Particle properties are functions of only one independent variable, the ~
particle size, and of the particle density which may change, due to varying particle composition. -
- Process coefficient (shape factors, boundary layers etc.) are assumed to be independent on particle size.
For core melt accidents these assumptions are considered to be valid. Internal mixing of species in one size class is quickly achieved by coagulation. The spatial homogenity in the control volume is accomplished by convection, especially in condensing atmospheres. So, presently no reason is known not to use the physical volumes of the reactor building as control volumes. Within a control volume the code calculates the following processes:
- Removal processes
. Gravitational settling
. Diffusional plat out
- Interaction processes
. Brownian coagulation
. Gravitational coagulation
. Steam condensation on particles
- Transport processes
. Aerosol sources
. Leakages These processes were chosen because they are dominant. Thermophoresis e.g. has been eliminated because in an LWR it may only occur in the early phases of the accident, when temperature gradients are big enough. During this phase, how-ever, all other processes are vigorous, too. So thermophoresis is considered to be negligible throughout the whole sequence.
Similar considerations were made for other processes that are not included in the model. Should one of them for any reason become important (e.g. in a new application) it could easily be incorporated into the code. In this sense, the code should be regarded as under continuing development. 1.1 REMOVAL PROCESSES The aerosol removal model in stirred conditions is well known /'4 ,7. The change
in differential number concentration d n (r) per time interval dt by deposition on a surface is d r) =
-N(r)u(r).h (1) where V is the volume in which the aerosol is contained, and A the process re-lated area onto which the deposition of particles takes place; u (r) is the deposition velocity u (r) = F B (r) (2) which a particle of radius r and mobility B (r) attains under the influence of an acting force F. For a spherical particle the mobility B (r) is giver by ;
Stokes' law (3) B (r) = 6 w n r i with n being the viscosity of the carrier gas and C (r) the empirical i Cunningham correction factor. In NAUA-Mod 4 the equation C (r) = 1 + 1.246 Kn + 0.42 Kn exp (- ) (4)
~
i is used (Kn = Knudsen number) / 5_/. l In NAUA-Mod 4 the size parameter r is the volume equivalent radius, which for normal applications is equal to the Stokes' radius (c.f. section " shape fac-tors" for detailed information). For gravitational settling the steady state settling velocity is obtained by substituting the gravitational force for F in eq. (2) u
* *8 r8 C (r) (5)
S (r) = 9n I, with g, the gravitational constant, and p,gg the effective density of the (spherical) particle. f = ! i . I i 1 l i i
< s
>1 '.* . - 5 -
For gravitational settling the surface area A in eq (1) has to be taken as the , sum of all upward facing horizontal projections of surfaces in the volume V. For all other processes the orientation of the surfaces is irrelevant. For diffusional deposition the equation u D (r) D
# ~ (
6 is used. D (r) is the diffusion constant D (r) = kT B (r) (7) with k the Boltzmann constant and T the absolute temperature. 6D is the diffusional boundary layer across which the particle concentration ' drops to zero with a constant gradient. NAUA-Mod 4 uses D6 as a constant input parameter, a value of 6 D= 0.01 cm is appropriate for most cases. 1.2 INTERACTION PROCESSES Interaction processes are most important for the behaviour of highly concentra-ted aerosols because they are responsible for the non linear behaviour of aerosol removal. They are calculated with very high time and particle size re-solution and this calculation consumes most of the computation time. The two interaction mechanisms considered in NAUA-Mod 4 are coagulation and steam condensation. Brownian Coagulation The collision frequency Kbdue to Brownian motion is given by K (r g, r y) = 4 x KT- (r; + rk ) (B (rg) + B (rk)) (8) r gand r are k the radii of the two particles involved in the coagulation pro-cess. The result of the process is the disappearing of the two particles and the creation of a new one with a mass and volume equal to the sum of the masses and volumes of the two original particles. Regarding the new particle as spherical too, one obtains the following conservation rules for the individual coagulation process
I - 6 - o m = m g+mk
- v. = v. + v (9) l j i k r1 = r1 + r5 i j i k l
where the subscript j denotes the new particles. Gravitational Coagulation The collision frequency K due to gravitational coagulation is given by K (r g, r k) " * * # * (#i+#k} */"s (#i} ~ "s (#k ! I The problem in this equation is the value of the collision efficiency c. .j
~
NAUA-Mod 4 uses the size dependent expression given by Pruppacher / 6_7 d s r. c = ( )' ; r; < rk (11) i k i t
'.I rather than a constant value, which might not represent the situation correctly [
over the whole real time period of a calculated case. ! j The calculation of steam condensation on particles is a very fast molecular pro- ; cess, it may also contribute much to the removal of particles through generation of large fast settling dropicts. G
! j The obvious solution, to expand eq (8) and (10) down to t 31ecular size values, ,
cannot be used for reasons of computing time. 'Iherefore, : more simple but yet [ precise enough description of the process had td be found. An iterative process of analytical studies and experimental investigations lead to the following j model. , The change in radius of a particle due to condensation or evaporation of steam !li can be described by ' I S - exp {2 cmp RT 1/r) I Lo LM p RT, (I , y ,
~
KT, RT, MDP, q . with S steam saturation ratio c(Tr) surface tension of water f, 4
-- 'T
- 7 -
M molar weight of water p,(Tr) specific density of water R universal gas constant T temperature of the droplet T, temperature of the carrier gas L(Tr) latent heat of water K(T,) heat conductivity of water vapor P,(T,) saturation pressure of water vapor It is assumed that the ieeal gas law holds and that T a r T,. o, p y, L, K and P, are calculated as functions of temperature every time step. This equation was derived for thermal equilibrium conditions in normal atmos-pheres. It was dthe objective of an experimental project to measure the possible deviations from.eq (12) in a simulated accident atmosphere with realistic pres-sure, temperatures and aerosols, the result of the experiments was that the equation can be used without any corrections [2,7. Obviously the steam concen-tration and the related changes at elevated temperatures are so high that no time delays or second order effects become noticeable. On the other hand, because the concentration changes and consequently the particle size changes are fast and big, a separate, much finer time integration scheme has to be used for evaluating eq (12). If the condensation process was calculated too coarsely (e.g. by using the time steps with which the other aerosol processes are calculated), over-shooting and instable oscillations about the saturation limit could occur. 1 I [ Consequently, the condensations routines consume much computation time and the I code contains control parameters to suppress them G enever not needed. i 1.3 TRANSPORT PROCESSES l Since the NAUA code is designed for closed containers, the only transport effects considered are aerosol sources and leakages. They are accounted for as volume sources and sinks, which means that no depletion in transport paths is calculated within the NAUA code. The aerosol source has to be specified as input data. Arbitrary time functions, size distribution parameters, particle densities and nuclide composition are acceptable. e
8 - l a
, e s ,
l The leakage is specified as input data, too. No size dependent effects are taken l l into account by the code. l ] I 1.4 SHAPE FACTOR CONSIDERATION t The most important by product of the experimental investigations was the confir-mation of the spherification process. Particles undergoing condensation - evapo-ration processes are compressed to almost spherical shape. The residual deviation f rom ideal spheres is sm.:11 and does not need to be taken into account in the model. The process also compacts coagulated aggregates of previously compacted f particles. The effect was experimentally established with n'on-hygroscopic i particles, it will certainly be stronger for soluble and hygroscopic particles. I It is unconceivable then, that the spherification effect should not occur in a water reactor accident for more than very short time periods or very small spaces in the building. Therefore, the spherical shape can be assumed throughout the calculation. Consequently all shape dependent shape factors could have been eliminated from the code equations. This has not been done in order to keep the code applicable for future problems of different nature, but the corresponding shape t' actors are input parameters whose value should be set equal to unity. These are the following three shape factors, which were not written in the above j equations: 4
- the dynamic shape factor e to be inserted in the denominator of eq (3)
- the coagulation shape factor f multiplied into eq (8) !
i,
- the condensation shape factor f , which is a factor in the exponential in eq (12). j i
,1 Defining these shape factors as only shape dependent requires the use of a 1 density correction for porous particles in eq (5) for gravitational settling and in eq (10) for gravitational coagulation. This is most conveniently done {
by using an effective density p,gg instead of the material density p Effective densities have to be determined experimentally, for UO aer2 sols l p /p = 0.47 eff jl was measured in [ 2,/ for the spherified particles. Therefore, unless special values are given, a density reduction by 50% can be recommended for all f i i l i ___ _ _ _ m
H o
,' _ 9 _
insoluble particle materials. The effective density is the density input para-meter for the NAUA code. 1.5 COMBINATION OF PROCESSES - In NAUA-Mod 4 the differenc aerosol processes are treated as additive. Occasionally the question was raised whether this assumption may be too rough or not [ 7_7. The second order correction in the combination of any two simul-taneous processes, however, depends on the rate of both of them. And the error in computed results increases with the length of the time step. The code, there-fore, automatically adjusts the time step duration such that the fastest aerosol process mostly (brownian coagulation) does not lead to too large changes in any s particle size class. This algorithm is controlled by an accuracy parameter, the value of which was determined in parameter studies. The accuracy parameter is included in the input list and has to be used with care. For the computation of condensation a separate similar procedure is used, which controls the condensation time step. It is smaller than or equal to the coagu-lation time step. These precautions keep errors within limits which are accept-able for the normal application of the code to core melt accident analyses.
- 2. THE NUMERICAL MODEL 2.1 THE MODEL EQUATION If all the mathematical expressions for the different physical processes dis-cussed in the previous chapters are combined, the complete model equation is obtained.
=
S (r, t) - ( D ( + s
+
T r 3~
!2 3 3 2
+/ K(/r3 - r'3, r') n ( / r - r' , t) n (r', t) dr' o
(r'-r'8)h
- n(r, t) / K (r, r') n (r', t) dr' + b at " (#' (13) 3r ,
l This integro-differential equation can be solved only numerically To facilitate this solution the particle size distribution expressed by n (r) is approximated l l r
. . . 8 1
by a number of monodisperse fractions (a kind of histogram). By this manipula-tion the given integro-differential equation can be transformed into a system ' of coupled first order differential equations. l 3n(k ,t) k at " S (#k' '} ~ ("D k + "S # k T( k L(#k ' "(# k' N
-E, (1 - 1/2 S g) K (r ,g r ) n (r g, t) n (r , t) 1=1 N N i I K (r , r.) S . n (r , t) n (r , t)
+ 1/2 I i=1 j=1 J IJ i j
('} k-1 k (t)
+ (1 - 61k) vk -V k-1
" (#k-1'*) ~ vk+1 -V k
* " (#k' ') l i
for k = 1, ...., N I Such an equation system can be treated much more easily and numerically more ! i stable than the complete integro-differential equation as it was found by t
~ I experience / 9,7. One difficulty arising by this reformulation is that the new particles formed by coagulation do not fit into the given classifica- [
tion of the particles if a non-mass equidistant classification is applied as usual. Therefore, an interpolation has to be performed whereby the new-formed particles are distributed between the two particle size classes nearest to l' ' l the new particles. This interpolation conserves the particle number and mass.
?
Another problem is the composition of the particles of the two components, { a non-volatile solid fraction and a volatile liquid fraction. On the one L I hand it is not possible to assume homogeneous mixing over the whole particle ! size distribution because the condensation and evaporation rates (eq 12) are strongly dependent on the particle size, on the other hand the introduction f of a second dimension describing the different composition of the particles
-l in addition to their size is too time-consuming and therefore not practi-J cable. The composition of the particles is, therefore, averaged in each size j class but varies from size class to size class. The variation of the composi- !
tion is calculated taking into account the influence of the coagulation of differently composed particles, sources of new particles and the condensation
~
of water on the particles. Fortunately, it can be shown [ 2_7 that the sedi- ~ mentation of the solid fraction (carrying the radioactivity) is underestimated # by this simplification and therefore justified. I L 6 4 r
II - 2.2 THE SOLUTION TECHNIQUE J The equation system (14) is solved by the Euler-Cauchy-method which is a standard numerical technique. To ensure the numerical stability, the time " step is calculated by the code itself using the ratio of the first to the zeroth derivative as a measure because, fortunately, it can be shown that this ratio is of the same order as the ratio of the second to the first derivative justifying this procedure. This method is used for the main part of the equation controlling the coagulation and the depletion processes. 'Ihe condensation, however, as the fastest and to num:-rical instabilities most sensitive process is inte-gratedi{Arpparateroutinewherethetimestepiscalculatedbythe difference between the result after the full time step and two half time steps. If the stability criterion is violated, the code goes back to the beginning of the time step and starts again with a smaller time step. The principle scheme of the code can be seen best in the block diagram (fig. 1). 2.3 PROGRAM UNITS
- 1. MAIN The most input data are read here and all the necessary starting values are set. To save computation time the temperature independent parts of the de-pletion coefficients are calculated before the entry into the time loop.
In the time loop the coagulation is computed except the coagulation frequen-cies. Taking into account the result of the coagulation and the removal pro-cesses the new particle size distribution is calculated after the time step DELZEI which is calculated in the subroutine DELTIM. All removal coefficients are temperature dependent. Some of the necessary values like the viscosity and the mean free path are provided as statement functions. At the same time the mass depleted by the different processes is calculated. Using the new distribution,the dry, liquid as well as the total airborne mass, the number concentration and the average radius are computed. Dependent on the number of time steps, a small or a large output including the particle size distribution are printed. To limit the amount of output the printing frequency is also restricted to a certain part of the total problem time.
o , , , L Therefore, the time between each printing is at least 1/1000 of the problem j time. Also data are written on predefined files to prepare plots of mass and
- number concentration, average radius and other values versus time, and of size k distributions at different times. ,
I At the end of the time loop the code enters the concensation subroutine ' NEWCON and goes back to the beginning of the loop if the stop condition is j not fulfilled. f If the stop condition is valid, all data necessary for restarting the code { are written on the restart file and the execution is terminated, j
- 2. CUN (subroutine) {
I The size and temperature (because of the temperature dependent mean free path) i dependent slip-flow mobility correction (Cunningham correction) is calculated according to the values given by Millikan and Fuchs.
- 3. KERN (subroutine)
The coagulation frequencies, brownian as well as gravitational, are calculated here. To save computation time in cases when only brownian coagulation is requested, two different entries for these two different requests are provi- j ded. As already mentioned, the Fuchs formulation is used for the collision efficiency.
- 4. DELTIM (subroutine)
The time step DELZEI is calculated due to the criterion explained in the previous chapter. The size of the time step and by this the integration accuracy can be influcenced by the parameter EPS.
- 5. SOURCE (subroutine)
The size dependent source of solid particles is computed from the given input data assuming for convenience a bomodal lognormal distribution because for nearly all applications only some average particle sizes are known and, there-fore, exact size distributions are not available. But in principle also an arbitrary table can be read as size distribution. This option is used in the next subprogram READL. The release of the particles can be given as a long time source or as a puff release. Also the time dependent particle density and the time dependent contents of
g
- 13 -
the different nuclides are calculated here, assuming that all material re-leased is instantaneously mixed over the whole solid part of the particle size distribution at the moment of its release.
- 6. READL (subroutine)
This subprogram provides the particle sources for secondary compartment runs. It reads the size and time dependent leakage of a proceeding run. Therefore, no further assumptions concerning the size distribution of the source particles for the secondary compartment have to be made. As in the subprogram SOURCE, the time dependent particle density, and the nuclide contents are calculated.
- 7. NEWCON (subroutine)
This subprogram performs the numerical integration of the growth rate due e to steam condensation according to the Mason equation. All important para-meters are given as temperature functions, mostly in the form of statement , functions. At the end of the subroutine the particles are interpolated into the predefined size classification according to their changed sizes.
- 8. DRYOUT (subroutine)
The liquid part of the particles, which leave the compartment through a leak, is removed and the particles are interpolated into the predefined size clas-sification according to the sizes of their solid part. The obtained size de-pendent leak rates can be read again by the subroutine READL in a subsequent job. The particles are reduced to their solid part since normally the thermo-dynamic conditions are different in the various compartments causing diffucul-ties predicting the behaviour of the volatile part. To be on the safe side it is therefore assumed that the particles become dry if they are transferred from one compartment to another.
- 9. TEMPCI (function) ,
The containment temperature is read as a time equidistant table. The actual temperature is calculated by linear interpolation between the given grid points. If the time is greater than the time of the last grid point, the temperature is assumed to be constant at this temperature for the rest of the calculation. The normal entry into the function (except the first one) is called TEMPC. 9 I
- 10. DRWURZ (function)
The third root of any number is calculated there.
- 11. STEAMI (function)
The steam source is read as unarbitrary table containing the time and the steam flow rate at a number of grid points. Between the grid points the steam flow rate is assumed to be constant. If the last steam flow rate is not equal to zero, the code produces an additional grid point at the end of the problem time with a flow rate equal to zero. The normal entry into the function (except the first one) is called STEAM.
- 12. TLEAKI (function)
The leak rate of the compartment is read as an arbitrary table containing the time and the leak rate at a number of grid pointe. The structure of the data is the same as that of the steam flow. Between the grid points the code assumes linear interpolation. If the last leak rate is not equal to zero, the code produces an additional grid point at the end of the problem time with the leak rate equal to zero exept that only one grid point is given. In this case the leak rate is assumed to be constant at this value for the whole calculation. The normal entry into the function (except the first one) is called TLEAK. 2.4 INPUT The complete input list is shown in fig. 2. It can be seen that the whole
" normal" input is read list-directed from a data set corresponding to the data set reference number 5 except the two text lines for the head line.
They are read by a format 18A4 (format statement 6031). All input data written by preceding runs of the code are provided as unformatted data. These are:
- 1. Restart data set (from uni 10)
This data set enables the user to start the calculation again if the exe-cution was terminated by reaching the problem time or the given CPU-time. The input from the unit 5 has to be the same as for the first run except I the parameter RESTRT. It has to be set on .TRUE. insted of FALSE . If RESTRT is equal to .TRUE., the code expects a restart data set.
g , I 15 - l 2. Leakage (from unit 2) If a run is performed for a subsequent compartment, the code expects the leakage written on a file by the preceding run as input. It contains the size y l of the preceding volume, the number and names of nuclides, contents of the i
; various nuclides, the density of the particles and the size dependent leakage at all time steps of the preceding run.
- 3. CPU-time (from unit 19)
)
If runs of several compartments are performed in one job but dif ferent steps 3, (see examples 3 to 5), the CPU-time needed for the stop condition has to be transferred from step to step because the code has to know the whole CPU-time , j consumed already by the previous steps. Therefore, only the CPU-time (CPUZT) read in the first step is important and should be equal to the time given on I the job card. In the ubsequent steps it is read over by the value from unit l 19. A list of the variables read by the card reader (unit 5) including the expla-nations of their meaning is viven in table 1.
]
l 2.5 OUTPUT The printed output (unit 6) comprises the input data, some informational out-put concerning the start of condensation and the preceding run for secondary f ( v compartment runs, a small output containing the values of the actual mass and i number concentration, the depletet masses and so on and a large output consis-i ting of the same data as the small one plus the particle size distribution.
] The printed output is selfexplanatory as it can be seen in the examples given in the appendix. Therefore, no further explanation is necessary.
1 In addition to the printed output some other output options are provided. l Since all these data are read during other runs of the NAUA code itself or by other codes like plot programs, they are all written unformatted. i
- 1. Restart data set (unit 9)
If the execution is terminated by reaching the problem time or by using up the given CPU-time, a restart data set is written to enable the user to continue the un by a new job. The consumed CPU-time has to be calculated by l the code using a machine-dependent routine. For further details see also the explanation of the parameter RESTRT in table 1.
r 16 -
- 2. Leakage (unit 1)
At each time step the code writes the size dependent leakage on a data set (variable ZLEAK), the actual contents of the various nuclides and the actual density of the particles onto this data s6t. At the beginning of the run (except for restarted runs) the size of the volume, the number of nuclides, the names of the nuclides and the duration of the particle source is written onto this data set. The data set is therefore very large in most cases, e.g. in the order of 10 MBytes.
- 3. CPU-time (unit 19)
For details see the description of the CPU-time data set in the previous chapter (description of the input). The data set is rewinded at the end of the run and the remaining CPU-time is written on it. It may be defined as a very small temporary data set.
- 4. Plot dat set (uni 8)
To enable the user to produce plots of the various time dependent values, the necessary data are written onto this data set. It can be read again by any suitable plot program. The variables contained in each record of the data set are listet in table 2.
- 5. Plot data set for the particle size distribution (unit 3)
Each time, the large output is printed, the particle size (number) distribu-tion is written onot this data set to enable the user to plot the size dis-tribution at this time. The variables written on each record are listed in table 3. If a problem is computed by a number of restarted jobs, the data sets 2, 4, 5 can be simply concentrated in the proper order, of course, since the starting records are written only in the first job and the structure of all other records is the same. W 4
. , 3. REFERENCES
/1/ W. Sch5ck, H. Bunz, M. Koyro Messungen der Wasserdampfkondensation an Aerosolen unter .
LWR-unfa11typischen Bedingungen KfK 3153 (August 1981)
/2/ H. Bunz, M. Koyro, W. Sch5ck NAUA-Mod 3 -
Ein Computerprogramm zur Beschreibung des d Aeroso1verhaltens in kondensierender Atmosph're KfK 3154 (September 1981)
/3/ Nuclear Aerosols in Reactor Safety CSNI SOAR # 1, OECD (June 1979)
/4/ N.A. Fuchs l
The Mechanics of Aerosols Pergamon Press, Oxford, p. 250 (1964) l
/5/ R.A. Millikan The general Law of Fall of a small spherical Body through a Gas
! and its Bearing upon the Nature of molecular Reflection from Surfaces Phys. Rev. Vol. 22, pp 1 - 23 (1923)
/6/ H.R. Pruppacher, I.D. Klett Microphysics of Clouds and Precipitation D. Reidel Publishing Company (1978)
/77 R.J. Williams, UKAEA l
SRD to W. Sch5ck: private communication
/8/ B.J. Mason The Physics of Clouds Clarendon Press (1971)
/9/ H. Bunz PARDISEKO IIIb - Ein Computerprogramm zur Beschreibung des Aeroso1verhaltens in geschlossenen Beh*ditern KfK 2903 (April 1980) ,
-- 18 --
Input Data u Initial Conditions u Differential Quotients due to Depletion Processes and Coagulation se Calculation of Time StepAt
%' Condensation and Evapo-
. . ration Dates for each Size
-Y Class "
Time Integration w New Partic:e Size Distribut . Calculation of Time Stepa t' u t =t+At u Time Integration v
=
Output t' r-At t' = t' + A t' t <t total and ves CPU-time < CPUZT yes , ,. no Stop N Fig. 1 Block Diagram of the NAUA-Code
__ i o __ C MAIN PROGRAM READ (5,6031) TENT 6031 FORf!AT (18A4) C C INPUT BY LIST-DIRECTED INPUT C VARIABLE NDUti?!Y TO READ THE LINE NUMBER IF NUMBERED RECORDS ARE . C USED AS INPUT C IF NON-LINE-NUMBERED RECORDS ARE USED, NDU:1!!Y liAS TO BE RE!!0VED C READ (5,*) VOL,FSED,FDIFF,NDUM31Y,
+ FORM , FORT!C , FORMKO , DE LD , NDU2151Y ,
+ RM I N , R!!AN , KM AN , EPS , CUT 0FF , NDU!!!!Y ,
+ SZEIT, tit!E , CPUZT, NDUMt!Y ,
+ RESTRT , ZWC0!!P , FO LGE , G R AVK , LE AK , NDU 1}iY ,
+ LKOND , T I ME K 1,TI t!E K2 ,TI t!EK 3 , T I MEK4 , NDU21MY ,
I
+ NWRITE,NPLOT,NDUMMY, j + NPHASE ,NUKLI D ,NDU!!}!Y ,
; + (SRATE(1,1),KONTIN(1,I),RG(1,I),
+ S IGL( 1,1 ) , RH01 ( 1, I ) ,NDU!!MY ,
+ SRATE(2,I),KONTIN(2,I),RG(2,1),
I + S I G L ( 2 ,1 ) , RH01 ( 2 ,1 ) , TQ ( I + 1 ) , NDU;151Y ,
! + (NA!!NUC(K) , AKTIVI '1,I ,K) , AKTIVI (2,I ,K) ,NDUMMY, I + K=1,NUKLID),I=1,NPllASE) ! C j READ (19,END=50) CPUZT j C k IF (ZWCOMP) READ (2) VOL1,NUKLID,SOUTIM,(NAMNCC(K),K=1,NUKLID)
I C 90 READ (10) SZEIT,ZCON,TCE,DELZEI,DEPDIF,DEPSED DEPD2,DEPD3,DEPS2,
+ DEPS 3, NTOT , SU:lLK , SUB!LKA , AKTIV ,Wi l , V , Z , RHOB , KULI!! , KOLI ri, j + S , AVMAS 1, AV!!AS2,ZNORM ,FIRST,DELTAT,DA!!PFr!,SGEWTR ,DEPDA ,
1 + DEPSA,SUMSOU,NUKLID,NAMSUC
! C i
120 IF (ZWC0!!P)
+ CALL READL (Z,V,SZEIT,DELZEI, CUTOFF,Rif0B,AKTIV,NUKLID,SCEWTR,
+ KULIM,K0 LIM, DILUTE,ZWCOMP)
C SUBROUTINE READL (Z,V,ZEIT,DELTAT, CUTOFF,RHOB,AKTIV,NUKLID,SGEW,
+ KULIf t ,KOLIM , DILUTE ,2WC0!!P)
C READ (2,END=90) ZEIT2,RH02,RLPV2,(AKTIV2(K),K=1,NUKLID),
+ (Q2(K),K=1,KMAN) l C
/ FUNCTION TEMPCI (T) j READ (5,*) N,DELT READ (5,*) ((TI(9*(I-1)+K),K=1,9),NDUMMY,1=1,NN)
C FUNCTION STEA!!I (TI) READ (5,*) N,NDUM31Y,(TD(K),D(K),NDUM:1Y,K=1,N) C g FUNCTION TLEAKI (T)
- RE AD ( 5 ,* ) N , NDU11MY , (TL(K ) , RL(K ) , NDUM51Y , K= 1, N )
FIG.2 INPUT LIST
. l t l
) !
I l
r- v
--.20 -
Table 1: Card-Input (Remark: R4 : Real 4 R8: Real 8 I: Integer 4 L: Logical 4) l Card . . Dimen- Recommen-Va r iable Type Explanation sion Number ded Value ! 1+2 TEXT R4 Head-line to be printed to - - (1..,36) identify the job j 3 VOL R4 Volume of the compartment em 3 - FSED R4 Floor area of the compartment cm 2 - FDIFF R4 Total surface area of the em 2 - compartment 4 FORM R4 Mobility shape factoi' for the - 1. FORMC R4 Coagul.ation shape factor l experimen- - 1. FORMKO R4 Condensation shape factor tal values - 1. _f see /1/ DELD R4 Diffusional boundary layer em 0.01 5 RMIN R4 Lower size limit of the particle cm < 1.E-6 classification RMAX R4 Upper size limit of the particle em 20. E classification 100.E -- 4 KMAX 1 Number of classes of the particle - 80 - 110 classification EPS R4 Accuracy parameter for time - 0.1 integration CUT 0FF R4 Cutoff value for particle size em 1.E - 15 distribution. If the number. of particles in a class is smaller than CUTOFF, the class is set to zero 6 SZEIT R4 Starting time of the problem, equals see - the time of first particle release
- in the accident to be analysed TIME R4 Total problem time. Af ter reaching h -
this time the execution is terminated , i 2 - l 6 '- - _ _ . _ . ._ . . . - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __
-- 21 --
(Table 1 cont') b I I CPUZT R4 Same CPU-time as inserted on the job min - card. If the CPU-time is reached, the execution is terminated and restart data are written onto the restart file. To do this a machine dependent routine has to be used to calculate the CPU-time already consumed 7 RESTRT L Flag to expect a restart file - - I
. FALSE. (= NO)
ZWCOMP L Flag to indicate subsequent com- - - partment run, dat ritten by a pre-ceding job are ex,2cted
. FALSE. (= NO) ,
FOLGE L Flag to indicate that a run will be - - made reading the leakage of the ac-tual run as input. Activates the subroutine DRYOUT . FALSE. (= NO) GRAVK L Activates the gravitational - - 1 l agglomeration i
. FALSE. (=NO)
LEAK L Activates the leakage - -
. False. (= NO) 8 LKOND L Activates the stean condensation - -
onto the particles
. FALSE. (= NO)
TIMEK1 R4 Lower and upper time limit of the see - l TIMEK2 R4 first period of condensation see - TIMEK3 R4 Lower and upper time limit of the see - TIMEK4 R4 second period of condensation sec- - l r i A
V
. * (Table 1 cont')
l l 9 NWRITE I (1): Number of time steps a small - (1,2) output has to be printed (2): Number of time steps a large - - output has to be printed in-cluding the size distribution NPLOT I Number of time steps data have to be - - written onto the plot data set 10 NPHASE I Number of release periods for the - 5 30 particle source NUKLID I Number of nuclides to be balanced - 5 50 (2 + NU- SRATE R4 Source rate for particle release for g or - KLID)* (1,1) the mode 1 in period I g/sec (1-1) KONTIN L Flag to indicate if SRATE (1,1) is - -
+ 11 (1,1) a puff or a continuous release
.TRUE.: continuous
. FALSE.: puff RG(1,1) R4 mean geometric radius of mode 1 cm -
in period I SIGL(1,1) R4 logarithmic variance - - of mode 1 in period I RH01(1,1) R4 particle density of mode 1 in g/cm 2 - period I SRATE(2,1) R4 g or - (2 + NU- KONTIN the same meaning as g/sec - KLID)* (2,I) L y the above values but for - - (I-1) RG(2,1) R4 the mode 2 in period I cm -
+ 12 SIGL(2,1) R4 RE01(2,1) R4 - j g/cm' -
TQ (I+1) R4 end of release period I sec (2 + NU- NAMNUC(K) R8 Name of nuclide K - - KLID)* AKTIVI R4 Content of nuclide K.in mode 1 of - - (I-1) (1,1,K) period 1 as mass fraction
+12+K AKTIVI R4 The same for mode 2 - -
(2, I,K)
i U i, . .
. . -- 23 --
(Tabl. I con') m These (NUKLID + 2)
- NPHASE (= NMAIN-10) cards define the complete ,
particle source including the information about the contents of tl.e various nuclides. The names of the nuclides may not be different , in the different release periods.
- in function TEMPCI -
NMAIN N I Number of grid points for the - 1 200
+1 temperature function DELT R4 Time interval for the time-equidis-tant temperature table NMAIN TI (K) R4 Table of the containment tempe- C -
+K+1 K = 1,N rature up to NMAIN
+N+1
+
= NT - in function STEAMI -
l NT + 1 N I Number of grid points of the steam - 5 200 source t NT+K+1 TD(K) R4 Time at the grid point K of the set - steam source 4 D(K) R4 Steam flow rate from time TD(K) g/see - to TD (K+1) up to NT+N+1
= NST - in function TLEAKI -
NST+1 N Number of grid points of the con- -
.1 200 tainment leakage NST+K TL(K) Tim'e at the grid point K of the see -
+1 leak rate
-1
, RL(K) Leak rate out of the compartment see -
. K = 1,N at the grid point K l
up to NST
+N+1
_y_
. . a Table 2: Structure of Plot Data Remark: All Variables are of type Real 4 Variable Explanation Dimension SZEIT time see CON particle number concentration em" 3
SGEW total mass concentration g/cm RAV average radius em DEPDIF accumulated deposition due to g/cm2 diffusion DEPSED accumulated deposition due to sedimentation g/cm2 SCEWTR solid part of SGEW g/cm 2 SGEWWS liquid "- g/cm 2 DEPD2 solid part of DEPDIF g/cm2 DEPD3 liquid "- g/cm 2 DEPS2 solid part of DEPSED g/e:m2 DEPS3 liquid _"- g/cm 2 SUMLK accumulated leakage out of the g ecmpartment AIRM total airborne mass g AIPMT solid part of AIRM g AIRMW liquid "- g NUKLID number of nuclides g O ___._m
e e
'. . -- 25 --
Table 3: Structure of Data to plot the size distribution Variable Type Dimension i NTOT I*4 number of the actual time step - SZEIT R*4 time sec h 5
-3 Z (K) R*4 particle number in class K cm K=1, KMAX i
Table 4: Structure of the first record of the data set to plot the ! size distribution f i Variable Type Explanation Dimension KMAX 1*4 total number of size classes - (see input list) R (K) R*4 radius of the particles in the cm ] i' K=1, size class K I KMAX l X (K) R*4 logarithm of the radius - K =1, KMAX I i 6 ? -
e
~ 26 - {
. . O ,
I I i Appendix , t i Sample Problems Case 1: Single compartment calculation with an instantaneous aerosol source I and without steam condensation. Simple and f ast running sample for first check after installation of the code. 1 Case 2: 4 Two compartment calculation taking into account a best estimate aerosol j source for a core melt down accident in a 1300 MW PWR and steam conden- I sation on the particles in the containment. The droplets are assumed l t to evaporate after leaking into the annular gap as the secondary compartment. l [ Technical remarks: Both examples consist of the following data:
- Complete set of job control cards for an IBM machine including the card input (Unit 5).
Input values as printed out by the code at the beginning of each run.
- Examples of calculated values as printed out at certain time steps.
- Plots of the total airborne mass, the total accumulated leaked mass of certain nuclides and the particle size distribution at different times.
For the second sample problem the last three items are given for each compartment corresponding to the two steps in this job. W 4
-- 2 7 --
//LAF295El JOB (0295,691,P4311),K0YRO, NOTIFY =LAF295, TIME =120 :
//* MAIN LINES =30
//El EXEC F7CG
//C.SYSPRINT DD DUMMY )
//C.SYSIN DD DISP =SHR,DSN=TS0657.NAUAMOD4. FORT
//G.FT01F001 DD DUMMY
//G.FT02F001 DD DUMMY
//G.FT03F001 DD UN I T=D I S K , VO L=SER=B AT00H , D I S P= ( , C ATLG ) ,
// DSN=LAF295.E13. DATA, SPACE =(TRK,(20,5)),DCB=DCB.VBS
//G.FT08F001 DD UNIT = DISK,VOL=SER= BAT 00H, DISP =(,CATLG),
// DSN=LAF295.E18. DATA, SPACE =(TRK,(20,5)),DCB=DCB.VBS
//G.FT09F001 DD DUMMY
//G.FT10F001 DD DUMMY
//G.FT19F001 DD DUMMY
//G.SYSIN DD
- NAUA SAMPLE CASE # I-----------------------------------------------
POINT SOURCE, NO STEAM, GRAVK, SINGLE COMP. ----------------------- 72000.E6 6460.E4 50540.E4
- 1. 1. 1. 0.01 0.002SE-4 20.E-4 101 0.1 1.E-15 0.0 100. 20.
.F. .F. .F. .T. .T.
.F. 5.E5 6 E5 10 100 5 (
2 3 1000000. .F. 0.4E-4 0.7 4. O. .F. 0.4E-4 0.7 4 5000.
'NUCL l' O.02 0.
'NUCL 2' O.04 0.
'NUCL 3' O. O.
500000. .F. 0.5E-4 0.8 3. 1 400000. .F. 0.6E-4 0.4 2. 6000.
'NUCL l' O. O.
'NUCL 2' O.05 0.08
'NUCL 3' O.03 0.04 1 1.
130./ i 1 h
- 0. O.
J 1
- 0. 1.16E-7
//
l l } l -
-- 28 -- ! . . s t.
l i NAUA - MOD 4 INPUT DATA LISTING
*********************************** {
t TEXT : NA UA S AM P LE C A S E # 1 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - POINT SOURCE, NO STEAM, GRAVK, SINGLE COMP. ----------------------- 1 CONTROL VOLUME PARAMETERS VOL : 0.72000E+11 FSED : 0.64600E+08 FDIFF : 0.50540E+09 AEROSOL PROCESS PARAMETERS FORM : 0.10000E+01 FORMC : 0.10000E+01 FORMKO : 0.10000E+01 DELD : 0.10000E-01 NUMERICAL PARAMETERS RMIN : 0.25000E-06 RMAX : 0.20000E-02 KMAX : 101 EPS : 0.10000E+00 CUTOFF : 0.10000E-14 PROGRAM CONTROL PARAMETERS SZEIT : 0.0 TIME : 100.00 CPUZT : 20.00 RESTRT : F ZWCOMP : F FOLGE : F GRAVK : T LEAK : T LKOND : F TIMEK1 : 0.50000E+06 TIMEK2 : 0.60000E+06 OUTPUT CONTROL PARAMETERS NWRITE : 10
- 100 NPLOT : 5 E
i
f...-._ . . . , . . , , . . , , . , _ _ , - . - _ _ , , , , _ _ l
~.
AEROSOL SOURCE F UNCTIONS CONTAINING 2 RELEASES AND 3 NUCLIDES UP TO 5000.0 SEC
- M00E1 1.00000E+06 (G) RG- 4.00000E-05 SIGL 7.00000E-01 RHO 4.000 .
CONTAINING THE FOLLOWING NUCLlDE FRACTIONS NUCL 1 2.00000E-02 NUCL 2 4.00000E-02 CUCL 3 0.0 M00E2 0.0 (C) RG 4.00000E-05 SICL 1.00000E-01 RHO 4.000 CONTAINING THE FOLLOWING NUCLIDE FRACTIONS I NUCL 1 0.0 ! NUCL 2 0.0 NUCL 3 0.0 UP TO 6000.0 SEC MODE 1 5.00000E+05 (G) RG 5.00000E-05 StGL 8.00000E-01 RHO 3.000 i CONTAINING THE FOLLOWING NUCLIDE FRACTIONS ! NdCL 1 0.0 NUCL 2 5.00000E-02 NUCL 3 3.00000E-02 M00F2 4.00000E+05 (G) RG 6.00000E-05 SIGL 4.00000E-01 RHO 2.000 CONTAINING THE FOLLOWING NUCLIDE FRACTIONS . NUCL 1 0.0 NUCL 2 8.00000E-02 NUCL 3 4.00000E-02 l CONIAINMENTTEMPERATURE IN TIMESTEPS OF 1 SEC UP TO O SEC l 130.00 l l STEAMFLOW INTO THE CONTAINMENT AT 1 GRID PolNTS n e l TIME (SEC) RATE (G/SEC) l 0.0 0.0 LEAARATE OUT OF THE CONTAINHENT AT 1 GRID P0lNTS TIME (SEC) RATE (1/SEC) 0.0 0.11600E-06 9 J
I TIME STEP # 100 1.5911 SEC PROBLEM flME 132.70 SEC = 2.2 MIN = 0.0 HRS SPECIES MASS CONCEN1 RAT 10NS ACC. SEDs. OEPOSIT ACC. DIII. DEPOSIT AIRBORNE MASSES ACC. LEAKED MASSES (C/CM**3) (C/CMa*2) (G/CM**2) (C) (C) COND. WATER 0.0 0.0 0.0 0.0 ORY PARTICLES 0.132427E-04 0.718488E-03 0.247219E-01 0.953476E+06 0.150285E+02 TOTAL 0.132427E-04 0.718488E-03 0.2472790-01 0.953*76E+06 0.150285E+02 NUCL 1 0.264854E-06 0.143699E-04 0.494555E-09 0.190695L+05 0.300569E+00 NUCL 2 0.529709E-06 0.281395E-04 0.989115E-09 0.381390E+05 0.601141E+00 NUCL 3 0.0 0.0 0.0 0.0 0.0 PARTICLE CONC. = 1.25119E+06 (1/CM**3) AV. RADIUS = 0.5386 (MICRONS) AVERACE DENSITY = 4.00 (C/CM**3) LEAK RATE = 1.10663E-01 (C/SEC) CONT. TEMP. = 130.0 (DEC C) SATURATION RATIO = 1.00100 ACC.AER. SOURCE = 1.00000Et06 (C) 1 8 l
-~~- .-m
I i >
, ~ 31 -
J i a E
'1 e
a e e e e e e e e e e a e e e e e e e e e a e ae eVm e<m 8 Z$ e es eCZ emo EON e c- C. C. C. C. C. C. O. O. O. C. O. O. C. C. C. C. C. C. C. C. C. C. C. O. C. C. C. C. O. O. O. O. O. C. C. C. C. C. C. C. C. C. O. O. C. C e e COCCCCOOCCCCCCCCCOOOCCCOOCCOCOOCOCOOCCCCCOOOOCCCCCO e e I e e e eO ~CcamNeOOCN%@cc:#3mmANNee-CCCootmo@cceeONNNNNNN@c4@ mnNNNNNNewe--we-----*-~~e-eweeOOCCOOOOOOCCCCCOOCOOO eM s<~ s e e e e e e e e e e e e e e e e e ee e e e ee e s e e e e a s e e e e e e e e e o e e e e e e e e 8 Im Ww.WWWWWWWWwWwWWWWWWWwWWWWWWWWWWWWWWWWWWWWWWWwWWwWW OOOONNC3CNNC Co&3N e e 1 08 ~ Norm *O-~~N-33C@@NCACCCcomS@~Occ4NNN43C3cN-NNcm-C:"NmMeccC m3NJeNOBOMM&eC&AN&mCMBCO
, e~Z CO4NNNONeNm?mCNJ.N"NMNOCOC@Nm@NM@NNO3co@ccccee&NOOOO NNm3ecNNCOCOCe@eNNN#NONMCNCoocamme&JNJNmecmN#COCmNa 8 wQ eCN -
~.e N N.3.C.4.:.m.e .N. m e m e n c.N.D.e N.D.e N.e.w.e. m
. . . . . . . . c.e. w
. m.D.C.w.N.m. e.k e e N.m 3.D.D.e
. . . . . . .e e N.
eeMO - CCCOCCOCCCCCOOCCCCCCCCCCCCCCCCCCOOCCOOOOOCCCCCOOCCC I e e e i
. e eN 3~7C@r3mNN="CO~~wNNNmmmmm:332:33necDochberberefecto
--COCOCCOCCCCCOCCCCCCCCCCOCCCCCCCCCOOOCCCCCOCCCCCCO
{ eM e e e e e e a e e e e e+++++++++++++++++++++++++++++++++++++++ f r e-eQ e WWswWWWwwWwwwwWWWWwWwWWwwwwWWWwWWWwwwWwWWWWWWwwwWWw CO:70cm<cNDNecuec~3m-ComeccommN~cc@mDemNe~owacoNNeN
; e e. ~ re"remNOCm40:memachNmeOecm:s&N43%N@2NNecDNC"emNmr3OODOONN3Ne@mcDe"N33" C=4 EMU DmD@ceNsme@NN ONmeNeornNw ce@OeNwNccomOch@
- t. ezo N
MCOOrm3G"m"nN@wmc~mceeme"N@cQCCOwscNNekNN~@@N3&PmeN
- c. ~. ~. @. e N. N. ~. ~. r. N. . . . . . . . . . . . . .c.
e g e . .
.N.m.c.@.w.e . .e
. e N.N.m
. . m c. .c.c.c.c.c.c.c.c.c.c.
f j e e - e CCOOCCCCCCCCCCCOOCCCCCCCCCCOCOCCCCCCOOCOCCCCCCCCCCC e . 8 e e . se *C@omNaCOONN@6A334mmNNNaeeOO
, ea eCa mNNNNNNN--we-eee-----wwwww-eOOommmOOOOCONNNNNNN@@@@ eOCCOCCOOOOOCCOCCCCCCC e e e e e e e e e eee a ee e e e ee e e e ee e e e e e e e e e e e e e e e e e e e e e e a e e e e-M eEt WWwWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWW CocONNOn@NNmes@@cNaks"CeccNNNc3C DNemNcmeO see - N O # m e c e - - f. - E c C C @ N e o C C o c c m a m 4 N J - N C & O m m m e O & a N" NoMCM&QO mmec@c eME e-U @CCNNNCNANmomCNNNeNmNcCCC@Nm@NmCNN73C@@Oc@Q@"@NecCO NNm3mfNNDO4 Ce@eNNN3NCNm4NOOcne@mbOsN3N>wDmNPOComNe EON 000 c. ~. ~. N N. e. @. 4.
- s. m e.N. .m.
e m e m c. h. o. . N. @. e N. e. ~. e m. @. ~. e .n.. c.
. . . . o. . N. m. c.. N . .e e N. m. c. D. e. e e w N I W-CCOCCCCCCCCCCCCCCOOCOOCCCOCCooCCCCCCCCCCCCOOOCCCCCO
- 1 e
e e e eC p<
@@@@@fbefffbPPh@bbreffbObfDfffD33333333333233333333 CCOOCCOCOCOCOCCCCCCOOOOOCCOOCCCCOCC00000C00000CC000 og eeaeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeaeeeeeaeeeeee WawwWWWwW6WWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWW e
I e. <34 CECO 304443N3N&eNNN#ACONNcOO&oCeNONN&e:#-4NNONO@c CO-NAmNOCOOeAe3@@@N#3cNCO3ecomCON#@CNem@c40m333@CNm 1 i sN e&~ ~ ? 7 te r & C % O b C N @ @ @ N @ C O P M b O O & b O N C b c O O N D O C 3 m 3 N N O C m & C N @ C Nm;NDO"NMDc%@emeDOmkCa@mCMODm"mo"m3DNGCN3DONDCerC3C l e<E 1 eLG @ N C. C. @. *. *. *. *. *. e e
- N. N. N. N m. m. m. 4 4 4. A. c. @ @. .N.O.
. ew . e
. *. *.* .". .N.N. N
. N N .M.m m M 3 e A @
l i - COCocooCCCOOccecceCoCococeoeceoCoCoccooceCOcCoCocco e l e e I e i e eM Nm3cCNe&C*NMae@NemC-Nm#e@NemO-Nmse@NemO-Nmne@Neme-N wee-wee-4NNNNNNNNNmmmmmmmmmm44###a34c4ccenceceAcc@c I I e e
w- - m, d 63 0.65T4:E-04 0.2562tE-06 0.53815E+05 0.25621E-06 0.0 . 64 0.71928E-04 0.30633E-06 0.49135E+05 0.30633E-06 0.0 . 65 0.786?2E-04 0.35983E-06 0.44076E+05 0.35983E-06 0.0 66 0.86091E-04 0.41523E-06 0.388t:1[+05 0.41523E-06 0.0 67 0.94181E-04 0.41012E-06 0.33626E+05 0.47012E-06 0.0 68 0.10304E-03 0.52425E-06 0.28598E+05 0.52425E-06 0.0 69 0.11273E-03 0.57364L-06 0.23898E+05 0.51364E-06 0.0 70 0.12333E-03 0.61679E-06 0.1962?E+05 0.61679E-06 0.0 71 0.13493E-03 0.651780-06 0.15835E+05 0.65178E-06 0.0 72 0.147620-03 0.677020-06 0.12561E+05 0.67702E-06 0.0 73 0.16150E-03 0.69136E-06 0.97961E+04 0.69136E-06 0.0 74 0.lT669E-03 0.6942TE-06 0.75123E+04 0.6942TE-06 0.0 75 - 0.19330E-03 0.68565E-06 0.56656E+04 0.68565E-06 0.0 76 0.21148E-03 0.66599E-06 0.420?5E+04 0.66599E-06 0.0 77 0.23131E-03 0.63631E-06 0.30665E+04 0.63631E-06 0.0 78 0.25312E-03 0.59788E-06 0.22003E+04 0.59188E-06 0.0
- 79 0.27693E-03 0.55242E-06 0.15526E+04 0.55242E-06 0.0 80 0.30297E-03 0.50179E-06 0.10770E+04 0.50179E-06 0.0 81 0.33146E-03 0.44794E-06 0.73419E+03 0.44794E-06 0.0 82 0.36262E-03 0.39278E-06 0.49164E+03 0.39278E-06 0.0 83 0.39672E-03 0.33811E-06 0.32320E+03 0.33811E-06 0.0 84 0.43403E-03 0.28554E-06 0.20846E+03 0.28554E-06 0.0 85 0.47484E-03 0.23636E-06 0.13tT7E+03 0.23636E-06 0.0 86 0.51949E-03 0.19159E-06 0.81562E+02 0.19159E-06 0.0 GT O.56834E-03 0.15188E-06 0.493T8E+02 0.15188E-06 0.0 88 0.62179E-03 0.I17610-06 0.29199E+02 0.11761E-06 0.0 89 0.68026E-03 0.88820E-07 0.16840E+02 0.888?OE-07 0.0 90 0.74423E-03 0.65312E-07 0.94567E+01 0.65312E-07 0.0 91 0.81421E-03 0.46660E-07 0.51594E+01 0.46660E-07 10 . 0 .
92 0.89077E-03 0.3?311E-07 0.27285E+01 0.32311E-07 0.0 l 93 0.97454E-03 0.21629E-07 0.13948E+01 0.21629E-07 0.0 0 94 0.10662E-02 0.13951E-07 0.68707E+00 0.13951E-07 0.0 l 95 0.11664E-02 0.86372E-08 0.32484E+00 0.86372E-08 0.0 96 0.12161E-02 0.51086E-08 0.14612E+00 0.51086E-08 0.0 97 0.13961E-02 0.28708E-08 0.62964E-01 0.28708E-08 0.0 98 0.15274E-02 0.15221E-08 0.25504E-01 0.15227E-08 0.0 99 0.16710E-02 0.75613E-09 0.96719E-02 0.75613E-09 0.0 100 0.18282E-02 0.34810E-09 0.34003E-02 0.34810E-09 0.0 101 0.20000E-02 0.16485E-09 0.12298E-02 0.16485E-09 0.0
- - - . , - - ~
-- ~ - - -
__n rr s:;n- - --- - eee**e ******eeeewoommenee==*eaea............ oeeee=======emee****************e=*****e=====ee**eneeseeeeeeeeeeee==========** 8.6900 SEC PROBL E M TIME 10086.10 SEC = 168.1 MIN = 2.8 HRS , TIME STEP # 1900 MASS CONCENTRAll0NS ACC. SEDl. OEPOSIT ACC. DIFr. del'OSIT AIRBORNE MASSES ACC. LEAKED MASSES SPECIES (c) (c) (c/ cme *3) (c/CMee2) (c/ cme =2) 0.0 0.0 0.0 0.0 COND. WAT ER 0.423063:+06 0.749920E+03 DRY PARTICLES 0.587588[-05 0.228155E-01 0.113156E-05 0.537588E-05 0.228155E-01 0.113156E-05 0.423063E+06 0.7t9920E+03 4 TOTAL 0.?97408E-07 0.27590 tat-03 0.133351f-07 4 0.21r 131 a E+0fs 0.860591E+01 4 NUCL 1 4 0.374572E+02 NUCL 2 f 0.3378442E-06 0.1123GSE-02 0.561121E-07 0.2t29580+05 4 0.151172E-06 0.310671E-03 0.160097E-07 0.1088'4 4 E +05 0.110124E+02 NUCL 3 PARTICLE CONC. = 1.96149E*05 (1/ cme *3) AV. RADIUS = 1.0969 (MICRONS) AVERAGE DENSITY = 2.72 (C/CM*e3)
= 4.91580E-02 (C/SEC) CONT. TEMP. = 130.0 (DEG C) SATURATION RATIO = 1.00100 LEAK RATE ACC.AER. SOURCE = 1.90000E+06 (C) 1 8
i I
O
~ 34 ~
e
,1 l
e e e e e e e e e 9 e e 4 e I e e
' I e -
e 8 l 6 s > e b 8 e e ee sea i e<m ' eke s ee 80K , eNU e0% \- e Q Ce 0 0 0s 0 0 0 0 0e 0 0 0. C. 0 0 0 0e e0 0 0 0 0 0 0 0 0e 0 0 0 0 0C.O O.O.O O O.O.O. e e e O.O O.O.C.O.O eO.C.CeCe e a e w C00000000003000000000000000000000000000000000000000 e e e ! I e e e@2NemeNN@e22mmNNepecOO@@@eceekh*NNN@@@@@@@@@@@@@@@ l eW NNNNNe*eweeeeeeeeeeeeeeOOOOOOOOOQ000000000000000000 l eM e ee eeeeeeee eeeeeoee ee ee eee e e e e e e e e e e e e e e eaee e e e a e e e , e a wwwwwWWWwwWWWWWWWWWwwwwwwwWWwwwwWwWWWwWwwwWWWWWWWWw e m Nme@CN%@@m@mmme@%e@2NEM@ONN@@@2OOOmmmO@m@meNeNNe2@m 5 e e @eme#2@20eOchmeeCece4Coc2NkewemkN2Oce2 Nee %Ce@@@@ec2 eCe @Oe240c@NNNe@NN@meNOAc%@@@@@m@ceNeme2@eNeeN@mmece@n e-K OOeNNONe@2%N@@~2eQN@@@C@~mecommeNN@%mhmench-mh@@Nne eWU g 80% @. e. r h e *mee e c e. N e n. e c. e. n. e N. @. e m. .@. e n. c. e N. e. m. e N. m. 2 4. @. e ..
****.** . . . . . eN
- Ne m. m .2 2 2 2. m. m. m. Ne N 840w 000CC000000000000000000000CC00000000000000C00000000 e
e a e e e NO&%@c2mNNwCOcee*NNNmemm222222222ecereceer222224qmm ew -~C0000C0000000000000 0000000000000000000000000C.00 eW e ee ee o e e e e e++++++++++ +++++++++++++++++++++++++++++ e= wWWWwWWWWWWWWWWWWWWWWWWWWWWWWWwwWWWWWwwWWWWwWWWWWWW eO QN@@@2@eNnnm@NoceNmece@N27eeN@Oweec=c@u27CnN@@m@@@c necemoceQNce@nnON7%+4newOeckon@Necenm@ nmo =@%@neN%ke e e. a &#@NeoeNenNoe*NmmemmenenNeca~N@@eh24emesN@OsmhOcNNN e eto CornmemenNNQesce&@nomme@eN22emec@eNm222mNOcem@emmen ezu w. e. .r.e m.2 2.m.e. n.m.e. m . e.N.2.@.e m.n . . e N.2 4.*.e u.m s.e.n.e.@.e.e e.e ...e. e..e e e c.@.e.m.h.e e.e.r. a % . . . . e e w C000000C00C0000000000000000000000000000000000000C00 e e e e e
- eh c42Ne@en%@e22mMNNeweOCO@@@eece%nNnNN@@@@@@@@@@@@@@c eO NNNNNeeeeeeeeeeeeeeeeeeCOOOOOOOOO000000000000000000 sea ee e ee e eeeeeeeeee eeeeeeee e e ee e e ee e e e e e e e e eeee e e e e eee e-m WWWWwwwwwwwwwwWWWWWw6WwWWWWWWWWWWWWWwwwWWwwwWWwWWWW eme Nme@ON%@@m@mmme@ne@2Nem4CNh@@e2OCommmomm@menence2@m one weece2@2OnoeNmemonew@Oce2NneecmnN20erekOen@e@@@@ee2 0WK @eca@OceNNne@NN@nemon@n>e@@@m@eeNemL2@eneeN@cmece@n eeu C5cNmONe@2nN@ces*Ckw@@omemee*m e0%
eMC @ e. c. N.e.
. *=**w*e *e eene e N e d* *e. r. e s. e N @. ewa m e. e n. c. e* .N 5 c. m e* **
* * . . N *%= e@ n n n m e s c h m eMw C000000000000000000C0000000000000000000000000000000 e
1 e e e e eeeeeeredece442222224422#242224422222mmmmMMmmmmmmmm e C000000C0000000000000000000000000000000000000000000 ee e ee ee ee eee oee eeeeee eeoee eeee e e ee ee a ee eeee e ea eo e ee e e WWWWWWWWWWWWWWWWWWWWWWWWWWWWWWwWwWWWwWWWWWWWWWWWWWW
- ee nNaceQNNacoecoeN@%N@ende2NNmNe@eceNenommmNO@on~Nmn4 eachcedwoome@N2@ohem@@@cmq24@ph CONm@@e@mame@@2 ee 2N@
emNm2n~@m-em@Ne eN OemcoQecONOce@CN40@2m2NNOOmmem@ N@@ ema ~CnememO@me@Cem2eneON24 Nece+02 een 20eNm44N@emenom e4K M.a.n.d e e e.@*@*M*e*@ * @. .e ee .e
- ee
- e
* *.N.N.N.N. . m. m m+ 2e 2. b. e. . @. k. k. e. m. e w . .e. e. e. e. v. e.N.N.N.N.m m.
e40 e w C00000000000000000000000000000000000000000000000000 M e e e e e eT M" N m n e @ h e @ C " N M a c 4 N e m o w N m n e @ N e @ O w N m n e @ N e @ C w N M a c W N e @ O emmmmmm e 4
- - - -, - -. .L = = = _ : : _~. r :: ~ -- - ~ ~ ~ ' - ~ ~ ~** ~ ~ .
82 0.36262E-03 0.19594E-06 0.36085E+03 0.1959'sE-06 0.0 . 83 0.396I2E-01 0.15617L-06 0.21965f+03 0.15611L-06 0.0 . 84 0. 4 3'*0 3 E-0 3 0.11927E-06 0.12810E+03 0.11921L-06 0.0 85 0. 4147t 4 E -0 3 0. 86t4 31 t -0 7 0.1089?t+0? 0 . 8 67 31I-07 0.0 86 0.51949E-03 0.58150E-07 0.36799t+02 0.58/50t-01 0.0 87 0. 568 3 r4 E-03 0.369561-07 0.11677t+02 0.36956L-01 0.0 88 0.62179E-01 0.21183[-07 0.77379t+01 0.21183E-07 0.0 89 0.68026E-03 0.10868E-07 0.30320E+01 0.10868L-01 0.0 90 0. 7 tats 23 E-0 3 0. 4 39r41 E -08 0.10's ?6 E + 01 0.48941E-08 0.0 91 0. 81 r.21[-o 3 0,18938t-08 0.3081oE+oo 0,18938t-o8 0,0 92 0. 890 7 7t-0 3 0.6160$[-09 0.76539E-01 0.616051-09 0.0 93 0.91f54E-C3 4 0.16492[-09 0.1568a /E-01 0.16492L-09 0.0 914 0.1066?E-02 0. 3 %?*;E - 10 0.25814E-02 0.3*>6?$E-10 0.0 95 0.11664E-02 0.61061[-11 0.33788E-03 0.61061E-11 0.0 96 0.12161E-02 0.81930f-12 0.34621E-04 0.81930E-12 0.0 91 0.13961E-02 0.85111E-13 0.21885E-05 4 0.85111E-13 0.0 98 0.152 i r4 E-02 0.68070E-14 0.167/6E-06 0.68070E-Its 0.0 99 0.16T100-02 0.f1585E-15 4 0. 78265E -08 0.41585f-15 0.0 100 0.18282E-02 0.193370-16 0.?T792E-09 0.19331E-16 0.0 101 0.20000E-02 0.70603E-18 0.71502E-11 0.70603E-18 0.0 I N I l1 2 L _ _ _ _ _ _ _ _ _ '
TIME STEP E 4688 811.7292 SEC PHOBLEM T IME 360024.19 SEC = 6000.4 MIN = 100.0 HRS SPECRES MASS C0hCENTRAft0NS ACC. SEDI. DEPOSIT ACC OlFF. DEPOSIT AIRBORNE MASSES ACC. LEAkfD MASSES (c/CM**3) (c/CM**2) (c/CMa*2) (c) (c) COND. WATER 0.0 0.0 0.0 0.0 DRY PARTICLES 0.194028E-09 U.293463E-01 0.165637E-05 0.139700L*02 0.110'sSOE+04 TOTAL 0.194028E-09 0.293463E-01 0.16563TE-05 0.139100E+02 0.110480E+04 leUCL 1 0.982061E-12 0.308685E-03 0.159930E-07 0.707083E-01 0.10402SE+02 IsuCL 2 0.111421E-10 0.149801E-02 0.862463[-07 0.8022T3E+00 0.578358E+02 seUCL 3 0.499185E-11 0.478509E-03 0.295391E-01 0.359413F+00 0.201353E+02 PARTICLE C0feC. = 1.67496E+02 (1/CM**3) Av. RADIUS = 0.4378 (MICRONS) AVERAGE DENSITY = 2.72 (C/CM**3) LEAK RATE = 1.63451E-06 (C/SEC) CONT. TEMP. . 130.0 (DEC C) SATURATION RATIO = 1.00100 ACC.AER. SOURCE = 1.90000E+06 (C) l M i
+
_. ~ . w
't l
l l b s : e 4 e i e i e 1 e e se e e
, e i e a
h e e e e e
, e e
o e f eW eMa e<m are eee
' eCK sNu eCN e o C.C. C. O.C. C. C. C.C.C. C. C. C. C.C.C.O.C.C. C. C. C. C.C. C. C. C. C. C.C. C. C. C. C. C. C. C. C. C. C. C.
e = e COCCCCCCCCCCCCCOCCCCCCCCCCCCCCCCCCCCCCCCC e e e e
; e @@2mecockc@ceamm eW eW N.NNNNN=*========NN=ee=CCOCCCCC"~NNm3+>@~2****=-=--*=e*****wwwwwwNN i
e<a eee e e e e eeeeee eeeeee WWWWWWWWWWWWWWWWWWWWWWWwWWWWWWWWWWWWWWWWW eeeeeee eeee eeeeeeee e ee etm N e n @ N e @ m N e 4 @ @ c m o - > @ % N N e @ N m use s N m a @ @ % @ c N e e m e e 24%:nc~~#eenm@~ceNam#cm@cONNnN~c@~~@m k ece kemaere 5~bmarts cc=C@ e-x N o c e -- e c N , e @ r occ ~ e m o m @e C n te s @ @ m e e C @ a c @ @C N N N e s eJu am>4~~ae-No@~cm-Cee@ eQs ~. ~. ~. 2. *. N. N. N. N. .. e m m. e. c.. ~. c. e n. ~. N. a. c. ~. ~. N. N. N. . .N. ~. ~. @. N. m w N. N. e m. m. =. m. ewC e - e COCCCCCCCCCCCCCCCCCCCCOOOOOOOCCCCCCCCCCCC e e e e ek eW *~~CCOOOCCCCCC X C = ~ a
- N r. N N N N N a * = = D" *-NmcN@*#
- e. C e % @ c # m m N = = = gCCCCCCCOOCCCCCCC CCCCCCC*a e= eeeee ee o ee e ee e++++++++++++++++++ee eeeeeee eC WWWWWWWWWWWWWWWWWWWwWWWWWWWWWWWWWWWWWWWWW c%@~Ner@@@@dekm a @ N e t n @ 2 ~ 7 ~ 2 = C4~@mam c422:eN@ e c t. c = N C Ccam%ceanc e
e e.
- N?cem@Cmm4No@m@@@*@~c@-@@e@=~@@b%@$eNcN=@@+ce
% k 2N#t4NecneeccaC J,
e20 m@@~mNmecc--sNacN@eNNACNNoae%N-cmacCwNeme ezu %..@ m.*.N. e.m.N. .N..e.w m.@.N. e.a e s e . . ........ . .. c.N ... t.s e *
- N N N a = @ a m e...
= = =.# . c N.. N 4.
e e e w CCCCCCCCCCCCOCCCCCCCCCCOCCCCCCCCCCCCCCCOC e
" a ' e .
e~ e@gm*Coc%@@edamm eO NNNNNN--=--==== NNewe*C.CCCCCCCw=NNm.ee>@*2
...... -e----ww.ww w eNN i
i i eCa eam eeee ee eeeI ee eeeeeeI ee eeeeeeeeeeeeeeeeeeee WWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWW eze N d e @ te c @ m N e @ @ @ o m C w w @ k te N e @ N m e c c N
- a @ @ % @ c N e e m I N a m o r O 4 e ~ te t i e~k NCore*@ec%m@b@e*maccaccaNadkNec eM e~u a@NC%emaece@ ~ ~~7m@CaeC@
@mee @sacCNNNea ,
eOw tabCcwemeNde@ecek@momcom@nkhee"@c@%cmeC@es eWQ =.=. ~. 4.=.N.N.N.N.=.% m. .c. .c.e d. .N.a.m.a. -.N.N.N.ee.*.*. .N.N.w.N.N.a m.m.e.m. 9Mw CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC ' l1 e e I e e . eO Weeeee@@##{################4&####MMmmMMmm CCOUCCCCCC CCCCCCCCCCCCCCCCCCCCCCCCCC3CCC I
#4 eeeeee eeeeeee eeoe eeeoeee e e e e eeeee e e eeTeee eK e WWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWW e
ew
- e. eCNmacc@c&-NeNNo-an-amNeNe@cceN-N4mmmNC@c ema-e*mem a@ Chem @@@cm an>CN@aN@@eCNmo@c@m CCheCCCcesCN4C@nmanNC eMa e<K @N m@eN>Cn@@C~mNmen~@m
~ con '
m e @. m 7@.@n.me.e@.@@.C~ ~. m~. a c nw.e~.Ce.NN.a N.
@ h. .mN. m m@ e n@b e @C @e ~ e @ a c e wN m n @ n e eEU b. c. . e e. . . N. ......n.n.e..~.~.~.~.~.~..e .
8 e e
=
CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC i i e i e i e ed c c k w @ C = N m o r e n c e C e N m e e @ k e @ C ~ t_ i m m e @ k es w o e h m e h mmmmmegadddatetahechbehebb@@w@@@@@@@hkkh>k ' e e -I I
--. as --
e s &
'o
~.
~ <
- ~
o
~.
m'a e-
" 3 m :
.d.'o' c .
r "c
~.
~
o o
. . . . " 10 ' ' ' ' ' ' ' " 10' ' ' ' ' ' ' " 10' T It1E C S )
LW1 RIRBORNE AND ACC. LEAKED MASSES M l
~ 39 ~
8 a R S R 4-m m z m 88 Ed
/ N w
m L
~.
MA
&d-8
{ d'
\
R g. , 8
- e ,,
10' ' ' '10' '10' '10' TIME CS) . Ju. F1 RVERAGE PARTICLE RADIUS I
._._.._._m_
_. w ._
. . a
~
S
- i
- ( nucl 2 nuci 3
~
o nuci1 _ m. ea w-lV . a v, i i P
. S ,
T S
"'10' ' ' ' ' ' ' " 10' ' ' ' ' ' ' " 10' 10' ' ' ' ' ' " ' 10' '
TIME (S) rgr4a L u, Al.W1 ACCUMULATED LEAKED NUCLIDES s 9 I
j
. . _u- ; D -
. Mm 9 ,G 2m
/
2h 58 m Ew x\\
/ .i s!
e-l @l l h, ' 11h 83 ,
\\
ll E' 6 fl
\
!. ~
- i tl 9 1 51 h fl .
\
i
'o 10-8 i i 466366 10 4 i i i 6 6 fi4 10' i i i 5 5I55 10' STOKES RADIUS
%, LRF1 e
PARTICLE SIZE DISTRIBUTIONS e
--
- 2 --
i
. . o I
//LAF295E2 JOB (0295,091,P4311),K0YRO. NOTIFY =LAF295,TI!!E=120 I
//* MAIN LINES =30 f
//E2 ENEC F7CG
//C.SYSPRINT DD DUMMY
//C.SYSIN DD DISP =SHR.DSN=TS0oS7.NAUA?!004 FORT ,
f
/ / G . FT01F001 DD DSN = LAF295 . L . LI ST ,UN I T= D I S K . VO L=SE R = B AT00ll , DC B = DC B . V B S ,
// DISP =(.CATLG), SPACE =(CYL,(20,5),RLSE)
//G.FT02F001 DD DUMilY
//G.FT03F001 DD N IT= D I S K , VOL=S ER= B AT00ll , D I S P= ( , C AT LC ) ,
// DSN=LAF295.E23. DATA. SPACE =(TRK,(20,5)),DCB=DCB.VBS ,
//G.FT08F001 DD UN IT= D I SK ,VOL=SER = B AT00ll , DI S P= ( , C AT LG ) ,
DSS =LAF295.E28. DATA SPACE =(TRK,(20,5)).DCB=DCB.VBS $
//
/ / G . FT09 F001 DD DS N= LAF295 . REST 2 , UN I T=D I S K , V01.=S Ek= B AT00ll , DC E = DC B . V B S ,
// DISP =( ,C ATLG ) , S PACE =(TRK , ( 1,1 ) ) l
//G.FT10F001 DD DUt!MY
/ / G . FT 19 F001 DD DSN =65C PUZE IT , UN I T=S YS D A , DC B =DC B . V B S , D I S P= ( , PAS S )
//G.SYSIN DD
- N AU A S A!! P LE C AS E # 2 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
CONTAINMENT, BEST EST!!! ATE AEROSOL SOURCE, STEA!!, GRAVK ----------- 72000.E6 6460.E4 50540.E4
- 1. 1. 1. 0.01 0.0025E 4 50.E-4 101 0.1 1.E-15
- 0. 100. 120.
.F. .F. .T. .T. .T.
.T. O. O. 1846. 12402.
20 200 5 5 1 170.94 .T. .1E-4 .29 8. O. .F. .1E-4 .29 8. 1170.
'NUCL l' .05 0 350.26 .T. .1E-4 .29 8.
O. .F. .1E-4 .29 8. 2026 L
' NUC L l ' .05 0 956.94 .T. .1E-4 .29 8.
O. .F. .1E 4 .29 8. 2340.
'NUCL l' .05 0
- 0. .F. .1E 4 .29 8.
O. .F. .1E-4 .29 8. 8580.
'NUCL l' .0 0 111.11 .T. .1E-4 .29 8.
61.11 .T. 1.66E-4 .77 2.5 10380,
'NUCL l' .05 0 1 1.
130./ 4 1846. 5088.38 4615. 36279.07 5607. -7295.01 12461. O. 1
- 0. .11574E 6
/*
i> _ o3
//E2S2 ENEC F7CG
//C.SYSPRINT DD DUMilY
//C.SYSIN DD DSN=TS0657.NAUAMOD4. FORT, DISP =SilR ,
//G.FT01F001 DD DUMMY
//G.FT02F001 DD USS=*.E2.G.FT01F001, DISP =SliR
//G.FT03F001 DD UNIT = DISK,VOL=SER= BAT 00ll, DISP =(,CATLG), *
// DSN= LAF 295 . E 2 S 23 . D ATA , S P AC E= (TR K , ( 20 , 5 ) ) , DC B = CC B . V B S
//G.FT08F001 DD UN I T= D I S K , VO L= S E R= B AT00ll , D I S P= ( , C AT LG ) ,
// DSN= LAF29 5 . E 2 S 2 8 . D AT A . S P ACE = (TR K , ( 20 ,5 ) ) , DC B =DC B . V B S
/ /G. FT09F001 DD DSN=LAF29 5.EEST2S2, UNIT = DISK VOL=SEReBAT00ll, I // DC B =DC B . VB S , D I S P= ( , CATLG ) , SPACE = (TRK , ( 1,1 ) )
//G.FT10F001 DD DUt!!!Y i //G.FT19F001 DD DSN=*.D4.G.FT19F001. DISP =(OLD. DELETE)
{ //G.SYSIN DD
- NAUA SAMPLE CASE 82-------------------------------------------------
ANNULAR GAP, AEROSOL SOURCE = LEAK OF CONTAINMENT, NO STEAM, GRAVK -- l 1.4675E10 2.564E7 2.36235ES f 1. 1. 1. 0.01
! 0.0025E-4 50.E-4 101 0.1 1.E-15 I 0. 100. 120.
.F. .T. .F. .T. .T.
.F. O. O. O. O.
10 100 5 1 1 4 0, .F. I 1 1 1 0. .F. I 1 1 0. 3 'NUCL l' O. O. il 1 1. li 130 / a i 0, 0.
]p 1
- 0. .5E-6
//
s k 4 4 9 e
I '
-- 44 -- . . a NAUA - MOD 4 INPUT DATA LISTING TENT : NAUA SAMPLE CASE #2-----------------------------------------------
CONTAINMENT, BEST ESTIMATE AEROSOL SOURCE, STEAM, GRAVK ------- CONTROL VOLUME PARAMETERS VOL : 0.72000E+11 FSED . 0.64600E+08 FDIFF : 0.50540E+09 i AEROSOL PROCESS PARAMETERS FORM : 0.10000E+01 l FOR!!C : 0.10000E+01 FOR!!KO : 0.10000E+01 DELD : 0.10000E-01 NUMERICAL PARAMETERS Rt!!N : 0.25000E-06 RMAX : 0.50000E-02 KMAX : 101 EPS : 0.10000E+00 CUTOFF : 0.10000E-14 PROGRA!! CONTROL PARA!!ETERS SZEIT : 0.0 TIME : 100.00 CPUZT : 120.00 RESTRT : F ZWCOMP : F FOLGE : T GRAVK : T LEAK : T LKOND : T Tit!EK1 : 0.0 TI!!EK2 : 0.0 T!!!EK3 : 0.18460E+04 TIMEK4 : 0.12462E+05 OUTPUT CONTROL PARAMETERS NVRITE : 20
- 200 NPIDT : 5 i
)
. g $l i
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0, 0 0 0 0 5 8 8 8 8 8 8 8 8 8 2 _ C E _ O O O O O O O O O 3 S S H H H H H H H H H H t R R R R R R R R R R O D l 1 1 1 1 1 1 1 1 1 1 L 0 0 0 0 0 0 0 0 0 0 C - - - - - - - - - - U E E E F E E E E E N 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 E 9 0 0 1 0 0 0 0 9 0 0 0 0 0 1 _ 0 0 0 0 0 0 0 0 0 0 9 9 9 9 9 9 9 9 9 7 P D U S _. N A 2 ? 2 2 2 2 2 2 2 7 C S E T I m l S L L L L L L L L L L S M O . E C C C C C C C C C C l P S I I I I I I I I I I 1 o - A S S S S S S S S S S P D i I t 5S S5 SS SS 55*gSS 55 SS 55 s D R _ E 0N O4 ON OM 0 O% 0 *4 ON rh 'o SN E C R O O O - C C C 0 O C - O R fl fI El EI E I EI El EI E8 El C 1 5 ) f H oT Di OT OT nlT r)T 1T O1 0i F wAC OC A OC OA CC A oC A iC oA 0C 0A h C M A mC A 0A0C O 4 E foR &R ' RG M R O. t i mR oR 0R h iR OR 6R S P _. I C. F 0. I t. t. T O. T O. T 0. F O. F O. F 6. F E T T A
% 1E 1 E 1E 1E 1E 1E 1E 1E 1E 1E T A 1
D D C D C C O C O D S ! A I I I I I I I I I I E T h T CL CL CL CL GL OL OL CL L CL M M E M RC RC O 3 U RC RC u U RC RC U U RC u FC U NCU FC U I T f w )
=
% )
C )h M )b N )M N h M )A )% A C 4S4 l C 6 C C C C C M I E nP0 A F 0 _ S E E G EG 0 C O f0 EG A S +++ S - SC SC I t M E N Sh 4 % % S4 Sh T / EEt M / E _ C /I 2)I /I2)I /l 2)l )i )i
/ l 2/' F 4 C t
1 CW0f.N (C-(C CW0CW ( O-(O GW0CW (O* C CW CW (O (C Cw0CW ( O-(O R t; 0 C ( %/5 329 9c c C 1 ( 4 7 5 C EE L LE L LE L L L LE L T E 062 E E 1 SC2LO L C2L0 C2L C C2L01 L A UI0CO 0 T 0O0 O E0CN L O L E L O L 0 E0C' 0C R E H I A 5 31 O. HI T A 1 f S+6 O E O f S+F 0 E 0 E S*iO E O F S E f S+F E c+F t t E P T P 0000
- F R 0. _
E000EO. E00 E0 t M O O _ E C 0. 4 N0 E O.E H O 5. 6 t0E0 H 4H H H E
) 0 0.10 E e. 6 HE n. E
+ f I T _
R09T5 I O62I5 T009T5 T00 T0 I001T51T0T 1 N ) :4 4 's S t ) _ U7O 20 46 5 31 T I C 0000u C _ O1TC 31G 4 E + + ++O _ S1 .% 0. 0l w 20 5. 3I E 0. G 0B N 2 3 5. % 9l
- 0. NG 850.C 0.E 0 .N 1. G% f w S EEEE E
S 1l k h N OI 0I 0I 11 I 6I M0o ( 0CU1 E ( _ L 4 h M N M M N L 6ST6T O l1 t 1 I 1 t 1 I1 1 I 1 I 1l % 0. T E 41 n 4A E S01 A 2A O1A 2Al 1 01A ?A 01 IA 2AI 1 OTA 2A A0M w O1 ET Lff LTEI LE1 L1ET LET L1fTLET LTET LEI L1 3A i S. 6. I t. S 1u2R
- I M 0 .
R CMC 0%C CMCD%C POOU0CUPOOUOcuP0OuOCUPCOt 0MCD4C CMCOMC CNC0MC41E T .
.A T O0OOE f T 0 AUMCMMCkUMChMChUPChPChUMCMNCMUMChPOUPOOu0OU0 9 CMC S L
s
. I i TIME STEP g 1355 1.1820 SEC PHCOL E M IIMI 1807.43 SIC = 30.1 MIN = 0.5 liitS SPEC 4ES MASS CChCfMIRAllCNS ACC. SE OI . OLPOSli ACC. Olft. DIPOS!I AlH80HNE MASSES ACC. LEAb!D MASSES (G/CM**3) (C/CM**2) (C/CH**2) (G) (G)
COND. WATER O.0 0.0 0.0 0.0 DRY PARTICLES 0.581512E-05 0.601892E-04 0.355691I-06 0.418132E+06 0.362856f+02 TOTAL O.581572E-05 0.601892f-04 0.358691E-Of, 0.418132E+06 0.362856E+02 huC1. 1 0.290144E-06 0.303904E-OS 0.lf9324E-07 0.?u9336E+05 0.181401E+01 AV. RADIUS AVERACE Of NSI TY
- PARTICLE CONC. = 1.2OOO6E+07 (1/CM**3) = 0.1769 [ MICRONS) =
8.00 (C/CH**3) LEAA RATE = 4.346521-O? (C/SEC) CONT. TEMP. = 130.0 (DEG C) SATURATION RAllO = 1.00100 ACC.AER. SOURCE = 4.23172E+05 (C) CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC START CF COMDEASATICM AT A IIHf CF 1.84642E+03 SEC. A 11 MP. OF 130.0 C . AND A STEAM CONTINT OF 1.49866E-03 C/CM**) CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC l I k
.g . . . . .- .,. ._. _-_ , _ - - _ , _ . - _
oo..........................................................................................................................
~
11MC STEP # 2758 88.5399 SEC l'ROBLEM TIME 3973.72 SEC = 66.2 MIN = 1.1 HRS k , SPEC t ts MASS CONCENTRA100MS ACC. SEDB. DEPOSIT ACC. Diff. DfPOSB1 AtELOME MASSES ACC. LEANED MASSES ( c/CM" 3 ) ( C/CM"? ) ( G/CM"2 ) (G) (G) COND. WATER 0.19 Fr463 E-04 0.386?l8E+00 u.837188E-0T 0.1421T3E+0T DRV PARTICLES D.716545E-05 0.369151E-02 0.10TT16E-05 0.559112E+06 4 0.19f253E+03 4 TOTAL 0.27311TE-04 0.189911E+00 0.116035E-05 0.1980884E+0 T
- 0.198s253E+03 seUCt_ 1 0.388218E-06 0.184548E-03 0.538534E-01 0.21951TE+05 0.9T1115E+01 i PARTICLE COneC. = 8.99920E+0$ (1/CM"3) AV. RADeus = 0.5823 (MICRONS) AVfRACE DENSITY = 8.00 (C/CM"3 )
EEAA RATE = 6.47711f-02 (C/SEC) CO*li . TEMP. = 130.0 (DEC C) SATURATION RATIO = 1.00089 4 ACC.AER.SCURCE = 8.00308E+0S (C) l l i O I
. _ .. , , , . W E "
~&~
e e . I e e e e e e e e e e a e e e e e
.e e
e e e e
?OGCCCCCNNAN e# 000000000000 eWa e e e e e e e e e (e wWWWWwwwwWwe ere Nc%m O m 7 & r4 0 % w e e
- JJO?O&e<Oa*W eOZ % O c N aa m .' c c N D a eNJ O#3*C.NNaa mCO EON 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0. C. C. 0 0 0 0 0 0 O. O. O. O. O. O. O. ". *. M *.M . .*.C. C. ". N. ' 1 4 e c .
e e w 000000000000000000000000000000000000000000000000000 e e e e e % P N O C k ? 3 m N N "
- O O ? ? @ C C D C % % % % D @ < @ 4 Q D e b ? Q k D C C C C C T C C C 9; k k e@ OOOQ G0000000 ODOODOOOp032DG300000000 cA Ne Ne t4 e "e *eweee= e"e "a "e *e *e *e e e e e e e e e e e e e e e e e e e e e e a e N eeeeeeeeeeeeee e(*8"" wwwwwwwwwwww6WwwwwwwwwwwwwwWWWWWWwwwwwwwwwwwwwwwwWW e E f O C O P 4 4 '* M J O M m O % Q O p fd O A Q N SI W @ m M m O O % O '* Q N S J 7 SJ*
- J 4 O e # * 'O" @O % # ? M ? O O C m N C O ] p S 7 0 0 "A @ 3 J m A m m O
- 3 4 C "N "O B O O N s) - ON O?}(e%0%
ece 4" 7 c 0 % ** ~ 2mem%re?-prca e >~~t- O p m t *me w e r m m m Q N .~ 4 t e w S 3 7 '- D e m m N e~X
.J 23m4BPN%*er.<%Gr>PCm=c?-?D4mkOpmDN34"emD24OO"m2CC?e3On
- C.N.N.a e.=.=.p eCN . . . . C. r 4 7. (4. .%. =. m e. . N. A. J. . N. #. 4. . . N. m. e. N 7. **. *. *=. m o. .%. -.- N. J. N o. c. c. e. %
. . . . . -. e e#O- 000000000000000000000000000000000000000000000000000 e
e e' e e N C h ? 2 N a C = ~ N m m m 3 4 2 3 d P P r e t c h e r e c r t & # r p 3 3 m te te N N m m m m m m m m "e ~ee OOOOOOOOCOOOOOOOOOOOOOOOOOOOOOOOOOOOOGOOCOOOOOOOO e e e e++++++++++++++++++++++++++++++++++++++++++++ e "M e e- wwwwwwwwWwwwwWmwWWWWwwwwwWWWWWWWwwwwwwWwwwWwwwwwwWW eO O O *J % f40 0 C C N O Nr4fJP.P Q A Q 3 Ofe 7*M3e N O N O m A N C O 3 0 4 J G P-
- N O m Q N J "
e e e. ** P4>~3N~D*kN.*De3%Awm7OOQ*=.=caQ%"t,40)<PNPeO*N"JA O N % 3 k m k N *. C O t* 4 O k e h e 3 D J % O * *= @ N te b e J N b N
- C3**3%A%e nm*"OsaDy23 ON?ew e2Q ~ r a tefa O O c e a e D 3 O O J c k edu 4
- 2 % 2 2
- 2 C % 7 N e @ m m t e r
- Q " eaONOPNNemC3"%c1&p*
e s .N. =. =. 0. .fa.N.a &.m. . .e. m O.*.N.e c.c.=.".N.
. . m. m.#.e. e. 4.N. o. o. c. . . . . . . N. e m . . v.=. .e .m . e.n. e
. .e e -
e = C00000000000000000000000000000000000000000000000000 e e e e . kt N O M%b3 mN Nm" O OTO 7 m n C C %%%%4D DD4 4 D bf ? DH CM m m mh kk k k % hk e e "O QOOOOCGOOOOOOCOLOO e e e e e e e e ee e e ee a 000000340000000000 eCa Ne Ne Ne Ne "e *e *e *e *e *e *e *e *e "e "e e e e e e e e e e e e e a e ee e eeee e~M WwwwwwWwwWwwWwwWWWWWwwwwwWwwwwwWwwwWWwwwwwWWWwwwwWW otmmN ez+ e=e e
*OON O 2J&mmon4msOca c c r y c o n m e m m m e nNS3ms-~~~Ocmk% O m " c~ t4B D% *e rc mntymmQ o m m m m m e m m m " ~ m N s D esc a r m h m m e - J c eA6 DenOnm"2m2m%c~ebr&&oe e- : s m e r r N ~ ~ N % e c e r e w- c e m e c o m ~ O c m eN s e ---w .t e-- e e O c % e% O m >
eOw . e. .N. o.&. .N.a. c. .e. N. m. o. .% .@
. .w. w - m. w.N . . .= - N. c.e. . -.- .- e.m.4. c.
eMC . o.N.N.~.c. . . . .@.N. m.N.% . .- m eew C00000300000000000000000000000000000000300000000000
't e
e e e
- P r e r t e r P r e f f e n d 3 q d s t e d .? e n d 3 q M A n d y; J *34mmmmmmmmmmmmmmm so000000000000000 C eQ e< 000000000000C000.
e e e e e e e e e e e e 0Q0000000Qreefe e e e e e e ee e e e C00Q e e e e e e e e e e e e e e e e e eee am wwwwWwwwwwWWwWwwuWwwwWwwWwwWwwwwwwwwwWwwwwWwwwwwWWw o ceN r &
- o e m
- c e # & - & o e & - ed o ee e
- c m e r4 - O N o m O r4 N m c m QOSP WMQ4073%AN7moeNa#@ O w e c - e m ed om** N "O O J e @ w M e W ew NON m e *- C a c O 4 A o *-rd 5 "4 A M Q G "e>~4N# .f m rd Q N + m & S? .94fvN31O e a: - MNJ"Q*Nc34a"5 oNo semsen c ~ O - m 3 c n w - m w n n r 5.m e e *)O N Qe -r4maoccea-Nac~Sd?
e
-mec~amN
' e4E N. m m. J. a. s. e. c. a. r . c. o. e. . .e .a r. . e.- .e4. N. N. N. m. -e m. a. d. c. o. w. * . r- e. m. . .. e m .m.
. e
. - - N. r . ev. N. m. e. n d.
- ekQ . . . . e - e 0000000000000000000000000000000000000000CC000000000 e e e e eM f4 N k fe D 7rg0m" Pa M mmm3 m b bmkm%m@mOm*4f 4 em 4 43 d? dD#k#n#&JOe "CN0m 0 c.1eboDc % e e@4 70 04 e@N@ A@ $ Po @d ek@MN WNb Ne bh bmbdhf D e b
l i ,
,. . l l ? m j
M h l e l* lP
!l" 'i i
I' t
%44t<404bf?Obefff@@@@@CCk
't e
'j w
.....COCOOOCOCC.CCCCCOOO e e a e e i e e e e e e e e e e wwwww.wwwwwwwwwwwwwwwwwww i e a e e e a e
' Cre%mNbC%CCSPPNC%@%C%NOOO M*4ca: 0#NemCNMC-4C#m&C3Cm EONP.0%Om#33C&NNeo&COpNmm3
.I MO %%CNCNNOC%@em*f4@r47 CON g.e. w . .e. e. e m. 4. m . e
. .=. =. w. .e. =. = w e. % c. p h. e. n
. . . c.
- CCCCOOCCOCOCOCOCCOOOCCC3C 16
.6
%%%%%%c%%%%%%%%%%%%eCCCec f_
' I' COO *******COOCOCCOCOCO000 eeeeeaeeeeaeeeeeeeeeeeaea wwwww.wwwwwwwwwwLeww6wwwww CancQCesm%C~e i, 207~mec~cr7"kC*PMs*3f.O2?%
OOC 244f% n
'6 dbN"000%
OOmreCOOce Cec @*%C4PNkmOffeN*
*%4
@ma .*N .
c A#w
- - - . . . C. . =. c..r.c, 6. f. N. =.. e. e*
a N g & f c. N. r w COCCOOCOCCCOOOOCCCCCCCCOC OOCOCOOCOCOOOCOOOO m m m m m m m m m m m m m fsP=fk N N "CCOOC "C C " C C C
+++++++++++++++++++++++++
wwwww.wwwwwwwwwwwwwwwwwww CNe~m72rOkecQwc~mern%"200
@ % f. S e > P, 3 N A % C 4 m
- C 7 3 O c %
- N & N 2NONmCCC4%2e7?C"2%dmetort 4&acmOm-@CecNo-COommcacC4
". a N. a .=.=.~. f4. N. M N. e. *. C. 4 3 0. w
. . . .@ . e. rv. a.@ . N. a.
COCCCCCCCCCOCOCOCOCCCCCCC
%@@@t<@pereffereb@tD@@@@k CCOOOCC.OCCOOCOCOCCCCOCCC e e e e ee e e e e e ee e e e e ee e # # e e a wwwwwwwwwwWWWWwwwWwwwwwww NOmCMP,m#MMgkCO%@ =mm?3O%*@f4P.ON%
C
- fe n f43f dCw c = > N m * @ "# @ 4 % N a & # @ %" C*CCcC@oe D M G C Q Pw #
M* 4 C c7f4&f4N @@ C4 emp %% @ m7 e om N. e. w. m . . ee . m. c. e...e n. w. e. e. w. .e .e @ N. e. g r4. e.e. m. COCCCCCCCCCCCCCCCCCCCCCCC
.i
& k k k $ k k k CCCCCCCCCCCCOCCOCOCOCOCCC ee e e e e e e e e e e e e e e ee oe e eea e 6wwwwwwwwwwwwwwwwwwwwwwww CPMNONNO M b4 Of4CMc @
N k *C Ce 4r fd rtQ @O @ mP tk W
%C N M N "# 7 C N R b " D C
# fd e # & w a e f a m 3 N N e e e 4 @ 4 # 4 = C N Q
$*@NMDBNCettmf4@QldkCmk*bQ
#. e. e. d. @. % C. @.
. m e * *
- e w r4. N. id. fe. m. m. m. #. #. c.
CCCCCCCCCCCCCCCCCCCCCCCCC N C @ C a camac4%C&C*Nmac@Ne@e NNACCCCCCCCCCP@@@@@@@@@6*C we
(
~
l l . I lME ST E P # 5619 1.1021 SI C PRont t M isMI 14414.62 SIC = 114.6 MIN = 2.9 HHS 1 SPECIES MASS CONCENTRAllONS ACC. SEDI. DEPOSil ACC. Diff. OfPOSli AlHBORNE MASSES ACC, LI AkED MASSES l (c/cH**3) (c/cH**2) (c/cM**2) (c) (c) l COMD. WAIER 0.0 0.384859f+00 0.211236f-06 0.0 DdY PARilCLES 0.28748tE-05 0.1393348-ol 0.160538I-O'2 0.206986f+06 0.350335E+03 TOTAL 0.28T481E-05 0.393736E+00 0.181662[-05 0.206986E+06 0.350335E+03 l MUCL 1 0.994523E-07 0.660458[-03 0.7782680-07 0.716056E+04 0.170910E+02 l PARTICLE CONC. = ti.71477E*06 (1/CH**3) AV. RADIUS = 0.2087 (MICRONS) AVERAGE DENSITY = 14.77 (C/CM**3) l l LEAR RAIE = 2.396820-02 (C/SEC) CONT. TEMP. = 130.0 (DEC C) SATURAllON RATIO = 0.96591 l ACC.AER. SOURCE = 1.11139E+06 (G) I 8 I l l _m e so- + e -. , - -
. ~ , . - - - -
+ - - - - - -- . . . - - - . . _ - _. _ _ . :::~L.r_ = r TL '_~ 7- ~ ~_T._~r_ _ _ -- _ . TIME STEP # 8533 858.2080 SEC PRORl l M flML 360538.19 SfC = 6009.0 MIN = 100.1 HRS SPECl(S MASS CONCINIRATIONS ACC. SEDI. OE POS 11 ACC. Delf. OEPOSII AIRHORNE MASSE S ACC. LEAKED MASSES (C/CM**3) (C/CM**2) (G/CM**?) (G) (C) COND. WAT[R O.0 0.38ta859E+00 0.211?36E-06 0.0 DMy PARilCLES 0.15518:1E-09 0.171110t-01 0.256358E-05 4 0.111106E+02 0.697268E+03 TOTAL 0.155141E-09 0.401845[+00 0.2 77f4 /8E-05 0.111/060+02 0.697268E+03 NUCt. 1 0.536721E-11 H.170444E-03 0.11093/E-06 0.386439[+00 0.2908270+02 PDRT ICLE CONC. = 1.83638E+02 (1/CM**3) AV. RADIUS 2 0.3279 (MICRONS) AVERACE DENSITY = 4.77 (C/CM**3) LEAA RATE = 1. 30524E-06 (G/SEC) CONT. TEMP. - 130.0 (DEC C) SATURATION RATIO = 0.96591 ACC.AfR. SOURCE = 1.11139E+06 (G) l S I
- i
- l l,m l 0
#e S
S) A3 M* OM
?C H/
G 000000000.O.0000000000000000000000000 ( 00O00000000000000000000000000000000
/4?198655f433221 111 00000001122346802 S ?2?21 1 1 11I1 11111111 11 1 1 11 11 11 111 122 S - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
A) EtLEtEE0ELEEE0EE00EEELf0EEEEELELELE M3 4ta1 6/T97lI9836929386469816961385455
?7796619ta5?532685878?8a 0664569664564 O'
I M 6619588146?1786494196535521 30496800016816021115033935290688 r464633564 541 LC 31 19371173151 41 24711 2221 1 7392216951 O/ . . . . . . . . . . . . . . .
~ SC( 00000000000000000000000000000000000 208654321 00011122222222111 001 735192 T 11 000000000000000000000000000000001 S - - - - - - - - - + + + + + + + + + + + + + ++++++ - - - - - - -
I E000LtLELLEEEEEEEE0EELEEEEEEE0EEEEE D 95930?246l622246710591769O1 43?742886
- ?8903902l2'4771062981 0492/2700059902 RC NC
.] 52111 8474 1331 011?2966532619'439412212312 51818581551231592 s48508331 878s506 691741 007030
/ . . . 4 1. 1. 5 2.I 3 9 2 4 1 .1. 1. 12221194161 1
( 000000000000000n0000000000000000000 T F 's 2198655433221 111 00000001122346802 U ?27211111 1 111111 111 1 1 1111 11 11 11 1 122 B) - - - - - - - - - - - - - - - . - - - - - - - - - - - - - - - - - - - I3 R E.E0EEEEEEEEEEEEEEE000EEELEE0EEE00EE 416//971I9836929386469816967085455 T" 8 21796639452532685878?4066456966t4564 SM 66F958814621788: 4 94196535521 46463356 IC 30496800076876021115033935290684541 D/ 311 93111731 4 SC . . . . . . . . 5 1.r. 1 2 4 7 1. 1. 2 2 2 1. 1. 7 3 9 2 2 1 6 9 5 1 S( 00000000000000000000000000000000000 A M
. 5555555544r44la444444ta444444444443333 D 0000000000O0O000000O000000000000000 A - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
I R 0E0EE0LEEEEEEEE0EEEEEEEEEEEEEEEEEEE
- 1 07830?2353006681 832611822191 059315
. 866594177930111 496025308415181 8001 4 T 18464035781506557200301550070215681 R) 83952087013467913692593728408650124 AM 45561889111111122223334455677891111 PC . . . . . . . . . . . .
1 ( 00000000000000000000000000000000000 123456189012l456/8901 23456789012345 K 33' 333333444ti 48e42;445555555555666666 V Tlllll1 (1. .Ili llJl llll ll llI
-w_
n
- ~ .
a Cont. 1 c
)
i -
~
S i /
/ /\
I _ ! !: mo e-i m :
- l. m, v
i c r - a. co n- >. o :: \ n -
~
'o 10' ' ' '10' '10'
''10
TIME CS) a utflF1 AIRBORNE AND ACC. LEAKED MASSES
l o , , 8 c; cont. N 8 N m E o 58 E: w E EN Ed-S / o. W d-
/ f 8
d 10' ' ' '10' ' '
' ' ' ' ' ' 10' ' ' '10' ' '
'10' T It1E C S )
1 ulRF1 . AVERAGE PARTICLE RADIUS i
[ - -
; o
~
l
~
6m L cont.
~
f f
~
h j a 3h 66 m
~
2h
~
I B's / l i 5i [ Y I .~ ( i i E 5-i G- ' U~ l !\ a ". 7 l )- -
%~
w :
/
~
S < j f 69 h
. I I l
. I
~
S 10 I $ 4 85)$$ 10-5 3 3 466365 10' i 3 4 $ $ t5510' STOKES RADIUS rfg iu,LulRF1 PARTICLE SIZE DISTRIBUTIONS
a -- 58 -- NAUA - MOD 4 INPUT DATA LISTING TEXT : N AUA S AM P LE C A S E # 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ANNULAR GAP, AEROSOL SOURCE = LEAK OF CONTAINMENT, NO STEAM, GRAVK CONTROL VOLUME PARAMETERS VOL : 0.14675E+11 FSED : 0.25640E+08 FDIFF : 0.23624E+09 AEROSOL PROCESS PARAMETERS FORM : 0.10000E+01 FORMC : 0.10000E+01 FORMKO : 0.10000E+01 DELD : 0.10000E-01 NUMERICAL PARAMETERS RMIN : 0.25000E-06 RMAX : 0.50000E-02 KMAX : 101 EPS : 0.10000E+00 CUT 0FF : 0.10000E-14 PROGRAM CONTROL PARAMETERS SZEIT : 0.0 TIME : 100.00 CPUZT : 120.00 RESTRT : F ZWCOMP : T FOLGE : F GRAVK : T LEAK : T LKOND : F TIMEK1 : 0.0 TIMEK2 : 0.0 TIMEK3 : 0.0 TIMEK4 : 0.0 OUTPUT CONTROL PARAMETERS NWRITE : 10
- 100 NPLOT : 5 B
M
--. - ...... _ _ ,, _ ~ ., _ . -...c _ _ . _ _ _--
e AEROSOL SOURCE FUNCTIONS CONTAINING 1 RELEASES AND 1 NUCLlDES UP TO O.0 SEC MODE 1 0.0 (C) RG 1.00000E+00 SIGL 1.30000E+00 RHO 1.000 CONTAINING THE FOLLOWING NUCLIDE FRACT10NS NUCL 1 0.0 MODE 2 0.0 (G) RG 1.00000E+00 SIGL 1.00000E+00 RHO 1.000 CONTAINING THE FOLLOWING NUCLIDE FRACil0NS NUCL 1 0.0 CONTAINMENTTEMPERATURE IN TIMESTEPS Of I SEC UP 10 0 SEC 130.00 STEAMFLOW INTO THE CONTAINMENT AT 1 GRID POINTS TINE (SEC) RATE (C/SEC) 0.0 0.0 ! LEAkRATE OUT OF THE CONTAINMENT AT 1 GRID POINTS l l 11ME (SEC) RATE (1/SEC) 0.0...........0.50000.E..06.................................................................................................. SUBSEQUFNT COMPARTMENT PRECEDING VOLUPE 7.200001E+10 (CM**3). DILUTION FACTOR 4.906 Wi1H THE 1 TOLLOWING NUCLIDES NU.C.L..1......................................................................................................................
. l I
m x- _ _____-m
- a n TIME STEP # 100 20.0000 SEC PROBLEM ilME 2000.00 SEC = 33.3 MIN = 0.6 HRS SPECIES MASS CONCENIRATIONS ACC. SEDI. DEPOSIT ACC. Diff. DEPOSIT AIRBOPNE MASSES ACC. LEAKED MASSES (G/CH**3) {C/CMa*2) (C/CMa*2) (C) (G) COND. WALER 0.0 0.0 0.0 0.0 DRY PARTICLES 0.308400E-08 0.220733E-07 0.177111E-09 4 0.8525770+02 4 0. 18: 0198E-01 4 TolAL 0.308800E-08 4 0.220733E-07 0.177111E-09 4 0. 84 525 7 7 E + 0? 0.180191E-01 4 4 NUCL 1 0.154180E-09 0.110351E-08 4 0.885616E-11 0.226259E+01 0.700898E-03 PARTICLE CONC. = 1.05776E+0re (1/CMa*3) AV. RADIUS = 0.1572 (MICRONS) AVERAGE DENSITY = 8.00 (G/CM**3)
- LEAK RATE = 2.26127E-05 4 (C/SEC) CONT. TEMP. = 130.0 (DEG C) SATURAil0N RATIO = 1.00100 ACC.AER. SOURCE = 0.0 (C)
I S I
_n.,- - , . - -,,n n. n . .-_ n - _ _
,- _. . s. - -
K PART.-RAD. MASSD I S T R i flH T . NR.-DlST SOLIO MASS 020-MASS (CM) (C/CH**3) (1/CC) (G/CN'*3) (G/CM**3) 14 0.90589E-06 0.1?992E-?9 0.52153L-13 0.1299?[-29 0.0 15 0.10002E-05. 0.34fs34E-28 0.102/0E-11 0.Tf434E-28 0.0 16 0.110t43E-05 0.80318E-2T 0.178110-10 0.80378E-21 0.0 17 0.12193E-05 0.1658sta E-25 0.27237L-09 0.16584[-25 4 0.0 18 0.13fs62E-05 0.30056E-24 0.36I64f-08 0.30056E-?4 0.0 19 0.1486taE-05 0.482330-23 0.43833E-07 0.48233E-23 0.0 20 0.16411E-05 0.68388E-22 0.r:6177E-06 0.68388E-22 0.0 21 0.18120E-05 0.85718E-21 0. f4 3001 E-05 0.85118E-21 0.0 22 0.20006E-05 0.94977L-20 0.35399E-04 0.94977E-20 0.0 23 0.22089E-05 0.93047E-19 0.25766E-03 0.93047E-19 0.0 24 0.2f388E-05 4 0.80587E-18 0.16579E-02 0.80587E-18 0.0 25 0.26927E-05 0.61699E-17 0.98s306[-02 0.61699E-17 0.0 26 0.29130E-05 0.41745E-16 0.f7t407E-01 4 0.41745E-16 0.0 27 0.32826E-05 0.2t 4961E-15 0.21060E+00 0.24961[-15 0.0 28 0.3623:3E-05 0.13189E- 18e 0.826T3E+00 0.13189E-14 0.0 29 0. 40016E-05 0.61529E-1a4 0.28658E+01 0. 615? 9 E- 184 0.0 30 0.44182E-05 0.25356E-13 0.87741E+01 0.25356E-13 0.0 31 0.48781[-05 0.9??95E-13 0.237?8E+02 0.9?295E-13 0.0 32 . 0.53860E-05 0.?96580-12 0.56649E+02 0.29658[-12 0.0 33 0. 59846 7 E-05 0.8'4211E-12 0.11951E+03 0.84211E-12 0.0 34 0.65658E-05 0.21135E-11 0.22284E+03 0.21 35E-11 0.0 35 0.72493E-05 0.46998E-11 0.36816E+03 0.a6998E-11 a 0.0 36 0.80040E-05 0. 9 301ta E- 11 0.54134E+03 0.93014E-11 0.0 37 0.88372E-05 0.16508E-10 0.71383E+03 0.16508E-10 0.0 38 0.91572E-05 0.26556E-10 0.85319E+03 0.26556E-10 0.0 l 19 0.10T73E-04 0.3922SE-10 0.93628E+03 0.3922SE-10 0.0 'g t0 4 0.11895E-04 0.53888E-10 0.95564E+03 0.53888E-10 0.0 41 0.13133E-04 0.69680E-10 0.91806E+03 0.69680E-10 0.0 l
*42 0.14500E-04 0.85747E-10 0.83936E+03 0.85747E-10 0.0 43 0.16010E-04 0.10152E-09 0.13833L+03 0.10152L-09 0.0 44 0.17676E-04 0.11691E-09 0.63172E+03 0.11691E-09 0.0
#45 0.19516E-Of4 0.13?33E-09 0.531?St+03 0.13?33E-09 0.0 as6 0.21548E-04 0.148520-09 0.48299E+03 4 0.14852E-09 0.0 47 0.23791E-Ol4 0.166?3E-09 0.36837E+03 0.166?3E-09 0.0
#48 0.26268E-Oi4 0.18582L-09 0.30596L+03 0.18582E-09 0.0
, 49 0.29003E-04 0.20107E-09 0.25331E+03 0.20701E-09 0.0 50 0.32022E-04 0.22875E-09 0.20790E+03 0.2?875[-09 0.0
- 51. 0.35356E-04 0.28817E-09 4 0.16757E+03 0.248110-09 0.0 52 0.39037E-04 0.26047E-09 0.13067E+03 0.26081L-094 0.0 53 0.43101E-04 0.25925t-09 0.96628E+02 0.259?5E-09 0.0 5 84 0.f T588E-Ots 4 - 0.23892E-09 0.66163E+0? 0.23892E-09 0.0
$$ 0.525t2E-04 0.20091E-09 0.da1336E+02 0.20091E-09 0.0 56 0.580120-084 0.18s8180-09 0.721843E+02 0.14818E-09 0.0
$1 0.61051E-04 6 0.61/04L-10 4 0.73481L+01 0.64708L-10 4 0.0 58 0.70719E-04' O.2790fE-10 4 0.23545E+01 0.279040-10 0.0 59 0.78081E-04 0.13636E-10 0.85488E+00 0.136360-10 0.0 60 0.86210E-04 0.82569E-11 0.38158E+00 4 0.82569E-11 0.0 61 0.95185E-04 0.62/60E-11 0.21118L+00 0.62/60E-11 0.0 62 0.10509E-03 0.45055E-11 0.11588E+00 4 0.8s5055E-11 0.0 63 0.11603E-03 0.28581[-11 0.54595E-01 0.28581E-11 0.0 64 0.12811E-03 0.16700E-11 0.23700E-01 0.16100E-11 0.0 u a
. I a 65 0.14145E-03 0.88277E-12 0.93079E-02 0.88277E-12 0.0 . 66 0.15618E-03 0.83990E-12 4 0. 3 tal:61 E-02 0.43990E-12 0.0
- 61 0.17214E-03 4 0.16052E-12 0 . 9 38:28E-03 0.16052E-12 0.0 68 0.19039E-03 0.50983E-13 0.22046[-03 0.50983E-13 0.0 69 0.21021E-03 0.16419E-13 0.52/5IE-Ot4 0.16319t-13 0.0 70 0.23209E-03 0.41529E-its 0.991280-05 0.415290-Us 0.0 71 0.25626E-03 0.871580-15 0.15563f-05 0.87758E-15 0.0 72 0.28293E-03 0.15210E-15 0.20010E-06 4 0.15210E-15 0.0 73 0.31?39E-03 0.18907E-16 0.18509E-07 0.1890/t-16 0.0 Tre 0.31t91E-03 44 0.22913E-17 0.16665E-08 0.22913E-17 0.0 75 0.38082E-03 0.33407E-18 0.18052E-09 0.33407E-18 0.0 76 0.42046E-03 0.30812E-19 0.12370E-10 0.30812E-19 0.0 77 0.16423E-03 4 0.218 09E-20 4 0.638580-12 0.21809E-20 4 0.0 78 0.51256E-03 0.1280$E-21 0.28378E-13 0.12805E-21 0.0
=
I 8 l
cc.......................................................................................................................... . T IME STEP # 4009 9.2335 SIC PROBLIM llME 16435.65 SEC = 273.9 NIN = 4.6 HRS SPECIES MASS CONCENTRAil0NS ACC. SEDI. DEPOSIT ACC. DIFE. DIPOS11 AIRB0HNE MASSES ACC. LEAKED MASSES (G/CM**3) (G/CM**2) (G/CH**2) (G) (G) COND. WATER 0.0 0.0 0.0 0.0 DRY PARilCLES 0.153803E-07 0.909837E-05 0.930041E-08 0.225705E+03 0.141005E+01 TOTAL 0.153803E-07 0.909837E-05 0.930041E-OS 0.225105E+03 0.147005E+01 NUCL 1 0.655249E-09 0.435647E-06 0.443218E-09 0.961578E+01 0.699808E-01 PARTICLE CONC. = 2.01572E+04 (1/CM**3) AV. RADIUS = 0.2058 (MICRONS) AVERACE DENSITY = 6.04 (G/CM**3) LEAK RATE = 1.12926E-04 (G/SEC) CONT. TEMP. = 130.0 (DEC C) SATURATION RATIO = 1.00100 ACC.AER. SOURCE = 0.0 (G) I 9 l y --. - - . -
- E a __ e T IME STEP # 7248 431.7363 SEC PROOLEM TIME 360025.50 SEC = 6000.4 MIN = 100.0 ilR$
- SPEC 1ES MASS CONCENTRATIONS ACC. SEDI. DEPOSIT ACC. DiIf. OfPOSII AlHBORNE MASSES ACC. LEAKED MASSES (C/CM**3) (C/CMa*2) (C/CM**2) (G) (G)
CONO. WAIER 0.0 0.0 0.0 0.0 DRY PARTICLES 0.723041E-10 0.264993E-04 0. 409116 E-0 7 0.1061060+01 0.591103E+01 TOTAL 0.723041E-10 0.264993E-04 0.409116E-0/ 0.106106t+01 0.591703E+01 NUCL 1 0.255599E-11 0.110626E-05 0.162141E-08 0.315092E-01 0.238259E+00 PARTICLE CONC. = 5,45158E*02 (1/CM**3) AV, RADIUS = 0.1662 (MICRONS) AVLHAGE DENSIIY = 4.81 (G/CM**3) . LEAK RATE = 5.3286 TE-07 (G/SEC) CONT. 1 EMP. = 130.0 (DEG C) SATURATION RATIO = 1.00100 ACC.AER. SOURCE = 0.0 (C) I 2 I
- - ^ -^
k >
. . i l
D D e e i e e , e a e e e e e a e e e e e a e e e a e e se eZa e<m sZe e es eCZ I NG 80% e O O. O. O. C. C. C. O. O. C. O. C. O. C. O. C. C. O. C. O. C. C. O. O. C. C. C. C. O. C. C. C. O. O. C. C. C. O. O. O. C. C. O. I e e
- OOOOOCOOOOOCOOOOOOOCCOOCOCOOCCCCOCCCOCOCCC 4
a e e e e OccN-Oc%cc :mmNNe,----e-e--e--~~NNNmmasccON g e NNNNN-----------------------------------NN
, eM e<a eeeeeeeeeeeeeeeeeaeeeeeeeeeeeeeeeeeeeeeeee ww-wwWwwwwwwwwww-wwWWwwwwWWWW6wwwWwwWwwWww l eEm %-mN?%mNmm%%:3c%cem=c N#%Nace-OCm%cNeoeCem e s r%N%cecN%cm7:N3O?---C:03%NenNeNNChr. Omcoc-l eCe -m-O%:nmN::OOOwm*Omerc%m: Omen-c3DO-OOm@ con
. e-z cese%->N=<=oNmNcNer%-c%=maC+--mec-3c~=>ce~ j e-C %. .-- N . N N. m. a- #- m A @ m O r. c o - N m e c % C c A c J m e, - C J N N - m m - O c m , eCN . . . - . . . . . . . . . . . . . . . . . . . . . . . . . . . . . , IMO e - OCOOOCCOCOOOCOOOOCOCCCCOOOCCOCCCCCOOOOOCCC. e f e s e e ew m-?%#3mN-OC---NNNNNNNNNNNNN---COC--NmachoN f eM
--OCOOOOOOOOCOOOOCOOOOOOOCCCCCOOeOCOOCCCC-e e e e e e e e e++++++++++++++++++++++++ e e e e e e e e e i e- -WwwwWww. W wwwwwwwWwwwwwWWwWwWwWwWwwW-wWW eC 2OmO*ODN: r%mc-OMAcmNewb%NOsm7c%-Comcockcr e
- e. a &%NcO--0-ccmocA&cOcOcccc%m% mNcO%e%: -NcNOc e
eTC NCN%c-cmenkcNcmemr> %cc::mACO:N?kccmcmm@~N
- OemONOO--N#->cecc O-c--CONN %:ceNmcmc*m-c ezu e N m c. c. m. -. N. m .N. N. . :. . e. c. . N. m a.
. . c. c. c. c. c. a. m. N. . c. m
. . o. m. . c. . N. m. N. . N. N. N.
e e - OOOCOOOOOCCCCCCOOCCOOCOOOOCOOCCCOOCCOCOCOC e e e e - e* Sc r.-@CN@OrgmmNNN---------e----NNNmm3ACCON eO GNNNN--------w--------------ew---------rNN eOm e-m e e e a e e e e e e e e e e e e WwwW-wwWWwwwWwwwe e e e e e e e e e e e e e e e e e a e e e e ea e wWWWWWWWWWWWWWWwWwwWwbWww ace N-mNONmNemNN :c%comcc3Neo%:ca-AOM%CNommCmm e~e ONN%cacN%cmoJA&C&---Oc&&NNONN-NNONNeomeo@- eME -m*ON NmN: CCa@mmmmaccNMa&c-3-CMcC-oem& CON e-O EON emeCN-SNOcCCNmNONAAN-ON3mCC@w-Mec"#CN ?ccN eme N. -. N. . %. N. m
. e. c. e r,
. . e. m. p. .r. a. m. e N. m c. c. N o. c. N c. a. m N. . c. c. N. N
. . . . . . .-.m . .m - 0. c. m em-e< OCOOOCCOOOCCCOOcCOOOOOCCCOOOCCCOCCOCOCCOOC.
I E e e a e
- Pbffffffbffbbfbt22323333322333%33::2333mmm eC OOOCCCCOOOCCCCCOCOCCOCCCCOOOCCCCCCOCOCCCCC e< seeeeeeeeeeeeeeeeeeeeeeeaeeeeeeeeeeeeeeeee eE WwWwWWwwWWW6WwWWWWWWW6WWWWWWWWWWWWWWwWWWWW ee &cNCemcN-ONomCNNmemCOcco-omNcN-cNN-o-Comme e
- ccNmN4-cccca& NNNOMO-N-J@cCNamOcc-e-C-CCO*
eN CmokDN0-Ncac20mbM@wcCchbNNOCmC-PDCCkON-DDD e=~ NacmNcO3cmocNoc%CemacN@-memNewmNNccCecrC~N 8 <K sku N.N.N.N.m m.c.a.a. c. c.c.N.e.c.m. . . . . . . .N.N.N.N.m m.m.4 3.O.c.c.N N
. . . .m.@. . . .
e - CCCCCCCCCOCOCCCCCCCCCOOCCOOCCCCCOOCOOOOOCO e e e e e M3#cNCSC-Nm:@cNcSC-Nm3OcNC&OeNM3#cNC@C-NM3 e2 e NNNNNNNMmmmmmmmmm34423:Ma#1ceccAOccecc@cce 9
_u-b"
- ann. gap
~
,, b t.o -
! / \
c - FE*o~
~
O : / e- :
'o 7~
o / m: a 10' ' ' '"10'
10' '"10' ' ' ' ' ' 10' TIME CS)
ILRF1 AIRBORNE AND ACC. LEAKED MASSES r I l
1 l 1 - t I h 8 ~ ! a ann. gap R . S i N u, 5 58 y: v v) D ER Ed-a d-W d-
' ~N 8
g ,,
, . . . . . . I 0' ' '10' '10' TIME CS)
O, LRF1 AVERAGE PARTICLE RADIUS 9
l a -u- ' -
~
9
.- ann. gap
- l
~
S l2h g ~_ a : ~ x . G - v 3.6 h U- 2.6 h db 43\ a ~_ 33 m su M 69 h ,
/T \ g
- [
I 10 i 3 4 6 4f 66 10" i i 4 6 6 f 66 1o* $ i ; g 5355 10' STOKES RADIUS LRF)
- PARTICLE SIZE DISTRIBUTIONS
- i l
??' .
/
.. ;- p:
v-n Sandia Nati0nal Lab 0 rat 0 ries Albuquerque, New Menico 87185 April 2, 1985 I Mr. Robert Minogue, Director Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Commission Washington, DC 20555
Dear Mr. Minogue:
Enclosed is the December 1984 - January 1985 Bi-Monthly
; port of work managed by the Sandia Rea,ctor Safety Research Department. This report covers: .
Fin No. Accident Energetics A1016/A1385/A1390 Molten Fuel-Concrete Interactions A1019 Molten Core-Coolant Interactions A1030 Core Debris Behavior All81 Debris Bed Coolability - EURATOM A1263 Debris Bed Coolability - PNC A1264 Core Melt Technology A1218 High Temperature Fission Product Chemistry A1227 Hydrogen Behavior Program A1246 LWR Debris Formation and Relocation A1335 Hydrogen Mitigative and Preventive A1336 Schemes LWR Core Debris Coolability A1340/Dll24 Melt Progression Analysis (MELPROG) A1342 Quantitative Uncertainty Estimation for the Source Term A1383 Sincerely yours, t /d V. Walker, Manager Ja actor Safety Research Department 6420 Enct As stated ,' g yy 514 J sh-) (: j' '
/0
x --
. . . . . . . . . . . . . . . . . . . m .-
=
Mr.RobhrtMinogue April 2,;1985 Copy to (w/ enc) R. W. Barber, USDOE J. A. Martin, Jr., NRC/RES R. Minogue, NRC/RES P. M. Williams, NRC/NRR D. F. Ross, NRC/RES J. Reed, NRC/NRR O. E. Bassett, NRC/RES J. L. Telford, NRC/RES R. Bernero, NRC/NRR M.'Cunningham, NRC/RES C. N. Kelber, NRC/RES F. P. Gillespie, NRC/RES W. M. Morrison, NRC/RES V. Benaroya, NRC/CHEB G. P. Marino, NRC/RES W. R. Butler, NRC/CSB M. Silberberg, NRC/RES R. L. Palla, NRC/CSB R. T. Curtis, NRC/RES K. I. Parczewski, NRC/CHEB R. W. Wright, NRC/RES J. Austin, NRC/DCM T. J. Walker, NRC/RES G. Quittschreiber, NRC/ACRS P. M. Wood, NRC/RES R. Savio, NRC/ACRS S. B. Burson, NRC/RES 'Z. R. Rosztoczy, NRC/NRR J. T. Han, NRC/RES L. C. Shao, NRC/D/DET {1.{K . ~ Chan ,*:llRC/RES/ K. G. Steyer, NRC/CEBR J. T. Larkins, NRC/RES T. M. Su, NRC/GIB T. M. Lee, NRC/RES R.,Vollmer, NRC/D/DE C. Allen, NRC/NRR P. Baybutt, BCL F. Rowsome, NRC/NRR R. Denning, BCL J. F. Meyer, NRC/DCM W. Kato, BNL C. Austin, NRC/OCH M. Stevenson, LANL R. O. Meyer, NRC/ASTPO T. Kress, ORNL J. A. Mitchell, NRC/ASTPO M. L. Corradini, University C. Ryder, NRC/ASTPO of Wisconsin W. C. Lyon, NRC/NRR J. H. S. Lee, McGill L. G. Hulman, NRC/NRR University R. Van Hooten, NRC/RES T. G. Theofanus, Purdue P. Worthington, NRC/RES University C. Tinkler, NRC/RES R. C. Vogel, EPRI C. Nilsen, NRC/RES L. Baker, Jr., ANL T. P. Speis, NRC/NRR R. Anderson, ANL l J. E. Rosenthal, NRC/NRR E. Warman, Stone & Webster ' G. R. Burdick, NRC/RES M. Fontana, TEC M. Jankowski, IAEA; Division of Nuclear Reactor Safety, Wagranerstrasse 5, P. O. Box 100, A/1400 Vienna, Austria . 4
.\, *
.i
%l 1 ' Technical Highlights / Administrative Report l l' for the Nuclear Regulatory Commission (NRC) l Reactor Safety Research Program ,
- December 1984-January 1985 l
1 l 4 I I \ , l Printed March 1985 t I i \ i hSandia National Laboratori , .\. 4
and another eight experiments with 6.5 v/o hydrogen in air.
. These experiments include control burns without water drops. ..
1.6.1.4 The Ef fects of Aerosols on flydrogen Combustion (L. S. Nelson, 6427; G. D. Valdez, 6427)
' We began new experiments to investigate further the effects of a reduced aerosol on a lean hydrogen burn. This was prompted by the single experiment performed last October )'
with a metallic iron aerosol in 6.5 v/o hydrogen in air; in that experiment, we measured an approximately 50 percent higher peak pressure with a faster rise time than the control experiments on those burns with oxidic aerosols present. In the new experiments, we improved our powder disperser to obtain denser aerosols and switched to a f.iner iron powder than used in the single experiment in 1984. We completed three control burns at 6.5 v/o hydrogen in air without the aerosol, and attempted one with aerosol. Ilow-ever, our sequencing apparatus malfunctioned, yielding a burn under unknown condition's. 1.6.1.5 Consequences of flydrogen Combustion in the Presence of Aerosols - (L. S. Nelson, 6427; G. D. Valdez, 6427) We 16.6, performed burns in the VGES chamber with the compositions formed 9.0, ar.d 29.6 v/o hydrogen in air; each burn was per-with 1 kg of 10 w/o CsI-90 w/o dispersed throughout the combustion volume. A1203 As in the three tests performed in November at 11.0, 16.7, and 6.5 v/o hydrogen in air with the same aerosol disposal, free molecular iodine was generated. Although the chemical analyses are incom-Plete, initial measurements indicate that as much as 75 per-cent of the iodide ion present in the Cs1 was oxidized to molecular iodine during the burn at 29.6 v/o (approximately the stoichiometric composition). 1.6.1.6 Nonpowered liydrogen Igniters (L. R. Thorne and J. V. Volpani, 8353) A platinum igniter has been prepared for evaluation in sev-eral static dry hydrogen-air mixtures in a 5 m3 combustion chamber. 1.6.2 Documentation, Meetings, and Presentations s on December 5, 1984, presentations were made in Albuquerque to C. W. Nilsen and P. Worthington, NRC contract managers e [, .. r .. I
y y. __ . . . _ _ -... . l,.~ w h e
!. i '
4 ,
,_s,
. . : < ,,, r '
. .:. .. ~
_ .- . 1
., 'T '
I,
- 1 INFORMAL REPORT t 1
on ; SOURCE TERM PRE 0!CTIONS '0R VARIOUS CONTAINMENT FAILURE ASCUMPTIONS to I U.S. NUCLEAR REGULATORY COMMISSION August 29, 1984 by J. A. Gieseke, H. Chen, P. Cybulskis, R. Freeman-Kelly, M. R. Kuhlman, and K. W. Lee O BATTELLE Columbus Laboratories ( - e 505 King Avenue Columbus, Ohio 43201
~
($0 ,f[ IN>i: .. . .
. H
]
+
)
( . TABLE OF CONTENTS I Zion -- Large. Dry PWR ....................... 3 S u rry -- S ub atmo s p h e ri c PWR . . . . . . . . . . . . . . . . . . . . . 12 Sequoyah -- Ice Condenser PWR . . . . . . . . . . . . . . . . . . . . 28 Grand Gulf -- Mark III BWR ..................... 32 Peach Bottom -- Mark I BWR ............. ........ 49 LIST OF TABLES Table 1. Total Leak Area Estimated as a Function of Containment Pressure ......................... 4 4 Table 2. Accident Event Times. Zion ................ Table 3. Locational Distribution of Fission Products for Various Containment Failure Modes of the Zion Plant (TMLB') . . . . 13 Table 4. Surry Leakage Model . . . . . . . . . . . . . . . . . . . . 14 Table 5. Acci dent Event Times , Surry . . . . . . . . . . . . . . . . 14 Table 6. Locational Distribution of Fission Products for Various Containment Failure Modes of the Surry Plant (TMLB') ... 27 Table 7. Accident Event Times, Sequoyah .............. 29 Table 8. Locational Distribution of Fission Products for Various Containment Failure Modes of the Sequoyah Plant (TMLB') . . 39 Table 9. Accident Event Times, Grand Gulf ............. 41 Table 10. Locational Distribution of Fission Products for Various Containment Failure Modes of the Grand Gulf Plant (S 2 E) . . 48 Table 11. Accident Event Times, Peach Bottom ............ 50 Table 12. Locational Distribution of Species of the Peach Bottom Plant . . . . . . . . . . . . . . . ............ 56 0
- . . .s
( ' LIST OF FIGURES Page Figure 1. Containment Pressure as a Function of Time (No Leak). . . 5 Figure 2. Containment Pmssure as a Function of Time (Medium 6
._ Leak) . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 3. Containment Pressure as a Function of Time (High Leak) . 7 Figure 4. Containment Pressure as a Function of Time (Isolation 8 Failure) ........................ Figum 5. Total Volume of Leaked Gases (Medium Leak) ....... 9 Figure 6. Total Volume of Leaked Gases (High Leak) ........ 10 Figure 7. Total Volume of Leaked Gases (Isolation Failure) .... 11
~
Figure 8. Containment Pressure Response for Surry TMLB'-6 Sequence ........................ 15 Figure 9. Containment Temperature Response for Surry TMLB'-6 Sequence ........................ 16 Figure 10. Total Volume of Gases Leaked for Surry TMLB-6 Sequence ........................ 17 Figure 11. Containment Pressure Response for Surry TM.B with High Leakage . . . . . . . . . . . . . . . . . . . . . . . . . 19 Figure 12. Containment Temperature Response for Surry TMLB with High Leakage ...................... 20 Figure 13. Containment Leak Rate for Surry TMLB with High Leakage . . . . . . . . . . . . . . . . . . . . . . . . . 21 . Figure 14. Total Gas Leakage for Surry TMLB with High Leakage ... 22 Figure 15. Containment Pressum Response for Surry TMLB with I sol ati on Fail ure . . . . . . . . .' . . . . . . . . . . . 23 Figure 16. Containment Temperature Response for Surry TMLB with I s ol atun Fail u re . . . . . . . ., . . . . . . . . . . . . 24 Figure 17. Containment Leak Rate for Surry TMLB with Isolation F a i l u re . . . . . . . . . . . . . . . . . . . . . . . . . 25 Figure 18. Total Leakage for Surry TMLB with Isolation Failure . . . 26 Figure 19. Containment Pressure for Sequoyah TMLB'-y Sequence ... 30
, 1
.r *+
e O LIST OF FIGURES { (Continued) Page Figure 20. Containment Temperature Response for Sequoyah TMLB'-y Sequence . . . . . ................... 31 Figure 21. Containment Pressure Response for Sequoyah TMLB'-6 Sequence . . . . . ................... 33
'~
Figum 22. Containment Temperature Response for Sequoyah TMLB'-6
, Sequence . . . . . . .................. 34 Figure 23. Containment Pressure Response for Sequoyah TMLB with I s ol ati on Fail ure . . . . . . . . . . . . . . . . . . . . 35 Figure 24. Containment Temperature Response for Sequoyah TMLB with I s ol ati on Fail u re . . . . . . . . . . . . . . . . . . . . 36 T- Figure 25. Containment Leak Rste for Sequoyah TMLB with Isolation
,. F a i l u re . . . . . . . . . . . . . . . . . . . . . . . . . 37 Figure 26. Total Leakage for Sequoyah TMLB with Isolation
, Failu m . . . . . . . . . . . . . . . . . . . . . . . . . 38 Figure 27. Containment Pressure with Nominal Pool Bypass for Grand C, Gulf . . . . . . . . .................. 42 Figure 28. Containment Temperature with Nominal Pool Bypass for Grand Gulf . . . . . .................. 43 Figure 29. Containment Pressure Response for Grand Gulf S2 E with I s ol a t i on Fai l u re . . . . . . . . . . . . . ...... 44 Figure 30. Containment Temperatum Response for Grand Gulf S 2E with Isolation Failum ................. 45 Figum 31. Containment Leak Rate for Grand Gulf 2S E with Isolation Fa i l ure . . . . . . . . . . . . . . .......... 46 Figure 32. Total Leakage for Grand Gulf S 2 E with Isolation
, Fail ure . . . . . . . . . . .............. 47 Figure 33. Pressures in Containment Volumes ............ 51
~
Figure 34. Gas Temperatures in Containment Volumes . . . . . . . . . 52 Figure 35. Reactor Building Pressure During TC-y' Sequence . . . . . 53 Figure 36. Containment Pressure Response for Peach Bottom TC i Sequence with Isol ation Failure . . . . . . . . . . . . . 55 R
~ , 9 SOURCE TERM PREDICTIONS FOR VARIOUS C.? CONTAINMENT FAILURE ASSUMPTIONS by J. A. Gieseke, H. Chen, P. Cybulskis, R. Freeman-Kelly, M. R. Kuhlman, and K. W. Lee August 29, 1984 The fission product release analyses presented in BMI-2104
,_ were based on a variation of the so-called threshold containment failure model. In the threshold failure model, loss of containment function is , assumed to take place when the containment loading reaches a preselected level; at the preselected failure load the containment is assumed to fail, frequently by assuming a large opening or even complete loss of the containment function. Since it was recognized that the loadings that may lead to containment failure are not well defined, the BMI-2104 analyses also examined the effect of containment failure for cases where predicted loads o'nly approached the given failure levels without exceeding C them.
In the analyses for the Zion design, a containment failure pressure of 149 psia was assumed based on earlier analyses. Since the predicted containment pressures for the Zion accident sequences considered were well Delow this failure level, it was not felt appropriate to consi-der containment failure due to structural overloading. Thus in the analyses in BMI-2104 for the Zion design, only nominal containment leak-ages were explicitly considered. In the analyses for the Surry PWR design, on the other hand, predicted accident containment loadings were frequently close to the assumed failure level of 100 psia, without
, actually reaching this level; in these instances the analyses were per-formed by both assuming that the containment fails and assuming that the containment does not fail. Because of the importance to source terms of containment failure times and leak rates, it was believed important to investigate these effects more extensively and in a parametric fashion.
The Containment Performance Working Group (CPWG) was formed by the Nuclear Regulatory Commission to address questions related to the
~ .
)
2 ability of reactor containments to withstand the loads that may be imposed during severe reactor accidents. While the work of the CPWG is still continuing, among the early outcomes of the work of this Group is the postulate that containments may lose some of their functional capability due to the deterioration of penetration seals before the structural
. integrity of the containment is challenged. This postulate has led to the " leak before break" model in the form of an increasing leak rate as
, a function of internal pressure. In this report, the effect on predicted fission product release to the environment of this leakeefore_treaX Joodel has~.been iexploredifor'-thelioni.and.~5urry )WR designs. In addition, the effect on subsequent fission product release of the failure.of the containment.to.isolatte in the event of an accident has been evaluated for~each4 theSlant1esigns ronsidered 'in;3MI-2104 (Surry, Zion,
. Sequoyah, Peach Bottom, and Grand Gulf). It is important to distinguish between the two types of containment failures considered in these analy-ses. The containment isolation failure is present at the start of the accident due to a system fault and is not related to the accident environ-ment. The leak before break type of failure that is being postulated is C- a direct consequence of the accident environment.
While failures of containment isolation are distinctly different from failures due to accident induced pressur'e and/or temperature loads for the purposes of consequence (source term to the environment) assess-ment, the former can be a useful point of reference. Failure to isolate
, is known to have a finite and nonnegligible likelihood of occurrence and is thus of interest in itself. Also, failure to isolate can be a useful surrogate for noncatastrophic failure due to accident loads. It can be viewed in both contexts for the purposes of this study. The analyses performed for each of the plants are discussed below.
The basis for the analyses reported here was consistent for all plants. The containment isolation failure was assumed to be in the form of a preexisting hole 6 inches in diameter. For the " leak before break" cases, the equivalent hole sizes as a function of containment pressure developed by the CPWG were employed for the Zion design and similar assumptions developed by scaling for Surry. The accident l l
i
}
1 e.
?
3 sequences used for each plant are as follows: Zion-TMLB', Surry-TMLB', Sequoyah-TMLB', Peach Bottom-TC, and Grand Gulf-S2 E. Zion -- Large, Dry PWR I Table 1 presents the equivalent hole size as a function of containment pressure developed by the CPWG for the Zion design and used
, in these calculations. The containment isolation failure was assumed to take the form of a hole 6 inches in diameter. The analyses to examine the effects of the above containment failure modes were based on the TMLB sequence in the Zion PWR design. , The accident event times for the Zion TMLB' sequence from ', BMI-2104 are given in Table 2. Figure 1 gives the containment pressure ,. history without any containment degradation. Figures 2 and 3 show the containment pressures for the medium and high leaks, respectively, as . defined in Table 1. It is seen that leaks of the magnitude considered here have a perceptible, but minimal effects on the predicted containment pressure response. Figure 4 gives the containment pressure response for C the assumed failure to isolate. The effect on the containment pressure , response is quite dramatic compared to the previous cases. This is oecause the hole size assumed for the isolation failure is larger than for the leakage cases and because the isolation failure is postulated to exist at the start of the accident rather than developing in response to the accident loads. Figures 5 through 7 present the total volume of gases leaked from the containment for each of the foregoing cases; the differences between the isolation failure case and the leak before break case is again quite obvious.
The above containment responses were used as the basis for [ evaluating fission product releases to the environment. For the fission product sources in the containment, the release from the primary system as given in Tables 7.3 and Figures 7.5 through 7.8 of BMI-2104, Volume VI, were used since the response of the primary system was essentially unaffected by the changes in the containment response. The source from the core-concrete interaction was as given in Table 6.11 also from BMI-2104, Volume VI. N 4
m ( .. 4 TABLE 1. TOTAL LEAK AREA ESTIMATED AS A FUNCTION OF CONTAINMENT PRESSURE l Containment Low Medium High Pressure Leak Area Leak Aj;ea Leak Area (psia) (in.2) (in.') (in.2)
$' Normal
- Operating
, (14.7) 0.1 0.5 1.0 38 0.1 0.62 1.48
- 62 0.1 0.62 1.84 120 0.1 2.13 10.96 e 149 0.1 5.33 23.72 i
TABLE 2. ACCIDENT EVENT TIMES, ZION Time. Event minutes Zi on-TMLB '-1 Steam generator dry 82.5 Core uncover 109.8 Start melt 130.5 Start slump 158.5 Core collapse 159.8 Vessel head dry 169.2 Head fail 169.5 Cavity dry 316.4 Concrete attack 389.1 End calculation 1001.8 i g $
u . ha& .
~.A A _-a _ _ _ _ _ _ ._.2 _m. __ ____ _ _4 --
4 TOTAL COMPARTMENT PRESSURE o B e 8 3 g
- o o o o o o ,
o a.* l 8-o a . A r m
- g. -
8 o .
. 5 '
2: a o A g-5 o . E S d a. . R x 8- - mo 3 2 I 2 m
.n n of-
~
C
% A R. -
- trJ 5 l
== Vo )
M
?o g.
t= O / 5
- g. .
o e" % a
,). ) .
9
)
e e 3..
-...-au-.,+-,
.. .n.-- .. -- - -- -- -- -- - - - -
% t' 6
(; o U
~
f sa O b _g 5 k o- ,_
"h 0 $5 Z S o U
-v 2 g m
C W eM w _g 2 g
*p 0; W
n. o g
~k I E
o 8 _g N W o S
-8 ~
T l C C l
- e o o o e i
E * =
! S 8
( 3HDSS3Hd IN3N1HVdNO3 'IY10J. e
- , - , - - - , , . - , . - - - - - , - - - - - - , , - . . , - - . - - - . - - , , - - . . ---.-,,-- -,,- ---nn, - - - . . - - - - - , . e-e----- --,-,,.-,-c.----
TOTAL COMPARTMENT PRESSURE o E $ 8 8 8
)
o o o o o O t + l w
,, 8-s o sii m
8 o 5 2 l i 5 5 k-e f I ' h
= d g *8_
l m ty o 3
> l 2 ^ -
E o M. O
- C
- A -
~ tn vo I
= 4 .
b 8" o
/
m A .. 6 '. g_ . e l L i 3
9 ie EO;$A
- U 4IE*x 4 E* aNMnnD(N
. 4 2 4 6 8 1
0 K 0 0 0 0 0 0 0 0 0 0 f G
) .
U . R . E 4 C O , N T A I N M E . N T P R E . S S U R T I E A M-S E F U A
- n.
N ( , C T I M* O I N O N . f U T T . I M ) E* ,. E ( I S O , L A T
- I O
N . F A I L U . R E
)
7 ~ o m
.' 3 4
1
Sv
/ l 9
o e
~
O C y i- - F 5, E M Ce o w
- E o e e =
2 v 0 I 5 a C o Ci! o
-82 y
*p S 9
e # _ y 1 a : o e I
< 8 l
-$ = t 1
C O N O (. F 03}IY31 sassyD do annloA gyloJ, _ __. . . _ _ _ _ _ _ _ _ , _ . . _ . _ _ _ _ _ _ _ . _ . . _ _ , _ _ _ . . . _ _ _ . . _ . - - _ . _ _ _ _ _ , ~ . - . _ , . _ . _ . _ _ . , _ _ _ _ . _ _ _ _ _ _ - . _ _ _ _ _ _ _ . _
m - _
-\ ,.
. e e'
10
- C 6 !
9
.g C
e a D
, e
=> s Z e
*2 5 s
-[ v g
( .m i 8
-b
*p >
t-o E
-l
' C2:
e 8 C _g
-8 O
C
+ + 6 6 . .
O A A e o o o e e v M N W o
,pT* G3XV3153SSVD 30 3MfrIOA 7V101
. ,, . , , , . . . . - . , . . - - . , --,,,.--.,.-,,c -- ,-- ,,---c-,---.- ~ - , . , - - - - - -
a -- _ _ - - 11 o e C., _ i l
'I
- o
.g
=
g
-g g
=
- a ,_ $
a e Z m o
- 5
_v 2 e w I $ c h3 e
- 8, -2 W
=
d o d ;
~k 5 a
o y
-k $
m l
- Q
_g o
, 6 4 6 i o
o o o o o C. O e d 4 6 N
- C
(.. pt. caxvan sassvo ao antnoA Mol 9
.- . _ - _ . - _ _ , _ . , - - . . - , , - - , . , , . - , - . . - - _ , ~ , - . - , , . , . . , , _ _ , , , , -
4 .+ 12
, i The results of the assessment of fission product releases to C.- the environment for the various cases considered are summarized in Table 3. The effect of containment failure mode assumptions is most dramatic for tellurium, where the predicted release is seen to be 22 percent for the case of containment isolation failure. This is because much of the tellurium release from the fuel is predicted to take place directly to the containment after vessel head failure. The effect of increasing leak rate on the predicted release of Csl and Cs0H is not as dramatic, though still noticeable. /"
Surry -- Subatmospheric pWR 9 The analyses for the Surry subatmospheric containment PWR were again based on the TMLB' accident sequence. The specific conditions and assumptions utilized in the analyses were very similar to those of the TMLB'-6 case treated in V'olume V of BMI-2104 in which the containment was assumed to fail as a result of the steam spike from the interaction of the core debris with water in the reactor cavity following vessel C> head failure. Two containment failure modes were explicitly considered for the Surry design in this study. The first corresponds to the high leak case as postulated by the CPWG. The leak area as a function of contain-ment pressure as shown in Table 4 was used for this case. The second case was an assumed containment isolation failure with an opening equiva-lent to 6 inches in diameter. [ Table 5 gives the accident event times for the Surry TMLB'-6 sequence as it was evaluated in Volume V of BMI-2104; as noted previously, in the latter it was assumed that containment failure was caused by the steam spike from the interaction of the core debrit with water in the reactor cavity following vessel failure. Figures 3 and 9 give the containment pressure and temperature responses for this sequence; Figure 10 shows the total volume of gases leaked as a function of time corresponding to the assumed large opening at containment failure. For the analysis of the alternate containment failure modes, the response of the primary system was found to be essentially the same
13 ' i; 1 TABLE 3. LOCATIONAL DISTRIBUTION OF FISSION PRODUCTS FOR VARIOUS CONTAINMENT FAILURES MODES OF THE ZION PLANT (TMLB') Unit: Fraction of Core Inventory RCS Containment Envirorsnent , f BMI-2104 (Normal Leak) e Cs! 0.98 2.5 x 10-2 1.9 x 10-6 Cs0H 0.98 2.5 x 10 2 1.9 x 10-6 Te 0.29 0.63 7.8 x 10-5 Medium Leak Cs1 0.98 2.5 x 10-2 7.1 x 10-5 Cs0H 0.98 2.5 x 10-2 7.2 x 10-5 l Te 0.29 0.63 5.5 x 10-3 , High Leak Cs! 0.98 2.5 x 10-2 9.0 x 10-5
- Cs0H 0.98 2.5 x 10-2 9.1 x 10-5 Te 0.29 0.62 1.6 x 10-2 Isolation Failure Cs! 0.98 1.8 x 10-2 7.0 x 10~3 1 $ Cs0H 0.98 1.8 x 10-2 7.1 x 10-3 Te 0.29 0.42 2.2 x 10"I 1
i h I
? A .- I 14 TABLE 4. SURRY LEAKAGE MODEL Pressure, Leak Area, in 2
,. psia Low Medium High 9.9 0.10 0.50 1.00 32.9 0.10 0.60 1.50 < 56.9 0.10 0.60 1.80 105.3 0.10 2.10 11.00
( TABLE 5. ACCIDENT EVENT TIMES _ C Ti me .
. Event minutes Steam generator dry 67.5 Core uncover 95.5 Start melt 118.3 Core slump 146.3
{ Core collapse 148.0 Bottom head fail 152.8 Containment fail 152.9 l Reactor cavity dry 177.2 l Start concrete attack 254.2 End calculation 1073.4 I e
y
.o - - - -
- r. . ,. ., , . .O ,
q s. 90 0 5 E Event Time, min. g 80 0 Steam Gen. Dry 67.5 M Core Uncover 95.5 D i U) '70 0 - Start Melt 118.3
~
hl Start Slump 146.3 60 0 Core Collapse 148.0 Bottom I! cad Fail 152.8 i h ' Containment Fall 152.9 Cavity Dry 171.2 M Concrete Attack 254.2 400 4 _, u, L 7 d U 300 t h b 200 t O
~ - ~ ~ ~
j 10 0 00- i i i i i i i i i , 00 MM)0 1.'fM) 0 *H N) 0 100 0 TdMif t UEM)f) 7tM10 (MMIC 900 0 MMM)O IHM)0 Tl MI': -- (M I NUTK) t FIEIRE 8. CONTAIPNENT PRESSURE RESPONSE FOR SURRY TitB'-6 SEQUENCE i
. , p . .. . .. . ., n, ,, . . , . , m .m -3 .
.o 1 ..3 . ., b
- 150 0-u.
N 3000 N D b h L 2500 W - Event Time, min $ 200 0 h Steam Gen. Dry 67.5 -
% .t *
'I. angg W
Core Uncover 95.5 4 Start Mel't 118.3 0
% 1 start 51wap 146.3 g Core Collapse 148.0 E Bottom llead fall 152.8 O
O 100 0 Containment fail 152.9 Cavity Dry 177.2 Concrete Attack 254.2 50 0 i i i i i i i i i i 00 100 0 2tM10 :MM10 4(M10 fdM10 tWM)() 7(X) O IMM)O 900 0 1000 0 1100 0 TI MI: - (M I NUTM) FIGURE 9. CONTAINMENT TEMPERATURE RESPONSE FOR SURRY TM.B'-6 SEQUENCE
- l, ,.
r 17 I e E (i. O Q
" C
. C
-g ce et D
m ce C o w _-= O 5 C ' E- t0 U $o m 6
\ C og
\
_E~* w
\
h I S w gw m C
- a?
ac co
\ ~h oW He C $
~
~E
=. E!
8
.- 1 C
.2 w
M
- s. C C
C C C C C C C C
'4 C v M N ** O
(' (90L XgM paaval sese9 30 u ntoA legal
?... s .. i.' j 18 ' s
'/
as in the earlier analyses; thus the previously evaluated primary system fission product releases were used. The containment was modeled as a single compartment identical to the approach previously used for this sequence. It is, of course, possible that the leakages would be initially l into secondary containment structures rather than directly to the environ-l ment. The effects of secondary containment structures have not been addressed in the present study since their possible effect can be inferred from the results of other sequences reported in BMI-2104. The high leak-age model as defined in Table 4 was used in the MARCH containment analysis; the resulting containment pressure and temperature responses ; are illustrated in Figures 11 and 12. It should be noted that the centainment pressure did not reach the level corresponding to the last entry of Table 4. The relatively small leak areas associated with the pressures that are predicted in this case have a minimal effect on the peak containment pressure predicted. Figure 13 gives the containment leak rate as a function of time, and Figure 14 gives the integrated leak-age for this model. The changes in the leak rate as the pressure rises are clearly seen in Figure 13. The failure of the containment to isolate was modeled as an opening 6 inches in diameter, existing at the start of the accident. The containment was again modeled as a single compartment without consi- i deration of the possible effects of secondary containment structures. For the cases of isolation failure it is likely that the opening in the primary containment would be to some portion of the secondary containment; the assumption of direct release to the environment is a simplification in the analysis. The containment pressure and temperature responses assuming containment isolation failure are illustrated in Figures 15 and
- 16. The corresponding leak rates and total leakage are given in Figures 17 and 18. It is obvious that the leakages associated wit'1 the assumed isolation failure are substantially higher than those for the "high leak" case. It should further be noted that the total leakage for j
the BMI-2104 case is even higher than that for the assumed isolat4on failure when Figures 10 and 18 are compared. ! Table 6 shows the results of the transport calculation for fission products under the co.*ditions described above. It is seen that
D . ., -.. _ .- l
**. .' l 1
e
- I l
19 l 1 e I 8 (. . C 9 5
, C R
5 A N
= g
. 5 D m
W O 6
/ _a m a f:: a m Z
5 a.
- k m I
Ow Z. dw 0 W:5 I :E S$ m-C
$~
- c. G C, i ~ $ . l. .
yEz E y$ Z c:
-2
*C gm v od u-9'*
_x W
- 2. . cz:
D C ee b
~
.J O
9 e O C O C C O c 8 8 8 ? 8 O (. VISd '3HOSS3Hd 1N3K1HVdK00
y .- 20 9 C (i 8 I C _g f' M C E
- 5
-a ,
r C
~
m S k Y
=b
- m
%d
~5 z -
g5 s ms f e s
~ 5 ,i , g=
m m 5 gO e -- E* ga
~E-
= aW uw w
e
~
N 5 e c _g L-Q I o O O C C C C C C C
$ $ .$ $ S 8
8 3 '3Hf11VH3dN31 IN3NIEVdWOD l L__________________
M e
- 21 I
e O (. e-
+
t o I ~ f-E
~
g a a E, v 1 A o eF g t"*
"w w sn D ~5 bZ N 5 ga v .J WX e I hi5 R 55 5::-:
2 5m g_. 8g C
~
- 8 c
. o
~
z = c C C C. b c c a 8 a
, a 8_
(
, NIN/ Eld '31VH MV310IH13Nn10A INWIN00
1 k 22 l
* )
l ( $ M
\
1 0 1 e e CJ E m ae d e e.s E g da3 Og T O-as g
.5 e5 Q5 Ei C a
-$a .-E N o 'f W
o W 8 _5W b i C
' R 3
o O l o o e o e o e e
% N O C e v eJ c l
l k (90Lxg3J)paMealowntoAtegei l l ,
s . 23 C E
=
C_
-5 I I C
1 C t
-g n E.
v t = 8 l _a - t 5o w
- =
L oa
.- mm E -.
.x z ce
-c a- w" s
EE 0 m= w a E2 tc _ 7. -. .
. 2 g5g e :-
t"* _ .E =c . EW v-
- m
_e.: I w W s C C _E
.5 C
-6 C
C
- c. O b c C C g 8 8 ? '8 YISd '380SS38d 1N3N18VdWO3
M e e ', e
- 24 o
9 (i !!! C
~ # O g
e
-g 8 8
m
-h^= $ - E-- W$
eD wd _e z gg 8 - O Q C, _ai - m Z g
- g" 2 EE C= ~3
,. E QCD 5 oY vr O
~
_E C " b F
' N :
( e C I C h b c b o c C b
! R R R H E B 3
( 3 '3Hn1V83dW31 IN3W1HVdWOD
( ,' . m 0 . 0 0 o 1 1 0 - 0
,0
, 0 1
0
, 0
,0 9
0
,01 M
f Y 0 R R
- 0 U
,0 ) SE 7 R E RU OL T FI A
0U
- EF
- T
. ,00N AN RO 6I I KT M
( AA EL
- LO 0 S 0 - TI n, ,0 5
E N EH MT P I M IW A
,I 0l T N B.
l
,0 0T OM CT 4
7 1 0 E 1
,X R U
S G I F . 0 0
,0 2
r 0 0
,0 1
0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0
- 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 7 6 5 '4 3 2 1 p
x O<g4hA ). o ~MHMyD2O> -
. E-2;7'E.2OU 1l Ill I
v 8 TOTAL VOLUME GASES LEAKED.FT3 *Kf ) o - ~ w - = o o o C O M 8-a 8 A g.
? $ u c
9-5 5 6
- s y 8-
,, -o "U
g t=: g.
\
< t 5' o
)
-4 e
=
a~ 5 25 y C* a m b P E v y_ ,
- o E '
h $~ E o m
=
O O
\
8 O 92 l 8, .
~
a . 27 I (. TABLE 6. LOCATIONAL DISTRIBUTION OF FISSION PRODUCTS FOR VARIOUS CONTAINMENT FAILURE MODES OF THE SURRY PLANT (TMLB') Unit: Fraction of core inventory Species RCS Containment Environment e BMI-2104 Cs! 0.85 0.11 4.6 x 10-2 Cs0H 0.86 0.10 3.9 x 10-2 Te 0.30 0.59* 1.1 x 10-1 High Leak Csl 0.85 0.15 1.9 x 10-3 Cs0H 0.86 0.14 1.1 x 10-3 Te 0.30 0.69* 1.2 x 10-2 Isolation Failure Cs! 0.85 0.13 2.2, x 10-2 Cs0H 0.86 0.13 1.3 x 10-2 Te 0.30 0.50* 1 1 x 10-1
*The number includes a fraction of 0.43 for Te not released from the core melt.
e
a _ _ _. J, , e-* 28 the fraction of core inventory that is captured in the containment is the highest in the "high leak" case primarily due to the relatively small leakage rates. As a result, the case registers the lowest fractions [ that are released to the environment. It is interesting to note that among the cases investigated here, the BMI-2104 case yields the largest _ release fraction suggesting that containment failure modes can affect the release of fission products in a rather subtle manner. The implica-
, tion is that the volume of the gases leaked depends upon the containment pressure and the hole size. A high pressure spike, such as the one seen , in Figure 8, tends to create a large hole size causing the containment to depressurize rapidly. Another consideration is, of course, the timing of the fission product source with respect to the leakage rate. For example, Te which tends to be released to the containment over a rela- . tively longer time is affected to a lesser degree by early leakage charac-teristics.
Secuovah -- Ice Condenser PWR For the ice condenser PWR, the alternate containment failure mode considered was failure of the containment to isolate at the start 3 of the accident. The isolation failure was modeled as a 6-inch diameter opening in the lower compartment of the ice condenser containment communi-cating with the secondary containment. The secondary containment was assumed to have a free volume of 1,000,000 cubic feet and a representative set of structures. The secondary containment was assumed to have a suffi-ciently large opening to permit leakage of the gases and vapors released to it without appreciable pressure buildup. The effect of the secondary containment was not considered previously in the BMI-2104 analyses for this containment design. The accident sequence selected for this study was again the TMLB accident previously considered. In the BMI-2104 analyse for this sequence, it was found that the containment integrity could be threatened by an early hydrogen burn as well as by long-term overpressurization. Table 7 presents the accident event times for the two variations of this sequence as they were derived in BMI-2104. Figures 19 and 20 illustrate
l 29 TABLE 7. ACCIDENT EVEtR TIMES Event Time, minutes r,- Sequoyah TMLB'-Y Steam Generator Dry 62.0 Core Uncover 97.8 Start Melt 121.5 Start Slump 143.5 Core Collapse 145.0 Vessel Head Dry 149.2 Bottom Head Fail 157.8 Containment Fail 157.8 Concrete Attack 158.9 Ice Melt Complete 428.5 End Calculation 761.2 Seouoyah TMLB'-6 Steam Generator Dry 62.0 Core Uncover 97.8 Start Melt 121.5 Start Slump 143.5 Core Collapse 145.0 Vessel Head Dry 149.2 Bottom Head Fail 157.8 Concrete Attack 161.5 Containment Fail 552.5 Ice Melt Complete 556.0 End Calculation 761.6
n 30 o o
-f v
o O _g 5
. E W
h r
- d
_g - E
~
E B o y e
"h I $
C i = E E m o . W n.
-g u W
;'5
- r:
. o O
-l
- 8 a C 8
1 $ i i i i i i e o o o o o o o o o a a e a S $ a 2 2 (. e tSd *iunSS3Jd $U3W4J9dC03 [v401
31 o o
-f w i
u eN uu 5 a 3 cr o w D m a o - 3 r- . m i , d r i .
. : .t o R t ' i - 8 o
w k Cr*
*- o s w i./
o w
- -. .u w.n.. .
~
. p
- . -...uw '
2 m c w c.w.... . . . ' . - a:
- w
. . .. . . . ... . . . .o. .I . . ::*g . . ...: )
i
/ -
p E w
- u.::*... ' a.
1 w
...s s w ****. [ g-N o
=
4
...............**> e - *.x l
l d
~
l ," 8 12
! S 1 o e s s a s o
o e o o o o o ( _ _ d '4Jnt Jadwal guawgJedwog
2r 32 the containment pressure and temperature responses for the case of early containment failure due to hydrogen burning as presented in BMI-2104. Figures 21 and 22 show the corresponding responses for the long-term overpressure case. Since the primary system response for the containment isolation case was substantially the same as previously calculated, the fission
. product release from the primary system to the containment was taken from the earlier analyses and only the containment transport and leakage were reevaluated.
Figures 23 and 24 show the containment pressure and temperature responses calculated for the containment isolation failure case. It is interesting to note that the pressures resulting from hydrogen burns in the primary containment are still high enough that they could lead to containment failure, even in the presence of the preexisting leak. The burning is seen to take place in the upper compartment of the containment,
} consistent with earlier observations. The containment leak rate and total volume of gases leaked are illustrated in Figures 25 and 26. The leak rate is seen to be rather discontinuous in' response to various tran-sient events, such as hydrogen burns, whereas the integrated leakage is quite well behaved. These containment responses were used as the basis for the evaluation of fission product release to the environment. It should be noted that the assumption of the isolation failure in the lower compartment means that the airborne radioactivity leaking to the secondary containment does not have the benefit of scrubbing by the ice condenser.
Table 8 sumarizes the results of the fission product transport calculations. It is seen that due to a large leakage rate and the ice condenser bypass, t'?c isolation failure case yields much higher release fractions of fission products to the environment. Grand Gulf -- Mark !!! BWR For the BWR Mark !!! pressure suppression containment design, the failure to isolate at the start of the accident was selected as the characteristic noncatastrophic containment failure mode. The isolation failure was assumed to be in the outer containment with direct leakage
.21 -
e , S 0
~
l 33 q ) F (- , f s . e 5 m
, - g m
mL Y o ~
- Q E
B
- M
*i E
-5 r - -
s
;: 2 /. 10
( o
=
y
- p- a a
2 5
- li
-l :E 8
I N
-g W 8
c 3 O O O O O O 8 5 5 # 2 2
. tie . anis.;a su.m2a emo3 t 2oi l(
J e
, , , , , _ _ , , ~ , _ - - , . . _ . , - _ _ , _ - , - - -
l O 9 34 0
.-g
-g
. oo a W !
. O y s, W
-E 5 m
.'.,l
.. M m goosoeeeeeeeeeees *
\
- t'
\ O W s' &
* \
R B
.\s =
\ a s O w
's,
-gs. e
.r . m
*\.s
' 5 a.
_3 ' W b (, % o a
.** " . w s
g
&m
- t w
*., C H l
-- * * . . . . . . . . . . -8 g
ll
/
. 8 L ,..*,.* U e
~3 8
E ci i . .
- O O O O o o 3 . e2 2.e 12o ,2, 3
(
e - A .
. _ , o k 80.0 COMPARTMENT 1
- COMPARTMENT 2
----- COMPARTMENT 3 70.0-80.0-
~
4-e U3 k 50.0-N ' R t/ 40.0-
)
M N ~ (4
% ~^
m-20.0-10.0- i i i i i i i 0.0 100.0 200.0 300.0 , 400.0 500.0 000.0 700.0 800.0 TIME - (MINUTE) FIGURE 23. CONTAINMENT PRESSURE RESPONSE FOR SEQUOYAH TMLB WITH ISOLATION FAILURE l - . . _ _ _ _ _ _ - - - - - - -- -- ww- - _ . - - -
o - - - - n .
. m 1
= ..
3000.0 COMPARTMENT 1 COMPARTMENT 2
----- COMPARTMENT 3 2500.0-Ize 4 2000.0-N i .E E4 ii !!
4 1500.0- ii !!
! w m
N 1000.0- !!!I E-e
- i.
- ii:
*.y
- 500.0- '-
"s
._ . ~~~~~~--.~~~-....~...... . , , , ,
3 7 F r_g.gy.sti".:"."."".4*____---------.------....----------------- 0.0 , , , , , , , 0.0 100.0 200.0 300.0 400.0 500.0 000.0 700.0 800.0 TIMN - (MINUTE) FIGURE 24. EONTAINMENT TEMPERATURE RESPONSE FOR SEQUOYAH TMLB WITH ISOLATION FAILURE v _ - __.
h , c* 0 0 0 T 8 m ' 0 0 i0 7 3 0
, 0 0
6 l l A Y 0 D 0 U Q 0 )E 5 E SE R T RU U OL FI N EF A
- 0. I
- k T AN 0M i0 (
RO XT I 4 AA
. EL
- LO S
~
TI A - E N EH
- 0. I M MT i
t I 0 lW A e 3 0T TB NL OM CT 5 _ 4 0 2 E e 00 R U 2 G I I _
. F r _
0 _
- 0
- i01 r
1 _ 0 J 0 0 0 0 0 0 0 0 o 0 _ 8 7 6 5 4 3 2 t O DgxoPN.NEh
- M<ga O~MENMDAO> E*ZE2O
^ c i !' ! i : ! ;) ll l 4!
- . , . . n .
, , .m. , ,
W b p a0 g 7.0 - E* N h 6.0-b a 5.0 -
- to 4.0 -
4 e w 0 =
$ 3.0 -
m M 2.0 - g a O . 1.0-RO i e i i i i 1 0.0 100.0 200.0 300.0 400.0 500.0 600.0 700.0 800.0 TIME - (MINUTE) . FIGURE 26. TOTAL LEAKAGE FOR SEQUOYAH TPt.B WITH ISOLATION FAILURE
TABLE 8. LOCATIONAL DISTRIBUTION OF FISSION PRODUCTS FOR VARIOUS CONTAINMENT FAILURE MODES OF THE SEQUOYAH PLANT (TMLB') Unit: Fraction of core inventory Lower Upper Secondary Species RCS Containment. Ice Bed Containment Containment Environment BMI-2104 (6) Cs! 0.82 8.6 x 10
-3 0.17 5.5 x 10 -3 --
3.9 x 10 -4 Cs0H 0.83 5.4 x 10-2 0.13 5.0 x 10 -3 -- 4.5 x 10 -4 3.8 x 10 -3 Te 0.25 0.91* 3.1 x 10 -- 2.0 x 10-3 , BMI-2104 (y)
-2 -3 -2 Cs! 0.82 6.1 x 10 0.10 1.5 x 10 --
1.7 x 10
-2 -3 -2 Cs0H 0.83 3.9 x 10 0.12 2.9 x 10 --
2.3 x 10
-2 Te 0.25 0.70* 3.7 x 10
-2 6.2 x 10 -4 --
1.4 x 10 Isolation Failure
-3 -3 Cs! 0.82 4.9 x 10-2 0.12 1.2 x 10 7.8 x 10 -3 4.0 x 10 ;
5.1 x 10 -2 1.2 x 10 -3 8.0 x 10 -3 4.3 x 10 -3 Cs0H 0.83 0.11 Te 0.25 0.69 3.7 x 10 -2 6.0 x 10
-4 5.9 x 10 -3 1.7 x 10 -2
*The number includes a fraction of 0.67 for Te not released from the core melt.
l
n -- -- - . ~ . . . . . . _
.t . 1 I
40 j to the environment; a 6-inch diameter opening was selected to characterize the isolation failure. The 52 E sequence (small break with failure of emergency core cooling) had been previously selected to investigate the possible effects of suppression pool bypass. This sequence was also chosen to investigate the consequences associated with containment isolation failure. For the present analyses, a normal suppression pool flow was assumed. Table 9 gives the accident event times for this sequence as they were determined in the BMI-2104 analyses. Figures 27 and 28 illustrate the containment pressure and temperature responses derived in the earlier analyses; the containment was predicted to fail due to the hydrogen burns taking place during the core slumping process. The reactor coolant system response for the containment isolation case was found to be substantially the same as had been previously evaluated. Thus for the present analyses, the release of fission products from the primary system was taken from the earlier analyses and only the containment transport of fission products was reevaluated.
~
Figures 29 and 30 illustrate the containment pressure and C- temperature responses determined for the present study. In the BMI-2104 analyses, the high containment pressures due to hydrogen burning were found to lead to containment failure; here, due to the presence of the preexisting leak, the pressures are not quite as high as previously observed. The hydrogen burning is seen to take place primarily in the drywell compartment of the containment. Figures 31 and 32 illustrate the containment leak rate and total leakage as determined in this study. The leak rate is seen to be sensitive to the pressures resulting from the hydrogen burns, as would be expected. The integrated leakage, how-ever, increases smoothly with time. These containment responses were used as the basis for evaluating fission product release to the environ-ment. The results of the calculated fraction of core inventory released for CsI, Cs0H, and Te are listed in Table 10. It is seen from the table,that the release fractions for Cs! and Cs0H in the isolation failure mode are approximately half of the fractions for the BMI-2104 case, and release for Te is slightly lower than that for the BMI-2104 (
- n. . . .
41 TABLE 9. ACCIDENT EVENT TIMES Event Time, minutes E Grand Gulf 52E b Core Uncover 5.6 Start Melt 27.8
, Hydrogen Burn 41.4 Start Slump 45.6 Hydrogen Burn 54.4 Containment Fail 54.4 Core Collapse 57.6 Hydrogen Burn 57.6
. Hydrogen Burn 59.6 Vessel Head Dry 80.7 Head Fail 119.2 C Concrete Attack 119.2 Hydrogen Burn 125.6 Cavity Dry 213.7 Hydrogen Burn 245.5 End Calculation 720.3 4-t
( t
. . . - _ , _ , . _ . - . . , . _ , . , _ _ , . . . . _ . . , - . , _ . . . , . . . _ . _ , _ _ . _ . ~ . . , . ,r__ . , . . --
b9
, ~ _
.N 0
0 0 O. 8 0
,00
,i 7
0 S 0 S
,0 A 6 P 3
Y B
. L O
O P 0 L 0 A
,0 )E 5
N I M) T OE U N 2 S N I H T I C
'E WN
- 0. M E
- ,00( EU RQ 4 UE SS
- S(
A E RF E PL U M TG N i
- 0. I 0 T ED
,0 MN NA 3 I R AG T
NR
" OO CF 0 7 0 2
,0 2 E R
U G
- I F
- r 0
r 0
,0 1
r
\ _
g 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 8 7 6 5 4 3 2 1 A 4m(Sl O Iz M4 l jjJi
A l r At - i 3 0, 4000.0 DRY 1mLL
........... WETWn 3500.0-3000.0-N 2500.0-E*l g 2000.0-Wr
- w y \
1 1500.0 - E* . 1000.0 - LY.
. 500.0- il. .i . .r
. 1
; ,1 i. ... .. ii
- 0.0 , , , , i , ,
- 0.0 100.0 200.0 300.0 400.0 500.0 600.0 700.0 800.0 TIME - (MINUTE)
I FIGURE 28. CONTAINMENT TEMPERATURES WITH NOMINAL P0OL BYPASS ] FOR GRAND GULF (SEQUENCE S2 E)
.-s ._ o____._.___.._'
! ' # I
- - A- ; ; . ,
, ,o, *<
4 60.0 COMPARTMENT 1 COMPARTMENT 2 50.0-m - g . 4, 40.0-Cd ' M - D r in A y 30.0-
~
20.0-
- ~ -
4 i . 10.0 - , , , , , , , l 0.0 100.0 200.0 300.0 400.0 500.0 600.0 700.0 800.0 j TIME '- (MINUTE)
! FIGURE 29. CONTAINMENT PRESSURE RESPONSE FOR GRAND GULF S2E WITH ISOLATION FAILURE 4
l 1
- l 1 e
0 0 0 - O . 8 0 12 . 0 TT
. i0 NN 7
EE . MM . TT RR .
. F AA .
. L PP .
. 0 U MM . G 0
OO . i0 D CC .
. 6 N
,. - A
- . R G
R
- O F
0 E
. 0 S N
i0 )E 5 O P
. T S EE U RR U
N EL RI
- 0. I UA 0M i0 (
TF A RN
. 4 -
EO
. PI
- MT O .
' EA TL E TS O
0. MI NI E Mi l 0 NT
- f i03 T II AW T
NE O2 CS 0 0 3 0 i0 E 2 R U
. G I
I 7 . F
. 0 0
,0 1
,S4.!.1
!.i
!.i 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 O 0 0 0 0 0 0 0 5 0 5 0 5 0 5
? 3 2 2 1 O.. 1 N NNpE4MN%MNE w
A* 0 _ 0 0 _ 8 0 , 0 _ 0 i 0 _ 7 _ 0
, 0 i
0 6 F
. L U
G 0 D 0 N i
- 0) A R
5E G E T RR U OU FL N EA I
- 0. I 0M TF A
RN i0( O KI 4 AT
- EA
- LL
_ ' O TS A E NI E M MH NT 0. 0 I II i 0T AW T 3 NE
- O 2 CS 1
3 0 E 0 R i 0 U 2 G I F 0 _ 0 _ 0 1
\
h - - 0 0 0 0 0 0 0 0 0 8 6 4 2 0 1
- e. bgxos%d94% M<d O cE C NypAO> EZ2eZoU
- :! ii! 4 1 ;:iI :
I ; ii
) !
! . 1 0 a 0 0 8 0 0 -
,0 7
E R U L I 0 A 0 F
,0 6 N O
I T A L O S I 0 0 H
,0 )E 5
T I W T E U S 2 N F
,00. MI L
U G 0( n 4 D N A R G E R O F
- 0. IM 0 T E
,0 G 3 A K
A E L L A T
, 0 O
,00 T
2 . 2 3
. E R
U G -
. 0 I 0 F i0 1
0 0 0 0 0 0 0 0 0 0 5 0 5 0 5 0 5 0 c 3 3 2 2 1 1 p O m 4.Ah<Nl-CE . rn N40 1O1 Nyn1_O>
48 TABLE 10. LOCATIONAL DISTRIBUTION OF FISSION PRODUCTS FOR VARIOUS CONTAINMENT FAILURE MODES OF THE GRAND GULF PLANT (52 E) Unit: Fraction of core inventory Species RCS Drywell Suppression Pool Containment Environment BMI-2104 Csl 9.1 x 10-2 1.2 x 10-2 0.89 9.6 x 10-4 7.0 x 10-3 Cs0H 0.16 1.1 x 10-2 0.82 8.6 x 10-4 6.3 x 10-3 Te 0.26 0.39* 0.11 6.5 x 10-3 2.4 x 10-2 Isolation Failure Csl 9.1 x 10-2 1.1 x 10-2 0.89 2.5 x 10-3 3.8 x 10-3 { Cs0H 0.16 1.0 x 10-2 0.82 2.2 x 10-3 3.5 x 10-3 Te 0.26 0.40* 0.30 1.5 x 10-2 2.0 x 10-2
*The number includes a fraction of 0 355 for Te not released from the core melt
- m. .
- z. .
49 case. Again, the decreased leakage rate to the environment in the
) isolation failure is largely responsible for the observed reductions.
( Peach Bottom -- Mark I BWR For the Peach Bottom BWR which has a Mark I pressure suppres- [ sion design, the failure of the containment to isolate at the start of the accident was selected as the representative noncatastrophic contain-ment failure mode. The isolation failure was assumed to be a 6-inch diameter opening in the drywell of the containment, and the leakage I through it was assumed to go the secondary containment building. The TC sequence (transient with failure to scram) was used as the basis for these analyses. The TC sequence was one of the accidents treated in BMI-2104, Volume II, for the Peach Bottom design, including consideration of the effect of the secondary containment. Table 11 gives the accident event times for this sequence as calculated in the earlier study. Figures 33 and 34 illustrate the containment pressure and temperature responses in C the primary containment from the earlier analyses. Figure 35 gives the pressure history in the secondary containment as predicted under the assumptions of the earlier study. With regard to the predicted secondary
~
containment response, it should De noted that it was based on a constant Standby Gas Treatment System (SGTS) exhaust rate of 25,000 cubic feet per minute and utilized the entire volume of the reactor building. Even with these assumptions the secondary containment was pressurized for a { considerable period of time, leading to substantial leakage from the building that did not pass through the SGTS. Under the BMI-2104 assumptions for the TC sequence, containment failure precedes core melting, with the imbalance between heat input and removal to the containment leading to pressurization and failure. In the containment isolation failure case considered here, the containment is " failed" at the start of the accident and the timing of the entire accident sequence could be different from the case previously considered. Thus there is some conceptual difficulty in doing a comparable evaluation. l In order to circumvent this difficulty, the failure of the emergency t !
50 TABLE 11. ACCIDENT EVENT TIMES T __ _ Event Time, minutes r-Peach Bottom TCY [ Containment Heat Removal On 10.0 Containment Fail ' 58.1 ECC Recirculation On 72.4 ECC Off 72.6 Core Uncover 73.0 Start Melt 93.6
~
Core Slump 124.6 Bottom Head Dry 136.6 Core Collapse 178.9 Bottom Head Fail - '216.6 C', Reactor Cavity Dry 216.7 Start Concrete Attack 216.7 End Calculation 816.9 I e . D
t
,~
9 - _ It LL Ez -
> Wu Yr
)
C T
- R Dw E C
. N
. E U
. Q
. E S
( S E M U L O V T N E. W I A T N n O C N I S E R U S S E R P 3 3
. E R
U G I F 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4 2 0 8 6 4 2 1 1 1 r 4 M4,N Wbm4
M - j o l a 52 i i k U (! C
'. I l
i
.g l
c : QQ l * = r E i -h d E i 5
- i 8 r . : w i i C U
- : m
, ! ! g N
- o
! ^ o
- M :-
C b $ i D w
. i -g -z a . , i a M
v $ o
- u i A z
( , i i
-8*wl a
S i 2
- E b
I b I i C W i
- w
!. 3
. \- C .
_m a N w Cr:
- s. \
C i O
.h-
.,*...' O
~
s s s s s s s C C C C C C C C C ( a _ 5_ i 8_ 8 8 6 i E2Hn1YHacDG1
h w* 0
. 0 0 ,
4 0 3 0 3 0 3 0 0 E
$ C 3 N E
U Q E S I 'y 0 - 0 C T
$2 )E T G
N I U R U N D E
,00. MI R
, 0(
2 U S S E R _ - P G A - E N I
-_ D 0.
0 IM L I U
,5 T B
' 1 R
O T C A E R 0
,0 5 0
1 3
- E
. R f .
U G I F
. 0
, 0
- 5 0
0 O 5 0 5 0 5 0 5 0 1 R D 6 1 6 1 5 1 5 1 4 1 4 1
^- MM% N4DiQ C
n % l ;l :l l
- . 1 ::
- 1 P-54
' I core cooling system for the containment isolation failure case was assumed to take place at the same time as in the earlier analyses. With this ~
assumption and primary system treatment consistent with the earlier I analyses, it was found that a 6-inch opening in the containment was not able to keep up with the steam input to the suppression pool and that [ the containment was predicted to pressurize above the failure level I assumed in the earlier analyses. The containment pressure response for the above containment isolation failure case but with no overpressuriza- [ tion failure allowed is illustrated in Figure 36. Additional containment calculations were performed utilizing larger initial isolation failures; the results of these calculations indicated that initial hole sizes of several square feet in area would be required to maintain the primary containment pressure below the failure level. The foregoing observations are quite consistent with containment venting investigations conducted in the past in which it had been found that rather substantial openings (3 feet diameter) were required to keep up with the steam generation in sequences of the type considered here. Based on these observations, it is concluded that for the nominal contain-ment isolation failure hole size of 6 inches employed in these analyses the containment still would overpressurize and fail with the result that the accident sequence is nearly identical to the TC case considered in BMI-2104. Of course, other hole size assumptions could be made but a large initial opening would be required to prevent excessive pressure loads. In either case the environmental fission product source terms would not be expected to differ appreciably from the previous results for this TC sequence in which the containment was assumed to be initially i intact but was predicted to fail by overpressurization. The particular case of interest is designated TC-y in Volume II of BMI-2104 and since the isolation failure case considered is essentially identical, the results for TC-y are shown in Table 12 to indicate this equality between the two assumptions. (
9 . .
=
0 0 O 0 9
- N i0 0
0 8 0 C 0 T i0 M 7 O T T
' O B
0 H C 0 A i0 E 6 P R
) O '. E FE R
T EU
- 0. U 0
SL NI OA
- i0 5 N PF S
I EN RO M ( ET I RA ^ 0 0 i0 UL SO SS EI R 4E PH M I TI NW T _ E T ME NC 0 IN 0 AE i0 TU 3 NQ OE CS
' 6 0 3 0 E i0 R 2 U G
I F 0 0 i01 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 5 0 5 0 5 0 5
- 3 3 2 2 1 1
<~N% dNPmbm%
E i l
c r ; - p i r, - - 3
- - ,A b
s. TABLE 12. LOCATI0tlAL DISTRIBUTION OF SPECIES OF Tile PEACH BOTTOM PLANT (TC-y and TC-Isolation Failure) Fraction of Core Inventory Reactor Species RCS Pool Drywell Wetwell Bldg SGTS Environment 1.5 x 10 -2 6.9 x 10 -2
~ ' ~ ~ ~ ~ ~
Cs! 0.06 0.69 0 6.8 x 10-2 0.10 Cs0H 0.22 - 0.56 1.4 x 10 -2 0 6.1 x 10 -2 5.8 x 10 -2 9.1 x 10 -2 Te 0.34 7.9 x 10 -3 0.29* 0 0.11 1.3 x 10 -2 0.25 I
*This includes a fraction of 0.13 for Te which is found not to be released from the core-concrete interaction.
E
>_---------J
( 4/ g'w l) STONE 6 WEBSTER ENGINEERING CORPORATION 245 SUMMER STREET. BOSTON. M ASS ACHUSETTS ADDRESS ALL CORRESPONDENCE TO P.O. BOX 2325. sOSTON. M A SS 02107 W U TELEX 94 0001 g can C %stawcT+0N
' N " ' " " '" ": "' ,,o S
* ;'u't,*o ' .." CAL
- "'t'a?, " '
w a s hi N G,O M O C. Dr. M. Silberberg January 10, 1985 Accident Source Tem Program Office U.S. Nuclear Regulatory Commission 7915 Eastern Avenue Silver Springs, MD 20910 RESULTS OF SANDIA/SWEC CODE COMPARISON As noted in my letter of December 26, 1984 to Dr. Walter Pasedag, I promised to send an improved figure which depicts the results of the analysis of iodine leakage as a function of containment opening size as calculated by Sandia National Laboratories and Stone & Webster Engineering Corporation SWEC. The enclosed figure presents these results. Several notes are in order pertaining to this figure. First, it was re-drawnin log / log format to better illustrate the important region of opening sizes less than 1.0 ft2 Secondly, it should be pointed out that these analyses are based on a single control volume representation of the Surry containment and there-fore do not include any effects of multicompartments within the containment or structures outside the containment. The Sandia data were provided by Dr. David C. Williams of Sandia, whose cooperation in this activity is gratefully acknowledged. Any questions on the Sandia analyses should be directed to him. Any questions on the SWEC analyses should be directed to the undersigned. Sincerely, f Edward A. Warman
- = ,
Enclosure s CC: DrDCWilliams-Sandia National Laboratorier
- DrWRStratton-Stratton & Associates ProfRWilson-Harvard University 96)
[Tos/9 . 5
e + 1.0 i z ' 9 Ni~%y
,s~s W- M"~
c: u. _ ,/
/
g% s
< s
\
l
- e w1 0 -
e \
/
~
O - / N s
;- 44 d
2 . 8 _ ..
' ' ' ' ' I ' ' ' '
o.o1 O.1 1.0 10 SIZE OF PRE-EXISTING OPENING (f12) LEGEND: CONTAIN ANALYSIS WITH HeBURNS CONTAIN ANALYSIS WITHOUT H SURNS
-- ANS/ SWEC STUDY THREED/NAUA (NO He BURNS PREDICTED)
SIZE ON CUMULATIVE LEAKAGE AS CALCULATED BY SANDIA AND ANS / STONE G WEBSTER SURRY AB-S SEQUENCE l
GEM 18sts i [jL[f[ v . ,.., BROOKHAVEN NATIONAL LABORATORY ft,i ! s-I i~t ASSOCIATED UNIVERSITIES, INC. Upton Long Isicnd. New York 11973 Deportrnent of Nuctect Energy (516) 282s 2620 FTS 666' February 3, 1983 Dr. M. Silberberg, Chief Fuel Behavior Branch Division of Accident Evaluation Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Commission Mail Stop 1130SS Washington, D.C. 20555
Dear Mel:
I was pleased to be able to attend the L % Source Term Peer Review meet-ing on January 25 and 26. I look forward to continuing participation in the review process which you have initiated. I believe that you successfully established a sense of scientific " openness" concerning the basic source term issues which should help in future months to resolve at least some of them. I would like to add my comments on two issues which I see to be funda-mental to evaluation of the radioactivity release rates from containment: Early Versus Late Containment Failure The timing of containment failure is the single most critical determin-ant of the quantity of radioactivity released to the environment, as predicted ' by the Battelle calculations. In the context of risk to the public, there-fore, methods will have to be developed to evaluate the probabilities of early versus late failure. Containment loadings which could lead to early failure involve the phe-nomena of steam spikes and hydrogen burns. Objective quantitative assignment of probabilities will, in my opinion, have to be made on the basis of extreme-ly limited data bases, i.e., our knowledge of the physical behavior of the systems will be incomplete. The real question is, will our knowledge be ade-quate enough to put numbers to the probabilities? In my opinion, it appears unlikely. I raise this now so that perhaps alternate strategies can be planned to perhaps circumvent the real uncertainties in physical phenomena that will probably persist. g7/./ b b
k 4 t Page 2 Letter to Dr. M. Silberberg , February 3,1983 l Attenuation in Containment For sequences with late containment failure, the expected release atten-untion due to holdup is a function of rather complex physical processes which would occur in a complex 3-D flow field. The model being used by Battelle assumes no convective flow patterns in containment. The simplicity introduced i by this assumption will have to be verified by relatively large-scale experi-
- ments. In addition, the individual models used to compute the agglomeration and deposition of aerosols and vapors in containment must also be assessed.
i From your comments at the meeting, I gather that you are going to be con-
- sidering the issue of containment loadings due to various mechanisms. Please 4 keep me informed as to developments here, so that I may contribute towards development of a position on the " steam spike" issue.
1 est regards, Theodore Ginsberg Group Leader 4
, Experimental Modeling Group j TG:1z cc: R. Carbone G. Greene W. Kato , H. Kouts 1 N. Tutu i
i l i ) . i I i i 1 _ 1
HMml6 TON l j #
- B O STO N EDISON COMPANY I
800 80VLaTON Stass7 WOsTON, MASS ACHUBsTTS O2199 ,! 7.'.'. Y.* *." ..'.'"." ' * "' " " February ll,1983 BEco. Ltr. #83-29 Mr. M. W. Jankowski i Fuel Behavior Branch i Division of Accident Evaluation Office of Nuclear Regulatory Research U. S. Nuclear Regulatory Commission 1717 H Street N. W. Washington, D. C. 02555 7 License No. OPR-35 Docket No. 50-293 Boston Edison Coment and Position of Draft NUREG-0956, Volume 1
Dear Sir:
Boston Edison expresses its appreciation for the opportunity to i participate, as an invited observer, in the peer review meeting of the ! first draft of NUREG-0956, Volume 1: A PWR Analysis, presented in i Washington on January 25 & 26, 1983. We also commend the NRC's recep-l tiveness to consider comments, before the NRC releases the final NUREG-0956 document. l Boston Edison supports the spirit of NUREG-0956 which we believe is to realistically predict best estimate fission product releases to the environs for a range of accident conditions and sequences, utilizing best available techniques to model complex transport and removal mechan-isms. It is Boston Edison's position that, due to the significant impact , NUREG-0956 can have on licensing practice, emergency planning, safety goals, and indemnification policy on revised source tenns, NUREG-0956 should not be j issued as a final document, as its currently exists in draft form, until: I 1) Existing pmblems and uncertainties identified at the peer review meeting, have been satisfactorily resolved. Specifically, the areas requiring further investigation are physical and chemical interactions related to release phenomena, and transport and re-moval mechanisms for certain fission product species. I In lieu of the development of more realistic models to predict re-4 leases associated with the uncertainties and problems identified i above, the first draft of NUREG-0956, Volume 1, implemented con- ! servative and unrealistic models which overestimate releases to the
- environment. As a result, the environs experience a significantly
- higher radiological impact.
r
EDCON COMPANY Mr. M. W. Jankowski t February ll,1983 Page 2
- 2) Transport models are developed which will realistically predict fission product releases from the primary system to the Reactor Building secondary containment atmosphere to the environment through smaller cpenings than what was considered to date.
Mixing, removal, and depletion mechanisms, if realistically modeled, could significantly reduce fission product releases to the environment. As a result, the environs would experience a smaller radiological impact. The current method employed in the first draft of NUREG-0956, Volume 1 document is overly conservative and unrealistic in pre-dicting releases to the environment and is essentially addressing class 9 type accidents (i.e., large scale breach of Reactor Build-
, ing secondary containment) for all intent and purpeses.
- 3) The office of Nuclear Regulatory Research has had the opportunity to examine and consider experimental data being generated through-out the scientific community related to the subject matter pre-sented in draft NUREG-0956. Some of this data, as mentioned at
, the peer review meeting differs significantly from results pre-sented in draft NUREG-0956.
The discrepancies in published data should be thoroughly investi-gated in order that the most realistic and mechanistic models can be developed prior to finalizing NUREG-0956. Very truly yours, I 9 1 0
- _ - - ._y , . - - -- ----.---
0 h006ENT
/ 94 EGcGa,_
T!LICorita T AAgSMIT1*L REQUEST
,,-*s.. ,
10 GHGAh!ZATION NRC 'n'ashingt on , D.C. ADDRESS . _ _ Mail Stop 11.1055 ATTEN T10N _ ._ ._ M.. S ilf>cr fiero TYPE OF TEL(COPIER TELECOPIER TELEPHONE NO._ 427-4403 VERifiCAT:0N TELEPHONE NO. 427 d287 _ __ _ FROM R. R. Hobbins -- TElf rHONF NO. 9545 ,, Corp 4Nf J.G&.G..fI f.aho . I nc . _ , , 1fPE OF lELLCDPIER _ ._. 1ELECDPIER TELEPHONE No 576-9591 , VERIFICATION TELEPHONE NO. 6-9000 l 101 Al PAGES lest udeng transmittal sheell _1 2 ~83 DATF X
- RfiUAk ORIGINAL Yts WO I
ag e,gg Joanne W1111asin, 9015 ___ (x y y. y.-eq.vs;;:.
, m n - c - n ,n.--.....
J . .
__ _ _ ___ - - ~ _ _ . _ - _ - . - __ _ _ _ _ . _ _ _ __ _ _ . 4 RivifW COMMENTS ON NUREG-0956 i This letter briefly presents EG&G Idaho's comments on NURLG-0956. We agree with many of the peer review comments at the recent meeting. We further recogaire the uncertainties identified in the report (such as those listed on page 3-5) and agree that further evaluation of them is warranted. Hence, only additional concerns not heretofore mentioned and previous conments which we f elt required emphasis are discussed. The objectives of the NUREG-0772 follow-on study are appropriate. The approach to the study as stated on page 3-1 of the report, "...that t mechanistic predictions of radionuclide release and transport are possible if proper modeling is performed to represent the physical and chemical processes d6 ring LWR'ac'cidents" (emphasis added), is also justitied. j' However, we disagree with the direction in the report r.tated on page 3-3:
"It has not been a part of this study to produce quantitative estimates of uncertaintics in individual parameters and hence the overall impnetance of such uncertaintics has not been assessed" (emphasis adoeo). As noted in the report, numerous assumptio'nT were~ nice uary and uncertainties are present in the analysis. While some uncertaintics were addressed, many are not; the final calculated values for the source term from containment have littic value without knowing the overall uncertainty.
Consideration should be given to delaying publication; several reasons l appear evident. First, the source terms to the environment for certain accident sequences (lMLB-8. AB-B) are similar to the predictions of
- WASH 1400 and NUREG-0772. Hence, NUREG-0956 predicts limited or no
! retention in the primary system during those sequences. However, there is significant retention predicted for other sequences (V.SpD). If significant (or no) retention is subsequently demonstrated by experiment, it's going to be difficult to be convincing when f aced with all these reports. Second the codes used in this study ere oli in a state of flux or infancy: MARCH 1.1 was used instead of MARCH 7; TRAP-Miti is under development; NAUA required a modification for the ESf's and CORCON is new l and unvalidated. Third, a comment during the peer review meeting suggested 1 that the models, desite their inf ancy, were outstripping the available data; many assumptions without technical basis are required to perform 4 the analyses. Data from the first few Severe fuel Damace Tests in the Power Burst Facility will become available in the near term. Thus it j would appear prudent to consider delaying this study until the eredels and codes become more st abilized, and near-term data are available to aid in ) reducing or bounding sume of the uncertainties in the analyses. I 1 l One basic problem with the report is the lack of detailed treatment of i important phenomena in calculating the fission product and acrosol source i term to the containment. Specific items of concern are the simplistic treatment of core melting in MARCH, assumptions concerning the fission I
=
p.mr .m.: n . .. n . m e... ::v ..- . r . . +~ e
product and aerosol source term at the time of core slumping, modeling of the upper plenum flow. uncertainties in the TRAP-MELT modeling. and the absence of radioactive decay during transport. Our concern is that the lack of detail in the analyses results in a lack of confidence in the
- calculated source term to the containment.
The study recognizes two simplifications in the MARCH analyscs: a singic melting temperature; and slumping of the enre when .s specific fra. tion of the core achieves that melting temperature. However. . it is f urther assumed that once the core slumps into the lower plenum of the vessel no further fission product and aerosol release, deposition, or resuspension occurs until core-concrete interaction begins. Based on data from Test SrD-ST in PBT, quenching and fragmentation greatly enhance fission product release. Similar effects may occur when the core is cuenched as it f alls into the lower plenum. In addition, no consideration is given to the Interactions of the molten core with the lower core support plate, structures in the lower plenum, and the bottom of the reactor vessel and the resultant potential for enhancement of fission product and acrosol source terms. Not considering these phenomena may be expected to lesd to an under-estimate the fission product and acrosol source term. The timing of fission product and aerosol release is important in determining to what extent fission products are transported by acrosols; the simpitstic treatment by MARCil and COR50R inadequately addressrs this important question. The lack of detailed modeling of the upper plenum structure. flow paths, and thermal hydraulics severely limits the credibility of the calculated source term to the containment. The likelihood of significant deposition of fission products in the upper plenum is not addressed in any meaningful
; way. Although the report acknowledges that the one-dincnsional flows would underestimate the fission product and aerosol deposition dee to convection currents in the upper plenum,1 RAP-Mfl1 sensitivity studies have not been performed. Nnw that upper plenum desions are beino provided to Battelle Columbus Laboratorics, this area of analysis should be improved.
in this study, one should note that the flow path in the V sequence in
! B&W plants is dif ferent from other PWR's due to the operation of vent valves in the upper plenum.
i
- The TRAP-MELT uncertainties should also be examined. Vapor deposition l velocities are uncertain at present and not all fission product species l are modeled. Ihc two region plenum in the study does not adequately l account for flow paths in the vessel. , Particle depnsit ion by thermo-l phores15, which is proportional to the temocr4 tun? and concentration gradients at a wall surface. is improperly estimated because these large
, control volumes are at uniform temperature and concentration. Vspor resuspension is by evaporation, but irreversibic deposition by cnemisorption is not modeled unless it is implicitly included in the vapor deposition velocity formulat ion. Particles deposited on surf aces are not resuspended by high flow rates. finally, temperatures of surf aces are not considered in full detail. The surface heating expected from the residual decay heat of fissinn products deposited on the surf aces is not accounted for. I Control volumes used in TRAP-MCLT are large and temperature gradients in l j the upper plenum and in pipes are averaged, reducing the fission product
- and acrosol retention by cold surfaces. Materials which resuspend f rom a l 2 l l
l s W:. y :~ .:a w. -
.m . ~., . - - ~ ~- ~ -
l
~- . l
hot region and quickly redeposit in a cooler dor.nstreom locotion ore not properly considered. The overall effect of these potential deficiencies on fission product and aerosol retention in the primary system shnud be addressed. Another important phenomenon which is not addressed in the cocument is the decay of radioisotopes. As has been demonstrated by our werk in pBF, decay chain effects may play a significant roic in the make up of the source term resulting from severe core damage accidents. Early release of volatile fission products having short-to-medlem half-lives results in I enhanced release of long half-life non-volatile species. If failurt? or core melt occurs late many of the high volatility, short-to-medium half-life species have decayed into less volatile species, thus reducing the source term. Radioactive decay also plays a role in the re-evolution of deposited materials. Deposited iodine in the upper plenum subsecuently decays to xenon which is quickly transported to the containment. Specific questions and comments are noted below.
- 1) The basis for the assumed gap releases is not given. Althos9h these releases are small compared to the releases et higher temaerotures, results from P0r Test SFD-ST and tha Halden ITA-430 experiment suggest values of 1% or less for noble gases, cesium, and iodine.
?) Assuming no heatup due to decay heat of structures upon *nich fission products and aerosols are deposited may affect re.<,uspension of the materials.
- 3) Aerosols are calculated to sett le on horizontal surf aces in tne core and be permanently removed from transport. What horirontal surfaces are these? Are the aerosols calculated to settic nn a mniten core and he removed from transport rather than vaporized?
- 4) The tables are confusing: the use of fraction of the inventory, fraction of species available. fraction of retention Iceds to mis-conceptions; in some tables the fractions add up to 1.0. in some they do not.
- 5) The AB sequence has been extensively studied despite beino nonrisk-dominant. The justification for studying this sequence wvs not 1
convincing. Let us restate the opinion that BCL did as credible as a job as is possible under the current constraints of time and state-of-the-art for the codes, but it appears that the state-of-the-art is not adequately developed to permit a reasonable calculation with def ined uncertainties.
~
For this reason we suggest the publication of this document should be delayed until the uncertaintics can be better quantif ied. 3
,;. 3p. - g - -r- m ---- .- - ,~- - -- --- ~ - - -
i
- ]Offf/S0lY ARCONNE NATIONAL LABORATORY 9X)0 South CAss Anu. ARcon. llhos 60439 Tdghn 312/972 7533 February 4, 1983 Dr. M. Silberberg, Chief Fuel Behavior Branch Division of Accident Evaluation U.S. Nuclesr Regulatory Commission Washington, D.C. 20555
Dear Mel:
You are to be commended for bringing the peer group together to review a problem of current interest to all that express interest in nuclear power. Certainly not everyone agrees that the report is ready for pblication, even after the June myiew, but there are, I am sure, institutional needs that must be served. Obviously the contractor needs to address the report with even greater fervor to reduce / clarify the varied points of disagreement. There are a number of problems dealing with chemistry that were highlighted and I shall not go over them again. However, a few additional items do need clarification: i (1) The contention that control rod material will vaporize con-gn2ently (its stated compositron) is in error. I would suggest that Cd will be vaporized first followed by Ag and then In as the expected result (this position is supported by ORNL). Recent post-irradiation examination studies at ANL of TMI-II filter material has identified particulate material of pure Cd, Ag, and In along with some alloys but never in the control rod (stoichio-9 metric) composition. (2) There must be boundary limits placed on the I 2 limits that can be observed in any accident sequence. With CsI as the species being evolved in an accident the opportunities for molecular iodine formation is very limited and characteristically outside the bounds of the environments found in the stated accident scenarios. (3) In spite of the discussions there is still significant problems in the manner in which fission product tellurium is handled. Complex species have to be introduced into the please/ transport scheme to ensure correctness. Further, Dana Powers is in error when he states that nickel telluride is the most stable species. There are others more stable and of greater importance, i.e., ZrIg, CrI 2-
~.} 2 0 6Y '
(15 mTd Ewqv Tit (Mtasity dOrxp - Ancpw (MtRsiTES AssoctArkw
Dr. M. Silborborg Februsry 4,1983 l l (4) Of a more generic nature is the problem of semantics in that ! non chemists are using chemical terms erroneously. In particular I am certain that the SANDIA work on chemisorption is really a corrosion problem that may have other materials / fission product behavior ramifications. This needs to be clarified. On p. 5 7 an equation is given for the fractional release rate coefficient K = AeBT (1) that does not have the proper form for the curves given in 5 2. This is demonstrated by the need to give three sets of values for A and B for the temperature range of interest. The given equations result in straight line segments . Since most of the data are for lower temperatures (Lorenz <t600*C, Parker (2000*C, and SASCHA <1800*C) there is practically no data to determine how these segments can be extrapolated to higher temperature. Most of the curves that are given in NUREG -772 and are used in this report fit an equation of the form K = A exp(B/T) (2) Fission product release from the fuel is generally assumed to involve diffusion to grain edges and to involve vaporization. Both of these processes can be represented by Eq. (2). Thus for most of the curves Eq. (2) can be extrapolated using a single set of A and B parameters. The Xe-Kr curve is more accurately fitted by the sum of two exponential curves like Eq. (2). Purther, in this section of the report it appears to have been made that release / vaporization / aerosolization of fuel, zircaloy cladding, and structural material can be treated in a manner similar to fission product release. In fact the coefficients used for Zr and Sn in the cladding are the same as that for fission product Zr and Sb, resrectively. Structural materials are assumed to " vaporize" at .01 of the zircaloy cladding rate. The process for moving fission products out of the fuel into the steam environment is com-pletely different from that for oxidation, vaporization, or aerosolization of cladding, structural material, and fuel into the environment as vapor or particle form. If the equations for fraction release K(T) = AeBT of Table 5 1 are used to fit K(T) = Ae3/T then the value for B in the latter case is the heat of activation divided by the gas ccnstant. Most of the fission product release data yield a heat of activation of #40-50 kcal. The numbers given in Table 51 for fuel " release" also give a heat of activation of
#50 kcal. However, urania vaporization UO2 (s) + UO2 (g) requires some 140 kcal and reaction of urania with water requires some 118 kcal. Similar deficiencies exist for sirconium. If therefore seems unlikely that the equations in Table 51 will accurately predict the " release" of fuel or cladding in an accident.
4 Dr. M. Silbarbarg Fabrusry 4.1983 A final thought reflecting on the meeting format that would allow more detailed discussion among the diverse group present for the peer review. Rather than debating specific issues (i.e., thermal hydraulics, chemistry, . . .) among all those assembled why not consider dividing up into smaller, more homogeneous groups to handle the specific issues. Debate is for a limited time and a " dis-cussion" leader reports back the groups findings. This format could be an efficient way of resolving / handling specific issues very expeditiously. A dis-j advantage to this approach is that " cross-fertilization / education" of each others concern. However, if one is to produce a high quality report in a short 1 period of time the focused activity may have some merit. I appreciate the opportunity for participation in the peer review and would be happy to help in refining the chemical issues should the need arise. Our own studies on fission product release and thermodynamics and aerosol characterization may provide relevant background information to the development of this report. Again plaudits for the good job under such a narrow time frame. Since rely, Carl E. Johnson Chemical Technology Division I CEJ:kk l i e l
,, - _ . . - - --. ~
KELLY
/ .
UNIVERSITY Ofr VIRGINIA SCHOOL OF ENGINEERING AND APPLIED SCIENCE v - C H A R LOTT ESVILLE. 23000 C EPA RTM ENT OF NUCLEAR ENGINEERING AND ENGINEERING PHYSICS T ELEPMON E: God-924 713e KEACTOR FACILITY February 10, 1983 Mr. M. W. Jankowski Fuel Behavior Branch Division of Accident Evaluation U. S. Nuclear Regulatory Conunission Washington, D. C. 20555
Dear Mr. Jankowski:
Accompanying this letter are my comments on NUREG-0956. I am not sure about which office is responsible for the financial details of my travel expenses and consulting fees. I have provided Neill 4 Thomasson in the Office of Policy Evaluation with the necessary information, and trust that sooner or later I will receive a check frem his office or yours. l I appreciate this opportunity to work with the N.R.C. t Yours truly, J. L. Kelly, Professor Dept. of Nuclear Engineering and Engineering Physics JLK:ph Encl. i w . - - - -- , - - - - - , - e,, ,, - - -
I COMMENTS ON REVIEW OF NUREG-0956 l James L. Kelly, University of Virginia A two-day peer review meeting was held at U.S.N.R.C. Headquarters on January 25 - 26, 1983. The subject was the draf t edition of NUREG-0956 (Vol.1), "Radionuclide Release Under Specific LWR Accident Conditions (Volume I, A PWR Analysis)." This report (NUREG-0956) represents a major i ef fort to address the shortcomings of NUREG-0772, " Technical Bases for My Estimating Fission Product Behavior During LWR Accidents" (June, 1981). comments on NUREG-0956 are presented below. i I. Comments about code work A major criticism of NUREC-0772 was that the codes that were used (MARCH, TRAP-MELT, HAARM-3, CORRAL-2, QUICK, NAUA) were neither comprehensive
'- nor sufficiently mechanistic. Each exhibited serious deficiencies with respect to the ability to analyze the conditions along the fission product transport path.
NUREG-0956 indicates that considerable progress has been made in the e.g., MARCH 1.1. MERGE, CORSOR, past two years toward developing better codes, CORCON, TRAP-MELT, CORRAL-2 (modified) and NAUA-4. A number of deficiencies 4 However, even evident in the code work of NUREG-0772 have been eliminated. more progress must be made before the system of codes may be considered to be sufficiently comprehensive and mechanistic. A few examples of points where improvements in the codes are needed are (1) a more realistic fuel melting model, (2) a more realistic treatment of core-slump, (3) a mechanistic code to replace CORSOR,. s
-,..p- , --r-e r,-n, , - - - - ,- ,n. --- ,,-r -.y . . , , , - - - - - . , .
i (4) an accounting for control rod silver as an aerosol source. (5) a more realistic pathway for core-concrete aerosols, (6) an accounting for decay heat from deposited nuclides, , (7) a more realistic treatment of flow in the upper plenum, (8) an accounting for aerosols of uranium and plutonium, (9) a mechanistic model for containment failure, 4 (10) an accounting for retention in the auxiliary building for sequence V,
> (11) an accounting for chemical reactions among volatile and aerosol species, I
(12) an accounting for attenuation along containment leak path, and (13) an accounting for water condensation on aerosols. II. Comments about experimental data In the absence of an adequate experimental data base, computational results, such as presented in Chapter 7 of NUREG-0956, are of questionable l value. Unfortunately, NUREG-0956 fails to discuss the status of the data base, leaving the reader to infer that no significant advances in the acquisition of experimental data have been made since the writing of NUREG-0772. If that inference represents the true status of the data base, a serious deficiency still exists in the ability to analyze radionuclide re-(NUREG-0772 presents an extensive, but incomplete, list of data leases. base 10mitations in Section 1.4.) III. Conclusions (1) NUREG-0956 lacks balance in that it discusses recent advances in code work but totally neglects the important subject of i I
~
i
-_. . - - , - ~ , , , -. - , . . - - - . - - , , , ~,.
t i experimental data. An up-to-date summary of the data base should be included in NUREG-0956, or should be published as a separate report. (2) It is emphasized in NUREG-0956 that the nature (i.e., timing, 4 size, and location) of containment failure represents perhaps the single most important parameter with respect to fission product release to the environment. However, this event (containment failu're) is treated in a completely arbitrary and _ non-mechanistic fashion. A large effort, both in the experimental and code work areas, must be directed at developing a credible method for analyzing containment failure modes. (3) There is a huge amount of experimentation remaining to be done [ k to generate an adequate data base. (4) The code work in NUREG-0956 is advanced relative to NUREG-0772, but is still deficient in a number of important areas. Some of these deficiencies cannot be properly addressed until the requisite experimental data are available. Development work should continue with the codes. (5) The NRC should exercise caution in publishing NUREG-0956. The 4 computational results presented in Chapter 7 have been generated by using an inadequate data base as input into a system of non-t comprehensive, partly non-mechanistic codes. These results, 2 therefore, should be regarded with a high degree of skepticism. If published, NURE3-0956 should be replete with qualifying state-ments in order to redt ce the likelihood of the results being misinterpreted and misuse *d.= 1 1
~_ -, ,._ .._. _ .__ _ .._ _.._ ., _,__ , .. _
l
$1'00($
E<t.=s.[ t j [ i. s BROOKHAVEN NATIONAL LABORATORY
] { {. {, ASSOCIATED UNIVERSITIES, INC.
Upton. Long istond. New York 11973 Department of Nuclear Energy (516) 282' 2918 FTS 666' January 28, 1983 Mr. Melvin Silberberg Chief, Fuel Behavior Branch Division of Accident Evaluation .: Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Commission Washington, D.C. 20555
Dear Mel:
Attached are my comments on Draft NUREG-0956, subsequent to the peer review meeting of January 25 and 26,1983, t i Sincerely,
$ _/-
Herbert Kouts Chairman Attachment HK:ns 1 9 g 5(j Q l 3 o b/t
Comments on Draft NUREG-0956 by Herbert Kouts The draft of NUREG-0956 shows a substantial amount of thought devoted to the source term, beyond previous work on the subject. Though the remarks I make below are generally on lines critical of the draft, I do not wish to give the impression that I reject the draft or that I do not respect the quality of work that went into it. Rather, I am only commenting on ways by which the work might be further improved.
- 1. It is important to stand back from the draft and view it as a whole. It generated a great deal of opposition and adverse comment at the peer 7eview meeting. Let me put in my words what I feel was the cause of the reaction. The analysis done for WASH-1400 neglected attenuation of the source term through plateout and deposition of airosols in the primary system, except for some iso-lated BWR sequences that were not discussed in the present version of the draft NUREG-0956. The analysis for WASH-1400 also included assumptions as to chemi-cal forms of cesium and iodine that have since undergone change: cesium was assumed to be released in elemental form, as was fodine except for a small or-ganic iodine contribution. The new analysis considers the two bound in Cs0H and CsI.
When the new analysis takes all this into account, for some sequences it arrives at the same numerical results on source term as WAsil-1400. There was a general feeling that this result violates common sense, and so there was a struggle over the two-day period to see how this might have been the result of assumptions made, method of analysis, or method of interpretation.
Page Two i 2. I am one of those who felt that the lack of change of some of the numbers was not believable. I felt that the ability to understand the result was ham-pered by the method of development. It did not appear to be physically real, but had more of the character of numbers plugged into codes, leading to gener-ation of other numbers, without physical or chemical solidity. What was there about the physical and chemical models that made the result come out this way? What is the sensitivity of the results to input assumptions? I particularly felt the need for mass balances and energy balances, to lay a basis for physi-cal insight. I also suspect that many parts of the calculation are amenable to analytic treatment, and do not require computer treatment. I know of no better way to build insight and intuition than analytic calculations. A better physical and chemical understanding might lead to important changes in emphasis and results. 3. The draft report verifies that delay in containment failure reduces the source term. There is no apparent attention paid to the effects of water or the effects of oxydation - reduction condi tions . It can be argued (and was argued at the peer review meeting) that there is no such thing as a completely dry accident at a PWR. The original primary system water would profoundly affect the source term through generating a very wet containment through which fission products must move in AB, S2 0, or TMLB' events, and having a similar effect in the auxiliary building in ,V events. I believe this is true. The assumption that in some sequences the fission products go instantaneously from the primary system to the outside environment is reminiscent of WASH-740, and is certainly non-physical. 4. I believe that if the analysis is to be concentrated on a few low-proba-bility, high-consequence sequences, its name should be changed. The name im-plies a study of the source term without it being clear that the analysis is confined to extreme situations.
4 Page Three (
- 5. A strong effort must be made to reconcile highly disparate views of the source term being generated in several places. An early meeting is needed at which the different groups tell what they are doing and what they are finding i
out. The source of differences in results must be worked out and understood. q At the end, it will not be necessary that all parties agree, but it will be necessary that they understand the reasons for any remaining differences. i i I f I -1 4 4 4
., y.,. . , - , - _ . . - - -
yy , _ , . - - _ , , , , _ . . , _ . , _ . - . - , , - , , - - , -- _- , , , _
Y LEVY 5.LEW,INC. sue. ns 1999 South Soscom Avenue Compbett Coldomeo 95008 2233 USA 408/377 4870 February 8, 1983 Mr. M. W. Jankowski Division of Accident Evaluation, RES U.S. Nuclear Regulatory Conmission Washington, D. C. 20555
Dear Michael:
The following are my comments on the draft document NUREG-0956, dated January 14, 1983.
GENERAL COMMENT
S
- 1. The approach taken in NUREG-0956 is in the right direction. It attempts to recognize some of the physical processes which work to reduce
( radioactivity releases during LWR accidents. It is hoped that the effort will be extended eventually to recognize all such physical processes, and to bring appropriate realism to source terms.
- 2. NUREG-0956 emphasizes computer code development. It is apparent that the development of analytical models is way outstripping validation of such models. Comparisons to tests need to be added and, in particular, to reactor accidents. The TMI-2 accident offers an unique opportunity to obtain an integrated check of the various models proposed and it is urged that such a comparison to TMI-2 be included in NUREG-0956 to verify that the proposed models are realistic.
- 3. While !!UREG-0956 has made good strides in the mechanistic approach, there are still many processes which are not treated mechanistically. For example, the fission products release from the fuel is semi-empircal (Figure 5.2). Similarly, an empi source term for volatile iodides.rical Theapproach (TMI-2) most serious lackis ofused to specify a mechanistic modeling is in the containment. The failure modes, times of failure,
. location of failure go back to WASH-1800 methodology. A containment mech-anistic model is most essential. It should attempt to define the type and location of containment failure, its timing, cause, and size, and to take into account the physical processes of flow and fission products from their release point to the containment failure point.
- 4. There is substantial inbalance in the sophistication of modeling in NUREG-0956. As noted previously, the containment failure model is very
- elementary. Another good example is the very sophisticated model used for melt / concrete interactions. Yet this model uses an empirically prescribed rate of heat transfer between melt and concrete. M/9/ Mb4
_ - . - . . ~ _ - _ _ - l } EME Februray 8, 1983 i j i !
! 5. The MARCH code still drf ves the entire _ set of calculations being
- performed in NUREG-0956; yet even the version March 1.1 used in the draf t j issue leaves much to be desired. For example, j once melting starts to occur. MARCH does not recognize 1 the associated changes in heat and transport areas.
l This could have a substantial impact on fission products i release and thermal-hydraulics.
+
core slump is assumed to occur when 75 percent of the jl core is molten
, i
. no release of fission products is assumed after the time of core slumping and until core-concrete interactions j begin. An improved model in the bottom head which provides for steam generation and sparging of fission products by such steam would be helpful.
- 6. It would be very useful_ to_ point out where NUREG-0772 and
,NUREG-0956 are the same and where they are different and to explain any
{ differences. )i 1 SPECIFIC COMMENTS ABOUT SEQUENCES i l. I 1 j 1. As noted before, a TM!-2 sequence should be added.
- 2. During the AB sequence, the break location was selected to obtain j a minimum pathway of release from the primary system. For the 5,0 sequence, 4
a long pathway through the steam generator was selected. In fact, the I break location for large and small breaks can vary and the most probable ! source terms for such breaks would result from averaging all such possible
! locations. A simplified way_to deal with this problem is to carry out cal-j culations for A and 59 sequences for minimum _and maximum pathways and to i average the source terms. Such an app ~ roach or an equivalent one must be l implemented or the comparisons to WASH T460 would not be_ _ appropriate. I l estimate that ali the values calculated in NURN 6B6 for AB sequences 1'
would decrease by a factor of about 2. On the other hand, the values cal-culated for 5 0 2 sequences will increase and may not compare as favorably to WASH-1400
- 3. The use of a flat power flux reguires the use of an average power density and in particular an average fuel bur _nup rate to be physically con-i sistent. Yet NUREG-0956 uses a 71at flux and a peak bundle exposure of i 33,000 MWD / ton (seepage 6-22). It is suggested that a non uniform flux -
) and burnup case be carried out for sensitivity purposes. ' 1 i
- i. .i i
1 i l
i
)
l S. LGVV, if0C. February 8, 1983 l
- 1 SPECIFIC COMMENTS ABOUT THERMAL-HYDRAULICS t
- 1. NUREG-0956 provides improvements in this area, but it is still a patch. As noted on page 5-3, a distributed reactor coolant system model ~
j for use within the MARCH code is preferable. j 2. In several sequences, melting of the upper plenum takes place j and its impact on creating bypass flow areas needs to be recognized and j may be just as important as the impact of upper plenum details, i i SPECIFIC COMMENTS ABOUT TRANSPORT MECHANISMS i l 1. The reduction in source terms in NUREG-0956 appears to come I primarily from improved aerosol treatment and dynamics. An Appendix dealing with such aerosol dynamics is recommended. The Appendix should j describe the models, bases, validation, and uncertainities. It is apparent that particle size distribution is important and it is not clear why the
. values used in NUREG-0956 are justifiable. Furthermore, this is one area j where separate effects tests can help reduce uncertainties and it may deserve attention.
i
- 2. When fission product aerosols settle in the core or primary ,
4 i system, they are presumed to remain there af ter pressure vessel failure. . l It is not clear w1y some are not released or rejoin the molten material ' l if they settle on areas which eventually melt. ) 3. The condensation model needs to be examined and to take into account that there may be heat production in some of the aerosols. EDITORIAL COMMENTS i i j 1. An introductory section with schematics as utilized at the i meeting might help reader understanding.
- 2. I was confused by code interfaces and by which code was used in fact.
l I am sorry that my comments could not reach you sooner but I have j teen out-of-town since the meeting in Washington, and I hope that they 4 satisfy yours needs. 1 t i Very )trul .yours, e j . . ! D j Salomon Lev l President j , SL:jm I cc:M. Silberberg NRC i L. S. Tong ) -
$.0$l5A i
j Coments on Draft NUREG-0956 USNRC Peer Review Meeting, Jan. 25 and 26, 1983 l (Reviewer: S. K. Loyalka) 2 ] The stated purpose of this report is to: i (1) Develop updated release-from-plant fission product source terms for I four types of nuclear power plants and for accident sequences giving a range ! of conditions. The estimated source terms are to be based on consistent step-by-step analyses of fission product release from the fuel, and transport and
- deposition by using improved computational tools and performing the analyses in a consistent step-by-step manner.
(ii) Determine the effects on fission product release associated with { major differences in input parameters associated with plant design and accident sequences. (iii) Provide in-plant time and location dependent distributions of fission product mass for use in equipment qualification considerations. It is not necessarily the intent of this work to provide an all
, encompassing definition of source terms, but rather to make best estimates of s source terms for a range of conditions. The purposes are to be realized i basing analyses on:
1 i (a) The best available techniques , j (b) A level of detail consistent with current knowledge of pertinent j physical processes. l i It is expected that the present analyses and estimates would provide an
; indication of the conservatism inherent in current source term assumptions and 4
guidance for the development of new source terms. My specific coments on the i clarity of the report (Volume I), its technical content, and the extent to which it has realized its purposes are given below. 4 i (1) Clarity of the Report: l . i The report is well organized and provides sufficient details regarding j the purposes and the general approach. Figure' 3.1 and the discussions on i uncertainties (lack of basic information, paces 3-3 to 3-5) provide good j perspectives of the report. Perhaps it should be clarified on pages 2-2 to 2-3 that the purpose of
- the report was to use not only the state of the art techniques, but also l improved models and computational techniques to the extent possible (see last paragraph on page 7-5).
t l Since the report has used some new computer programs (MERGE, CORC0R, i substantially modified versions of TRAP-MELT, TRAP-COND), the authors have I
~. . , _ _ _ , . . _ _ , , _ , . _ _ _ . . _ , , , , _ , - . _ . . , , _ . . . . _ _ _ . . - _ . , . . _ , . . _ , _ , . . . _ . _ . _ _ - . , _ . , _ _ . . _ . , _ . - . _ , - _ . . _
attempted to describe details of these programs. In general, the descriptions are quite appropriate, and it would be useful to indicate when and how these programs would become available for distribution and verifications. Are these programs " frozen" or would they undergo revisions in the future? I think the results chapter (7) discusses the purpose (1) in detail, the purpose (ii) reasonably well, and the purpose (iii) rather indirectly. The authors may wish to include in this chapter some additional results corresponding to the purposes (ii) and (iii).
- 2. Technical Content (1) The choice of the plants and the accident sequences (page 3-2, chapter 4) appears appropriate. It would be useful to indicate here the probabilities (per year of reactor operation), and associated uncertainties, of the sequences also.
(ii) On page 3-5, some significant uncertainties that warrant further evaluation are listed. It might be emphasized that in some cases not only the experimental data are lacking, but the t goretical (mechanistic) bases are not adequately developed either. The authors might reference the report NUREG/CR-2629 in this conjunction. It may be clarified that quantification of some of the uncertainties would require long range and sustained research efforts. It is possible to include several additional items in the list on page 3-5 (for example: Hydrogen combustion, models for physical processes such as agglomeration, condensation, evaporation, reaction rates, etc.)--but since the list is not supposed to be exhaustive--this is not too important. (iii) Chapter 5 contains good descriptions of the MARCH, MERGE and CORSOR Programs, and their capabilities and limitations. The transportprograms(TRAP-MELT, QUICK, TRAP-COND, CORRAL,NAVA-4) are described in considerable detail regarding their computational philosophies. While these programs certainly represent the state of the art, and even considerable improvements upon it, there is a
" contraction of description" in that many assumptions and
" averaging" processes are used. The extent to which these approximations affect the results would be determined only after long-range and careful future analysis and experimentation.
Recently, I had the opportunity to write a paper " Mechanics of Aerosols in Nuclear Safety: A Review" which is to appear in the journal Progress in Nuclear Energy (May, 1983). I am including a copy of this paper with the present review--the authors might find this assessment of the state of the art and corresponding observations of some interest. ~ - . _ -. . . _ . - . . .--. -
.I 1
! (iv) . Chapter 6 essentially provides results of the MARCH, MERGE and ,
- CORCOR calculations (used as drivers for the transport codes). '
i These results are subject to uncertainties that the authors have
- discussed in considerable detail. Perhaps it might also be 4
mentioned that in the present calculations, the coupling of the transport codes with the above codes is one way and not mutual.
- Concerning sources of radionuclides (section 6.2); while the j ORIGEN calculations should be quite good, it is difficult at this
, time to know if the aerosols produced can really be characterized by a " fixed composition" corresponding to the releases. l Assumptions regarding the primary particle size distributions 4 appear reasonable--and while improved models will become available ! in the future--the assumptions are representative of the st te of the art. I ! The discussion on page 6-31 regarding the volatile /non-volatile i iodides is of interest. Assumptions appear reasonable, and ] reflective of the current knowledge.
! (v) The chapter 7 contains results for a variety of cases of interest
! (I believe some details--7.3.3, 7.3.4, etc. would be worked in at a later date). The tables 7.14 and 7.15 indicate that for some
. sequences the WASH-1400 results are overestimates of Iodine and i , Cesium releases by a factor of 10 or so, while for others the l present estimates are not so different from the WASH-1400 i estimates. It appears reasonable that containment is the main
! retention mechanism in the AB, TMLB' and S2 0 sequences and that in { the S2D and V sequences the Primary also contributes to retention, i i I would like to suggest that the " attenuation" and " retention"
; factors be clearly defined so as not to cause any confusion ! regarding their precise meaning. Also, is it planned to include I results from the TRAP-COND Code in these tables?
{ 3. . Realization of the Purposes l The report provides good accounts of the computational programs and procedures used in the calculations. The results reported are consistent with
, the stated purposes. The autho.rs have based these results not only on the I
state of the art, but also several improvements carried out during the course of their work. l Quite clearly, the reported estimations are subject to uncertainties and
; are affected by lack of basic knowledge regarding the basic processes relevant i to physical, chemical and mathematical aspects of the accident sequences. It
!~ would take sustained, long range efforts to resolve many of the questions, and obtain quantifications (and qualifications) of the uncertainties. The present , i a basis for a consistent ! report, step by however, step approach, should
- andbe very useful in understanding of providing'e the ro of approximations in the I l WASH-1400 estimates.
1 I .
, - . - , . __~ , , . . . . , _ _ _ . , _ _ - _ _ , , . . . _ , , _ _ _ . _ , _ . . _ , _ _ _ . , . . . _ _ . , _ _ , . __ ._,, _.._...,,_._~,. -v _ - -
/CEA1CZYK
, OAK RIDGE NATIONAL LABORATORY optmAtto av UNION CAR 81DE CORPORATION NUCLEAR DIVl310N O
POST OFFICE soE X OAK RIDGE. TENNESSEE 3703s February 3, 1983 Mr. Melvin Silberberg, Head Fuel Behavior Branch U.S. Nuclear Regulatory Commission 1 7915 Eastern Avenue Mail Stop 1130 S.S. I Silver Spring, Maryland 20910
Dear Mel:
The draft of Volume I of NUREG-0956 includes some valuable new work by BCL and SNLA on LWR accident source term estimation methods. In particular, the work on the thermal-hydraulic and fission product transport considerations 3 in the primary coolant system and the updated core-concrete interaction model represent substantial improvements over previous available procedures.
, However, there are some potentially important source term issues which the draf t does not address.
Among the technical issues which deserve to be treated more completely are the following: 4 1. The variations among nuclear plants, e.g., the variations existing among large containment PWRs, need to be taken into account. As was noted at the recent review meeting for the subject draf t, the impacts of differences in both plant design and operator procedures upon the probabilities of various accident sequences can be important. But possibly just as significant, and not mentioned, are the large poten-i tial differences in source term magnitudes for any specific sequence caused by some plant-to plant variations. Many of these effects could be estimated by application and/or adaptation of currently available procedures to a variety of conditions, e.g., different core sizes and compositions, various containment sizes, and a range of concrete com-positions.
- 2. The effects of high pressure in the primary coolant system upon the release rates from the core materials should be considered. It seems inconsistent to worry about some of the other effects of high pressure while ignoring that effect. The basic effect should be treatable by consideration of appropriate processes in certain metal and chemical production industries.
- 3. The model sensitivities and other sources of uncertainty in the esti-I mated source terms need to be consistently and thoroughly addressed.
Lack of such consideration in the subject report hampers interpretation and evaluation of the results. f7()Q(3 D5Y
Mr. Melvin Silberberg February 3, 1983 J , 1 4 In addition to the technical issues, there are some other issues which are important. If the goal of the subject report is to produce source terms which can be factored into regulation, then the perspective in which the source terms will be used needs to be clearly stated for the reviewer for at l 1 east the following reasons:
- 1. Source terms appropriate for realistic or real-time purposes and source
- . terms appropriate for probabilistic risk assessments can frequently be expected to be somewhat dif ferent. Whereas a best estimate in a mecha-l nistic framework is indicated for realistic purposes, e.g., the deci-sion on whether or not to evacuate during an accident, a source term involving much less detailed analysis but more emphasis on balance and consideration of the rest of the accident spectrum may of ten be adequate i for risk assessment purposes.
- 2. By definition, any source term depends in part upon the population-at-1 risk, e.g., the public or the equipment inside the containment. Inas-much as even "best estimate" source terms for risk assessments typically j have some biases with respect to the public incorporated into them, it should be noted that models and assumptions which result in best esti-1 mates of source terms for the public do not necessarily result in best j estimates of source terms for other populations-at-risk.
I 4
- 3. As was noted previously, accident source terms are plant-dependent.
]' , Regulations, however, have not taken these differences into account but probably should. l To a large extent, the foregoing problems in the area of LWR accident source q term estimation are long-standing ones and cannot be solved overnight. How-ever, the potential impact of the "0772" follow-on effort on regulation seems to require that at least the effects of plant-to plant variations and
] an assessment of the uncertainties be addressed by that effort.
Yours truly,
,~~)
y f.J.Niesc& i zyk Chemical Development Section Chemical Technology Division SJN/bcd cc: A. P. Malinauskas R. R. Wichner ' R. G. Wyser (2) W. F. Pasedag, NRC R. C. Denning, BCL J. A. Gieseke, BCL , SJN File l l 4 1 l l e- . - - ~ - - , - - . . , - - ,- -. , . - . v. ... . . L. , ., --..,,.,,,-- - - .-.---.-,n.,-- -n- ,,--. .-7.,- , - . , 2
PEWIM
, , l l
I
. M February 15, 1983 Scionee Apphcetiene, gas, 1
I Mr. Nel Silberberg U. S. Nuclear Regulatory Commission 1717 H Street, N. W. Washington, D. C. 20555
Dear Mr. Silberberg:
Thank you for inviting me to participate in the Peer Review of NUREG-0956. The area that I feel most qualified to comment on is the behavior of radio-iodine in containment. I have also been involved with measurements and analysis
~
at TMI-2. Specific connents are the following:
- 1. From the PRA point of view, it is doubtful that volatile iodine species would be a major contributor to risk. However, they should not be
- I ignored as was suggested by some at the Review, because these source terms are used for accident planning and equipment qualifications.
- 2. It seems to me a matter of faith that all iodine in CsI will remain in the ionic form. A fraction will be converted to volatile forms. The fraction 0.054 (page 6-31) may be correct. At least it has some basis in measurements.
Contrary to the belief expressed on page 6-31 there is a model available to estimate volatile species frcen one source at least. It is the model used in reference 6.7, which is our analysis for DOE of iodine behavior at TMI-2. The analysis is based on iodine being dissolved in the condensation layer and subsequently transferred to building surfaces using empirically determined trans-ferred coefficients. The iodine is then converted by surface reactions to volatile forms. The best descriptien of the conversion process is that on page ! 492 of the U. K. report, "PWR degraded core analysis", ND-R-610(s) April 1982.
- 3. The origin of the value of the generation rate fgr " volatile iodides" is not clear. Based on our experience, a rate of 2 x 10' fraction / hour given on page 6-31. The value determined from TMI-2 measurements is 6.8 x 10-9 i sec . The inventory on which this rate constant operates is a function of the quantity which is assumed to be made airborne during the accident. The smaller the amount made airborne the smaller the inventory involved in generating volatile 3 iodine.
- 4. Dave Campbell of ORNL suggested that tellurium had been measured in sump samples at TMI-2. I was unaware of any reliable measurements of telluritan in the sump but it had been some time since I'd seen any data from TMI. However, 850H 3 N N Nuc Environmentsi Services, a division of Sc/ence App //cetions, Inc. 3 Choke Cherry Rd., Rockville, Md. 20860, (301) 977 4400 oth , sAs o+n : Aen . Aan A,mer, A,nage.a. Ais.a., sees a. chicese, Hone.vene, i.e Jen , i. Aness . Pee Atee, seni. e., mere, sveny se. and Tues.
\ _- _ - - - . . . . - .
_ _ . . -- . . _ - _ -__ __m . _ _ _ _ _. 4 i February 15, 1983 Mr. Mel Silberberg 4 Page 2 the day after the Review, I had occasion to be at THI, and asked several persons in a position to know about such measurements. They are unaware of any such results. In conclusion I would like to add my vote to those who want to see an analysis made of the accident at TMI. The data from the accident are not the best, but if used properly should provide some verification of the models. Sincerely, SCIENCE APPLICATIONS, INC.
'LA : h.
Charles A. Pelletier CAPavr cca R. Ritzman, EPRI R. Vogel, EPRI F. Tooper, DOE i I I i i l 4 t j 1 I r i 6 I l
..-- -. . - _ - . _ _ - . . - - - , , - - . ,---_ .,-,,,-,,,,...-n,__, ,-, , _ - , - . _ - , . - - , . , _ - - - , - , - - - - . _
. - . _ . -= . _- .- -- .-. . - .
i (&& 0$ ' gg l j , %% t 4'M I Westinghouse Water Reactor ausss PlffWChPennsylvan6a ls230
; Electric Corporation Divisions
- NS-EPR-2732 March 16, 1983 1 Mr. M. W. Jankowski Fuel Behavior Branch Division of Accident Evaluation U.S. Nuclear Regulatory Consnission 3 Nicholson Lane Building 5650 Nicholson Lane Rockville, Maryland 20850 Dear Mr. Jankowski-I We have reviewed the NRC document, "Radionuclide Release Under LWR Accident Conditions," NUREG-0956, Draft, as requested by your letter of Janue-v 14, 1983.
Generally, the document represents a useful update of the fission pi :uct transport and deposition estimates for core melt accidents with respect to those presented in the earlier NRC Report, " Technical Basis for Estimating Fission Product Behavior During LWR Accidents," NUREG-0772,1982. However,
; the draft document reveals many important omissions as a result either of
. calculations that were not completed at the time the draft was issued, or failure to consider some of the important deposition steps. We urge that these omissions be treated and appropriate estimates for them be included in 3 the next draft of the document.
I Generally, the document considers the potential for radionuclide retention during the two principal transport steps, transport of released radionuclides
- from the core region through the reactor control system, and transport and i
deposition in the containment. Westinghouse has recently produced estimates I of fission product retention during transport in conjunction with UKAEA as i; part of the estimates of risk from core melt accidents for the Sizewell-B plant (proposed for construction by the Central Electricity Generatinq Board). These estimates are reported in the UVAEA document, SRO-R256, November,1982. Generally, estimates of fission product retention are similar to those in NUREG-0956 for sequences where the same transport steps are considered (note that the results in SRD-R256 are reported in a different format than those inNUREG-0956). The results in SRD-R256 show that there is a significant dependence of the magnitude of reactor system deposition on the location of_ breaks in Dioina systems (hot leg or cold leg). Results in NURifG-0956 also show this dependence even though such a conclusion is not drawn. For example, the large hot leg break sequence (AB) shows little retention while there is substantial deposition for the small cold leg break (S) sequence. Parallel consideration of both hot and cold leg break retention for large break (A), small break (S), and containment bypass (V) sequences is needed in NUREG-0956. I i
Mr. Jankowski Page Two With regard to the TMLB' sequence, the sequence does not appear to have been treated in enough detail. Sweeping of the vessel controls by steam when the core debris material slumps into the water in the bottom of the vessel needs to be considered with concurrent potential for deposition within the reactor system. The assumption of essentially instantaneous vessel melt-through at core slumping is unduly conservative. Potential for fission product retention in the relief line downstream of the pressure relief valves and in the quench tankalsoneedstobeestimated(seeSRD-R-256). With respect to containment transport and deposition, the improved aerosol models employed in NUREG-0956 show retention in the containment to be markedly greater than earlier predictions with the CORRAL code for sequences in which containment failures does not occur for a number of hours following fission product release to the containment. These more realistic calculations showing greater retention are encouraging. Large releases to the environment are still predicted in NUREG-0956 for sequences in which containment failure occurs soon after fission product release to the containment. However, recent studies for large dry pWR contain-ments in which Westinahouse has participated (Zion, Indian Point Sizewell-8), avihoitne probability of early containment failure to be markedly lower than Diose for delayerfaTTure. We urie that thiNRc report provide tTe needed-- pTrspective with respect to the relative magnitude for early and delayed , containment failures so that the significant role of the containment in reducing radionuclide release to the environment and minimizing risk will be properly recognized. The role of the auxiliary building or other secondary containment structures in reducing radionuclide release to the environment has not been adequately treated in NUREG-0956. The treatment of auxiliary building retention for the V-sequence in NUREG-0956 is brief, difficult to follow and the assumed building parameters do not appear to be proper. Potential for holdup and retention in secondary structures also needs to be treated for the failure to isolate containment sequences (8 failure mode). We urge NRC give needed attention to this important radionuclide retention step. In summary, NUREG-0956 provides a useful update to earlier NRC estimates of fission product retention for core melt accidents. However, significant addi-tional work is needed to place the limited calculations in perspective. We urge a careful review of the recent*radionuclide transport study done for Sizewell-B. While the estimates in that study are not based on an extensive set of computer calculations, they do represent a more complete consideration of the effects of overall system, containment, and secondary structure configurations, : and of break locations on the potential for radionuclide retention during ! radionuclide transport following a core melt accident than presented in NUREG-0956.
l Mr. Jankowski ! Page Three ;
. l 1 .
j Attached are specific comments developed during the review of NUREG-0956. The comments transmitted include those of both our assigned peer reviewer, Dr. D. H. Walker, and of the Westinghouse observer at the peer review meeting,
- G. T. Rymer.
! Very truly yours, i
N l E. P. Rahe, J . , Manager Nuclear Safety Department t
)
i /kk cc: M. R. Hayns - UKAEA l - i ) i 5 }, i i 1 i l i 1 i i
.__ -- . - -_ __ ~ -- . .-_ __- -. - - - --- -
I j NUREG-0956 COMMENTS li SPECIFIC COMMENTS ] p.3-1, last paragraph - Recent risk studies have shown that the Surrey plant, utilized as the PWR for analysis in the RSS(l), is not typical of current generation PWR's, with large dry containments, in a number of ways important to the radionuclide containment ' function including design and reliability of containment heat removal systems and pressure containment capability (2,31 While the desire to retain Surrey in these analyses for comparison to the RSS results is commendable, similar analyses for a current generation large dry PWR should also be performed. p.3-3, second full paragraph -
! Current risk studies have shown uncertainties in the magnitude of the radionuclide release to the environment for core melt accidents is one of the major uncg tainties associated with overall estimates of risk for a nuclear power station (2,3,s(i.
i Certainly some estimate of source term uncertainty is an essential next step to permit meaningful appitcation of the NUREG-0956 results. i 'l I p.3-5 and 3-6, list of uncertainties e
! Item (1) -
need to indicate which way could the results be biased? i ' j Item (6) - integral experiments to validate primary system transport models do not appear to be forthcoming, unless perhaps LOFT is used for this purpose. What does NRC believe is the implication of there likely being no experimental validation of the primary system transport models? 4 ftem (10) - TRAP-MELT parametric calculations to determine the importance of deposition velocities to overall release to the environment for important sequences are needed. i If the effect on release is large (orders of magnitude). TRAP-MELT calculations are not very 1 useful or neaningful until deposition velocities are determined. , p.4-10, last paragraph - 4 i The degree V-saquence. of retention in the auxiliary building should also be considered for the {i p.5-12, 1st paragraph -
~
Statements at the end of the paragraph need to be softened. For some PWR's, V-sequences are dominant contributors to risk. For such sequences, source material . released ex-vessel is not efficiently transported to the environment. Also for 4 I i
l i I i { some PWR designs, the ex-vessel debris will likely be cooled and certainly covered by water.
- For such designs, release to the environment may not be dominated by fission product failure sequences. released after vessel melt-through even for delayed containment Section 7.2.2 Deposition for TMLB' -
f 1. Studies done for the Sizewell-8 PWR indicated steam generation, as the hot i core debris migrates to the lower plenum, is sufficient to sweep the radio-i nuclide inventory from the vessel, into the pressurizer and out the pressurizer relief valve before vessel failure (see reference 4). Judgment would also lead one to believe that migration of core debris material into the water pool in the lower head would occur over a period of time such that sufficient steam would be generated and time available (of the order j of 10 minutes or more) for sweeping of the pressure vessel and pressurizer of released fission products and aerosols prior to vessel melt-through.
- 2. For such a sequence, radionuclide deposition in the relief line between the relief valves and the quench tank needs to be considered. Deposition i
could be substantial (see reference 4). i 3. Deposition in the quench tank should also be estimated for the TMLB' sequence both for cases where water is still present in the lines and for I- cases where the tank is emptied (tank surface temperatures are likely to i be cool). I i j V-Sequence, Tables 7-11 and 7-12 - l } Tabulated reactor system fission product retention values are for the cold
- I leg (discharge) side V-seque.1ce which was the type of sequence considered in the RSS. With the system changes incorporated in most designs to enhance check i
j valve testing, hot leg (suction) sit'e failures exhibit higher frequencies for i many current generation PWR's (see references 2 and 3). For such sequences, the steam generators, with their cool surfaces and potential for radionuclide retention, are not part of the flow path and hence predicted reactor coolant system retention is reduced. 1 Section 7.3.2, TMLB' sequence - t This section discussed the containment failure modes, they being an early failure i at 69 psi (y) and a delayed failure mode (42 hours) at 100 psi (o failure mode), i The phenomena producing the 89 psi pressure spike is not identified. If the pres-i ! sure spike is due to hydrogen burn, the occurrence of the spike is unlikely since concurrent release of steam and hydrogen in this sequence is likely with i rosultant steam inerting of the containment. at 89 psi is low. The report should clearly Further, the probability of failure ' i identify that the likelihood of the j' y failure mode is much less than that of the 6 failure mode for this sequence. l l 1
~~ ,r---,--.,wn n-. - - -r , -------w---,--,e-.. e--, -- . - - - - - - - -- w----- , , - - - - , - - - - --,,,---,--,,re,e,ee.,-e,- en -n.. e ~~--r--c------
4 Section 7.3.4, V-sequence, 1st paragraph - This section indicates that vaporization release will be treated in more detail in the next set of calculations. When this is done retention of the vaporization source within containment should also be considered since this source material will likely mix with the containment volume at large and be deposited rather than immediately migrating back through the melt hole in the reactor vessel and being transported to the break. EDITORIAL COMMENTS
- p. 6-33 -
Meaning of entries under slab at bottom of page are unclear.
- p. 6-36 .
j Were HIM and H10 intended to be the same ? l Equation 'for e (tensile stress of vessel) is unclear. i ) p. 6-37 - l t i For entries (2) and (3), Surrey is not typical of current large dry containment i PWR's. i
- p. 6-42 -
Quoted volume for auxiliary building seems too small by about a decade. Quoted areas also seem low. Was only part of a building or a compartment considered?
- p. 7-26 and 7-27 -
Containment leak rates of 1%/ day are used in the conparative calculations. Design 3 J basis leak rates in the range 0.1%/ day or more are typical of large dry containments. i p. 7-28, 2nd paragraph - i Words should reflect the need to compare Tables 7-16 and 7-17. Generally writeup 4 on this page needs editing. GENERAL
- 1. Recommend on temperature system. *C or *F, be selected and used throughout
- the report.
i i
REFERENCES
- 1. " Reactor Safety Study, An Assessment of Accident Risks in U.S. Commercial Nuclear Power Plants," U.S. Nuclear Regulatory Commission, WASH-1400,
; (October 1975).
- 2. " Zion Probabilistic Safety Study," copywrited by Commonwealth Edison
, Company, Chicago,(1981).
- 3. Indian Point Probabilistic Safety Sutdy, copywrited by Power Authority of the State of New York and Consolidated Edison Company of New York, Inc.,(1982).
- 4. M. R. Hayns. F. ' Abbey, P. N. Clough, I. H. Dunbar, and D. H. Walker, "The Technical Basis of ' Spectral Source Terms' for Assessing Uncertainties in Fission Product Release During Accidents in PWR's with Special Reference to Sizewell-8," SRD-R256, UKAEA, November, 1982.
- 5. cident J. T. Larkins Research Plan,"and M. A. Cunningham, NUREG-0900, " Nuclear V83, p. 5-55, Power U.S. Nuclear Plant Severe Regulator, *; Commission i
?
l 4 d j 2 Y
RETzNnAt February 2, 1983 A Dr. Mel Silberberg U.S. Nuclear Regulatory Commission 7915 Eastern Avenue Silver Springs, MD 20910
SUBJECT:
Comnent on NUREG-0956 (DRAFT)
Dear Pel,
The Commission and Staff are to be congratulated for holding the peer review of the referenced draft report last January 25 and 26. I appreciated the opportunity to serve as an invited member of the review group. Although connents were offered for the record at the meeting, it seems appropriate to submit further comments of a clarifying or amplifying nature in writing as follows: (1) The amount of work accomplished in assembling the calculational tools and in perfonning the analyses presented in the report is very impressive. Battelle-Columbus Laboratories staff who participated in the effort are to be commended for their diligence. l (2) The mass of ntsnerical results generated during the analyses and given in the report require considerable time to examine, digest, and compare with other calculations. This process could not be completed in the brief review period and thus the comments given here are incomplete. Consequently, the opportunity to review the work again in April is welcome. (3) The additional analytichi method and newer computer codes applied to the source term problem are noteworthy and of course determine the results obtained. I concur with other reviewers at the meeting who recomnended that a detailed description of this new methodology should be a part of the project documentation. (4) The core / concrete ra'd ionucl id e release and aerosol generation model developed by Sandia staff for this study appears to be a clear advance over previous methods used to specify source material releases for this portion of severe accident phencmenology. Early publication of the details of this work should be encouraged. l ggy s g ,cience Applications, inc. . . .... s,. . s.. 2oo. . . ... c, i. . . 2432.
, 1 Dr. Mel Silberberg February 2, 1983 l l l (5) A coment made with respect to the NUREG-0772 report should l be re-emphasized here. The empricial fission product and structural release model from NUREG-0772, which is used in the CORSOR code, is of very doubtful validity when core melting has progressed to the stage where slumping is occurring. The model will lead to an erroneous prediction of the composition of released structurals and probaby sn overestimate of the amount of aerosol generated within the reactor pressure vessel. A thermodynamic based vaporization model would be preferable which would include contributions fram control rod materials as well. (6) Use of thermal-hydraulic information from the MARCH 1.1 code in these more mechanistic source term release and transport analyses places undersirable limitations on the work and leads to over conservative and even erroneous results. An example of this latter tendency is the magnitude of the steam spike that is predicted for the so-called TNLB' sequence at reactor pressure vessel failure as shown in Figure 6.8 of the report. It is suspected that the leakage rates calculated by MARCH 1.1 for the containment in sequence AB are also artificially high. Improvements in t' this important area of accident analysis is needed if the source tenn definition effort is truly to be regarded as a "best estimate" endeavor. (7) It should be noted that SAI analyses of upper plenum temperatures for a PWR during core uncovery, heatup, and fuel melting in severe accidents, which have been performed for EPRI, produce results that generally lie between the Case 1 and Case 2 predictions obtained by Battelle using the MERGE code. Since we also used MARCH 1.1 to define the characteristics of the gas flow that would be entering the upper plenum in such accidents, it is not too suprising that similar heat-transfer results were obtained. (8) The listing and description of the numerous uncertainties in analyses and procedures presented in the report in Section 3 needs to be highlighted' for all readers of the report. I personally agree that each of these items represent significant sources of uncertainty or error. Perhaps readers could be directed to this section in the abstract, sumnary, or introduction to enLnce the probability that proper qualifications on the results will be registered in a timely manner. 4 s
. - - .,. , - --- _y
Dr. Mel Silberberg February 2, 1983 i I trust that the above coments will be of help in completing the next iteration on the report. If questions arise or clarification of any particular point contained in the comnents is needed, please contact me. { Very truly yours, Y. Y ;': (t-t r R.L. Ritzmen RLR:kar cc: Z.T. Mendoza (
. l I
l . . - - -
REYhetos COBtENTS ON PEER REVIEW OF NUREG-0956 j A. B. Reynolds ! January 31, 1983
- 1. Need for Simultaneous Development of Four Topics Related to the Source' Term _
I see a serious danger in NRC's present plan to publish a large new report on calculated radioactivity releases in relative isolation from several essential interrelated topics--danger to NRC's potential future position in the area of saurce term and danger to the creation of a reason-able perspective in the minds of the public in this area. I view NUREG-0956 as such an isolated report. I recommend that four topics be developed simultaneously to avoid this danger, namely
- 1. Data Base
- 2. Radioactivity Releases
- 3. Containment Failure
- 4. Probability Risk Analysis NUREG-0956 covers only the second topic. Either separate reports on each topic should be issued simultaneously or each topic should comprise one I volume of a four-volume report.
The data base, which was the dominant topic in NUREG-0772, is hardly mentioned in NUREG-0956. An updated review of the data on which Battelle's calculations are based needs to be published. A dc-scription of experiments in progress and planned and the data expected from those experiments should be included in the data-base reports. The Battelle-type analysis of radioactivity release is complex enough to warrant a separate report devoted solely to it. (This is in disagreement with some views expressed at the meeting to add data-base material--like chemistry--to NUREG-0956.) Each input quantity in the Battelle analysis should be referenced specifically to data provided in the data-base report. The peer review meeting brought to light the enormous importance of The early containment failure in developing a perspective on the source term. question is crucial enough to require a study and review comparable to the data-base review and the radioactivity-release analysis. Some industry observers at the meeting apparently believe that the early failure in the AB and TMLB' sequences (i.e. A:-y and TitLB'-y--or TMLB'-early 6) cannot occur. Or if they can, there will surely be a vigorous debate on the proper pro-bability to attach to such failures. The radiation release report should not be published until NRC developes a position, also subjected to peer review, of the probability of early containment failure. It should be possible for NRC to develope an interim position based on current knowledge, with the 1
recognition and acknowledgement that full resolution of this issue may not occur for several years--perhaps until after the Sandia containment experiments and further work on hydrogen burn pressures in AB and steam spikes in TMLB' sequences. NRC's position on the entire source term question must remain interim to the same extent that NRC finds their position on containment failure to be interim. The fourth topic is the PRA which, in addition to probabilities of events in each sequence, must use the results of the radioactivity-release /- analysis and the containment-failure information. The peer review committee did not have " Chapter 8" of NUREG-0956 which we were told would be forth-coming and would contain PRA. I am very uncomfortable with this situation. The PRA is being treated like an afterthought, not subject to the same peer review as the release calculations. I share the concerns of some of the industry observers that NUREG-0956 is concentrating only on release results, and that publication of these results without appropriate accompanying PRA results will create an unrealistic perspective of the source term issue in the minds of the public.
- 2. Status of Mechanistic Approach
' The mechanistic approach being proposed by the NRC for calculating fission product releases for specific accident scenarios is commendable.
The methods being used by Battelle appear to be the best currently available. The NUREG-0956 results are interim values in a field that is still evolving, however, both with regard to data base and methods development. This status should be emphasized more strongly in the report. The two sentences devoted to this caveat in the middle of page 2-2 represents insufficient emphasis. If for reasons unexplained to the peer-review members the Commission insists on publishing new release values by the summer of 1983, the status of more advanced codes like SCDAP, MELCOR, etc. (which were not discussed at the peer review meeting) and experimental programs like Marviken, DEf10NA, etc. should also be discussed in the report to provide perspective on the continuing evolution of the field. e 8
i
- 3. Disagreement Between NUREG-0956 and Industry Results A number of representatives from industry appeared to disagree that radioactivity releases as high as some reported by Battelle for the AB-S and y, TMLB'-y, and V scenarios are possible. I recommend that NRC request the appropriate computer codes from industry and have NRC contractors compare the methods and results of Battelle and industry groups at each possible step in the accident scenarios. Requests should be made of EPRI, IDCOR, and Stone and Webster. If these groups do not provide NRC with their codes, or if alternative methods are not really available, then NRC could more effectively discount the objections to NRC results.
This comparison is especially important with regard to condensation l phenomena in the auxiliary building in the V sequence. The Battelle methods show superheated steam entering the auxiliary building with the fission pro-ducts and little attenuation there. Stone and Webster calculations show prolonged condensation (i.e. a " rain forest" for 3.5 hours) in the auxiliary building with an accompanying high rate of fission-product attenuation. The Stone and Webster calculation has been well publicized and 4-has an effective forceful proponent in Ed Warman. There is enough plausibility the Stone and Webster scenario that, providing Stone and Webster will makv their . methods available to NRC, I think it is imperative that NRC have its con-tractors make the necessary comparisons to settle an issue of this importance before publishing results of its own. A second important area of disagreement is the magnitude of a steam spike in the TMLB' sequence shortly after vessel melt-through. Reference was made several times to Henry's analysis. NRC should initiate a detailed comparison of Henry's model and the models in MARCH due to the high importance of the steam-spike phenomenon on potential early containment failure with accompanying high release rates. 9 I
- 4. More Specific Comments e The Battelle models should be modified to include fissien product decay heat on structures from fission products deposited on them--
in the upper plenum and in the piping, e I was unconvinced that flow patterns in the containment building have little effect on fission product retention or behavior-- which was the impression I got from Lee's response to Ginsberg's question. Also the discussions of diffuseophoresis--both the lack of its treatment in the fiAUA code and the reference to German data indicating its importance--left me with the impression that this phenomenon had not been adequately investigated. I e Behavior associated with the high-pressure ejection of the molten debris upon vessel melt-through in the TMLB' sequence was not , specifically treated. Is the subsequent scenario influenced by the uncertainty in the aerosol particle size distribution from this ejection? e Regarding the steam spike in the TMLB' sequence, consideration should be given to the melt falling into water in the reactor cavity (or sump) in addition to the reverse. The reverse case, treated in tiUREG-0956 as I understand it is water from the vessel and accumulators falling onto a melt which has fallen into a dry cavity or sump. Had the TMI core continued to melt and melt through the vessel, my understanding is that there would have been water present below the vessel at the time of melt-through due to overflow from the drain tank. An observer at the meeting suggested that the basis for Henry's argument against rapid enough steam generation to produce a steam spike large enough to threaten containment was that the core debris particles would be too large for fast enough heat transfer to water falling on the bed. The reverse situation--molten material falling into water--would result in fine l fragmentation of the melt so that subsequent heat transfer may be , l faster. (Here I am not concerned with the steam produced during a steam explosion but instead with the steam produced during several minutes after contact between the melt and water.) If the case treated in fiUREG-0956--i.e. accumulator and primary system water falling on hot debris--is the more important, then the peer review members should be presented with an assessment of Henry's model or with arguments why such a large steam spike can be
~
produced. s During the discussion of the effect of containment sprays on purging fission products, there appeared to be a lack of under-standing by the Battelle representative of the difference in mass median diameter and number median diameter. The position was then taken that it doesn't matter anyway since purging is effective
i regardless of spray droplet size. I question this position. The equations added to NAUA for the analysis were carefully described in NUREG-0956, and one wonders why if droplet size is irrelevant. Moreover, while "most" fission products might be purged, one is concerned with the fraction not purged, i.e. which is available for leaking, and I would like to see some parametric variation of spray droplet size to be convinced that the fraction not purged is insensitive to droplet size. First, e Regarding containment failure, two items are of concern: the justification for assuming a 7 ft2 hole appeared weak. How sensitive are the results to this parameter? Second, how much attenuation might take place through a containment failure? It appeared that no credit for further attenuation was made in NUREG-0956. Several observers suggested that some attenuation could be justified so it would be useful to elicit from them how this might be done. These two items reinforce the notion that containment failure is itself a complex question involving failure mode, location, etc., and that a detailed review of containment failure must be an integral part of a source-term review. e I took special note of Levy's remark about the need to combine FARCH and MERGE type analyses on the way toward significant improve-ments needed for FARCH. Do any of the new codes like SCDAP and I MELCOR move in this direction? Will MARCH 2 be much better, and what is the schedule for MARCH 27 I am not against using MARCH 1.1 as in NUREG-0956 as long as a proper perspective on its shortcomings and plans to improve or replace it are appropriately emphasized. e I agree with the need expressed to provide figures showing Also, the Burns's control volumes used in MARCH, MERGE, and TRAP-MELT. concern about the need for more volumes in long systems in which mixing is occurring should be considered. e Concern was expressed several times about the small contribution of silver to aerosols. The explanation of the findings of Parker of ORNL together with thermodynamic arguments regarding the vapor pressure and dispersal of cadmium and silver soundet logical to me, and my confidence in Parker's ability leads me to feel that Battelle is right in. basing.their analysis on his results. The finding of silver and cadmium in TMI in proportion to the proportions in the control rods appears to have nothing to do with silver aerosol production. e Battelle says that mass balances are calculated; and to a limited extent, results are reported in NUREG-0956 for materials like Cs0H and Cs1 on retention in various parts of the system. Warman of Stone and Webster claims that he and his colleagues could not 4
- , , ~ - -
' follow where the important fission products ended up in each scenario after trying to do so. I have not tried to do this, but I think it would be useful in the report to be sure that it is easy to follow the final location of all the important fission products for each scenario, both to demonstrate the mass balance and to make comparisons with other calculations more effective.
(Perhaps this is done sufficiently in Table 7.18 for the TMBL'-y sequence. I rather like that table although I can't help but be surprised that condensation does not cause more of the Csl to settle prior to containment failure. See next comment.) e How good is the calculation of 0.8 for the iodine containment re-lease fraction in TMLB'-y, as shown in Table 7.207 Why doesn' t enough steam condense after failure to cause enough rain inside the containment to dissolve most of the Csl before it has a chance to escape from the rupture? It is difficult for me to believe that all but 20% of the Cs1 aerosol that enters the containment escapes. I think that if flAUA is telling us this, some intermediate results should be added to the report to convince the reader of its validity, especially since this is probably the most pessimistic result in the report (other perhaps than the 0.2 release of sequence V). e Tables 7.16 and 7.20 showI aquestion factor ofwhether about 10thisincrease increaseinshould iodine release of AB-S over AB-y.
! be so high. First, Fig. 7.17 shows only a factor of 3 increase in aerosol mass released in AB-S over AB-y. Since the CsI in AB-S enters a containment full of condensing steam--a rather thoroughly wet atmosphere--and since Csl is so soluble, I would expect the difference between Csl release between AB-S and AB-y to be less than the difference in total aerosol release, hence less than a factor of 3. Second, there must be some_ time between Cs1 release from the primary system and release through the opening in the containment, and I would expect that much of the Cs! should be swept out in the wet atmosphere during this transition. flowhere in the report is the opening in the unisolated containment of AB-B discussed (neither in Section 6.1.1 nor in Chapter 7). Surely the release from the For opening must depend on the size and location of the opening.
these reasons I am very suspicious of the validity of a release i as high as 0.5 for Iodine as reported in Tables 7.16 and 7.20. e In Table 7.16 I noted a factor of 3 release for the AB-y Cold Te and Hot Te over the AB-y Cold and Hot sequences. Why does the Te release increase so much? It does not appear to me that retention of Te in the primary system in AB-y could account for this large a reduction relative to AB-y Te, hence. I question the large release result (0.12) of AB-y Te. Unless there is strong evidence that no Te is released from the fuel in the primary system, it may be ' misleading to report the AB-y Te result.
e I question whether the CORRAL-2 results should be included in the report. Since CORRAL-2 treats iodine entirely as elemental instead of Csl, it is clearly incorrect. I expect that this error leads to the factor of 10 increase in release of iodine for the AB-y calculation for CORRAL-2 as compared to NAUA (about 0.04 in Table 7.16 versus 0.4 in Table 7.17). The high CORRAL-2 result is almost surely unrealistic for this sequence; hence it's in-clusion in the report is detrimental to the perspective being transmitted by the report, e In general I was pleased to see the use of NAUA which treats steam condensation on aerosols and the emphasis the report places on this effect. In fact the lack of this effect in CORRAL-2 on Cs! is the main reason that CORRAL should not be used. If another code is to be used for comparison that treats condensation, I would suggest MAER05, or perhaps some of the work of Loyaika. I also am aware of how weak the data base is for demonstration of the extent of Csl removal in steam containment environments. About the only steam environment containment test with a soluble aerosol like Csl was the AB-2 test at CSTF at HEDL, in which Na 02and Na 022 were the soluble aerosols. The major steam containment facility in the U.S. is the ORNL NSPP facility--which is using non-sgluble 0 and 0 0 aerosols--and which has been analyzed by Schock with Fe$ th $AUA co$e but with success only by using artificially high 8 ' diffuseophoresis. The larger German steam containment facility, DEMONA, is not yet in operation. I expect that wet steam environ-ments get rid of Cs! rapidly (as various industry spokesmen are arguing) so that proper modelling of the containment environment is essential to NRC's revision of its source term position. e Application of the Battelle methods to THI-2 was brought up several times at the meeting. Is it obvious such a calculation would predict the small measured iodine releases and large noble gas releases from the auxiliary building? Likely the TMI-2 release through the auxiliary building was enough milder than the V sequence analyzed in NUREG-0956 that analysis of TMI-2 by Battelle would shed little light on the validity of their methods for the V sequence. However it would be useful to be assured that the Battelle methods applied to TMI-2 do not show higher iodine releases than measured while simultaneously showing the rather large noble gas releases observed. e The containment failure Greek letters--8, y, 6, c--should be identified in Table 4.2 in addition to the Roman letters.
ELECTRIC POWER RESEARCH INSTITUTE NM February 22, 1983 EPRI l 4 MR. M. SILBERBERG Assistant Director Source Term Program U.S. NRC Uashington, D.C. 20545
Dear Mr. Silberberg:
SUBJECT:
REVIEW OF DRAFT NUREG-0956, "RADIONUCLIDE RELEASE UNDER LWR SPECIFIC ACCIDENT CONDITIONS" We were pleased to review NUREG-0956 and congratulate you for embarking on the peer review process for this important document. We appreciate the NRC efforts towards a sound technical basis for the source terms NUREG-0956 is an effort in the right directicm. We also appreciate the efforts of the staffs of the Battelle Columbus Laboratory and Sandia National Laboratory for performing this work. GENERAL COFM NTS The most limiting factor in our review ef fort was the limited time allotted. The draft document (first volume) was made available to us only one week before the review meeting on January 25 and 26, 1983. It reports results for the Surry PWR computed for several accident sequences with various codes. The numerical results presented could not be digested in the brief time before the peer review meeting. A thorough review will require more time to carefully examine the methodology and results and to compare them with earlier studies. Such an effort is difficult, since the draft docu-ment does not describe the methodology adequately. The computer codes em-ployed are either considerably modified older codes, or new codes with no documentation. This lack of documentation should be corrected as sc.on as possible. It should provide, as a minimum, a description of (1) the physi-cal basis of the models employed, (2) the relationship between the various models in a traceable manner, 'and (3) any model validation and/or bench-marking performed on a separate-effect or integral basis. The methodology in NUREG-0956 is based on the MARCH 1.1 code. This vers-ion of the code, even as modified, has significant deficiencies, many of which have been documented by J. R. Rivard in NUREG/CR-2285. Employing the MARCH 1.1 code for the accident scenario studies throws doubt on the accuracy of the reported results. [ TO'A/3 g dH a Quarters 3412 Hilly ew Avenue. Post Othee Box 10412. Pato A!to, CA 94303 (415) 855 2000
/ w,sningten omee 1800 uassacnusetts nenue nw suae ico wasn,naron oc 2co3e c202) er2 9222
-_ -.m - - . _ .. . - ~ + -
Mr. M. Silberberg Page 2.
' The NUREG-0956 methodology is described as deterministic and best-estimate.
However, at certain places in the analysis, very conservative (or non-deterministic) arguments are employed which radically affect the results of the analyses. A case in point is the catastrophic failure of the Surry PWR containment one minute af ter the molten corium is discharged into the con-tainment cavity during the postulated TMIM accident. The MARCH 1.1 calcu-lations generate a steam spike of 85 psia when the malten corium interacts with water in the containment cavity. The Surry containment design pres-sure is 45 psig; the containment could hold for much higher pressure levels; for example, WASH-1400 assumes a failure pressure of 85+15 psig (100tl5 psia),with failure defined as cracking of the concrete and breaching of the liner. According to WASH-1400, 0.03 inch cracks would forms enough for gas release at a catastrophic rate. (N.B. not catastrophic structural failure). Thus, the containment failing at 85 psia is neither a best-estimate nor a deterministic assumption. Even if it were to fail at this low pressure, the fine cracks through many feet of concrete would greatly attenuate any aerosol in the gas, an important mechanism shown in the CSE experiments but not reflected in the NUREG-0956 analysis. The accident sequences analyzed in NUREG-0956 are the so-called " dry" se-quences, where no water or water vapor in the transport path of . e fission l products is assumed. Certainly the water initially contained in the pri-mary system will be somewhere along the fission product path. The presence of steam greatly increases the retention of the aerosols as demonstrated in numerous experiments. Thus, the " dry pathway" assumption is not a valid t assumption. We recommend that the location and state of water during the course of the accident be traced with the same attention that is given to the fission products.
; Lastly, the NUPEG-0956 methodology should be validated against the TMI-2 data obtained during and after the accident. Reasonably accurate predic-tions of the TMI-2 accident during the accident phases,where the codes are applicable,will generate confidence in the codes used in NUREG-0956.
SPECIFIC COMMENTS
- 1) High values of source term have been derived for some cases where the containment fails catastrophically. That this occurs is not clears a mechanistic model for the containment failure should be employed.
This model should predict the location and mode of containment failure, . the leak rate to the atmogphere and the fission products and water vapor transport through the f ailed containment. Similar analyses ' should also be performed for the case of containment isolaticri failure, where the fluid and fission product transport through the open valves, piping or penetrations should be treated mechanistically. Such mechan-istic descriptions could increase the predicted retention of fission products, for those cases where containment is assumed to fail or not ; ' I isolate. i ( l \
Mr. M. Silberberg Prge 3. 2) The calculated release from the primary system for the V sequence i appears to be too high. and the steam generator. The fission products must traverse long pipes Supersaturation ratios for the predominant fission product species in the cold primary system are very high. The aerosols in formed will rapidly agglomerate and settle at various points the system. Thus, it does not seem reasonable to predict that only 60% of the CsOH and CsI formed will be retained. This prediction may be due to the modeling of the long pipes as a single "well-mixed" com-partment, instead of a number of serially connected "well-mixed" com-partmentsswith differential deposition of aerosols from one end of a pipe to the other. 3) The V sequence calculation does not assume retention of the fission products in the auxiliary building. The auxiliary building failure should be modeled mechanistically and leakage rates determined. The effect of the wet atmosphere prevalent in the auxiliary building on the scrubbing of the fission product has been ignored. We believe that very substantial reduction in the fission product source will oc-cur, if mechanistic and realistic treatment of the transport of the fission product sourca in the auxiliary building is employed. servation in this respect is that the Windscale accident (dry, An ob-oxidiz-ing atmosphere) released only 12% of the iodine and 5% of cesium in-ventory in two days under forced flow conditions. 4) The hydrogen combustion pressures generated by the MARQi 1.1 code do not account for the mitigating effect of steam in the containment atmosphere. i 'Ihe Whiteshell tests (under EPRI sponsorship) showed a 30% decrease in the peak pressure at steam concentrations of 30%. The Accurex tests (also under EPRI sponsorship), with hydrogen injec-tion rates corresponding to the MARCI code predictions, showed inter-mittent burns with pressure peaks of approximately 30 psi, when steam was present in the containment atmosphere. Thus , the current data does not support the failure of a PWR dry containment due to hydrogen combustim. 5) The fission product sources used in the various accident scenarios are based on the empirica.1 models described in NUREG-0772. The magnitudes of the fission products appear to be reasonable however, the core ma-terial source seems too high. A thermodynamics-based model would be preferable for vaporization of core materials and their transport, since various chemical compounds are being formed during this time.
- 6) There is evidence that, at the temperatures prevalent in the upper plenum, chemical reactions occur between stainless steel and CsOH, and stainless steel and tellurium compounds. Penetration of cesium into stainless steel has been observed in laboratory experiments (fast reactor fuel experiments) . Analysis of the activity on the control-rod lead-screw from the TMI-2 vessel showed much cesium, some of which could be removed only after etching with nitric acid. This mode of fission product retentim in the upper plenum was not considered in NUREG-0956.
Fission products chemically fixed in the upper plenum will not appear in the containment even if the vessel ruptures.
Mr. N. Silberberg , Page 4. 7) The MERGE code does not consider natural convection flow patterns, in the upper plenum; such flow increases the residence time (in the TMLB' 52D and V sequences) and the aerosol deposition rate on the cold sur-faces by impaction. The probability of chemical reactions is also increased. We appreciate the invitation to comment on NUREG-0956 and recognize the very significant efforts put forth by the BCL, Sandia, ORNL and the NRC staffs. If you have please any either contact further one questions of us. concerning our coments to NUREG-0956, Sincerely yours, g f l/ R. C. VOGEL B. R. SEHGAL 1 Nuclear Safety & Nuclear Safety & Analysis Department Analysis Department RCV/BRSans cc: J. Taylor I. Wall W. Loewenstein D. Ros sin i n , , - - .- _ e --- , - - - - - .
THMPSoM
/
;[)(nion of' Comcenmeo February 3, 1983 SCIENTISTS M. Silberberg Chief Fuel Behavior Branch Division of Accident Evaluation U.S. Nuclear Regulatory Commission Washington, DC 20555
Dear Dr. Silberberg:
Re: peer Review Meeting, 25-26 January 1983, for the NUREG-0956(Draft) Study on Accident Source Term (Volume I: pWR Analysis) Thank you for the invitation to this meeting, which was conducted with your usual courtesy. I regret that I was obliged to leave early. For the moment, I have just a few comments, as follows: (i) Collaboration Among Research Teams A number of speakers at the meeting expressed concern tha't different teams (NRC, IDCOR, etc.) might produce different results. It was suggested that these teams should compare their results before publication and try to eliminate differences, thus avoiding embarrassment for the nuclear industry. Such collaboration would be inconsistent with the principles of scientific objectivity you enunciated at the beginning of the meeting. In fact, to achieve such objectivity (both in actuality and in appearance) it will be necessary to avoid collaboration, rather than to encourage it. Furthermore, it will be necessary to expose the research results to detailed scrutiny by teams of reviewers who have not participated in the research efforts themselves. (ii) Uncertainty The "best estimate" approach is adequate for the preliminary stages of this research. However, before the research results are used in the regulatory arena, there should be a systematic treatment of uncertainty. It should be possible to see how uncertainty in any area propagates through the analysis. 1384 Massachusetts Avenue
- Cambndge, Massachusetts 02238
- Tet (617) 547 5552 1346 Connectcut Avenue, N.W.
- Suite 1101
- Washington, O C. 20036
- Tot (202) 296 5600 ywy :3
. .. - . - - . _ - - =.. - . _ - . . . - . -
M. Silberberg 3 February 1983 page 2 4 4 ( (iii) Experiments l Several speakers suggested that your analysis l should be tested against experiments where possible (the TMI-2 event was mentioned several times). I 2 heartily concur with this view. , There should be a systematic treatment showing the degree to which different parts of the analysis 4 are backed up by experiment. This treatment should
; parallel the treatment of uncertainty as proposed in j comment (ii) above. .
i I repeat a suggestion from my comments on the
- draft of NUREG-0772.(see my letter to you of 31 March j 1981), namely: e j "This report should include estimates of i analytic and experimental programs which might help to resolve uncertainties. Such estimates should cover both the time scales l and resource requirements of such programs."
In view of the regulatory importance of the source term issue, a substantial experimental program is justified if it will help to reduce uncertainties. 1 i~ (iv) Research Schedule i.. The schedule you have outlined (publication of a final version of NUREG-0956 in mid-June) is tight. I It will not be possible to resolve the issue by this j point. Although you have indicated that elements of the j research will continue beyond June, it is not clear that the entire source term issue will remain open. It is appropriate to provide a summary of the status ] of knowledge by June 1983, provided that uncertainties are fully addressed. It is not appropriate to pretend
.that the issue will be resolved at that point.
i i Thank you for your attention. Sincerely
)
Gordon Thompso %n Sn l 1 i 9
/ lVMAfAN STONE 6 WEBSTER ENGINEERING CORPORATION I 245 SUMMER STREET, Boston. MASSACHUSETTS moonass act conasspomosNcz ro p o. som asas. soston. unss. ozio7 W W TsLsX soooot C ses,#wCTION w onn
*"'f."?." **".",O ""/ .
;**'s,^.0 wasplMGTON *". "O C.
Mr. M. Silberberg February 2, 1983 Office of Nuclear Regulatory Research U.S. Nucletr Regulatory Commission 7915 Eastern Ave. Willste Bldg. Mail Stop 1130SS Silver Springs, MD 20910 PEER REVIEW COMMENTS ON NUREG-0956 Vol. I (Draft) As requested during the January 25 and 26,1983 peer review meeting on the draft copy of NUREG-0956 Volume I, I am pleased to submit a number of written comments. ( The NRC staff and its contractors should be complemented for the great deal of effort which has been and continues to be expended toward the goal of publishing a best estimate interim source term. My colleagues and I at Stone & Webster are committed to providing whatever technical support is required to assist you in meeting your goal. As the time for review and comment for this first draft has been very brief, our comments are limited to the accident sequences with the highest quoted releases to the environment, i.e., TMLB'-y, AB-8, and the V sequence. The attachment to this letter includes a number of comments on the draft report. These comments are briefly summarized below along with some additional comments and observations. (1). No analyses are presented to support the postulated contain-ment failure for the various sequences. (2). The mode of containment failure is not described for any sequence. (3)., For the AB-S sequence, the postulated unisolated penetrations in the containment lead into contiguous buildings (e.g. aux-iliary or safeguards buildings). Fission product transport and deposition in these buildings must be considered in this i sequence in contrast to the draft report analysis which assumes release through such penetrations are directly to the environment. G$62/ 6 0D ( U i
(4). Cesium iodide and tellurium are erroneously assumed to be released only after core melt begins. When the fuel temperature data in the draft are combined with the well documented release rate data in Figure 5.2, the release of Cs1 and Te from the fuel prior to core melt is directly demonstrated. Thus the data in the report do not support the assumptions of the time of release of these fission products from the fuel. This error has important ramifications in the remainder of the analyses. (5). The graph presented by Dana Powers of Sandia at the peer review meeting should be incorporated into and used in the report.* This graph addresses the release rate of aerosol as a function of fuel temperature. Use of this graph in conjunction with the stated core melt temperature and duration of core melt, results in release of a substantial amount of aerosol before core slump begins. Aerosols released while the reactor pressure vessel is intact must be analyzed relative to their behavior in the reactor vessel internals and other portions of plant systems and buildings. (6). The release of specific radionuclides, in addition to CsI and Te, should be analyzed using the release rate fractions presented in Figure 5.2. , (7). The simplified model of the reactor vessel internals does not adequately represent the physical situation as it affects fission product transport and deposition. l (8). A single zone representation of the containment is not adequate to realistically address fission product transport and deposition in the containment. As an example, the AB- 8 sequence analysis assumes a hot leg break. Such a break would have to occur within the reactor cavity or the steam generator compartment. As no subcompartments have been analyzed, the practical effect, in the present draft, is to have a direct leakage path out the postulated unisolated containment penetration. Such modeling is unrealistic and does not lead to "best estimate" releases. (9). The Cs1 and Te from the fuel can be readily demonstrated to be released into the reactor vessel internals, the reactor coolant piping, the pressurizer and the quench tank, in the TMLB'-y sequence. In the draft report, such is not the case. A set of very questionable assumptions are combined to result in the conclusion that these fission products are released during core slump. The entire release scenario for this sequence is questionable. (10). The solubility of the approximately 50 pounds of CsI in terms of the thousands of pounds of condensed water in the containment atmosphere has not been adequately treated.
- See Attachment 3 ,
1 I (11). The solubility of CsI in the liquid present in the RCS and ECCS piping is not addressed in the V sequence. Similarly, the transport and deposition of aerosol in the reactor vessel internals, RCS piping and circuitous ECCS piping, including reaction of Te with RCS metals are far from adequately analyzed. The effects of aerosol depletion in the circuitous piping with right angle bends is not analyzed in the model. Based on the above comments and those included in the Attachment 1, we strongly urge that more rigorous analyses be performed for each of the major release sequences. We do not believe effort should be expended on the other sequences at this time. As noted above, we at Stone & Webster Engineering Corporation share your commitment to the goal of developing best estimate interim source terms. We would be pleased to offer any technical information which we can, to further 4 your investigations in a timely manner. If we can be of any assistance please contact the undersigned at (617) 589-6510. 1 Sincerely, Q &R FoL i <d. A. Warman Chief Engineer Nuclear Technology Division Attachments EAW: met e
,-.,m- -- ,-- - ,- ,-
w
t i ATTACHMENT 1 COMMENTS ON NUREG-0956 DRAFT VOL. I
- 1. TMLB'-y Sequence A. Cs1 Transport and Retention The Cs and I core release rates and fractions of retained Cs1 in the reactor coolant system, as reported in Tables 6.10, 7.7 and 7.8 l unrealistically shew that no CsI is considered to have been released before the start of core melt. Core melt is reported to start at 201 minutes (3.35 hours). This apparently correlates with the fuel region temperatures in Figure 6.5 which remain unrealistically low for up to 180 minutes (3 hours) for region TRO (1,7) and up to 200 minutes (3.33 hours) for region TRO (1,2). The data in Figure 6.5 appear to be in conflict with the temperature data in both Figures 6.6 and 6.7, which show upper grid structure and gas temper-atures significantly higher than the fuel region temperatures, for
. the same time period. In addition Table 6.3 indicates core heating commences at 60 minutes.
Core melting is reported to start at 201 minutes (3.35 hours). During this 201 minute period, the release of CsI from the core has been unrealistically neglected. The well substantiated data reported in Figure 5.2 indicate that the release of CsI would be completed during approximately the first half of this time period. In any best estimate analysis, with the copious amounts of liquid present in the RCS, it is unrealistic that CsI remains undissolved as assumed in NUREG-0956. For the TMLB'-y sequence the location of release is the pressurizer relief valve discharge tank. A best estimate analysis of the trans-port and retention of releases from the containment would have to include the effects of significant release reductions in the piping from thepressurizer relief valves to the tank, in the tank, and along the pathway volumes and surfaces from the tank to any point of release. B. Te Transport and Retent,1on The temperature profiles shown in Figure 6.5 are consistent with the 4130*Fcore melting temperature in Table 6.1. These temperature pro-files show that for the 69 minute period from the start of core melt to core slump (i.e. 201-270 minutes) the core remains at 4130*F or higher. When combined with the release rate data shown in Figure 5.2, the release rate for Te is indicated as approximately 0.1 fractions / minute. Thus, all of the Te should.have been released in less than 10 minutes. For the next 59 minute period sensible heat is added to the l
, reactor coolant system from the decay heat in the core, and this heat drives the Te into the various regions of the reactor vessel internals, reactor coolant system piping, pressurizer, and water filled quench tank.
l To assume that the 25.4 kg (56 lb) of Te are not substantially reacted with ! these various metal surfaces is grossly unrealistic. ,
=,
' - - - - ~-- , _ , , , , , _ _ ,
I (
- 2. A3-8 Sequence A. Cal Transport and Retention i
The masses of CsI injected into the containment, as reported in Figure 7.3 and 7.4, unrealistically show that no CsI is considered to have been released before the start of core melt. Core melt is assumed to start at 1620 sec (27 min.) as reported in the material provided at the peer review meeting. This apparently correlates with the fuel region temperatures depicted in Figure 6.1, which remain unrealistically low for up to eighteen minutes for Region TRO (1,7) and up to thirty five minutes for Region TRO (1,2). The data in Figure 6.1 also appear to be in conflict with temperature data in Figure 6.2, which shousupper grid place structure and gas temperatures significantly higher than the fuel region temperatures, for the same time period. In addition. Table 6.3 indicates core heating commences at time 0.52 minutes. 4 Core uncovery is reported to start at 30 seconds (0.5 min.) and core melt is reported to occur at 1620 seconds (27 minutes) in the material provided at the peer review meeting. In this 0.5 to 27 minutes period, the release of CsI from the core is unrealistically neglected. The well substantiated data reported in Figure 5.2 indicate that the release of CsI would be completed during this time period. Concurrent with the release of CsI from the core during heatup sev-eral hundred thousand pounds of coolant inventory are also released. As the 50 lbs of CsI are highly soluble in water, it will substantially , dissolve while still in the reactor coolant system. It is also noted ! thatin anybestestimate analysis, two phase conditions will exist in the vicinity of the postulated pipe break. Our calculations indicate that approximately 33,000 lbs of water con-dense in the containment during this 0.5 to 27 minute period as shown in the Figure labeled Attachment 2. Aprroximately 53,000 lbs of CsI are soluble in 33,000 of hot water (approximately 14,000 lbs CsI are soluble in 33,000 lbs of cold water). As the total inventory of CsI is only approximately 50 lbs, thermis over 1000 times as much relatively hot water available to dissolve all the CsI in the core than required. , i The large 1,0CA pipe break location, for the AB sequences, is the re-i actor cavity or stesa generator cubicle. A best estimate analysis of l' the transport and retention of releases from thc containment would have to include the effects of significant release reductions along f pathway volumes and surfaces from the reactor cavity or steen generator j cubicle to any point of release. L l 9
l i In Table 7.1 the containment failure time is reported to be 0 minutes. This apparently is based on the S designation which is failure to isolate the containment. Containment penetrations lead into con-tiguous buildings (e.g. auxiliary and safeguards buildings). There-fore any analysis of releases via these penetrations must consider the aerosol transport and depositions in these buildings. When these effects are properly incorporated in the anlaysis, substantial reductions in the releases to the environment result. B. Te Transport and Retention The temperature profiles shown in Figure 6.1 are onsistent with the 4130*F core melting temperature in Table 6.1. These temperature profiles show that for the 30 minute period from the start of core melt to core slump (i.e. 27 minutes to 57 minutes) the core remains at 4130*F. When combined with the release rata data shown in Figure 5.2 the release rate for Te is indicated as approximately 0.1 fractions / minute. Thus it would take only 10 minutes to release all of the Te inventory at that temperature. As the core is at that temperature for 30 minutes all the Te inventory is releared in 10 minutes. For the next 20 minute period, the reactor coolant system and reactor vessel system remain intact with a heat source to drive the aerosol into regions where deposition of Te is known to react with steel. As there are over a hundred thousand pounds of these materials in the upper plenum alone, a best estimate analysis must address these re-actions of Te. Also, as there are only 25.4 kg (56 lbs) of Te and there are over 100,000 lbs of metal and the design of the reactor vessel internals is such that the surface area to volume ratio is large, any reasonable analysis of the face of the 25.4 kg (56 lbs) of Te will show they substantially reset with the steel. e h 1
. e J
1 i ! i 1 2 1 9
- 3. V Sequence 4
A. Cs1 Transport and Retention 1 The mass of CsI injected into the auxiliary building as reported in i Figure 7.8 for the V sequenceu 'nrealistically shows that no CsI is l released before the start of core melt. The core was reported to be uncovered at 4.9 minutes, core melt is assumed to start at 37 minutes, and core slump is assumed to start at 64 minutes, as reported in the material provided at the peer review meeting. a ' For the 32.1 minute period during which the core is uncovering, the
! release of CsI from the core is unrealistically neglected. The well i substantiated data reported in Figure 5.2 indicate that the release , of CsI would be completed in this period, for reasons similar to those ; we stated for the AB-8 sequence.
2 It was not possible for us to perform similar calculationsof Cs1 release for the V sequence because the necessary data have not been provided in NUREG-0956. i [ As this period of core uncovery, of 32.1 minutes, is longer than the 4 I 26.5 minute period for the AB-8 sequence, more opportunity exists for t CsI release from the core. The 50 lbs of highly soluble CsI are released during this period together with copious amounts of coolant inventory. These transit the Ibng and circuitous RCS & ECCS piping together providing ample opportunity for the CsI to dissolve. 1 i Upon release from the core region, the Cs1 (and the other fission products) must transit through the reactor vessel internals, reactor
' coolant system piping and interfacing small diameter ECCS piping. As the temperature of all this piping is well below the 1148' melting point of CsI. the physical form of any Cs1 that may escape being dis-solved during this transit period will be that of solid particles.
1 The piping in question is wetted surfaces containing a two phase mix-ture. In addition, the piping has many bends (frequently at right angles in most plants) whic't would have a major reduction impact dur-
; ing the transport of these CsI particles. The release fraction for i CsI for the reactor coolant syntes, reported in Table 7.20 is only ! 0.4. As there are only approximately 50 lbs of CsI available from the core, is is unrealistic that a rigorous analysis would support that 20 lbs of CsI inventory was released and that only 30 lbs were i retained in the reactor coolant system. ^
Furthermore, our analysis of the auxiliary building pressure capability shows that the data presented in Table 6.3 which indicate- the immediate structural failure of the auxiliary building are incorrect.' Structural l' f e 9 i e. y l
integrity of the buildina.vhich our analysis shows persists indefinitely. vill serve to entrap air-borne fission products entering from the ECCS piping by the various mechanisms of aerosol depletion, including condensation of water from the primary system. B. Te Transport and Retention As the V Sequence is stated to be very similar to the AB sequence with regard to thermal hydraulics and core temperature behavior, the arguments above for early Te release will not be repeated. In addition such arguments would have been difficult to pose in that no temperature profiles have been provided for the V sequence comparable to Figures 6.4 and 6.5. From the data that were provided at the peer review meeting, the core is reported to be uncovered at 4.9 minutes, the core melt is reported to start at 37 minutes, and the core is reported to slump at 64 minutes. During the period of core melt (27 minutes) the core temperature is i assumed to be 4130' F or higher. Using the release rate data fr/2 Figure 5.2 For Te, the release fraction of approximately 0.1 fractions / minute will result in the complete release of the available Te in less than 10 minutes. During the more than 17 minutes that remain of the core melt period, the energy from the decay hear generation in the core would act t t to drive the released Te to interact with the complex reactor vessel internals, reactor coolant system pipins, and interfacing small diameter ECCS piping. As there are over a hundred thousand pounds of steel in the upper plenum alone and the surface area to volume ratio is large, a best estimate analysis must address the reaction of Te. The temperatures of reactor coolant system piping and interfacing small diameter ECCS piping play important roles in determining the fate of any fission products released from the core. The transit path would include approximately 10 feet of reactor coolant system piping and approximately 100 feet of small diameter (e.g. , 6 inch) interfacing ECCS piping. A realistic analysis must address the transit of 25.4 kg (56 lbs) of Te in the reactor coolant system and interfacing ECCS piping which contain large surface areas and hundreds of thousands of pounds of steel. e e 1 l t l
=
l l
. 7
'd i
LETTERS ABOUT THE PEER REVIEW MEETINGS May 24, 25, 1983 e 9 4
CAMPGall OAK RIDGE NATIONAL LABORATORY CPERATED av UNION CARBIDE CORPORATION NUCl[ AR DIVISION O POST OFFICE sox X OAK RIDGE, TENNESSEE 37830 February 3, 1983 Dr. Mel Silberberg Division of Accident Evaluation U.S. Nuclear Regulatory Commission MS 1130 SS Usshington, DC 20555
Dear Dr. Silberberg:
I' find the ;iUREG-0956 draft to be a major disappointment, and I strungly recommend that it not be poblished in its present form. There are several reasons for this, based lardely on ny opinion that the report blends some fact, sone modeling, suae speculation and assumption, and a few very specific scenarios to arrive at conclusions that will be taken out of context, over-generalized, and accepted by some as far more defi-nitive than justified. There are occasions when it is better to admit the answer is unknown than to insist in presenting an apparently precise, but perhaps erroneous, answer. At the same time, I agree that there has been progress that should be reported, although I am concerned there has been so little progress. My objection here is with the degree to which definitive impressions are likely to be derived from this report (unjustifiably in my opinion), par-ticularly from the first and last few pages which constitute the part , that the majority might read.
; The minimal attention to chemistry (and to some extent physics), along with the superficial and confusing mechanistic description that is given, is indicative of a very serious shortcoming - namely that the principal authors not only do not take such factors seriously, but they don't appear to be really aware that they exist. Chemistry is relegated to some simplistic assumptions, either buried within the codes or supplied as undocumented input, without adequate definition in either Case.
The treatment of physical and chemical adsorption described at the review meeting suggested a lack of understanding of this subject, and some con-fusion of the adsorption processes with possible subsequent processes such as diffusion into an oxide film on the plenum. tkchanisms going beyond adsorption seem to be grouped under the name of chemisorption. J 0
Dr. Mel Silberberd Page 2 February 3, 1983 Tne TMI data, although limited, should be examined for indications of chemical behavior. There is significant cesium associated with surfaces of the plenum, even after nearly four years innersion in water. There is significant tellurium outside the primary system, both on painted con-tainment wall surf aces and in precipitate on the floor. This precipitate also contains significant fission product iodine and control elements. There are separate particles of different control rod constituents in solids recovered f rom the reactor coolant, suggesting selective evaporation of cadmium. I will make little specific couaent about the couputer codes and nodels because others are better qualified. I will add my support to the conten-tion that (1) they include hidden and unrealistic assumptions about the behavior of nature and (2) whether intended or not, their results tend to assume an aura of respectability and rigor that is not warranted. It seems to me that the computer codes are sometimes used to " launder" opi-nion (input assunptions, for example). I don't think the overd <ining opposition to the codes, expressed at the review meeting, is in .' 4e nature of sour drapes; you would be advised to take it seriously. I would suggest that a much core satisfactory approach would be to issue a series of reports, the first including part of this one and consti-tuting essentially an update of NUREG-0772. There have been improvements in our understanding of the chemistry of some of the fission products, especially aqueous iodine chemistry. There has been a lot of work on aerosol generation and behavior, and enaputer ,nodels have been improved. All this sort of thing should be grouped in a single report as the basis for subsequent calculations, and there should be periodic updates; this report would contain what science there is, along with the technical state of the art. This material could then be referenced, and not repeated in toto time and time again, as it is now in the excessively long and repetitious safety studies. There should then follow a series of reports on the results for designated reactors, including all the site- and scenario-specific fac-tors, the first being taken f rom the bul'< of this report. These should exclude the information common to all the studies, by referencing the revised 0772, so they would be shorter and more canageable. I think it would be greatly advantageous to separate the specific studies, which tend to get bogged down in a myriad of decatis, from the more general considerations that apply across the board. Comments more specifically directed to the text follow: Last paragraph, p. 2-2, says physical and chemical processes are incor-porated on a " mechanistic basis". This term is not just vague; it is inaccurate, because in many cases the input assumptions, not mentioned here, are more important than mechanistic models. Later on (p. 5-13)
Dr. Mel Silberberg Page 3 February 3. 1983 thers is mention of the input assuoptions. Howavar, there is no Jood indication of the significance of errors la either the codes or the assuuptions. The inadequacy of the approach used is not properly described. On p. 3.3 and elsewhere these calculations are refarred to as "best estimates", but there are various places where clearly consarvative esti-mates were taken, and where ef fects were omitted froo treat.nent because there was no readily apparent way to codel then [ paragraph (8) on p. 3-5. for example. but most cases not overtly noted). This is a best estimate only in some respects, and conservative in others. Bottom paragraph, p. 4-5: The distinction here, conpared to TMI, is not clearly drawn. In all cases, fuel heatup begins only af ter the water level drops below the top of the core. Page 5-8 and vicinity: The text in this area is unacceptable. It is stated that control rod material melt release is simulated as tin and steel. Other caterials known to be present (cadmium, indium) are ignored. Although cocpounds, alloys, and such will have different properties, it is of soce use to coasider the boiling points of some niements that are pre-* sent. The order of increasing boiling point, after water, is cadmium-765'C; tellurium-987; (CsI-1280); antimony-1637; indium-2050; silver-2193; chromium-2642; tin-2687; nickel-2335, and iron-23d5. The control rod ele-ments boil at 765, 2050, and 2193'C, while the stand-ins (tin and iron) boil in the 2700-2900*C range, several hundred degrees higher. This sort of modeling appears coupletely erroneous to me. Apparently it is assumed that control material, or at least silver, is not released; this is based on a single preliminary report that is not referenced. In contrast, there are repeated observations of csdatum boiloff. Alloying can reduce volatility, but tin is certainly alloyed with zirconium and it has a much higher boiling point than silver, yet it is assumed to be released wnile silver is not. Similarly, it is assumed tellurt2m is not released, based on an undocumented observation that is contradicted by TMI. Also, note that the ef fects of alloying or cocpound formation may be eliminated if the zirconium is completely oxidized. Models that depart substantively from known properties of materials must be well documented and explained at least qualitatively. In such respects the present text is totally inadequate. Near the middle of p. 5.8, it is stated that control rod silver is not available for aerosol generation, but no reference is given. This pro-bably refers to recent German work, and ignores contradictory work elsewhere. Unless such behavior can be unambiguously documented and rationally explained, it would be advisable to relegate it to a footnote, at best. Meanwhile, the analyeis should not be based on this assumption. Q
Dr. Mel Silberberd Page 4 February 3.1983 Page 5-10 and Fig. 5.2: The curve for structural saterial (f rom Table 5.1) should be added. Other curves aid ht be better labelled, such as adding "Sn f roo cladding" to the Sb curve. Middle of p. 5-15: Are the " particles" referred to in the first bullet the same as the ones in the third bullet? Are ONLY CsI, Cs0R, and Te con-sidered as constituents of particles (first bullet), with no other material? Top of p. 5-16 is also unclear. Terms in equation are poorly defined. The terms " deposition mechanism" in line 4 and " deposition terms" below the equation ara ambi Juous. The last two sentences of the first paragraph are not directly related to the foregoing, and are unclear. Perhaps these ideas should be in a separate paragraph. The cousents on " particle agglomeration" should be separated from the others since this process does not have a shorter time constant. Middle of p. 5-20: 1s a steam explosion an oxidation? Bottom of p. 5-22: Is steam condensation (and particle growth) treated as steam condensing on inert surfaces or on a chemically active surface? Salt i particles can taka up water, grow, and " melt" when the relative humidity is well below 100% and water would not condense on inert surfaces. Page 6-3, line 15: Should " existing" be exiting?. Bottom p. 6-20 and top p. 6-21: This is another example of a consnevative assumption. You permit radiative heating from below, but not radiative cooling from a structure around 1400*C to the relatively cool structure above. Middle paragraph of 1 6-21: Maybe the thermal-hydraulic conditions inside containment can be ptedicted with confidence, but are they? I nave been surprised at the extent of compartmentalization in the lower part of the TMI containment. Are sequential control volumes used? Is the containment ventilation system included because, even if not operating, it can provide one leg of a thermal convection loop? (Ditto stairwells, elevator shaf ts.)
. At TMI the cooling-ventilation system ended up nearly the highest radiation source (except 'the sump) and the top end where it discharged suggests high containment surface temperatures (scorched paint) and high radiation levels. (The reactor for this study does not have containment coolers like TMI, but does it have a containment ventilation system?)
Page 6-22: Apparently 33 i W MW day / ton is an error. Middle paragraph an; V.sie 7 are in error because of omitting cadmium and indium. The text s ;y uta .he silver release was not, but not what it ,. was; was it zero? See also consent above re. p. 5-8. -It is stated again J
- I 1
. . .. .- . . ~ .. _..- - - _ _ _ - _- .~ .. .
J Dr. Mel Siloerberg
,. Page 5
! ; February 3, 1983 l that control rod silver is not a source of aerosol, but this is not docu-mented and, in ny opinion, not firm enough to base this study on. Page 6-31: Iodine section says iodine is present as " nonvolatile iodide", but Cs1 is really rather volatile compared to nany other coristituents; suggest "relatively nonvolatile". In last half of paragraph, was the
- 2x 10E-7 fraction per hour based on reappearince of iodine af ter Kr venting at TMI? It is not clear whera it caine f rois. The 0.05?. figure appears conservative, since no TMI measurements were quite that high. It is stated to be conservative; it is not a "best" estimate.
Page 6-32: Tnis seems to repeat sich of what was just said. It midht be a worth sumarizing by pointing out that volatile iodine is aleays a very small fraction of the total. Page 7-27 and -23: Tables appear to be misnumbered. Page B-2: 'Ihe statement of a 25% variation in temper.sture, using the 1 Centigrade scale, causes concern about the scientific insight of the i authors. j Since rel/ , j l s I !O b. jl.[ [ D. O. Campbell .[ Chemical 1)3velopment Se ction Chemical Technology Division DOC / bv t i i i i i 2 5 I i - i i i k 9 i 1 4 +.~. . . . _ . . . . . . _ . . . . - .. . _,.,s .. .. - . - . _ . . . . . . _ . - _ _ _ . , _ _ . . . _ _ _ . . _ . . _ . . . _ _ , . . . . . _ - _ . . , ~ . . , . . . . . _ , . _ .
i l woMR COUGLAS WINSLOW COOPER. PM, D, 27e wAmtecnovow svage ecstoN wass caste 1 June 1983 Dr. Michael W. Jankowski Fuel Behavior Branch 1830 SS, Div. Accident Technology Off, Nuclear Reg. Res. U.S.N.R.C. Washington, DC 20555 Ref: Contract No. NRC-04-83-Oll
Dear Mike:
It was good to meet you at the review group sessions a week ago. I will reiterate a few comments, already on the record; enclose a reference on diffusiophoresis; and enclose a worksheet and receipts for beginning the payment process. Major comments:
- 1. Core temperatures are crucial for the source term during melting,and the choice of nodes is proving significant.
- 2. Gradual versus sudden slumping is another major issue.
- 3. I think the pool scrubbing model is much more nearly correct than I did when it was first presented to me.
- 4. The spray scrubbing model can be bmproved by using more conventional expressions for impaction efficiency, by in-corporating the initial velocity of the droplets, and by improving the estimated Sauter mean diameter. (I am sending Ken Lee of BCL some references to assist here.)
k
- 5. I think that condensation to the walls will be more significant than currently believed: roughly exp(-V*/V) where V* is volume condensed and V is containment volume, when aerosol is present during entire period of condensation.
ftd t's
- 6. It was not clear that droplet growth was modeledidue to the supersaturation expected to exist once containment starts leaking.
- 7. Where the aerosol encounters HEPA filters, the flow rates will be greatly changed once the filters fail, and the flow through a charcoal bed could lead to sufficient deposits in the bed to cause both plugging and efficient
/ aerosol removal.
/
c c-6J ars//f / 4 SW.L- t t
D.W. Cooper 1 June 1983 p.2
- 8. Particle charging due to differential ionic mobilities may still prove significant. The mass concentrations are so great that small charges per mass could produce large electric fields.
- 9. The 100x differences between NAUA and CORRAL-2 code results need some explanation, with ,the choice of NAUA justified.
- 10. The sensitivity analysis could easily get very compli-cated: the important parameters should be given distribu-tion estimates, then the distributions sampled, and the worst and best,5% tile and 95% tile, results for release fractions selected and displayed.
Enclosed is a copy of a brief paper by Dunbar (1983) of the UKAEA, called to my attention by Ed Warman of Boston's Stone and Webster office: Stefan flow can be important. The enclosed worksheet gives my expenses for the meetinas. I don't know how to calculate the totals correctly in the light of the per diem requirements.*Please have someone do it for me as per NRC regs, and I'll sign their version. Enclosed are the air fare and hotel receipts, just in case this can be processed without returning it to me. To the worksheet totals for the meeting should be added $9.35 for Express Mail for my February memo and a fee of $1288.14 for 42 hours (at $30.67/hr) for the work done in connection with this meeting and the one before it. I notice that my contract allows two meetings, this one just past and the next. Later ones, if I'm needed and IBM 4 has no problems with it, may need some contractual adjust-ments. I'm looking forward to the next meeting. Perhaps I'll ) win a bet at that one.... Take care. Cordially, 7 G
*I stayed in DC an extra day, not charged in any way to NRC, returning Thursday, instead of Wednesday the 25th.
UNIVERSITY OF CALIFORNIA, LOS ANGELES KMTEWBMG
?
RERLELE) O o is In t l% E Los 4%ELEs Rit ERstDE 44% DIE (.as 44% F R t%Cist u
%%%T484RB484 4 4%i4 8 RL Z u, .v -
g.y.
\ LECH ANICs 4.ND STHt"CTt'HE% der 4HT51ENT
%Cil(MH. OF ENGINEEMING 4%D AltLIED %CIE\t E I.0% ANGELEE CALIFOR\l 4 'mus2 4 May 31, 1983 Dr. Melvin Silberberg
, Accident Source Term Program Office Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Commission Washington, DC 20555
Dear Mel:
I have reviewed the handouts and have read the material given to us at the Peer Review on May 24 and 25th and have the following comments.
- 1. Time, location and mode of containment failure still appears to be as important as fission product chemistry in determining a source term for many of the sequences. At the first Peer Review we saw the importance of
" containment attenuation" in moving from the early 6 containment failure mode to the late 6 containment failure _ mode for the PWR large dry (Surry) contain-ment.
It was only through questioning at this meeting did it become clear that failure pressure and location differences between this work and WASH-1400 were the two dominant factors. For the AE sequence in Peach bottom, the failure pressure and location were changed from 178 psia in the wet well (WASH-1400) to 125 psia in the drywell, and this apparently overshadowed all the other " improvements" in models, phenomenology and computation. For the July meeting some parametrics on containment failure location and pressure (for the Mark I) should be completed. Furthermore, I believe the Project's (and its Contractor's) efforts on containment analysis should be integrated into the program as soon as possible, and should become part of the peer review.
- 2. For BWRs the suppression pool, plays a key role in the determination of the source term. I raised the question as to whether or not in a Mark-III, there were sequences which could bypass the suppression pool. The answer, "that they would be of such small frequency or be incredulous", doesn't sit f()}l3J0h
l l Page 2. Letter to Dr. Melvin Silberberg (con't.) May 31, 1983 right in view of the comment we heard from Bill Hopkins of Bechtel, regarding a flow path back to the dry well from above the suppression pool, for re-circulation of steam. I believe he called it an opportunity for a "re-rinse". Looking at it in reverse, a valve failure would imply a pool by-pass. Although this apparently is unique to Grand Gulf, the standard Mark III might have other subtle flow paths for by-pass. Moreover, only transient events are being considered in the study for Mark III's. There may be some bypass sequences with pipe breaks which are potential risk contributors if they lead to early failure. In any event, Battelle, with its vast PRA experience should not only rely solely on the industry for the definition of its sequences; it should use some initiative and decide for itself whether or not the pool can be bypassed.
- 3. In reading over the draft material for the Mark III containment (Vol. III), it states that for the TC sequence, the suppression pool will boil (pg. 4-7). What does this imply? Will the release be to a saturated mixture or saturated pool; or will steam evolve. If the pool is saturated, it will re-release any scrubbed fission products when it boils or vaporizes.
On page 6-7, it states, "after leaving the vessel, fission products are carried down the steam line and relief line to the suppression pool". If it has boiled, will there be water present? Or will it be merely a saturated pool? At the July meeting, I would like to see a better definition of the thermodynamics of the pool. Perhaps on a P-V or T-S diagram. Other questions arise. Under what conditions could the pool flash? And, for example, could it flash at containment failure?
- 4. The addition of the Zion Plant to the Element 2 Work Scope raises a whole host of new issues. Although it was shown that seismic events were not a major contributor to core melt frequency, seismic events did dominate risk at Zion because of concurrent seismic containment faulure. Will the study focus on internal initia'ingt events only; or on both? All of the sequences studied so far (for SURRY, Peach bottom, Grand Gulf, and soon Sequoyah) are based on internal initiators. In determining dominant sequences, Battelle has looked at frequency and consequences, which is appropriate. But at Zion, the presence of an external initiator changed the consequences for some sequences dramatically. For Grand-Gulf, only transients are being con-sidered. If external events were considered, or if an ATWS fix were made, -
pipe breaks might become the risk dominant sequences. More than likely the time and location of failure would change for a seismic initiator. Other initiators, might also have similar effects (e.g. fires, missiles etc.).
Page 3. Letter to Dr. Melvin Silberberg (con't.) May 31, 1983 Before initiating the Zion study, several ground rules should be developed by the Project and presented at the July or August meeting.
- 5. There appears to be some prejudging with respect to dominance, vis a vis the sequences and with respect to the risk itself, i.e., "end of spectrum" events. Estimates o calculated to vary between 10 fand core-mejt 10- perfrequency for some reactor year. Whileplants all is core-melts may not lead to public health risks, this project is supposed to assess the sorce term, not pre-determine this. Element #4 should resolve the risk question, after frequency and consequence have been determined, not before them.
All in all the meeting was informative and interesting. I hope to see more of BCLs initiative and innovativeness with respect to BWRs at the July meeting. Battelle is doing an excellent job in model and phenomeno-logical improvement given the time and financial constraints. I would like to push them a little into some global thinking. I hope these comments prove useful. Sincerely, W.E. Kastenberg, Professor WEK/shm P.S. Thanks to the pipe smoke, I ended up with an upper respiratory ( infection and spent the Memorial Weekend in bed. I would appreciate your banning pipe and cigar smoke at the next meeting, cc: M. Jankowski R. Benero l L. . . .. .
Lay 5.LEVV,INC. Suite 725 g &y $ y 1999 South Boxom Avenue y\ CompbeH, Cohfomio 95008-2233 7 USA g\ 408/377 4870 May 20, 1983 Mr. M. W. Jankowski Division of Accident Evaluation, RES Office of Nuclear Regulatory Research V. S. Nuclear Regulatory Commission 7915 Eastern Avenue Silver Spring, MD 20910 ,
Dear Michael:
I am sorry that I shall not be able to attend the Peer Review Committee Meeting because my wife is to have hand surgery on May 25, 1983. I have, however, reviewed the draft Volume II (BMI-2104 Report) material sent to me; my coments are as follows: Accident Sequences
- 1. The accident sequences do not recoanize the newly approved NRC BWR Emergency Procedures Guidelines (NEDO-24934). The procedures to be implemented at all BWR's will have a substantial impact on the consequences of the selected accident. For instance, the operators are instructed to vent the containment and the presumed containment failures in BMI-2104 may not_ occur. Also, as shown by Table 2, there is substantial time between the start of core melt and reactor vessel melt-through (varying from 2 to 5 hours) which should gi/e operators substantial opportunities to remedy the situation, such as activating containment spray and develop means to add ater to the react.>r core.
- 2. Under sequence TC, it is important to realize that it presumes not only failure of the control rods to insert but also failure of the liquid standby control system to add borated water to the reactor. Here, again, the new Emergency Procedures Guidelines provide means for the operator to reduce plant power by lowering the reactor water level and cutting back on high pressure addition. Also, it should be made clear that it is assumed that the containment failure stops all means of pro-viding water to the reactor, including the use of balance of plant pumps.
- 3. Clearly, the containment level failure mode and its location i and timing are very important. As previously noted under my previous ,
PWR comments, no mechanistic containinent failure model was provided and i this is, again, true for BWR's. It is all the more important for BWR's due to the fission products scrubbing the suppression pool provides. On page 4-15, the uncertainty about failure pressure and location is noted, I and the need for a good mechanistic BWR containment and reactor building model cannot be overstressed. Because of the uncertainties at this time, it is recomended that fission oroducts release be carried out for various containment failure modes. l gjV n/M
i
-2~ May 20, 1983
- 5. LGVV, WK.
.A ANALYTICAL METHODS
- 1. No experimental validation of analytical methods is provided,
- l. .
While the changes made to MARCH 2 appear reasonable and in correct direction, no proof is provided that MARCH 2 is superior to MARCH 1.1. i
- 2. . It is not clear what MARCH 2 does with: lower plenum of BWR and all the hardware present there.
- l. Two new computer codes are introduced: VANESA and SPARC and I 3. :
again no experimental data are offered to judge their capability. i BASES FOR TRANSPORT CALCULATIONS
- 1. In Figure 6.3, fission products-release is always uowards i
.through shroud heads, steam separators and lower outer annulus. They i' can also travel in the opposite direction. This will become important '
j as molten material falls below the bottom core plate. The escape path ! would now be through the bottom plenum and upwards through the jet pumps. The structures in the bottom plenum may provide substantial areas for fission products deposition especially those that are still at low I temperatures.
- 2. It is not clear whether as each radial zone melts, it is allowed to find its way into the bottom plenum or whether one must wait until the entire core slumps. This may have a different impact on BWR's versus PWR's.
On page 6-21, it is not' clear why the ECC makeup switches from 3. the condensate storage tank-to the suppression pool and the pumos fail _. In this sequence power is available and one could continue to provide water to the condensate storage tank.
- 4. It is hoped that the homogenizing of the fission products is coupled with a homogeneous power production for consistency purposes.
f RESULTS AND O!SCUSSION i i 1. Under sequence AE, it is important to point out that this is - not a best estimate case for the following reasons: , 3
- minimum fission product release path is assumed and other break locations may provide increased opportunity 4
for fission products deposition j_
- very rapid blowdown is assumed. There are many large breaks (e.g. in steam lines) where coolant escape will be much l slower; .also, more water would be left in bottom of vessel i
and the release fractions may change due to time delay and l
- more water being available.
l
4 SLLinMP,NM(; May 20, 1983 i As noted for PWR's, a variation of large break locations needs to be considered to get a best estimate for such accident sequences.
- 2. The impact of actuating containment spray should be evaluated as this is an important option available in BWR's.
I hope the above comments are helpful. Please feel free to , read them and enter them into the meeting minutes. Also, I shall appreciate receiving a copy of additional material provided or utilized for presentation at the meeting. 1 Best wishes for a good meeting. I Sin rely urs, d ! Salomon LevgT-i -President i [ SL:jm cc:M. Silberberg, USNRC ? l i 3 1
C00MR DOUGLAS WINSLOW COOPER PM, D, 276 M ARLGORouGM STREET SOSTON. M Ass 021 t o 22 May 1983 To: Dr. M. Silberberg, USNRC From: Dr. D.W. Cooper, Consultant Re: Review of the Transcript of the Meetings 25-26 January 1983 on RADIONUCLIDE RELEASE UNDER SPECIFIC ACCIDENT CONDITIONS In preparation for our meetings 24-25 May, I reviewed the transcript of the previous meeting and reviewed my n.emo to you of 6 February on the report, and I summarize here my comments related to the aerosol science and tech-nology aspects of the work, the first twelve relating to the transcript, the second twelve recapitulating my memo:
- 1. Heating of the surfaces by deposited radioactive .erosol particles could reverse or reduce thermophoresis and reduce deposition.
- 2. The degree of vaporization of silver remains important i and uncertain. The more silver vaporized, the less radio-active aerosol eventually released.
- 3. Uncertainties in core temperatures and in the vapor pressures of the alloys' constituents are very important.
- 4. The aerosol particles are likely to be mixed hygroscopic (CsI, CsOH, ...) and non-hygroscopic materials. They may retain water even when the volume is not saturated.
- 5. Whether or not the particles are resuspended will depend not only on their size and on the gas flow velocity, but on the nature of their surfaces and the surfaces to which they attach: " stickiness" is crucial. ,
- 6. Is gravitational deposition in the core taken into account?
- 7. Turbulent deposition and resuspension in the LPIS pipe could be crucial for sequence V. Is plugging likely?
- 8. Where might the deposition of the aerosols be sufficient to change the flow geometry appreciably?
- 9. Do we have any more information on containment failure geometry? What are the results of parametric studies?
I D.W. Cooper 22 May 1983 p ga 2 l
- 10. Are thetutcircumstances where there will bc " rain" in the containment,even without the sprays 3 that could lead to significant aerosol capture?
- 11. Convection currents in the region of the core could be important in the slow-flow cases such as TMLB' and S 20*
- 12. Aerosol concentrations greater than 1 g/m 3 could begin to produce hindered settling, which is slower than that expected for individual particles.
- 13. The work is generally at the state of the art and shows a high degree of technical competence.
- 14. Turbulent deposition and resuspension need more considera-tion.
- 15. Condensation of water vapor at the walls produces a cleaning effect, almost like flowing the gas through a filter, and needs consideration.
- 16. High radiation fields should produce charged particles, due to the difference in mobility of the bi-polar ions produced.
- 17. Same as 11: convection in the core should be important.
- 18. Condensation will occur on particles much smaller than 0.6 um diameter (and condensation tends to produce a fairly monodisperse aerosol, geometric standard deviation of 1.5 or less) . Mixed nuclei will have condensate even without sa turation. [ Lea d.s w.li +.6lu.tu. M ,es4 M ley ad,4/w .6 syg%m a.)
- 19. Equation 7.2 has an unconventional impaction term.
- 20. Concentrations of 1000 g/m 3 are not likely to persist for 103 s.
- 21. Some releases predicted by CORRAL-2 are 100x different from those predicted by NAUA in amount.
- 22. 1 pm seems small for sparged metal particles.
- 23. Equation B-9 would better be in terms of the important dimensionless groups with a coefficient determined by the experiments.
- 24. Cumulative size distributions are easier to use and interpret than are differential size distributions.
Look forward to seeing you again. . D
,/~ ,
I-- V COMMENT TRACKING SYSTEM FOR THE . SEVERE ACCIDENT SOURCE TERM PROGRAM 0FFICE Christopher P.Ryder April 1984
}$1fl0
a - - - - - - - - - .- - . - - - - - . - 4 J -* 4 a. cSe. w.. l i l l INTRODUCTION The NRC held five meetings to review and critique the source term estimates done by the Battelle Columbus Laboratory. ASTP0 requested written comments. The letters were studied to determine technical issues (Section 2); of these issues, the significant issues were arranged to formulate general issues 1 (Section 3). 4 Section 1 - List of issues Section 2 - Specific issues with references to letters and transcipts 3 Section 3 - General issues composed from specific issues Section 4 - Letters
! Section 5 " Review of Computer Models of Containment Aerosol Deposition," 5. L'eal,. ACRS/NRC.
Section 6 - Affiliations l i
- MEETING DATES 1
January 25, 26 1983 , May 24, 25 1983 July 28, 29 1983
- October 12, 13 1983 January 26, 27 1984 i
l 1 l l 1 1I 4 j
-..____....m... . . , . _ . . . _ . . - ,_ __._..._-- . ,.-._ _,
SECTION 1: Summary of the Resolutions of the Source Term Issues t
}
ISSUE RESOLUTION A. ACCIDENT SE0VENCES A.1 Alternate Pathways
- a. Complete bypass of a suppression pool
- b. Penetrations
- c. Pipe break
- d. PWR reactor cavity
- e. Reactor plenum
- f. Partial bypass of an ice condenser
- g. Complete bypass of an ice condenser
- h. Reverse flow
- i. Reactor cavity annular space A. 2 Systems Behavior
- a. Operator action
- b. Jet pump
ISSUE RESOLUTION B. FISSION PRODUCT RELEASES FROM FUEL B.1 Prior to Core Melting B.2 Core Temperature B. 3 Relative Timing of the Release of Species B.4 Calculated Core Inventory 8.5 Fission Product Release Equation I.
= _ _ _ _ _ . _ _ ... ____.
1 I l J i i i j ISSUE RESOLUTION j C. THERMAL HYDRAULICS i ! C.1 Subcompartments
- C. 2 Conservation of Mass and Energy j C.3 Upper plenum
} C.4 TML8' Sequence Pressure Spike i i C. 5 Adiabatic Gas Expansion in a j Containment i r ] C.6 Steam Sweeping i C. 7 Heat Transfer Coefficient C. 8 Core / Concrete Interaction j i i f. I i i - } I ! t f I i I i ) 4 i 1 1
ISSUE RESOLUTION
- 0. CHEMISTRY 0.1 Chemisorption
- 0. 2 0xidation Reaction 0.3 Hydrated Inorganic Species 0.4 Cesium Iodide
- 0. 5 Radiolytic Effects
- a. Transmutation
- b. Other isotopes
- c. Reactions 0.6 Tellurium D.7 Concrete
- 0. 8 Tellurium / Concrete Reaction 0.9 Boron Carbide 0.10 Corium Composition 0.11 Silver I
i
ISSUE RESOLUTION E. REACTOR COOLANT SYSTEM-TRANSPORT, DEPOSITION, AND REENTRAINMENT OF AEROSOLS E.1 Turbulent Deposition and Resuspension E. 2 Plugging in Pipes E. 3 Gravity Deposition in the Vessel - E. 4 Decay Heat From Released Fission Products E.5 Evaporation / Condensation E.6 Fission Product Removal in Steam Generators E. 7 Hindered Settling E.8 Deposition Mechanisms E.9 Upper Plenum l 1 i
- l
- l 4
l ISSUE RESOLUTION F. CONTAINMENT - TRANSPORT, DEPOSITION, AND REENTRAINMENT 0F AEROSOLS F.1 Aerosol Behavior
! a. Diffusiophoresis i
- b. Steam condensation 1
l
- c. Charging i
I d. Particle density / t e. Photophoresis, Thermophoresis i F. 2 Interaction of Aerosol Mechanisms j F. 3 Upward Draft F.4 Heterogeneous Atmosphere i j F. 5 Air Scrubbing l 4 F. 6 Homogeneous Nucleation 4 ! F. 7 Browton and Gravitational Coagulation i ! F.8 Ice Condenser Fog i 4 i i I i . I i f I i i l i
, s. w-. , , - , , .-.~,,---,.c-,.. .- ,,,,-, ,,-- ,,,,,,,, -, n - -,, , , , ,a_.-v.,,., ,,,,-..n.n,---r-.a--.,,,.,,-,,.,,ccv,.-.-,,,,~,. or,,,,,~,,
l ISSUE RESOLUTION
/
G. CONTAINMENT LOADS AND FAILURES G.1 Failure Mode, Time and Location G. 2 Hydrogen Generation t G.3 Hydrogen Igniters G.4 Hydrogen Burning Mode l G.5 Hydrogen Overpressure l j G. 6 Containment Leak Rate G.8 Modelling and a Probability I ] Distribution i G.9 Fission Product Retention l i
- i
! i I
i j i f i . 4 I i , i i ! l I i i i t l 1 i I
ISSUE RESOLUTION H. SUPPRESSION POOL l H.1 Water Flashing H. 2 Bubble Shape H.3 Bubble Movement H.4 Carbon Dioxide Solubility l l l l t l 1 l I I
ISSUE RESOLUTION I. ICE CONDENSER I .1 Shear Stress I.2 Ice Compression
- 1. 3 Decay Heat I.4 Layering I.5 Aerosol Deposition 1
e
s -~ .- a _ ISSUE RESOLUTION J. FILTERS l 4 ) d 9
.m - _ __. _-
- . . x .-
ISSUE RESOLUTION K. POLICY K.1 Sequence Selection K.2 PrejudgingResults K. 3 Use of Methodology K. 4 Sequence Probability K. 5 Comparison Study K. 6 Presentation l K. 7 Additional Analyses i K. 8 Review Process K. 9 QUEST l l l l l l l 1 l l
l 4 s ISSUE RESOLUTION L. MODELS L.1 GENERAL
- a. Core geometry degradation i b. Static. control volume i
a c. Number of control volumes j and nodal points
- d. Arrangement of control volumes
- e. Heat capacity
- f. Two phase flow
- g. Assumptions
- h. Aerosol particle size
- i. Iterative models
! j. Mixture levels
- k. Atomization j L. 2 NAVA
- a. Aerosol mechanisms
- b. Spray drop impaction term j L. 3 TRAPMELT 1
L.4 SPARC 0 1 4 l I
t i ISSUE RESOLUTION M. CODE RELIABILITY M.1 Validity
- a. PWR model for BWR calculations
? ] f i
- b. Reactor building 1
1
! M.2 Consistency in Modelling f
M.3 Uncertainty .
- a. General i
j b. Study M.4 Representatives ! a. Plant i
- b. Scaling up experiments l M. 5 Data i M.6 Interfacing Codes 3 M.7 QUEST i
3
- a. Type of study
- b. Tellurium releases I
{ s i ) i j i { 4 i i
{ l SECTION 2: Description of the Source Term Issues Key: 12345 refer to the first through the fifth peer review meetings x = reference to letter
- o = reference to transcripts Page numbers refer to the transcripts NOTE: Three of the five meetings were reviewed and summarized for this document: -
January 1983 - No summary written. May 1983 - summarized. July 1983 - summarized. October 1983 - No summary. Most issues were discussed at the fifth meeting. January 1984 - summarized. l l
. . -- l
f j Issue A. ACCIDENT SEQUENCES i j 1. Alternate Pathways . a. Complete, bypass of a suppression pool 1 The MARK-III containment design is less vulnerable to an inadvertent bypass of the suppression pool than the MARK-1 and MARK-II designs. However, in MARK-III containment, conduits communicate to the suppression pool; these conduits are potential bypasses. .Because a suppression pool removes significant amounts of l fission products, a bypass has adverse consequences. The definitions of the sequences neglect the bypasses, hence, the computer codes fall short of pre-dicting some important consequences.
- 1 2 3 i s
Reference:
Group - W. Kastenberg x A. Reynolds x Observers - none Contractors - R. Denning o P. Cybulskis o j Pages - 412 30 l Section 3 references: Page 26
.:- . i ~
i e t i 6 4 i i 1 l l
~-
Issue A. ACCIDENT SEQUENCES
- 1. Alternate Pathways
- b. Penetrations The electrical penetration of a containment or a drywell may fail before the concrete structure fails; the penetration seals and insulation may degradate under intense heat, radiation, and pressure. A release path through failed penetrations may be either diffuse or concentrated, depending on how many penetrations fail.
The release path would likely pass through the auxiliary building because that is where most of the penetrations lead. The analyses have focused on the failure of the containment structure and neglected a failure 4 of the penetrations. ~ Conducts leading to a drywell may vent a drywell overpressure. 1 2 3 e s
Reference:
Group - A. Reynolds x S. Levy x Observers - G. Petrangeli x Section 3 references: Page 27 4 2 i
Issue A. ACCIDENT SEQUENCES
- 1. Alternate Pathways 4
- c. Pipe break '
i The location of a pipe break will determine both the release pathway and the fission product removal mechanisms acting along the pathway. Hence, the point at which a pipe breaks has an influence on the consequences of a reactor accident. In some cases, the influence may be minor; that is, a break anywhere along a section of a pipe may give rise to similar pathways. In other cases, the influence may be significant; that is, different break points lead to i different pathways. The sequence definitions inadequately account for a variety of break points. ' In the Surry PWR, a break in a large diameter pipe of the ECCS system may occur in a section of the pipe in a water tight compartment. The compartment would flood. The fission products exiting the break would pass though about 3 ft of water. Fission products would be partially scrubbed by the water. 1 2 3 i 5
Reference:
Group - S. Levy x - R. Hilliard x W. Kastenberg x i A. Reynolds x R. Vogel x E. Warman o Observers - E. Rahn x Pages 218 i Section 3 references: Page 24 4 i k 4 . 3 y er g- e + e - ,+ ,es ,_. .c- .- -,*4.-~,-,--me..w- --%-- -,.g- w. r- -m-, er-- , - ,
Issue A. ACCIDENT SEQUENCES
- 1. Alternate Pathways
- d. PWR reactor cavity In some PWR containment designs, the reactor cavity may communicate with the containment. Water, accumulating on the floor of a containment, may find its way into the cavity. When the core melts through the vessel, a large volume of water will turn into steam and raise the containment pressure enough to cause an early containment failure. The possibility of such a pathway and its consequences have been given little attention.
1 2 3 i !
Reference:
Group - A. Reynolds x Observers none Section 3 references: Page 26
%.w. %
i 4 1
,, - - . - . , n , y - ., , .
Issue A. ACCIDENT SEQUENCES
- 1. Alternate Pathways
- e. Reactor plenum In many of the BWR and PWR sequences, the fission product release pathways begin at the core and pass through the upper plenum. However, eddy currents may produce a pathway through the lower plenum, especially when the core slumps and melts through the vessel. The release pathways through the lower plenum are not taken into account.
1 2 2 a s
Reference:
Group - S. Levy x x Observers - none Section 3 references: Page 26 5
l Issue A. ACCIDENT SEQUENCES l
- 1. Alternate Pathways
- f. Partial bypass of an ice condenser Steam and fission products are dispersed throughout the ice baskets of an ice i
condenser by turbulent flow. No' structures force the steam and fission pro-ducts through the stacks of baskets. Thus, a major fraction of the steam and fission products would flow through the ice condenser but around the ice baskets. The ice condenser would be effectively bypassed even though steam and fission products enter the volume of the ice bed. i 1 S d $ 5
Reference:
Group - D. Cooper xo Observers none 1. 1 Pages 176 Section 3 references: Page 27
- I e
6
Issue A. ACCIDENT SEQUENCES
- 1. Alternate pathways
- g. Complete bypass of an ice condenser An ice condenser may be inadvertently bypassed bec;use a' cicaligned valves or vents. The alternate routes are not considered i.i any of the sequences.
1 2 2 d 5
Reference:
Group - W. Kastenberg xo Observers none Pages 440 Section 3 references: Page 26 I
^'
l l 7 f _________a__ - + - r - . - ,
Issue A. ACCIDENT SEQUENCES
- 1. Alternate Pathways
- h. Reverse flow When containment recirculation flow fail, reverse flow through the ice beds may occur. This is not considered in the sequence.
1 2 3 3 }
Reference:
Reviewer - D. Cooper o i S. Levy x i D. Walker o Observers - None Pages 161 Section 3 references: Page 27 f e mammuma f 4 4 8 8
Issue A. ACCIDENT SEQUENCES
- 1. Alternate Pathways
- i. Reactor cavity annular space The annulur space in the reactor cavity may become a fission product pathway when the steel containment fails at some location and the shield building fails at a different location. A significant pathway arises where fission products could be removed. Furthermore, the reactor cavity is vented through an air cleaning system. Thus the shield building need not fail for a pathway to arise through the reactor cavity annulur. These alternate pathways are not accounted for in any sequence though they are significant.
The current models treat a BWR plant as merely the steel shell containment. The annular space between the steel shell and the concrete supporting struc-tures is ignored. This concrete cavity is lined with layer fiberglass and foam. Should the steel containment rupture, the fiberglass / foam liner would likely be damaged. The damaged surface and the remaining surfaces of the annular space should be effective towards removing fission products. 1 2 } i 1
Reference:
Group - none Observers - F. Rahn o E. Warman o Pages 167 240 Section 3 references: Page 26 9
J ] Issue A. ACCIDENT SEQUENCES
- 2. System Behavior
- a. Operator action By the definition of a particular sequence, a specific combination of events is assumed. The events represent the most likely behavior of the reactor hardware with little account of human factors. Human factors include operator intervention and operating procedures. An operator may relieve excess pressure in a drywell by venting the drywell pressure through the standby gas treatment system. The sequences define the most likely behavior of unattended hardware.
1 2 3 i s i
Reference:
Group - R. Hilliard x W. Kastenburg x
- 5. Levy x x i Observers - G. Petrangeli x 4
R. Sehgal x 1 Contractors - P. Cybulskis o Pages 27 ! Section 3 references: Page 25, 29 1
Response
' The sequences define only physical phenomena. Operator intervention or operating procedures are accounted for by the probability of a given sequence occurring.
Reference:
P. Cybulskis (Battelle), July Peer Review, p27 of transcript. i j l l i 10 1 _ . . , . . . _ _ _ . _ . , , . - ,- _ _ - ~ . - . . . . . . . ,
1 Issue A. ACCIDENT SEQUENCES
- 2. System Behavior
- b. Jet pump i
In some sequences a failure of the containment building leads to a failure of jetpumps. Some of the pumps may survive a containment. failure and continue to operate. Often, an extensive failure of a safety system is assumed, even 4 when at least a portion of the safety system should remain functional. These i assumptions lead to conservative source term prediction. 1 2 3 i s i
Reference:
Group - None Observers - D. Hankins o i Contractors - P. Cylbulskis o R. Denning o Pages 29 1 433 4 c Section 3 references: Page 27 i 4 I f 11
- - - , . . - . - - _ , . , - , , - - . , - , . ,-- ~ _ . , . . - . . - , - .
b Issue B. FISSION PRODUCT RELEASE FROM FUEL
- 1. Prior'to Core Melting Because a core begins to melt at high temperatures, the volatile fission products are likely to evaporate before the core melts. The modelling of.
early fission product releases is complex because the compositions of mixtures are unknown, the vapor pressures of mixtures are unknown, and the mobility of
' chemical species are unknown. Many of the volatile fission products may be out of a core prior to a meltdown, and hence, escape the melting phenomena.
The 'early fission product releases are significant and should be incorporated into the models. 1 2 1 1 1
Reference:
Group - A. Castleman x Observers - R. Hobbins x E. Warman x e 12
i t Issue B. FISSION PRODUCT RELEASE FROM FUEL
- 2. Core Temperature The release of fission products, both as vapors and as aerosols, is dependent on temperature. The maximum core temperature that can be achieved and the core temperatures that are achieved have not been determined.
1 2 2 1 5
Reference:
Group - A. Castleman x i D. Cooper x x J. Kelly x D. Rowe x Observers - R. Hobbins x
\
l i 1 1 , 13
Issue B. FISSION PRODUCT RELEASE FROM FUEL
- 3. Relative Timing of the Release of Species t-Both the fuel and the control rods contain chemical species having a range of volatilities. Some of the species are significant in themselves because they are fission products. Other species are significant because they serve as condensation nuclei. Chemical reaction may influence the release of fission products. Alloys, that form from fission products and cladding, may influence fission product releases. The vapors from a degraded core will likely have a composition that differs from the core (including control rods). The differential release of fission products may lead to different types of aerosols.
1 2 3 i s
Reference:
Group - W. Kastenberg x x C. Johnson x i Observers - D. Campbell x S. Niemczyk x 4 Contractors - M. Kuhlman o Pages 124 221 Section 3 references- Page 9 , 14
] Issue B. FISSION PRODUCT RELEASE FROM FUEL
- 4. Calculated Core Inventory The amount of radioactivity calculated to remain in a core is excessive for two reasons; first, the high temperature in a molten core will force fission product emission; second, the surface area assumed for retention should be lower than estimated due to melting.
1 1 2 $ 5
Reference:
Group - 5. Levy x Observers - none Section 3 references: Page 9 s A 15
Issue 8. FISSION PRODUCT RELEASE FROM FUEL
- 5. Fission Product Release Equation The fission product release equation used in CORSOR and in NUREG-0772 has the following form:
' BT K = Ae where K = release rate A = constant B = constant T = temperature A more conventional form of the equation is as follows:
K = AeQ/(QRT) where Q = activation energy R = gas constant This is the Arrhenius equation. 1 2 a i 1
Reference:
Group - C. Johnson x , j .A. Reynolds x l 1 Observors - J. Kelly x 1 0 l. 1 n. I 16
4 Issue C.: THERMAL HYORAULICS
- 1. Subcompartments The thermal hydraulic characteristics of fluid flow through the subcompartments of a containment do not appear in the models. The pathways through a containment
'l are oversimplified. The thermal hydraulic behavior of aerosols and fission products along pathways in a containment may be inadequately known for modelling. The same arguement may apply to the thermal hydraulic and deposition in the auxiliary buidling. Significant deposition may occur in these areas of a plant. I 1 S 2 S 5 Referencc: Group - A. Reynolds x Observers - G. Petrangeli x ]; E. Warman x l 4 e f
..=m-==_,
j 17
i i i l l Issue C. THERMAL HYORAULICS
- 2. Conservation of Mass and Energy It is unclear whether the codes simulate the movement of mass and energy through a reactor system or strictly account for finite amounts of mass and energy. For example, given that_ vapor has condensed on an aerosol particle and later evaporates, the models should allow only a finite amount of mass to evaporate. An accounting of the location of water is not kept.
1 2 } } }
Reference:
Group - A. Reynolds x o D. Rowe x x R. Vogel x x Observers - H. Kouts x R. Sehgal x Contractors - R. Denning o Pages 101 142 1 1 18
Issue C. THERMAL HYDRAULICS
- 3. Upper Plenum In the Surry (PWR) analysis, the flow through the upper reactor plenum is inadequately modelled. Descriptions of the models are unclear about what surfaces in the upper plenum are available for aerosol deposition. It is unclear how the surfaces are modelled. The models oversimp'ify the deposition phenomena in the upper plenum.
The internals in the upper plenum may melt and cause a change in the flow pattern. In the BWR sequences where fission products flow out of the reactor vessel ' through the steam dryers in the upper plenum, the flow bifercates such that 15% of the fission products go through the steam dryers an 85% of the fission products bypass the steam dryers. The bases for this flow is unclear. 1 2 3 4 5
Reference:
Group - S. Levy x Observers - J. Kelly x R. Hobbins x
- 8. Sehgal x Contractors - P. Cybulskis o R. Denning o Pages -
32 54
Response
The upper plenum of the reactor vessel at Surry is considered to be a well mixed volume with three regions, the top support plate, the control rod drive tables, and the core barrel. Westinghouse described the upper plenum in terms of surface areas, flow areas, and structure thicknesses.
Reference:
P. Cybulskis, January 1984 peer review meeting, p 32 of the transcript. The "15%" and the "80%" figure are based on literature provided by the General Electric Corporation.
Reference:
R. Denning, July 1983 Peer Review Meeting, p 54 of the transcript. 1 I 19 l l
\
i 1 i Issue.C. THERMAL HYDRAULICS
- 4. TMLB' Sequencer Pressure Spike An 89 psi pressure spike is predicted by the TMLB' sequence in the Surry (PWR) analysis. The cause of the pressure spike is unidentified.
j If the cause is a hydrogen burn, then the occurrence of such a spike is questionable because steam will likely attenuate the burn. 1 2 1 1 E
Reference:
Group - A. Reynolds x Observers - R. Sehgal x e I f s b f 4 i 1 20 l.
Issue C. THERMAL HYORAULICS
- 5. Adiabatic Gas Expansion in a Containment When a containment fails, the rapid pressure drop should lead to an adiabatic gas expansion. The adiabatic expansion should enhance the condensation of steam and lead to enhanced fission product removal. This phenomenon is not modelled.
1 2 3 1 1
Reference:
Group - D. Cooper x o A. Reynolds x Observers none Contractors - P. Cybulskis o Pages 29 49 l Section 3 references: Page 7
Response
As a containment failure is currently modelled, material exists the failure at a constant enthalpy and then expands. A two phase flow does not necessarily arise. Instead, a superheated state may result. If an isotropic expansion is assumed, then a two phase region is predicted.
Reference:
P. Cybulskis (Battelle), July 1983 Peer Review Meeting, 5 p 29 of the transcript. 21
- - - , , -r- - - - - - , -,,s y
,-,,.v. -- - - - - -
! Issue C. THERMAL HYDRAULICS
- 6. Steam Sweeping Rapid quenching of the debris of core may cause a steam pulse that reentrains aerosol particles.
1 Experiments done at the Brookhaven National Laboratory on core water interaction show a series of steam explosions occurring. The steam explosions tend to 4
' disperse the core debris, making the debris coolable. But experiments done at the Sandia National Laboratory show no such explosions. Hence, the debris is , not dispersed and may remain uncoolable. The core-water phenomenon is poorly 1
understood and is not well modelled. 1 2 3 4 5 I
Reference:
l Group none ! Observers - G. Greer.e o R. Hobbins x E. Rahn x { S. Nienczyk x i Contractors - D. Powers o i Pages 364 l Section 3 references: Page 10 - i I l l a 2 J 1 i 3 22
fi Issue C. THERMAL HYDRAULICS
- 7. Heat Transfer Coefficient.
The calculated heat transfer coefficient for a steam generator is erroneous.
.The calculation is based on normal flow through a steam generator. During an accident, when one side of the steam generator may be stagnant, the heat transfer will be much less than during normal operation. The predicted steam generator heat transfer rate should not depend on the core heat transfer model.
Battelle assumed that the heat transfer between the debris of a degraded core and the water in a reactor cavity was efficient. The heat transfer calculations include thermal resistances. 1 2 3 3 5
Reference:
Group - R. Ritzman x o Observers - None Pages 28 Section 3 references: Page 3 i i 4 i i i i l 23 l r
.,A4e--__ams. +--.# "su--+ ~.A -e.I+ m -- .- -.-2 h*- E-2 -<a+ -.~.- -4h ~- - ,n I
i 1 i Issue C. THERMAL HYDRAULICS
- 8. Core / Concrete Interaction The postulated corium/ concrete / water interaction is based on metal / concrete /
water interaction experiments. The proposed scrubbing efficiency is larger than that used for calculation on the suppression pool at Beach Bottom on Grand Gulf. Film boiling should occur at the molten corium/ water interface. Bubble sizes should be described by a Taylor initability calculation taking into account the changes in surface tension from impurities in the water. 1 2 3 i 1 j
Reference:
Group - S. Levy 1 x Observers - None i Section 3 references: Page 16 l i
}
'I a
-w m %
4 24
Issue D. CHEMISTRY
- 1. Chemiosorption The definition of chemisorption is unclear.
Chemisorption is not a strong interaction; " reaction" was being confused with
" adsorption. " The composition of metal surfaces will influence the retention of fission products; a deposition velocity for a given fission product should be a function of both the surface area and the surface composition.
At high temperatures and pressures, the distinction may not exist. Furthermore, the type of sorption may vary from one chemical specie to the next. Fission . products may sorb onto surfaces, including aerosols. I 1 2 3 5 3 i
Reference:
Group - A. Castleman x C. Johnson x o A. Reynolds x R. Vogel o j Observers - D. Campbell x i J. Cobble o R. Hobbins x i J. Kelly x S. Loyalka x G. Petrangeli o R. Sehgal x E. Warman - x Pages 124 236 133 221 289 324 Section 3 references: Page 18 1 25
Issue D. CHEMISTRY
- 2. Oxidation Reaction The changes in the composition of aerosol and vapors when hydrogen burns has not been determined. Also, heat of reactions are unaccounted for.
I' The composition of the aerosols is considered to be constant. ) A hydrogen burn may influence vapors as well. One fission product that could be effected.is cesium iodide: CI s
+ Cs 0H + HI (unbalanced) 1 2 3 3 1
References:
Group - R. Hilliard x
- 5. Levy x Observers - S. Niemczyk x
- S. Loyalka x x D. Wren x Contractors - M. Kuhlman o Pages 60
~
Section 3 references: Page 11 1 f t a 26
Issue D. CHEMISTRY
- 3. Hydrated Inorganic Species At high temperatures and pressures, an inorganic molecule may be hydrated.
Because the hydrated molecule has a lower vapor pressure than the unhydrated molecule, the hydrated molecule is more mobile than the unhydrated molecule. The mobility of some fission products is underestimated because the hydrating phenomenon is unaccounted for. 1 2 } d 1
Reference:
Group - A. Castleman x Observers - J. Cobble o o D. Wren x Pages 140 219 Section 3 references: Page 19 mae e -w % _, 27
Issue D. CHEMISTRY
- 4. Cesium Iodide In the Surry (PWR) analysis, cesium iodide is considered to be undissolved in the primary system. Given the volume of water in the primary system, cesium iodide should be dissolved.
In the Surry (PWR) analysis, the report states on page 6-31 that iodide is present as nonvolatile cesium iodide. However, cesium iodide in relatively volatile. The chemical form of iodine was considered to be entirely I . This was changed to C I Other forms of iodine, such as iodomethane2 have not been considered. s .Little is known about the chemistry of iodine. For the BWR AE sequence, cesium iodide and cesium hydroxide are predicted to be released in equal fractions. This is surprising since C 0H is much more reactive than Cs I* s Cesium reacts more rigorously with Inconel than it does with steel. In the reaction with steel, cesium is found associated with silicates. Below 1000 C the cesium is loosely bound. Above 1000*C the cesium is tightly bound. Cesium iodide will likely react with structural metals such that cesium is bound to the metals and iodide is released as HI and/or I2-1 2 3 5 3
Reference:
Group - A. Castelman - o S. Levy x A. Reynolds x R. Hilliard x C. Johnson x Observers - D. Campbell x E. Warman x C. Pellitier x Contractors - D. Powers o Page 201 207 125 215 Section 3 references: Page 2, 14 28 l 1:
l l Issue D. CHEMISTRY
- 5. Radiolytic Effects
- a. Transmutation Isotopes transmute into different fission products. For long-lived isotopes, the contribution of transmutation is insignificant. For short-lived isotopes, the gain and loss of chemical species may be significant.
The fate of cesium after the decay of iodide is unaccounted for. The mobility of cesium may be much different than predicted because a significant fraction of cesium may be liberated from iodide by isotope decay. 1 2 3 4 {
Reference:
Group - W. Kastenberg x o Observers - R. Hobbins x D. Wren x 1 Contractors - T. Taig . o Pages 283
- 480 i
Section 3 references: Page 17 . i
Response
The bulk of the isotopes appear to be long-lived isotopes.
Reference:
T. Taig, p480 of the January 1984 meeting.
)
a 29 i i
Issue D. CHEMISTRY
- 5. Radiolytic Effects
- b. Other isotopes Other isotopes, such as strontium, barium, ruthenium, and rubidium, may also contribute to risk. Consequences cannot be accurately predicted by considering many such isotopes as a category "other."
Zirconium enhances the release of barium and strontium while retaining tellurium. 1 2 3 e s
References:
Group - A. Castleman x
'V. Kastenberg x x A. Reynolds o Observers - J. Kelly x E. Warman x Contractors - M. Kuhlman o Pages 274 393
Response
The fission products that are modelled are the volitile species - cesium, iodine, and tellurium. Less volatile fission products, such as barium and strontium are also modelled - these isotopes are considered of secondary importance.
Reference:
M. Kuhlman. May 1983 Peer Review Meeting, p 393 of the transcript. 30
J Issue D. CHEMISTRY
- 5. Radiolytic Effects
- c. Reactions Radiolytic reactions are not considered because little is known about how they might occur in a post-accident containment.
1 2 3 4 5
Reference:
Group - R. Hilliard x '. Observers - D. Campbell x J. Kelly x
- 0. Wren x i
Section 3 references: Page 17 i i 1 i l i 1 31
t . Issue D. CHEMISTRY
- 6. Tellurium
- Models have tellurium leaving a core in the elemental form. Compounds of tellurium, such as gaseous H2 Te, are not considered.
1 2 3 1 1
References:
Group - C. Johnson x A. Reynolds x
- R. Vogel x Observers - D. Campbell x J. Kelly x E. Warman x
- 0. Wren x Section 3 references: Page 2
't I e E 4 l 1 l 4 4 I 32 4
, . . - , .---.e-., ,, .. -, . . , . . . . , ,,,-,-..-..-....-.na, .-,,,m,,-. -..,-e.,,. -
1 Issue D. CHEMISTRY
- 7. Concrete ,
The composition of the concrete in a reactor cavity determines the extent to which many important reactions occur. One such reaction is the conversion of carbonates to carbon dioxide. When the concrete composition is unknown, the limestone content is assigned a value of 80%. This leads to 36% carbon dioxide in the concrete. In the Grand Gulf plant, the reactor cavity concrete is 50% to 60% carbonates. This leads to about 22% carbon dioxide in the concrete. There, the codes overpredict the amount of C07 generated by a core melt to challenge a containment. The predicted consequences are overestimates. In the Grand Gulf analysis, the water content of the concrete was assumed to be about 6%. The water content is about 10% to 15%. The water content influences the amount of hydrogen and carbon dioxide from a core-concrete interaction. The steel content should also influence the core-concrete interaction. Rebar itself has a minor effect; some reactors (i.e., TMI) have a steel plate in the concrete reactor cavity. 1 2 3 4 5
Reference:
Group - D. Campbell o R. Vogel o D. Walker o Observers - none Contractors - P. Cybulskis o o A. Reynolds x R. Sehgal x
- 0. Wren x Pages 153 42 27 159 300 307 Section 3 references: Page 21
Response
The difference between 36% carbon dioxide and 22% carbon dioxide would make a difference in the challenge to a containment. The difference should not be very large.
Reference:
P. Cybulskis (Battelle), July 1983 Peer Review Meeting, p 42 of the transcript. Discussion 33
Issue D. CHEMISTRY
- 8. Tellurium / Concrete Reaction Current experiments indicate how tellurium reacts as it is released from fuel and how it reacts during a core-concrete interaction. Reactions of tellurium and concrete have not been studied. Tellurium-concrete reactions may have a
, significant influence on tellurium release. 1 2 _ _3 _4 5
Reference:
Group none Observers - G. Petrangeli o Pages 232 233 l 1 Section 3 references: Page 16 i .( i 4 l 4 4
$MO+D I
i I r l l l 4 i i l 34 i
. . , , . - _ _ . , . , , - . - , - . . - , . . - , - , , . - . , ,,...n~-.-...-,--,._,---.~-r-,.,-
,,. . --, ,-.,- -.n..,,,.,.
. .. - ~ .- .- .. - - . _ _ _ . .- .-
1 i Issue J. CHEMISTRY
- 9. Boron Carbide Boron carbide is omitted from consideration in the simulate BWR core inventory.
This is questionable for two reasons: (1) Because control rod boron can be oxidized to volatile boron com-pounds, boron carbide can contribute to the aerosol mass. I ' (2) . Boron compounds can react with radionuclides; baron oxide reacts with cesium iodide to form cesium borates and iodine. Steam reacts with boron carbide forming a boron oxide layer that has a high affinity for cesium.
~
Cesium iodide reacts with boron oxide. Iodine is liberated. ! Boron compounds may have a significant influence on the chemical reactions of radionuclides such as cesium and iodine. 1 2 3 5 3
Reference:
Group - L. Zumwalt x Observers - J. Cobble o S. Niesczyk x x x Contractors - D. Powers o Pages 212 i 478 213
- 215 4
i Section 3 references: Page 15 1 1 l i ,
}
i 35
Issue D. CHEMISTRY
- 10. Corium Composition The composition of corium should depend somewhat on the type of sequence that occurs.
This is due to the rate at which the U02 and structural materials melt and mix. The corium composition ultimately influences the fission product release. 1 E 2 i 5
References:
Group - D. Rowe x Observers - S. Niemczyk x Section 3 references: Page 9 l ! 36 i i
4 i 1 Issue 0. CHEMISTRY I
- 11. Silver Little is know about the release of silver. Silver may serve as a source of condensation nuclei onto which aerosols form.
I Silver may bind a significant amount of iodine as silver oxide, a precipiate. 1 2 3 4 5
Reference:
Group - A. Castleman x
' D. Cooper x A. Reynolds x x Observers - D. Campbell x J. Kelly x Section 3 references: Page 2 1
e f I l i 1 5
-emr***%
4 s 37 k _ _ _ _ . _ , _ , ._. . ~ _ . _ _ _ _ . _ _ . , _ . _ . . . _ . - .
Issue E. RCS - TRANSPORT, DEPOSITION, AND REENTRAINMENT OF AEROSOLS
- 1. Turbulent Deposition and Res'uspension When aerosols pass through pipes at high speeds, turbulent deposition and resuspension should occur. These phenomena do not seem to be adequately modelled. During post accident conditions, the turbulent flow through pipes may not be the same as turbulent flow during normal conditions. For example, one thought is that the aerosols and vapors spiral through the pipes without touching the pipe walls;,this flow is extremely complicated.* Not enough is j known about-this turbulent flow to model it.
) 2 1 3 i 1 i
Reference:
Group - D. Cooper x A. Reynolds x Observers - S. Beal o i R. Hobbins x 1 Section 3 references: Page 3 j i 4 l i a i i a i Discussion with S. Beal not as on record. 4 L 38
Issue E. RCS - TRANSPORT, DEPOSITION, AND RESUSPENSION OF AEROSOLS
- 2. Plugging in Pipes
- Aerosols deposit in the pipes of a reactor coolant system. Little is known about how the aerosols accumulate and less is known about how the aerosols resuspend.
Furthermore, resuspension mechanisms are ceglected from the models. Thus, the way in which accumulated aerosols modify flow is unknown. A calculation of the volume of fission product material rather than the mass of fission produce material may be necessary to determine the significance of plugging. . The models fail to accurately predict flow through pipes because plugging from aerosol deposits is not considered- .
Reference:
1 2 3 i { Group - D. Cooper x x x A. Reynolds x i Observers - G. Petrangeli x Section 3 references: Page 3 1 1 1 2 J l 39
Issue E. RCS - TRANSPORT, DEPOSITION, AND REENTRAINMENT OF AEROSOLS
- 3. Gravity Deposition in the Vessel When aerosols are released from a degrading core, the aerosol concentration in the vessel is likely to be high. Aerosol particles should rapidly agglomerate.
Particles should deposit rapidly by gravity deposition. A result of the Surry (PWR) analysis indicates a highly concentrated aerosol persisting to (location ? ) for thousands of seconds. An aerosol of any appreciable concentration is unlikely to be stable. Agglomeration occurs. Concern is that modelling is incorrect or the results were misinterpreted. 3 An aerosol concentration of 1 kg/m is surprising, almost without an analog in experience; a highly concentrated aerosol is unlikely to persist. A cloud of particles having a concentration in the range of kilograms per-cubic meter may behave as a fluidized bed rather than as an aerosol. 1 2 3 i s
Reference:
Group - D. Cooper x x R. Vogel xo Observers none Pages 74
-W Mw+ wo, l
1 I 40
)
Issue E. RCS - TRANSPORT, DEPOSITION, AND REENTRAINMENT OF AEROSOLS
- 4. Decay Heat From Released Fission Products Hardware surfaces should be warmed by the decay heat from fission products.
The heat should create convectior currents and modify the deposition both aerosol and vapor fission products. The decay heat may also melt pipes. Fission products may resolve and form aerosols. Neglecting decay heating from the source term calculations can result in an underestimated source term. If fission products revaporize from the primary a system while an areosol is in a containment, the aerosol may act as conden-sation nuclei and remove fission products. If fission products revaporize after the aerosol has settled, then the fission products may remain suspended. To determine the sensitivity of the predicted retention by the reactor coolant system, Battelle estimated the heat load from the fission products and calcu-lated how this would modify deposition. According to the calculations, decay heat reduces the deposition of vapors but not of aerosols in the reactor coolant system. The models underestimate the fission product release because too much credit is given for retention by the reactor coolant system. 1 2 3., i 5_
Reference:
Group - D. Cooper. x R. Hilliard x S. Levy x A. Reynolds x x
- 0. Rowe x x R. Vogel .
o L. Zumwalt x Observers - R. Hobbins x J. Kelly x G. Petrangeli x E. Warman o Contractors - R. Denning o M. Kuhlman o Pages 56 79 57 82 58 87 59 178 67 Section 3 references: Page 1 41
Issue E. RSC - TRANSPORT, DEPOSITION, AND REENTRAINMENT OF AEROSOLS
- 5. Evaporation Condensation .
The decay heat from fission products may cause deposits to vaporize. The vapors may condense and form an aerosol. The aerosol may then settle to start the evaporation - condensation process repeatedly.
;. 1 2 3 4_ 5_
s
Reference:
Group - A. Castleman x x
- 0. Cooper x R. Hilliard x W. Kastenberg x S. Levy x R. Vogel x Observers - H. Kouts x R. Hobbins x l
i Section 3 references: Page 1 i ( t e I I J
,+
I i 42 i _, _, , , , --7.-.---em- = e-- *^e - ' - ~ ' - ' - ' ' " - * -~'"~*"" -'-'* ~ ~ -" ^ ' '
Issue E. RCS - TRANSPORT, DEPOSITION, AND REENTRAINMENT OF AEROSOLS
- 6. Fission Product Removal by Steam Generators During a severe accident, the steam generators may remain as an effective heat sink. Significant amounts of steam may condense within the steam generators and remove fission products. These phenomena may be a significant mechanism for removing fission products.
1 2 3 i }
Reference:
Group - S. Levy x
- 0. Rowe x xo Observers - F. Abbey x E. Rahe x E. Warman o Contractors - P. Cybulskis o pages 234 34 Section 3 references: Page 3
Response
At least in some sequences, the steam generators may not receive fission products.
Reference:
P. Cybulskis, January 1984 peer review meeting, p 34. 43
d 4 Issue E. RCS - TRANSPORT, DEPOSITION, AND REENTRAINMENT OF AEROSOLS
- 7. Hindered Settling A high concentration of aerosols may lead to hindered settling. The importance of this phenomenon is open for discussion. At a concentration of 1 kg/ma , the settling velocity of aerosol particles is reduced by about 15%. Such concen-
' trations are extremely high and unlikely to persist. The phenomenon should be { given adequate consideration. 1 1 g } d 5 t
Reference:
Group - D. Cooper x x t
- Observers none l
L 4 h j l i J i l l I I, l j l i . ]. 44
Issue E. RCS - TRANSPORT, DEPOSITION, AND REENTRAINMENT OF AEROSOLS
- 8. Deposition Mechanisms 6
Deposition in the reactor coolant system is not well understood. Moisture or bulk water in the pipes will influence deposition. The models do not adequately account for Stefan flow, electrostatic interactions, and turbulent reentrainment.
~
A consistent methodology needs to be developed; general space dependent forms of aerosol equations are needed; unambiguous deposition coefficients are needed; recented turbulent deposition correlations (i.e. DeMota, Friedlander) should be used. Deposition rate expressions and correlations seem to be ad hoc. 1 2 3 i i t
Reference:
Group - D. Cooper x x x R. Vogel x Observers - R. Hobbins x H. Kouts x S. Loyaika x G. Petrangeli x l Section 3 references: Page 3 5 I 45
+
Issue E. RCS - TRANSPORT, DEPOSITION, AND REENTRAINMENT OF AEROSOLS
- 9. Upper Plenum The assumed homogenous mixture in the upper plenum needs to be reconsidered.
The temperature of the upper plenum and fission product in the composition in the upper plenum will likely be heterogeneous. Some high temperature streams will transport fission products more than the current models predict. 1 S 2 i 5
Reference:
Group - S. Levy x x Observers none M
~
. - - - - ~
1 i i 4 46 i
Issue F. CONTAINMENT - TRANSPORT, DEPOSITION, AND REENTRAINMENT OF AEROSOLS
- 1. Aerosol Behavior
- a. Diffusiophoresis The significance of diffusiophoresis has not been determined. At first, this phenenomenon was not modelled. Refined models included diffusiophoresis.
However, the consensus among the code developers is that diffusiophoresis is insignificant for two reasons; (1) the steam in a containment should condense before aerosols enter the containment; (2) the surface / volume ratio of a containment is small. Diffusiophoresis may be significant because it is driven by three gradients - a temperature gradient, a concentration gradient, and a pressure gradient. 1 2 3 i s References- Group - D. Cooper- x o A. Reynolds x Observers - S. Beal
- o o S. Loyalka x Contractors - J. Gieseke o T. Kress o K. Lee o l Pages 44 22 390 48 142 198 Section 3 references: Page 5 i
Response _. Though large amounts of steam and fusion products are expected in a post accident containment, the steam condenses before the fusion products are released.
Reference:
J. Gieseke, July 1983 peer review meeting, p 22 of transcript. Diffusfophoresis can be a dominant mechanism for the removal of small particles. In an ice condenser model, it is accounted for by steam removal calculations.
Reference:
A. Postma, July 1983 peer review meeting, p 198 of the transcript.
*S. Beal, " Review of Computer Models of Containment Aerosol Deposition,"
June 1983. Advisory Committee on Reactor Safeguards, Nuclear Regulatory Commission, Washington, DC 20555. I 1 1 47
Issue F. CONTAINMENT - TRANSPORT, DEPOSITION, AND REENTRAINMENT OF AEROSOLS
- 1. Aerosol Behavior
- b. Steam condensation Steam can condense on both aerosols and on walls. The steam condensation phenomenon is not well understood for two reasons; (1) the affinity of aerosols for water is unknown; (2) a quantitative description of a boundary layer at a wall is unknown.
Mason's equation is used to describe condensation and evaporation; this equation. when used to describe heat and mass transfer, is invalid for small particle sizes. The heat changes during condensation and evaporation may have a significant influence on the thermodynamic behavior of the aerosols. Condensation readily occurs on particles as small as 0.6 pm. If the particles are hydroscopic, then water can condense on the particles when the atomsphere is unsaturated. Steam condensation is a significant way to remove fission products, and, therefore, deserve a rigorous study. 1 2 3 3 5
References:
Group - W. Castleman x D. Cooper x x S. Levy x A. Reynolds x x - Observers - S. Beal
- R. Hobbins x
- 0. Campbell x S. Loyalka x 4
- Section 3 references
- Page 7
*S. Beal, " Review of Computer Models of Containment Aerosol Deposition," ~
June 1983. Advisory Committee on Reactor Safeguards, Nuclear Regulatory Commission, Washington, DC 20555. 48
d 1 I Issue F. CONTAINMENT - TRANSPORT, DEPOSITION, AND REENTRAINMENT OF AEROSOLS
, 1. Aerosol Behavior
- c. Charging a
j Inside a post-accident containment, radiation fields are likely to exist. The
- radiation fields should charge aerosols. Though the entire aerosol would be i
electrically neutral, the positive and negative ions would have different j mobilities. Thus, an electrical charge should modify the behavior of an
- aerosol. This phenomenon has been given inadequate attention.
l
- 1 2 3 4 5
Reference:
Group - D. Cooper x xo i' Observers none i Contractors - J. Gieseke o
- K. Lee o Pages 142 137 37 t
Section 3 references: Page 17 9 1 4 1 i [ 1 1 4 i i 49 i 4
, - - . , . - , ,. - --. . ~ , . - . - _ - , - _ , - . - - . . , , - , . , ,
Issue F. CONTAINMENT - TRANSPORT, DEPOSITION, AND REENTRAINMENT OF AEROSOLS
- 1. Aerosol Behavior
- d. Particle density The density of aerosol particles has not been determined. The particle density is a major factor in the behavior of aerosols. The currently used value of 10g/cma has been assumed with little scientific information.
1 E 1 1 5
Reference:
Group - D. Cooper xo R. Hilliard o Observers - T. Kress o Contractors - M. Kulhman o Pages 111 67 388
Response
By the January 1984 meeting, the calculations were revised, a bulk density of 3 g/cm3 is used.
Reference:
M. Kulhman, January 1984 peer review meeting, p 67 of the transcript. 50
l l
, Issue F. CONTAINMENT - TRANSPORT, DEPOSITION, AND REENTRAINNENT OF AEROSOLS i
- 1. Aerosol Behavior
- c. Photophoresis, Thermophoresis i ,
Photophoresis has gone unacknowledged. Its importance should be established. i- The aerosol mechanisms have been inadequately assessed. i, 1 2 _ _3 _4 _5
Reference:
Group - none } . Observers - R. Hobbins x il S. Loyaika x i i i i i i 1 j l i i ) I i . l 1 i i s i-I l i
- 51 i
i 1
, , . . - . , - - - . , . , - - ,,-.-wn..,
. . , , . - ,.e.,~ . ,,, ,, ,,,. ,, -. n .,,,,,,, ---,, ,,,,,+..._-,nm.,,,,,.-_ _.,.,.,_,,._,,--n,.,-p.m,.-,-,. ,
i Issue F. CONTAINMENT - TRANSPORT, DEPOSITION, AND REENTRAINMENT OF AEROSOLS
- 2. Interaction of Aerosol Mechanisms Some aerosol mechanisms, such as gravity settling and impaction, are independent of one another in a thermal hydraulic sense. Other mechanisms, such as thermo-phoresis, diffusiophoresis, and turbulent deposition, operate on an interaction of heat, mass, and somemtun transfers. The interaction is neglected from the modeling. The thermal hydraulic behavior of many aerosol mechanisms is
- incorrectly modelled.
Aerosols of cesium hydroxide and cesium iodide are treated as being separage and independent. Actually a single aerosol composed of varying amounts of each compound will likely exist. 1 2 3 e s
Reference:
Group none Observers - S. Beal
- o o S. Loyalka x D. Wren x COMMENT Coagulation coefficients may be altered by orders of magnitude when ambient i
conditions are changed by small amounts. This observation is attributed a strong synergistic relationship among the aerosol mechanisms (thermophoresis, diffusiophoresis,etc.)
"Mechamics of Aerosols in Nuclear Reactor Safety: A. Kesrew.
S. Loyalka. Progress in Nuclear Engineering, Vol. 12, No. 1 p 4, 10, 22-3, 26 l Section 3 references: Page 6 i
*S. Beal, " Review of Computer Models of Containment Aerosol Deposition,"
June 1983. Advisory Committee on Reactor Safeguards, Nuclear Regulatory Commission, Washington, DC 20555. 52
Issue F. CONTAINMENT - TRANSPORT, DEPOSITION, AND REENTRAINMENT OF AEROSOLS
- 3. Upward Draft Convection down the walls of a conta'inment is likely to produce an upward draf t in the center of the containment. The draft may hinder the gravity settling of aerosols. This assumes that the atmosphere in a containment behaves as a homogeneous mass.
I Sedimentation on horizontal surfaces should not be effected by an upward air velocity when an atmosphere is homogeneous. Aerosol particles will still deposit within a boundary layer. Settling would be prevented if gas emenated from a surface, such as during evaporation. i The models neglect the upward draft when predicting gravity settling. 1 3 3 $ 5
Reference:
Group - D. Cooper x R. Hilliard x
' Observers - S. Beal
- Contractors - T. Kress o -
- Pages 373 374 i
Section 3 references: Page 8 , 1
*S. Beal, " Review of Computer Models of Containment Aerosol Deposition,"
June 1983. Advisory Committee on Reactor Safeguards, Nuclear Regulatory Comission. l 53 l
Issue F. CONTAINMENT - TRANSPORT, DEPOSITION,.AND REENTRAINMENT OF AEROSOLS
- 4. Heterogeneous Atmosphere The thermal hydraulic conditions in a containment during post-accident condi-tions is uncertain. One thought considers the atmosphere in a containment to be homogeneous; the atmosphere behaves as a single volume. Convection currents and steam should keep the atmosphere well mixed. Another thought considers the atmosphere to be heterogeneous; the atmosphere behaves as many small volumes. Hydrogen may stratify a containment atmosphere.
1' 2 3 i !
Reference:
Group - A. Reynolds x
- 0. Rowe o Observers - S. Beal
- T. Ginsberg x R. Henry o R. Hobbins x Contractors - P. Cybulskis o o i
Pages 162 164 Section 3 references: Page 8
*S. Beal, " Review of Computer Models of Containment Aerosol Deposition,"
June 1983. Advisory Committee on Reactor Safeguards, Nuclear Regulatory ' Commission. 54
4 l i I
. I Issue F. CONTAINMENT - TRANSPORT, DEPOSITION, AND REENTRAINMENT OF AEROSOLS ! ; 5. Air Scrubbing i
i Water drops falling through the atmosphere of a containment can scrub the air of fission products. The mechanisms are well understood. Also, the distribu-1 tion of water drop sizes from a spray system has not been rigorously determined. 1 Therefore, the modelling of scrubbing by spray water drops seems to be unneces-
, sarily inadequate.
l . j The water drops from melting ice are expected to remove an insignificant j amount of fission products. Only aerosols, no vapors, are assumed to be going j through an ice condenser. } _l _2 _3 _4 _5 'i
Reference:
Group - O. Cooper x
! 5. Levy x ; A. Reynolds x j R. Ritzman o 1
j Observers none j
- Contractors - K. Lee o o Pages 51 176 177 !
~
Section 3 references: Page 5 i i i i 1l 1 7 i
Response
j A distribution of spray drop sizes makes the calculations unnecessarily tedious ' and difficult.
Reference:
K. Lee, May 1983 peer review meeting, p 51 of the transcript. l . i, 1 1
- 55 i
.__...-__,,.,..-_-_____._..___,...,_,._..__,,.,_..._,._.,_,,,_.,_,..,,.._-_,_,_,__._._.m__..__ . , . - , _ , . ~ . , , , -
i Issue F. CONTAINMENT - TRANSPORT, DEPOSITION, AND REENTRAINMENT OF AEROSOLS
- 6. Homogeneous Nucleation Homogenous nucleation is described using outdated equations. Old textbooks are readily available. Nevertheless, in recent years, many advanced equations describing nucleation have been developed. The equations in the current source , term models are being related on the basis of availability rather than on the basis of appropriateness.
Ionic nucleation could occur in the reactor coolant system (or in the containment). 1 $ $ $ 5
Reference:
Group - A. Castleman x Observer'- S. Loyalka x Section 3 references: Pages 7, 17 N f i e 0 i l 56 l l t - J
Issue F. CONTAINMENT - TRANSPORT, DEPOSITION, AND REENTRAINMENT OF AEROSOLS
- 7. Browian and Gravitational Coagulation All expressions describing Browian coagulation and gravitational coagulation have limitations. The expressions used in the NAVA code should be examined.
1 E 1 $ E
Reference:
Group - none Observers - S. Loyalka x 1 1 3
*" Mechanics of Aerosols in Nuclear Reactor Safety: A Review,"
- 5. Loyalka. Progress in Nuclear Engineering, Vol.12, No.1.
57
Issue F. CONTAINMENT - TRANSPORT, DEPOSITION, AND REENTRAINMENT OF AEROSOLS
- 8. Ice Condensor Fog The impact of the fog model on the calculated source terms is unclear. Instead, a model based on the carryover of water with steam might be preferable. The validity of the bag.model is questionable.
1 2 3 i s
Reference:
Group - S. Levy x Observers - none. Section 3 references: Page 5 I s8 l
Issue G. CONTAINMENT LOADS AND FAILURES
- 1. Failure Mode, Time, and Location The mode, the time, and the location of a containment failure needs to be determined.
1 $ $ d 5
References:
Group - D. Cooper x W. Kastenberg x x x x S. Levy x x A. Reynolds x R. Vogel x Observers - D. Campbell x E. Rahe x R. Sehgal x
..m.new.,
59
4 Issue G. CONTAINMENT LOADS AND FAILURES
- 2. Hydrogen Generation The models for hydrogen generation appear to predict an excessive amount of hydrogen. This may be caused by the following: ;
an inadequate modelling of the zircalloy channels, 1 - a failure to recognize depleted steam, or a single mode used to repr.esent the lower reactor plenum The excessive amount of hydrogen that is predicted implies an early containment 3 failure; this is conservative. In the Sequoyah analysis, the 52HF sequences describes a series of hydrogen burns challenging the containment. The last burn, after the ice is melted, causes a containment failure. questionable. Nearly all of the The severity of the simulated hydrogen burns is 1 zirconium should be oxidized; the depleted
; oxygen should limit hydrogen burning. The challenge to the containment may be 1 overestimated.
The volume of core zircalloy simulated by the MARCH 2 code accounts not only for the cladding but also for the channel boxes. The extent to which zircalloy and steam react depends both on the amount of zircalloy and the surface area; this'is important because the steam-zircalloy reaction' produces hydrogen that pressurizes a containment. The modelling of core zircalloy is inaccurate.
~
l 2 2 i 2 !
Reference:
Group - S. Levy x 1 i D. Rowe x R. Vogel x Observers none Contractors - P. Cybulskis o ____ Pages 155 Section 3 references: Page 13 F 60 rrw- -- ,----e- r- -, -- 4 yy g
Issue G. CONTAINMENT LOADS AND FAILURES
- 3. Hydrogen Igniters An assumption in the Sequoyah analysis is that the containment hydrogen igniters are evenly distributed. The igniters are not evenly distributed;
' they are more numerous above the ice beds. The close packing of hydrogen igniters may sustain a hydrogen burn at relatively low hydrogen concentra-tions. Thus, hydrogen would burn and not explode. The assumption of a uniform distribution of hydrogen igniters is conservative.
. Major effects of hydrogen igniters have not been considered. These effects included the following:
- degraded operation alter hydrogen burning due to variations in the atmospheric chemical composition Only some of the accident sequences agree with the design basis of the hydrogen igniters.
1 S 2 1 5
Reference:
Group - W. Kastenberg o Observers - S. Niemczyk x Pages 156 l + 61 9
, -~-
Issue G. CONTAINMENT LOADS AND FAILURES
- 4. Hydrogen Burning Mode Two modes of hydrogen burning have been proposed for BWR's. In the first mode, hydrogen accumulates in a containment, ignites, and explodes. In the second mode, hydrogen burns as a diffusion flame. The correct mode must be determined to accurately predict a pressure spike in a containment.
1 2 3 1 1
Reference:
Group - R. Hilliard x Observers - R. Sehgal xo Contractors - R. Denning o Pages 305 64 67 Section 3 references: Page 13-b l ~ l 62
s l 1 Issue G. CONTAINMENT LOADS AND FAILURES
- 5. Hydrogen Overpressure The AE sequence (LOCA) in the Peach Bottom analysis has 40% of the zircalloy in the vessel reacting with water. Enough hydrogen is produced to overpres-surize the containment and cause a containment failure. That a 40% zircalloy reaction could produce enough hydrogen to cause a containment failure is ques-tionable. This calculation is conservative.
1 g 3 1 5
Reference:
Group - none Observers - E. Fuller o R. Sehgal o Pages 305 57 Section 3 references: Page 13 O
Response
The basis for this calculation is as follows. Hydrogen and steam enter the containment where the core slumps into the lower vessel head. A major frac-tion of noncondensable gases are swept from the drywell to the wetwell and fuel the wetwell. This pressure causes the wetwell to fail. l
Reference:
P. Cybulskis (Battelle), July 1983 Peer Review Meeting, p 57 of the transcript.
- - . ~ .
63
Issue G. CONTAINMENT LOADS AND FAILURES
- 6. Containment Lesk Rate In the Surry (PWR) analysis, the containment leak rate is assumed to be 1%/ day. The design basis for a large and dry PWR containment is 0.1%/ day.
1 S 2 4 5
Reference:
Group none 1 , Observers - E. Rahe x = 1 e 7 f I 1 64
. - , ~ . - , , -
, -,, - , <, - - ,,,c -
w e
, =. ._ - _ - -. -
1 Issue G. CONTAINMENT LOADS AND FAILURES
- 8. Modelling With a Probability Distribution All of the cases shown in the Zion study use a point value for the containment failure pressure. If the containment pressure is portrayed as a distribution, a probability for containment failure from a slow overpressurization arises.
This view has not been addressed. A spectrum of failures may be identified and a probability assigned to each failure. Risk is calculate. The approach is similar to the Sizewell study in Britain. 1 2 3 d 1
Reference:
Group - D. Cooper x W. Kastenberg x x Observers - F. Rahn o Pages 432 O t S i 65 l . . .
Issue G. CONTAINMENT LOADS AND FAILURES
- 9. Fission Product Retention i
The tortuous path formed by a failed (cracks or open penetration) containment may retain fission products. 1 $ 2 d 5
Reference:
Group - none Observers - R. Hilliard x 4 J. Kelly x k [m
--me e
. m *w ee .
i 4 i 66 x,. - _. - - , _ . _ . , . - _ . - . - - . . , . . _ - . , -
J Issue H. SUPPRESSION P0OL
- 1. Water Flashing During an ATWS, the temperature of the water in the suppression pool may rise above the normal boiling point. Given this condition, the suppression pool would flash if the containment failed and depressurized. Flashing during a core meltdown might impair the capability of the suppression pool to remove fission products. A sequence has not been defined to account for flashing in the suppression pool.
1 S 2~d 5
Reference:
Group - W. Kastenberg x Observers - W. Arciers o Page
- Section 3 references: Page 4 M'An
*No page reference available.
67
4 Issue H. SUPPRESSION POOL
~
- 2. Bubble Shape The SPARC spherical code models gas scrubbing in the suppression pool as a gas entering bubbles. The calculations then change the shape of each bubble from a sphere to an ellipse. This approach predicts no retention of small particles.
When a model has gases entering an elipical bubble, calculations show that small particles are removed. The calculations beginning with spnerical bubbles are conservative. ! 'l $ 2 4 5
Reference:
Group none Observers - J. Holtzclaw o Page 108 Section 3 references: Page 4 e 68
--n. , , . -
Issue H. SUPPRESSION P0OL l
- 3. Bubble Movement The bubble density in a suppression pool may be high enough to cause bubbles to interact, such as by foaming.
Surfactants may alter the movement of bubbles. Substances may be leached from the concrete and form an alkaline solution. This can cause frothing even if the substances are not surface activents. i Currently, it is assumed that the bubbles move independently. 1 2 3 i 1
Reference:
Group - A. Castleman x
- 0. Cooper x Observers - J. Cobble o Contractors - P. Cybulskis o Pages 326 32 Section 3 references: Page 4
Response
Many of the suppression pools are lined with an epoxy resin.
Reference:
R. Hobbins at the May 1983 Peer Review Meeting, p 326 of the transcript.
===m%
69
Issue H. SUPPRESSION P0OL
- 4. Carbon Dioxide Solubility The solubility of carbon dioxide in a suppression pool was not modelled. The water in a suppression pool should absorb at least some carbon dioxide to reduce the pressure is a containment. Without considering carbon dioxide solubility, the source term models are conservative.
1 2 3 4 5 Group - none
Reference:
Observer - none Contractors - R. Denning o Pages 433 Section 3 references: Page 4 e i 70 J
d Issue I. ICE CONDENSER
- 1. Shear Stress Ice exposed to shear strcss would change shape, hence, the retention of fission products would be altered. This may be a significant phenomenon.
1 2 3 i 5
Reference:
Group - S. Levy o Observer - none Pages 194 i
! Response Calculations indicate a flow velocity of 1 ft/second through the ice bed. Any deformation of the ice by shear stresses would be minimal.
Reference:
A. Postma (Battelle Northwest), July 1983 Peer Review Meeting, ' p 194 of the transcript. ! 1 71 ! 1
Issue I. ICE CONDENSER
- 2. Ice Compression Ice may compress as steam and fission products pass through an ice bed. The ice may become a solid plug and lose its effectiveness in removing steams and fission products. This phenomenon may significantly influence the consequences of an accident.
1 2 d
} 1
Reference:
Group - D. Cooper o Observers none Pages 190 O O 72
Issue I. ICE CONDENSER
- 3. Decay Heat The decay heat from fission products would likely accelerate the ice melting during an accident. The availability of ice to suppress steam would then be reduced. Ice melting from decay heat is currently not in the models. The models underestimate the consequences of an accident.
1 $ $ $ 5
Reference:
Group - none Observers - P. Laughlin o Contractors - A. Postma o Pages 190 195 l [ Section 3 references: Page 1 l 2
; Response When modelling the ice bed, the decay heat from fission products is considered to be insignificant compared to the heat from steam.
Reference:
A. Postma (Battelle Northwest), July 1983 Peer Review Meeting, p 197 of the transcript. , 73
Issue I. ICE CONDENSER
- 4. Layering Deposits of fission products may build up on the ice in an ice bed and reduce both steam condensation and fission product deposition.
1 2 a i s
Reference:
Group - S. Levy o Observers none Contractors - A. Postma o Pages 195 a Comment Layering may insignificant. The ice will be melted during an accident. Thus, I deposits will be washed from the ice and new ice surface will be exposed. i 74
i Issue I. ICE CONDENSER
- 5. Aerosol Deposition Aerosols deposit in an ice condenser by impaction and interception mechanisms.
The collection efficiency from the ice baskets is given by the following equations
- E= stk2 where E = collection efficiency (stk+0.5)"
- 0*04 + 2 d2de stk = Stoke's number dp = particle diameter impaction interception dc = collecting cylinder diameter stk = Vi Podp 2 Cm where Vi = gas velocity 90 de Pp = particle density Cm = Cunningham correction U = gas viscosity The steel backets have horizontal strips on which aerosols are expected to deposit. The strips are 6.35 mm long, 1.91 mm thick, and 254 mm wide. Impac-tion should be a dominate collection mechanism when ice is no longer present.
The decontamination factor should be about DF = 10 because the air will be recirculated through the ice condenser. Impaction will change the flow area and geometry. The equations describing the deposition of aerosols on the baskets need to be modified to account for the change in flow area and geometry. 1 2 3, 4 }
Reference:
-Group - R. Hilliard o Observers - none Contractors - M. Kulhman o K. Lee o P. Owazarski o Pages 190 196
*NUREG/CR-3248, " Studies of Fission Product scrubbing in Ice Compartments,"
p4.10-1, equation 18-21. 75
Issue J. FILTERS Fission products entering a reactor building will be removed by charcoal bed filters and HEPA filters. Large amounts of aerosols and steam will cause the filters to plug. The HEPA filters may fail. The change in the flow patterns when filters plug and fail is not adequately considered in the models. 1 S 2 S 5
Reference:
Group - D. Cooper x P Observers - none ) i 4 e O i a 2wem. 76
Issue K. POLICY
- 1. Sequence Selection The criteria for selecting sequences should be reexamined. Examples show that the selection criteria are inconsistent and incomplete.
No small break LOCA sequences are considered for Mark-I BWRs. No LOCA sequences are considered for MARK-III BWRs. No emergency. safety feature power failures are considered for MARK-I BWRs or MARK-III BhRs. Including the Zion plan in source terms analysis raises the issue of including external event initiators into all analysis. Operator actions may play a significant role in initiating and propagating sequences. A V-sequence is not considered for Zion because the risk is assumed to be low. A " complete" range of sequences should be clearly defined. Some sequences are selected on the basis of risk while other sequences are selected because they involve plant-specific features. The selection criteria result in analysis that are incomplete and that cannot be compared. To determine if the containment failure has been adequately studied, the sequences should be grouped according to the relative time of a core melt and a containment failure. Two groups are readily evident; one, a core melts before a containment failure; two, a core melts after a containment failure. In addition, a slow core / containment failure and a rapid core / containment failure should be considered. When the sequences are categorized, they can be evaluated to determine if the consequences of reactor accidents have been comprehensively studied. 1 2 3 i 5
Reference:
Group - D. Cooper x W. Kastenberg x x x S. Levy x A. Reynolds x R. Vogel x Observers - S. Niemczyk- x x Section 3 references: Page 29 I 77
Issue K. POLICY
- 2. Prejudging Result =
Sequences should not be selected on the basis of consequences. The consequence of a sequence is what a source term analysis is to estimate. By selecting sequences according to consequences, the estimates are prejudged; that is, an analysis is braced. 1 2 1 1 E
Reference:
Group - W. Kastenberg x Observers - none e 78
i J J
- Issue K. POLICY
- 3. Use of Methodology 4
The methods developed to estimate source terms will likefy to suited to a specific applications. .Two applications are emergency planning and equipment
- qualification. The conservative approach used in developing the source term models will make the methodology inapplicable for equipment qualifications.
1 { 3 i 1
Reference:
Group none Observers - S. Niemczyk x 1
. -- ~. ,
'i . i a d 79
. m , _ . _ . , . . ,,c - - . , . , , p - , - - , - - - , , - - - , , - - , , - . , -
t l Issue K. POLICY
- 5. Sequence Probability The probability of any the selected sequences should be assoicated with the sequences.
1 _2 _3 _4
- _5
Reference:
Group - none Reference - S. Loyalka x i P I Z O i 1 4 3 d i l~ 1 80
Issue K. POLICY
- 5. Comparison Study Comparison studies should be done for the following subjects:
I WASH-1400, NUREG-0772, NUREG-0956, Codes Fission product retention in the reactor coolant system of both the Grand Gulf station and the Peach Bottom station.
.The data from the Three Mile Island accident may not be useful because the core was reflooded.
1 2 3 i 1 1
Reference:
Group - A. Castleman x D. Cooper x x W. Kastenberg x S. Levy x x
' A. Reynolds x R..Vogel o Observers - J. Kelly x Contractors - D. Powers o
- Pages 349 273 i
Response
The Reactor Safety Study (RSS) predicts a 2 hour tellurium and the VANESA code predicts an extend tellurium release; this occurs because the VANESA code accounts for tellurium diffusion into metal. The RSS predicts a lower barium t and strontium release than the VANESA code; this occurs because the RSS accounts for fewer chemical species than the VANESA code.
Reference:
D. Powers, July 1983 peer review meeting, p 273 of the transcript. 81
Issue K. POLICY
- 6. Presentation In the review process, the emphasis should be on the phenomena that were modelled rather than the codes that perform the calculations.
1 2 } d }
Reference:
Group - W. Kastenberg x A. Reynolds x
- 0. Rowe x Observers - H. Kouts x S. Nienczyk x R. Sehgal x e
O 82
i Issue K. POLICY
- 7. Additional Analyses Additional ana' lyses should be done to include sequences that involve a containment isolation failure, a containment bypass, and retention by the outer containment (Grand Gulf). A sequence similiar to the Three Mile Island accident should also be analyzed.
1 2 2 $ 2
Reference:
Group - S. Levy x , Observers - E. Rahe x R. Sehgal x G. Thompson x
- ~= .- ~
83
,-,,,,---r, -- - - , ,-
l l I Issue K. POLICY !
- 8. Review Process The way in which the coments from the review meeting and the findings of other groups (i.e., containment loads group) will influence the source term research is unclear. ,
~
1 2
~
3
~
4
~
5
~
Reference:
Group - W. Kastenberg Observers v H. Isbin x 84 l . -
i Issue K. POLICY l
- 9. QUEST The wide range of results found by the QUEST analyses cast doubt on the value of the BMI-2104 studies. The doubt is greater than is warrented.
1 2 3 4 5
Reference:
Group - A. Reynolds x Observers none2 4 .I i . b f 4 a 3 85 _ _ _ _ _ _ . . - _ . _ _ . . _ . ~ - ._. .
Issue L. MODELS
- 1. General
- a. Core geometry degradaton In the modelling of a core melt down, a gradual slump is assumed. No geometric changes are assumed when accounting for surface areas and flow areas. The modelling of a gradual slump is inaccurate or incomplete.
However, no check to make sure that a melted core can be pnysically accommo-dated in a lower node (i.e. one that has not yet melted). The calculation may be physically impossible. Current models have a core melting homogeneously. Cladding melts at a lower temperature than UO2 . The models do not appear to consider a failure of the core structural material prior to the melting of fuel. A collapse of the core may influence the melting process. In the modelling of a core melt down, a gradual slump is assumed. No geometric changes are assumed when accounting for surface areas and flow areas. The modelling of a gradual slump is inaccurate or incomplete. Three scenerios of a core melt have been developed; (1) Coherent drop - the core instantaneously slumps as one unit when 75% of the core is liquified. (2) Downward gradual slump model. radial sections of a cone are calcu-lated to melt independently. Upper nodal regions in a radial section heat lower nodal regions. The melt progresses downward. (3) Upward gradual slump model - radial sections of a core are calculated to melt independently. Lower nodal regions in a radial section heat upper nodal regions. The melt progresses upward. The three scenerios reflect.an evolving process of developing the computer models. The coherent drop model is no longer used. Instead, the gradual slump models are used. However, these models were developed more on intuition than on facts. Thus, even the gradual slump models are arbitrary. 1 2 1 $ 1
Reference:
Group - A. Castleman x S. Levy x x x x D. Rowe x Observers - D. Campbell x D. Cooper x R. Denning o W. Harrington x 86
Issue L. MODELS
- 1. General i
- a. Core geometry degradaton (continued) i 1 2 3 3 }
Observers - R. Hobbins x J. Kelly x S. Niemczyk x G. Petrangeli x x A. Reynolds o x L. Zumwalt o Pages 64 334 128 Section 3 references: Page 28 ) o 87
- .- - . ~ . - - , , . - - . - - . . .- , - . . - . . . - . .
Issue L. MODELS
- 1. General
- b. Static control volume The control volumes used in any of the codes should respond to changes in fission product temperature or pressure when time intervals are comparable to the following ratio:
control volume volumetric flow rate Studies have shown that the number of control volumes and the placement of compute modes influences the predicted results. The' predicted source terms may be influenced by the static control volumes. 1 3 2 $ 5
Reference:
Group - D. Coooer x Observers none 88
Issue L. MODELS A
- 1. General
- c. Number of control volumes and nodal points The predicted source terms are influenced by the number of control volumes and/or nodal points used to mathemathically represent various parts of a
- plant; i.e. , core, reactor coolant system. . containment.
The difference in retention of cesium iodide in the Grand Gulf analysis j relative to the Peach Bottom analysis may be due to the differences in the number of core nodal points. I 1 2 3 i }
Reference:
Group - S. Levy. x l
- 0. Rowe x x x i Observers - R. Sehgal x x 1
R. Hobbins x ]
- R. Sherry o E. Warman x Contractors - P. Cybulskis o i
Pages 217 148 1 1 Section 3 references: Page 20 'i- - I r 1 4 i i 't 89
4 Issue L. MODELS
- 1. General
- d. Arrangement of control volumes The control volumes are cennected in series to represent the reactor systems.
No parallel connections are made. Though the natural circulation within a control volume can be adequately represented, the natural circulation in a reactor system (many control volumes) cannot be easily represented. 1 2
~
3 4 5
' ~
Reference:
Group - D. Rowe x S. Levy x l Section 3 references: Page 20 I
.ee 90.
- , , - 4 +. -. _ w - - --
Issue L. MODELS
- 1. General
- e. Heat capacity The heat capacity of the reactor coolant system and the containmer.t have not been accounted for. Not only will these structures serve as a heat sink but also as a fission product sink.
1 2 3 4 5
~
Reference:
Group - A. Reynolds x Observers - J. Kelly x Section 3 references: Page 1 I ~ I
+-
91
1 Issue L. MODELLING , 1. General i f. Two phase flow i Two phase flow is expected at pipe breaks. This is not in the models. _1 _2 _3 _4 _5
Reference:
Group none i I Observers - E. Warman x I t t 1 9 1 l i ? t 4 f I 5 i i 92
Issue L. MODELS
- 1. General
- g. Assumptions Some assumptions are conservative even though the numerical results are said to be best-estimate.
The ideal solution assumption is not always applicable. The dry pathway assumption appears to be invalid considering the amount of water in a reactor system. The containment sprays are unaccounted for in the analyses of the Grand Gulf station. The instantaneous vessel failure assumption is unrealistic. Interactions of a molten care with the lower vessel internals are ignored. The Peach Bottom analysis has the containment failing in the drywell such that a maximum amount of fission products is released. The mode of a containment failure is uncertain. Therefore, three sets of source term calculations should be done, namely, (1) assuming a failure in the drywell (2) assuming a failure in the wetwell above the suppression pool, and, (3) assuming a failure in the wetwell below the suppression pool. An assumed failure leading to a maximum fission product release leads to conservative source term predictions. The assumptions need to be justified. 1 2 3 i }
Reference:
Group - A. Castleman x S. Levy x x
- 0. Rowe x Observers - D. Campbell x S. Loyalka x R. Sehgal x R. Hobbins x Section 3 references: Page 22, 23 93
~~ Issue L. MODELS
- 1. General
- h. Aerosol particle size The initial particle sizes are questionable.
1 2 } d 1 References,: . Group - R. Hilliard x Observers - none Contractors - M. Kuhlman o Pages 71
Response
A sensitivity analysis showed that the predicted aerosol behavior is sensitive to the assumed distribution when the aerosol particle concentration is low. When the aerosol particle concentration is high, the predicted aerosol behavior is insensitive to the assumed distribution. This occurs because a concentrated aerosol coagulates rapidly; an error in the initial distribution is cancelled. Dilute aerosols coagualte slowly; an error in the initial distribution persists.
Reference:
M. Kuhlman. January 1983 Peer Review Meeting, p71 of the transcript. i 94
Issue L. MODELS
- 1. General
- i. Iterative models The codes should be structured to perform iterative calculations.
1 $ 2 $ 5
Reference:
Group - none Observers - S. Loyalka x e O i 95
. _ _ , .. _. ,. , . - ~ ,, -
.--r - - --- -- - ~ - - ~ ^ ~ - ' - ~ ' ~ ~ ~ '
i Issue L. MODELS
- 1. General
- j. Mixture levels The model for the mixture flow in the reactor coolant system is based on a uniform verticle flow area. One mixture level is used to represent both the hot leg and the cold leg.
1 E E 1 E
Reference:
Group - R. Ritzman . x Observers - none .i i 0 4 96
Issue L. MODELS ! 1. General ' l
- k. Atomization i
i In the MARCH code, the calculations for core modes dropping into the bottom head assumes a spherical shape for core debris. The diameter of the sphere is
- an input variable. The number of spheres is calculated using the sphere i volume and the node volume. Each sphere has an assumed morphology; the center j
contains all uranium oxide and may contain some zirconium and/or zirconium i oxide; presumably the shells contain only zirconium and zirconium oxide.*- The cladding may melt before the U02 melts. i A significant aerosol source term may arise from a high pressure ejection of molten fuel from a reactor vessel. D. Powers conducted experiments where 4 simulants of molten fuel were ejected from a 1" orifice under 600 psi. A l large amount of aerosols were generated. This phenomenon of aerosol formation A is not modelled. It appears to be a very significant source of aerosols. ' 4 Diameters of
! Particles (microns) Source 0.5 Condensing vapors 5.0 Freezing droplets
) 50.0 Break of melt ligaments 4 l When the bottom head of a reactor pressure vessels melts, the core debris will ' be released into the reactor cavity and the containment building. Depending l on the characteristics of the vessel-failure, the core debris may remain in j the cavity or be swept into the containment. 1 2 3 i }
Reference:
Group - W. Kastenberg x A. Reynolds x D. Rowe x x l i __. Observer - D. Powers o S. Niesczyk x x ~ j Page 294 J 3 Section 3 references: Page 10 ' e
- *0RNL/TM-8842, " Status of Validation of Codes Used in the Accident Source Term i Reassessment Study (BMI-2104); Status of Validation of the MARCH 2 Computer Code" (Draft: ~ July 12, 1983), p 47.
4 j 1 97 1
. _ _ _ _ . _ _ _ _ , . . _ _ _ . - - _ _ _ _ . _ . . _ _ _ . - ~ . . . _ _ _ . ., _ -. _ ._ _
l 4 4 ] Issue L. MODELS
- 2. NAVA
- a. Aerosol mechanisms The NAUA code lacks the following aerosol deposition mechanisms:
1 turbulent deposition, turbulent coagulation, gradient coagulation, electrostatic effects. boundary layer thickness 1 2 3 4 5 j
References:
Group - D. Cooper x
, Observers none i
t 1 i f t il \ . i i i !I i i i I k 4 98 l
l l i Issue L. MODELS
- 2. NAVA
- b. Spray drop impaction term In the equation describing how spray drops collect aerosols, the impaction term is expressed unconventionally. The equation is equation 7.2, p. 7-67, of the BMI-2104, volume 1 report.
2 ' E= 1 + 0.75 in (2 Stk) 1.5 (r/R)2 Stk - 1.214 (1 + r/R)1 /3 inertial impaction interception where Stk = stokes number r = aerosol particle radius R = water drop radius A more conventional form of the inertiel impaction term is as indicated: (0.05)(Stk2) (Stk + 0.35)2 1 2 3 1 5
Reference:
Group - D. Cooper x o
~
Observers - none Pages 388 I a i a
. - ~ . . , .
99
Issue L. MODELS , 3. TRAPMELT Convective transport of fission products is modelled assuming a steam medium. A more accurate model would be based on a r. team / hydrogen medium. 1 2 3 s i
Reference:
Group - R. Hilliard x Observer - none 4 1 4 I I 100
F Issue L. MODELS
- 4. SPARC The SPARC code predicts a lower cellection efficiency for 0.6 micron particles than expected, considering a large pressure drop along a 10 foot depth of water in a suppression pool., The code may be inaccurate is some calculations.
1 2 2 S 5
References:
Group - D. Cooper x Observers - none e
- 101
Issue M. CODE RELIABILITY
- 1. Validity
- a. PWR model used for BWR calculations In the current models describing the reactor vessel failure, the core melts through the vessel and falls directly onto the concrete of the reactor cavity.
This model is accurate for a PWR, where the reactor penetrations enter the vessel from the top. However, the penetrations enter a BWR vessel from the bottom. The control rod driver and the instrument penetrations cover the cavity concrete; then, a direct core-concrete interaction should not occur, at least when the core first enters the cavity. The models do not consider the BWR penetrations. The models are be inappropriately applied to BWR analyses of the reactor vessel failure. 1 2 3 i i
Reference:
Group - W. Kastenberg o Observers - R. Sehgal x E. Warman o r. Pages 312 333 Section 3 references: Page 21
.m .. .
l . i 102 l 1
i 4 Issue M. CODE RELIABILTY
- 1. Validity
- b. Reactor building The reactor building is considered to be ineffective towards removing fission i
products. The models are conservative by ignoring the reactor building. 1 2 } } }
Reference:
Group - none Observer - J. Kelly . x G. Petrangeli- x E. Rahe x E. Warman x o Pages 242 1 l 1 s 4 i t 103 _,,. __ , . , , ,y , . , - * - - - +e, ' - = ' - ' " - - - ' ' - - - ~ ' * * ' " ' ' " ' ' * " ^" ' ' * '
Issue M. CODE RELIABILITY
- 2. Consistency in Modelling Some portions of the models are detailed and have a firm scientific basis.
Other protions are based on assumptions. This is analagous to measuring with a yardstick and a micrometer. The predicted consequences are only as good as the weakest significant assumption. 1 2 a i 1
Reference:
Group - S. Levy x x Observers none 1 4 j } 104
Issue M. CODE RELIABILITY
- 3. Uncertainty
- a. General The results of the analyses are reported without indications as to the uncer-tainty associated with each prediction. The predicted consequences are meaningless without measures of uncertainty.
1 2 $ $ 5
Reference:
Group - R. Ritzman x Observers - R. Hobbins x S. Loyalka )( G. Petrangeli x ' E. Rahe x G. Thompson x l l 105
Issue M. CODE RELIABILITY !
- 3. Uncertainty
- b. Study Before an uncertainty analysis is done, the models should be developed to a point where further changes are necessary. Uncertainty studies are meaningless as long as the models are evolving.
_1 _2 _3 _4 _5
Reference:
Group - S. Levy x - ! Observers - none 1 j t
~
I l 4 l 1 i
. ~ , ~ ~ .
i + < 106
Issue M. CODE RELIABILITY j 4. Rep'esentativeness r l
; a. Plant
) An analysis of source terms is necessarily plant specific. The analysis of a 1 given plant are inapplicable to another plant. i
.The upper dome region of the Surry reactor vessel is considered to be isolated
] from all flow; Westinghouse indicated that 1/2% of the coolant flow enters this upper done. A greater flow may occur in vessels made by other manufacturers. i The "as-built" plant may be different than the "as-drawn" plant. Fire protection systems are accounted for only for Mark-I reactors. , 1 2 3 4 { 4
Reference:
Group - W. Kastenberg o A. Reynolds x
, Observers - J. Kelly x G. Petrangeli x . S. Niemczyk x x
) E. Rahe x Contractors - P. Cybulskis o l Pages . 414' 24 i I Section 3 references: Page 21 1 1 i J l 107
. . _ _ . . _ _ _ _ _ . _ . _ . - . . __ . - _ . . _ . _ _ . , _ . , ~ _ . _ . _ ._,_ .__ . _. -_
Issue M. CODE RELIABILITY
- 4. Representativeness
, b. Scaling up experiments 4
Small scale experiements may not represent large scale phenomena. Fission product release estimates are based on experiments using spent irradiated fuel. The predicted behavior of tellurium is an extrapolation of the observed behavior of sulfur. 1 2 3 3 5
Reference:
Group - A. Castleman x R. Vogel o Observers - D. Wren x Contractors - P. Cybulskis o - Pages 164 52 Section 3 references: Page 12 ' I L i l l l t i i 108 t
, , . . . . _ _ , _ ..-_ -.-4 . - . _ _ .. . - _ ,r--- . . - - , - . ,_-,._,.v, - _ . . . . - . . . , . . . , _ . , . , _
1 I Issue M. CODE RELIABILITY
- 5. Data Data are scarce. Data are needed o'n subjects such as deposition velocities and high temperature thermodynamics.
Some data are conflicting: TMI and UKAEA data (Levy). 1 - ORNL and German data (Vogel). WASH-1400 cannot supply adaquate input data. 1 The temperature data for the TMLB' sequence is inconsistant (Warman). The TMI data about tellurium is inconclusive. The models should be validated with experimental data. While many of the i individual phenomenon are based on data, the phenomena together forming a model, have not been validated. 1 2., 3,, 4_ 1
Reference:
Group - C. Johnson o A. Castleman x x R. Hilliard x S. Levy x x x A. Reynolds x R. Ritzman x
- 0. Rowe -
x
.R. Vogel x Observers - P. Clough o H. Isbin x S. Loyaika x x S. Niemczyk x E. Rahe x ' C. Thomas x E. Warman x Contractors - D. Powers o T. Kress o o R. Wichner o Pages 332 229 286 285 324 450 Section 3 references: Page 12 109 1_---__---_--_____--_--___------___---_-___-----
Issue M. CODE RELIABILITY
- 5. Data (Continued)
Experiments with small sections of fuel rods were done to determine the release of tellurium. At 300*C to 400 C tellurium was released. At higher temperatures, tellurium reacted with ziralloy and was not released. At still higher tempera-tures, tin telluride was released (L. Johnson, July 1984, p 229).
Reference:
C. Johnson. July 1983 Peer Review Meeting, p 229 of transcript. r Some disciplines outside of the nuclear energy may have relevant data; the VANESA code may be verified with data from the. carbon blows in steel mills.
Reference:
D. Powers. January 1984 Peer Review Meeting (no page reference i for transcript available). 6 i
)
i i e. i i i 110
Issue M. CORE RELIABILITY
- 6. Interfacing codes Interfacing the codes is cumbersome and leads to errors. In the MERGE /TRAPMELT calculation, the condensation of fission products appear to be controlled by mass transfer. Heat transfer factors may be important.
1 E 1 $ 5 i
Reference:
Group - R. Hilliard x Observers - F. Abbey x i i i i
\
i e I i f ( l
.a - "
1 111 1 i
Issue M. CODE RELIABILITY
- 7. QUEST A. Type of study The Quantitative Uncertainty Evaluation of Source Term (QUEST) study in a sophisticated sensitivity study, not an uncertainty study.
1 1 S $ 5
Reference:
Group - R. Hilliard x W. Kastenberg xo D. Walker o Observers - none Contractors - C. Leigh o R. Lipinski o D. Williams o Pages
- 394 446 448 l
l
Response
The QUEST study is an uncertainty study because the inputs were carefully selected.
Reference:
C. Leigh, R. Lipinski, D. Williams. January 1984 peer review meeting, p 394, 446, 448 of the transcript. _ _ . _ _ . 112
y - - %.-- - - - ,-. s - -* _ - - .a-JJA* i Issue K. CODE RELIABILITY
- 7. QUEST
- b. Tellurium release The effect of oxidized and reduced zerconium on the release of tellurium was ignored.
The delta p analysis leads to an ancertainty in the release of tellurium similiar to the uncertainty in the release of antimony, barium, and molybdium. These fission products are non volitile fission products. Tellurium is a. volittle fission product. 1 S S i 5
Reference:
Group - A. Reynolds x Observers - none I 113 i
i Section 3: Major Technical Issues Taken from Section 2 ('
-0 Index DECAY HEAT 1 FISSION PRODUCT RELEASE 2 AEROSOL DEPOSITION IN THE RCS 3 FISSION PRODUCT DEPOSIT 0N IN A SUPPRESSION POOL 4 AEROSOL DEPOSITION IN A CONTAINMENT 5 INTERACTION OF AEROSOL MECHANISMS 6 STEAM CONDENSATION 7 MORPHOLOGY OF CONTAINMENT ATMOSPHERE 8 MORPHOLOGY OF CORIUM 9 MORPHOLOGY OF CORE DEBRIS . 10 MORPHOLOGY OF AEROSOL PARTICLES 11 DATA 12 CONTAINMENT LOADS FROM HYDROGEN 13 CESIUM IODIDE 14 BORON CARBIDE 15 CORE / CONCRETE INTERACTION 16 RADIOLYTIC EFFECTS 17 CHEMISORPTION 18 HYDRATE INORGANIC MOLECULES 19 STRUCTURE OF MODELS 20 REPRESENTATIVENESS OF MODELS 21 CONSISTENCY OF MODELS 22 ASSUMPTIONS OF MODELS 23 ALTERNATE SEQUENCES FROM PIPE BREAKS 24 ALTERNATE SEQUENCES FROM OPERATOR ACTION 25 ALTERNATE SEQUENCES FROM PATHWAYS 26 ALTERNATE SEQUENCES FROM EQUIPMENT PRFORMANCE 27 SELECTION OF CORE MELT MODELS - 28 SELECTION OF SEQUENCES 29
l DECAY HEAT Transport The decay heat from fission products may cause deposits to vaporize. The vapors may condense and form an aerosol particles. The particles may then settle to start the evaporation - condensation process. To determine the sensitivity of the predicted retention by the reactor coolant system, Battelle estimated the heat load from the fission products and calculated how this would modify deposition. According to the calculations, decay heat reduces the deposition of vapors but not of aerosols in the reactor coolant system. The models underestimate the fission product release because too much credit is given for retention by the reactor coolant system. Hardware Hardware surfaces should be warmed by the decay heat from fission products. The heat should create convection currents modify the deposition of both an aerosol cloud and fission product vapors. The decay heat may also melt pipes. Fission products may resolve and form aerosols. Ice melting The decay heat from fission products would likely accelerate the ice melting during an accident. The availability of ice to suppress steam would the3 be reduced. Ice melting from decay heat is currently not in the models. Heat capacity
- The heat capacity of the reactor coolant system and the containment have not been accounted for. Not only will these structures serve as a heat sink but also as a fission produce sink.
Issues E.4, E.5, I.3, and L.1.e Responsa When modelling the ice bed,-the decay heat from fission products is considered to be insignificant compared to the heat from steam.
Reference:
A. Postma (Battelle Northwest), July 1983 Peer Review Meeting, p 197 of the transcript.
- This comment was not directed explicately at the decay heat issue.
i I 1 l l
FISSION PRODUCT RELEASE Iodine The chemical fore, of iodine was considered to be entirely I2 . This was changed to C I. Other forms of iodine, such as iodomethane, have not be considered. Little is known about the chemistry of iodine. Tellurium Models have tellurium leaving a core in the elemental form. Compounds of tellurium, such as gaseous H2 Te, are not considered. Silver Little is known about the release of silver. Silver may serve as a source of condensation nuclei onto which an aerosol cloud may form. Issues 04, D.6, and D.11 am e 6%u 2 l c_
AEROSOL DEPOSITION IN THE RCS General Aerosols deposit in the pipes of a reactor coolant system. Little is known about how the aerosol particles accumulate and less is known about how the particles resuspend. Furthermore, resuspension mechanisms are neglected from the models. Thus, the way in which accumulated aerosols modify flow is unknown. A calculation of the volume of fission product material rather than the mass of fission produce material may be necessary to determine the significance of plugging. The models fail to accurately predict flow through pipes because plugging from aerosol deposits is not considered. . .
- Deposition /resuspension i
When aerosols pass through pipes at high speeds, turbulent deposition and resuspension should occur. These phenomena do not seem to be adequately modelled. During post accident conditions, the turbulent flow through pipes may not be the same as turbulent flow during normal conditions. For example, one thought is that the aerosols and vapors spiral through the pipes without touching the pipe walls; this flow is extremely complicated. Water Aerosol deposition in the reactor coolant system is not well understood. Moisture or bulk water in the pipes will influence deposition. The models do not adequately account for Stefan flow, electrostatic interac-tions, and turbulent reentrainment.
~
A consistent methodology needs to be developed; general-space dependent forms of aerosol equations are needed; unambiguous deposition coefficients are needed; recented turbulent deposition correlations (i.e. DeMota, Friedlander) should be used. Deposition rate expressions and correlations seem to be ad hoc. Steam Generator ' During a severe accident, the steam generators may remain as an effective heat sink. Significant_ amounts of steam may condense within the steam generators and remove fission products. These phenomena may be a significant mechanism for removing fission products. The calculated heat transfer coefficient for a steam generator is erroneous. The calculation is based on normal flow through a steam generator. During an accident, when one side of the steam generator may be stagnant, the heat transfer will be much less than during normal operation. The predicted steam generator heat transfer rate should not depend on the core heat transfer model. l Issues C.7, E.1, E.2,.E.6, and E.8 ' l 3 {
-- - {
FISSION PRODUCT DEPOSITION IN A SUPPRESSION POOL Flashinq During an ATWS, the temperature of the water in the suppression pool may rise above the normal boiling point. Given this condition, the suppression pool would flash if the containment failed and depressurized. Flashing during a core meltdown might impair the capability of the suppression pool to remove fission products. Particle deposition Gas scrubbing in the suppression pool is modelled in the SPARC code as a gas entering spherical bubbles. The calculations then change the shape of each bubble from a sphere to an ellipse. This approach predicts no retention of small particles. When a model has gases entering an elipical bubble, calcu-lations show that small particles are removed. The calculations beginning with spherical bubbles are conservative. Bubble movement The bubble density in a suppression pool may be high enough to cause bubbles to interact, such as by foaming. Surfactants may alter the movement of bubbles. Currently, it is assumed that the bubbles move independently. Solubility The solubility of carbon dioxide in a suppression pool was not modelled. The water in a suppression pool should absorb at least some carbon dioxide to reduce the pressure is a containment. Issues H.1, H.2, H.3, and H.4 4
I AEROSOL DEPOSITION IN A CONTAINMENT Diffusiophoresis The significance of diffusiophoresis has not been determined. At first, this phenomenon was not modelled. Refined models included diffusiophoresis. However, the consensus among the code developers is that diffusiophoresis is insignificant for two reasons; (1) the steam in a containment should condense before aerosols enter the containment; (2) the surface / volume ratio of a containment is small. Scrubbing Water drops falling through the atmosphere of a containment can scrub the air of fission products. The mechanisms are well understood. Also, the distribution of water drop sizes from a spray system has not been rigorously determined. f29 The impact of the fog model on the calculated source terms is unclear. A model based on the carryover of water with steam might be preferable. Issues F.1.a. F.5, and F.8
. 1 _ __ ._
a 5
INTERACTION OF AEROSOL MECHANISMS Some aerosol mechanism, such as gravity settling and impaction, are independent of one another in a thermal hydraulic sense. Other mechamisms, such as thermo-phoresis, diffusiophoresis, and turbulent deposition, operate on an interaction of heat, mass, and momentum transfers. Th: interaction is neglected from the modeling. Aerosols of cesium hydroxide and cesium iodide are treated as being separate and independent. Actually a single aerosol composed of varying amounts of each compound will likely exist. Issue F.2
- COMMENT Coagulation coefficients may be altered by orders of magnitude when ambient conditions are changed by small amounts. This observation is attributed a strong synergistic relationship among the aerosol mechanisms (thermophoresis, diffusiophoresis, etc.)
" Mechanics of Aerosols in Nuclear Safety: A. Kesrew. S. Loyalka. Progress in Nuclear Engineering, Vol. 12, No. 1 p 4, 10, 22-3, 26 6 l
STEAM CONDENSATION Adiabatic gas expansion ! When a containment fails, the rapid pressure drop should lead to an adiabatic gis expansion. The adiabatic expansion should enhance the condensation of steam and lead to enhanced fission product removal. This phenomenon is not modelled. Particles Steam can condense on both aerosols and on walls. The steam condensation phenomenon is not well understood for two reasons; (1) the affinity of aerosols for water is unknown; (2) a quantitative description of a boundary layer at a wall is unknown. Mason's equation is used to describe condensation and evaporation; this equation, when used to describe heat and mass transfer, is invalid for small particle sizes. The heat changes during condensation and evaporation may have a significant influence on the thermodynamic behavior of the aerosols. Condensation readily occurs on particles as small as 0.6 pm. If the particles are hydroscopic, then water can condense on the particles when the atmosphere is unsaturated. Steam condensation is a phenomenon that removes fission products from a gas. Nucleation Homogenous nucleation is described using outdated equations. Old textbooks are readily available. Nevertheless, in recent years, many advancec equations describing nucleation have been developed. The equations in the current source term models are being related on the basis of availability rather than on the basis of appropriateness. Issues C.5, F.1.b, and F.6
Response
As a containment failure is currently modelled, material exists the failure at a constant enthalpy and then expands. A two phase flow does not necessarily arise. Instead, a superheated state may result. If an isotropic expansion is assumed, then a two phase region is predicted.
Reference:
P. Cybulskis (Battelle), July 1983 Peer Review Meeting, p 29 of the transcript. 7
MORPHOLOGY OF CONTAINMENT ATMOSPHERE Upward draft Convection down the walls of a containment is likely to produce an upward draft in the center of the containment. The draft may hinder the gravity settling of aerosols. This assumes that the atmosphere in a containment behaves as a homogenous mass. Settling Sedimentation on horizontal surfaces should not be effected by an upward air velocity when an atmosphere is homogenous. Aeorosol particles will still deposit within a boundary layer. Settling would be prevented if gas emanated from a surface, such as during evaporation. Composition The thermal hydraulic conditions in a containment during post-accident conditions is uncertain. One thought considers the atmosphere in a containment to be homogenous; the atmosphere behaves as a single volume. Another thought considers the atmosphere to be heterogenous; the atmosphere behaves as many small volumes. Because the containment is significant in attenuating the release of fission products, the thermal hydraulic conditions should be thoroughly understood. Issues F.3 and F.4 F I 8 i - - -
MORPHOLOGY OF CORIUM Relative volatility Both the fuel and the control rods contain chemical species having a range of volatilities. Some of the species are significant in themselves because they are fission products. Other species are significant because they serve as condensation nuclei. Chemical reaction may influence the release of fission products. Alloys, that form from fission products and cladding, may influence fission product. releases. The vapors from a degraded core will likely have a composition that differs from the core (including the control rods). The differential release of fission products may lead to different types of aerosol clouds. Inventory i The amount of radicactivity calculated to remain in a core is excessive for two reasons; first, the high temperature in a molten core will force fission product emission; second, the surface area assumed for retention should be lower than estimated due to melting. Sequence The compositon of corium should depend somewhat on the type of sequence that occurs. This is due to the rate at which the UO and structural materials both 2 melt and mix. The corium composition influences the fission product release. Issues B.3, B.4, and D.10 em ph 9
MORPHOLOGY OF CORE DEBRIS Steam exolosion Experiments done at the Brookhaven National Laboratory or, core-water interacton show a series of steam explosions occurring. The steam explosions tend to disperse the core debris, making the debris coolable. But experiments done at the Sandia National Laboratory show no such explosions. Hence, the debris is not dispersed and may remain uncoolable.
. Initial melt In the MARCH code, t.he calculations for core melt dropping into the bottom. head assumes a spherical shape for core debris. The diameter of the sphere is an input variable. The number of spheres is calculated using the sphere volume and the node volume. Each sphere has an assumed morphology; the center contains all uranium oxide and may contain some zirconium and/or zirconium oxide; presumably the shells contain only zirconium and zirconium oxide. The cladding may melt before the UO2 melts.
Spraying A significant aerosol source term may arise from a high pressure ejection of molten fuel from a reactor vessel. D. Powers conducted experiments where simulants of molten fuel were ejected from a 1" orifice under 600 psi. A large amount of aerosols were generated. This phenomenon of aerosol formation is not modelled. It appears to be a very significant source of aerosols. Atomization has been given little attention. Sweeping When the bottom head of a reactor pressure vessels melts, the core debris wi'l be released into the reactor cavity and the containment building. Depending on the characteristics of the vessel failure, the core debris may remain in the cavity or be swept into the containment. Rapid quenching of the debris of core may cause a steam pulse that reentrains aerosol particles. Issues C.6 and L.1.k 10
MORPHOLOGY OF AEROSOL PARTICLES The changes in the composition of aerosol and vapors when hydrogen burns has not been determined. Also, heat of reactions are unaccounted for. The composition of the aerosols is considered to be constant. A hydrogen burn may influence vapors as well. One fission product that could be effected is cesium iodide: Cs I + C s0H + HI Issue 0.2 l e O S 4 11
. - . - - . , .,. - , , , , , . - - - c--- ,-;- - - -
DATA Extrapolating data Small scale experiments may not represent large scale phenomena. Fission product release estimates are based on experiments using spent irradiated fuel. The postulated corium/ concrete / water interaction is based on metal / concrete / water interaction experiments. Scarcity Data are scarce. Data are needed on subjects such as deposition velocities and high temperature thermodynamics. Agreement Some data are conflicting: TMI and UKAEA data (Levy). ORNL and German data (Vogel). WASH-1400 cannot supply adequate input data. The temperature data for the TMLB' sequence is inconsistent (Warman). The TMI data about tellurium is inconclusive.
,,se--=,
Issues _M.4.b and M.5 12
CONTAINMENT LOADS FROM HYOROGEN Mode of burn Two modes of hydrogen burning have been proposed for BWR's. In the first mode, hydrogen accumulates in a containment, ignites, and explodes. In the second mode, hydrogen passes through a suppression pool and burns as a diffusion flame. The correct mode must be determined to accurately predict a pressure spike in a containment. Extent of overpressure The AE sequence (LOCA) in the Peach Bottom analysis has 40% of the zircalloy in the vessel reacting with water. Enough hydrogen is produced to overpressurize the containment and cause a containment failure. That a 40% zircalloy reaction could produce enough hydrogen to cause a containment failure is questionable. The models for hydrogen generation appear to predict an excessive amount of hydrogen. This may be caused by the following: an inadequate modelling of the zircalloy channels, a failure to recognize depleted steam, or a single mode used to represent the lower reactor plenum. The excessive amount of hydrogen that is predicted implies an early containment failure; this is conservative. The volume of core zircalloy simulated by the MARCH 2 code accounts not only for the cladding but also for the channel. boxes. The extent to which zircalloy and steam react depends both on the amount of zircalloy and the surface area; this is important because the steam-zircalloy reaction produces hydrogen that pres-surizes a containment. In the Sequoyah analysis, the S2HF sequences describes a series of hydrogen burns challenging the containment. The last burn, after the ice is melted, causes a containment failure. The severity of the simulated hydrogen burns is question- '- able. Nearly all of the zirconium should be oxidized; the depleted oxygen should limit hydrogen burning. The challenge to the containment may be overestimated. Issues G.2, G.4, and G.5
Response
The basis for this calculation is as follows. Hydrogen and steam enter the containment where the core slumps into the lower vessel head. A major fraction of noncondensable gases are swept from the drywell to the wetwell. This : pressure causes the wetwell to fail. l 1
Reference:
P. Cybulskis (Battelle), July 1983 Peer Review Meeting, p 57 of I the transcript. j 13
CESIUM IODIDE Solubility In the Surry (PWR) analysis, cesium iodide is considered to be undissolved in the primary system. Given the volume of water in the primary system, cesium iodide should be dissolved. Volatility In the Surry (PWR) analysis, the report states on page 6-31 that iodide is present as nonvolatile cesium iodide. However, cesium iodide is relatively volatile. Form The chemical form of f odine was considered to be entirely I2 . This was changed to CSI. Other forms of iodine, such as todomethane, have not been considered. Issue 0.4 I 14
BORON CARBIDE Boron carbide is omitted from consideration in the simulate BWR core inventory. This is questionable for two reasons: (1) Because control rod boron can be oxidized to volatile boron compounds, boron carbide can contribute to the aerosol mass. (2) Baron compounds can react with radionuclides; boron oxide reacts with cesium iodide to form cesium borate and iodine. Boron compounds may have a significant influence on the chemical reactions of radionuclides such as cesium and iodine. Issue 0.9
% ew ,,
15 i
.. . . _ . . .I
CORE / CONCRETE INTERACTION l General The postulated interaction of a molten core, concrete, and water may be l optimistic. The proposed scrubbing efficiency is larger than that used for calculation on the suppression pool at Peach Bottom or Grand Gulf. Film boiling should occur at the molten corium/ water interface. Bubble sizes should be described by a. Taylor initiability calculation taking into account the changes in surface tension from impurities in the water. Tellurium Experiments indicate how tellurium reacts as it is released from fuel and how it reacts during a core-concrete interaction. Reactions of tellurium and concrete have not been studied. Tellurium-concrete reactions may have a significant infloence on tellurium release. Decomposition products
- Carbon dioxide is a decomposition product of the core-concrete interaction.
The carbon monoxide and hydrogen also should be listed as decomposition products. Issues C.8 and 0.8 l l l
- No reference available.
16
RADIOLYTIC EFFECTS Reactions Radiolytic reactions are not considered because little is known about how they might occur in a post-accident containment. The fate of cesium after the decay of iodide is unaccounted for. The mobility of cesium may be much different than predicted because a significant fraction of cesium may be liberated from iodide by isotope decay. Transmutating Isotopes transmute into different fission products. For long-lived isotopes, the contribution of transmutation is insignificant. For short-lived isotopes, the gain and loss of chemical species may be significant. - Charoina Inside a post-accident containment, radiation fields are likely to exist. The radiation fields should charge aerosols. Though the entire aerosol would be electrically neutral, the positive and negative ions would have different mobilities. Thus, an electrical charge should modify the behavior of an aerosol cloud. This phenomenon has been given inadequate attention. Nucleation Ionic nucleation could occur in the reactor coolant system (or in the containment). 1 Issues 0.5.a 0.5.c, F.1.c, and F.6 ] 17
CHEMISORPTION The definition of chemisorption is unclear. The literature defines chemi-sorption as binding forces as strong chemical bonds and physiorption as binding forces as strong as Van der Waals forces. But at high temperatures and pressures, the distinction may not exist. Furthermore, the type of sorption may vary from one chemical specie to the next. Fission products may sorb onto surfaces, including aerosols. Issue D.1 1 1
~
I i 18
HYDRATED INORGANIC MOLECULES At high temperatures and pressures, an inorganic molecule may be hydrated. Because the hydrated molecule has a lower vapor pressure than the unhydrated molecule, the hydrated molecule is more mobile than the unhydrated molecule. The mobility of some fission products is underestimated because the hydrating phenomenon is unaccounted for. Issue D.3 O O e 19
STRUCTURE OF MODELS Arrangement of control volumes The control volumes are connected in series to represent the reactor systems. No parallel connections are made. Though the natural circulation within a control volume can be adequately represented, the natural circulation in a reactor system (many control volumes) cannot be easily represented. Number of control volumes The predicted source terms are influenced by the number of control volumes and/or nodal points used to mathematically represent various parts of a plant; i.e., core, reactor coolant system, containment. Issues L.1.c and L.l.d i 1 d ' 20
----.7 r _ - , - .~ . 7y-,-,-r-- , - - - - w --, -.- . ---- - .
REPRESENTATIVENESS OF MODELS ' An analysis of source terms is necessarily plant specific. The analysis of a given plant are inapplicable to another plant. Vessel dome The upper dome region in the reactor vessel of the Surry reactor is isolated l from most of the flow in the vessel. The same may not be true at other plants. i Equipment Fire protection systems are accounted for only for Mark-I reactors. Concrete The cornosition of the concrete in a reactor cavity determines the extent to which many.important reactions occur. One such reaction is the conversion of carbonates to carbon dioxide. When the concrete composition is unknown, the 1 limestone content is assigned a value of 80%. This leads to 36% carbon dioxide i in the concrete. In the Grand Gulf plant, the reactor cavity concrete is 50% to 60% carbonates. This leadr. to about 22% carbon dioxide in the concrete. There, the codes overpredict, t' he amount of CO 2 generated by a core melt to challenge a containment. 7he predicted consequences are overestimates. In the current models describing the reactor vessel failure, the core melts through the vessel and falls directly onto the concrete of the reactor cavity. This model is accurate for a PWR, where the reactor penetrations enter the vessel from the top. However, the penetrations enter a BWR vessel from the bottom. The control rod driver and the instrument penetrations cover the cavity concrete; then, a direct core-concrete interaction should not occur, at least when the core first enters the cavity. The models do not consider the BWR penetrations. The models are be inappropriately applied to BWR analyses of the vessel failure. i i Issues D.7, M.1.a, and M.4.a
Response
The difference between 36% carbon dioxide and 22% carbon dioxide would make a j difference in the challenge to a containment. The difference should not be
- - very large.
l
Reference:
P. Cybulskis (Battelle), July 1983 Peer Review Meeting, p 42 of the transcript. 21 i 1
l i CONSISTENCY OF MODELS Some portions of the models are detailed and have a firm scientific basis. Other portions are based on assumptions. This is analogous to measuring with a yardstick and a micrometer. The predicted consequences are only as good as the weakest significant assumption. Some portions of the models are " conservative," yet the source term estimates are considered "best-estimates." The ideal solution assumption is not always applicable. The dry pathway assumption appears to be invalid considering the. amount of water in a reactor system. The containment sprays are unaccounted for in the analyses of the ! Grand Gulf station. The instantaneous vessel failure assumption is unrealistic. i Interactions ignored. of a molten core with the lower vessel internals are Issue L.1.g i T e me me . . 22
4 ASSUMPTIONS OF MODELS The assumptions need to be justified. Issue L.1.g i 1 I i 23
l ALTERNATE SEQUENCES FROM PIPE BREAKS
)
The location of a pipe break will determine both the release pathway and the fission product removal mechanisms acting along the pathway. Hence, the point at which a pipe breaks has an influence on the consequences of a reactor accident. In some cases, the influence may be minor; that is, a break anywhere along a section of a pipe may give rise to similar pa+' ways. In other cases, the influence may be significant; that is, different .eak points lead to different pathways. In the Surry PWR, a break in a large diameter pipe of the RHR system may occur in a section of the pipe in a water tight compartment. The compartment would flood. The fission products exiting the break would pass through about 3 ft of water. Fission products would be partially scrubbed by the water. Issue A.1.c l l l 24 1
ALTERNATE SEQUENCE FROM OPERATOR ACTION l
.By the definition of a particular sequence, a specific combination of events is assumed. The events represent the most likely behavior of the reactor ,
hardware with little account of human factors. Human factors include operator intervention and operating procedures. An ope' rator may relieve excess pressure in a drywell by venting the drywell pressure through the standby gas treatment system. The sequences define the most likely behavior of unattended hardware. Issue A.2.a
~
;:n;___ '
Response
The sequences define only physical phenomena. Operator intervention or operating procedures are accounted for by the probability of a given sequence occurring.
Reference:
P. Cybulskis (Battelle), July Peer Review, P 27 of transcript. 25 I L
- - - - - ~ ~ -
ALTERNATE SEQUENCES FROM PATHWAYS Reactor Cavity The annular space in the reactor cavity may become a fission product pathway when the steel containment fails at some location and the shield building fails at a different location. A significant pathway arises where fission products could be removed. Furthermore, the reactor cavity is vented through an air cleaning system. Thus the shield building need not fail for a pathway to arise through the reactor cavity annular. These alternate pathways are not accounted for in any sequence though they are significant. In some PWR containment designs, the reactor cavity may communicate with the containment. Water, accumulating on the floor of a containment, may find its way into the cavity. When the core melts through the vessel, a large volume of water will turn into steam and raise the containment pressure enough to cause an early containment failure. The possibility of such a pathway and its consequences have been given little attention. Suopression Pool The MARK-III containment design is less vulnerable to an inadvertent bypass of the suppression pool than the MARK-I and MARK-II designs. However, in MARK-III containment, conduits communicate to the suppression pool; these conduits are potential bypasses. Because a suppression pool removes significant amounts of fission products, a bypass has adverse consequences. Ice Condenser An ice condenser may be inadvertently bypassed because of misaligned valves or vents. The alternate routes are not considered in any of the sequences. Plenum In many of the BWR and PWR sequences, the fission product release pathways being at the core and pass through the upper plenum. However, eddy currents may produce a pathway through the lower plenum, especially when the core slumps and melts through the vessel. The release pathways through the lower plenum _are not taken into account. Issues A.1.a, A.1.d, A.1.e, A.1.g, and A.1.i 26
1 i ALTERNATIVE SEQUENCES FROM EQUIPMENT PERFORMANCE
, Penetrations The electrical penetration of a containment or a drywell may fail before the concrete structure fails; the penetration seals and insulation may degradate under intense heat, radiation, and pressure. A release path through failed penetrations may be either diffuse or concentrated, depending on how many penetrations fail.
The release path would likely pass through the auxiliary building because that is where most of the penetrations lead. The analyses have focused on the failure of the containment structure and neglected a failure of the penetrations. Steam and fission products are dispersed throughout the ice baskets of an ice condenser by turbulent flow. No structures force the steam and fission products through the stacks of baskets. Thus, a major fraction of the steam and fission products would flow through the ice condenser but around the ice baskets. The ice condenser would be effectively byapssed even though steam and fission products enter the volume of the ice bed. When containment recirculation flow fail, reverse flow through the ice beds may occur. This is not considered in the sequence. In some sequences a failure of the containment building leads to a failure of jet oumps. Some of the pumps may survive a containmet failure and continue to operate. Often, an extensive failure of a safety system is assumed, even when at least a portion of the safety system should remain functional. These assumptions lead to a conservative source term prediction. Issues A. l. b, A.1. f, A.1. h, and A. 2. p - i l 1
,. -- l I
1 l 27 l
SELECTION OF CORE MELT MODELS Three scenarios of a core melt have been developed; (1) Coherent drop - the core instantaneously slumps as one unit when 75% of the core is liquified. (2) Downward gradual slump model radial sections of a cone are calculated to melt independently. Upper nodal regions in a radial section heat lower nodal regions. The melt progresses downward. (3) Upward gradual slump model - radial sections of a core are calculated to melt independently. Lower nodal regions in a radial section heat upper nodal regions. The melt progresses upward. The three scenarios reflect an evolving process of developing the computer models. The coherent drop model is no longer used. Instead, the gradual slump models are used, However, these models were developed more on intuition than on facts. In the modelling of a core melt down, a gradual slump is assumed. However, no check is made to ensure that a melted core can be physically accommodated in a lower node (i.e. one that has not yet melted). The calculation may be physically impossible. Current models have a core melting homogeneously. But, cladding melts at a lower temperature than UO2 . The models do not appear to consider a failure of the core structural material prior to the melting of fuel. A collapse of the core may influence the melting process. No geometric changes are assumed when accounting for surface areas and flow areas of a degraded core. The modelling of a gradual slump is inaccuate or incomplete. Issue L.1.a 28
SELECTION OF SEQUENCES The criteria for selecting sequences should be reexamined. Examples show that the selection criteria are inconsistent and incomplete. No small break LOCA sequences are considered for Mark-I BWRs. No LOCA sequences are considered for MARK-III BWRs. No emergency safety feature power failures are considered for MARK-I BWRs or MARK-III BWRs. Including the Zion plan in source terms analysis raises the issue of including external event initiators into all analysis. Operator actions may play a significant role in initiating and propagating sequences. A V-sequence is not considered for Zion because the risk is assumed to be low. A " complete" range of sequences should be clearly defined. Some sequences are selected on the basis of risk while other sequences
- are selected because they involve plant-specific features.
The selection criteria result in analysis that are incomplete and that cannot be compared. Issues A.2.a and K.1 - 29
\
SECTichl 4: LETTERS ABOUT THE PEER REVIEW MEETINGS January 25, 26, 1983 . E e h
llEY$tOS UNIV *dRSITY OF VIRGINIA SCHOOL OF ENGINEERING AND APPLIED SCIENCE
\
- CH AR LOTTESVILLE. 22901 V
E EP ARTM ENT OF NUCLEAR ENGIN EERING AND ENGIN EERING PHYSICS TELEPHON E: 804 924 7138 EEACTOR FACILITY June 6, 1983 R. Bernero and M. Silberberg J TO: FROM: A. B. Reynolds
SUBJECT:
Comments on Source Term Review Second Meeting I am sending you several com.nents and general impressions gained from the source term review meetings of May 24 and 25. ABR:sb 6 M* Enclosures 4 h/3
Comments on Source Term Review from Second Meeting i A.B. Reynolds i University of Virginia i 1 ' NRC Approach 1. I was exceedingly pleased with the revised approach to source term evaluation being taken by NRC, as described by Silberberg. Elements 1 through , 4 effectively identify the important parts of a comprehensive source term I review. Despite the ambitious nature of the program, a less comprehensive one could easily result in misleading conclusions. On Silberberg's chart for the scope of Element 4, he did not specifically list the probability analysis. He showed " uncertainty and sensitivity analysis", but this differs, it seems to me, from probability analysis--which, when combined with the consequence analysis of Element 2, would be required to give the overall risk appraisal. I expect the absence of probability analysis from the list was an oversight; but if not, where is it to be included? The Battelle team is doing an impressive job ef performing and coordinating Element 2. My hat is off to them. The review process is useful despite its limitations. It forces Battelle to justify their results against both expected and unexpected questions. It allows input from others; some of the remarks from the , observers indicates to me that these is of ten insufficient communication between Battelle and some industry groups knowledgeable about the specific reactor designs being analyzed. I sense that the peer review opens the door to increased communication between knowledgeable parties between review meetings. On the other hand, the peer review committee is limited in its effectiveness since we are not actually performing calculations or exercising the codes which is really the way to understand the models more completely. ! This was underscored by some rather perceptive comments by Bob Burns which he is able to provide because of his experience with detailed independent analysis being done by'IDCOR. For this reason the review / appraisal of IDCOR results (as shown by Silberberg as, part of Element 4) will be particularly important.
- 2. Further Documentation It is important to make available to the review members a description of the NAUA code. Our present lack of understanding of the details of the NAUA models or their verification is a serious ' gap. We hardly know what to pursue in detail or how to assess the reliability of one of the important stages in the scenarios. Regarding MARCH and MERGE, I think a more complace physical description of the models would be useful; and I agreed with Don Rowe in his plea for more diagrams of the flow volumes and more flow and heat transfer transient results in the various volumes during the scenarios.
There is a limit to how much information should be presented--which is a problem to Battelle--but I think they have erred so far on the side of not presenting enough. j i
l
~
I look forward to receiving information from Sandia on Te, C 0H, s and Csl surface chemistry--and from others if others are working on this. Surface chemistry appears to be important and we are in the dark about whether Batte11e's models are reasonable--and several of the review members (Ritzman, Johnson, Castleman, for example) appear to be capable of judging the reasonableness of the data.
- 3. MARK I Drywell' Relief Paths I recommend that Battelle follow up on a remark made by Bernero with regard to alternate relief paths through the drywell of the MARK I containment.
In conversations with Steve Hodge of ORNL, I gained the impression, later addressed by Bernero, that the conduits for the electrical leads which penetrate the drywell represent large potential flow areas. These pathways may become available for pressure relief when the drywell pressure rises so that drywell failure may be prevented altogether. Perhaps without sudden rupture of the drywell the SGTS could handle the flow from the drywell to reduce greatly the probability of a 6' RCB failure. In addition to accounting for the drywell relief paths, accounting for the reduction in power label below 30% for the TC sequence may help reduce the likelihood of drywell failure. In the TC sequence the power
- eventually drops from 30% to about 20% since the ECCS flow is - 20% of full
- flow. This may reduce the rate of steam or hydrogen pressure buildup enough to allow relief through the electrical lead conduits to prevent sudden drywell failure.
- 4. Design Features Not Covered by Selected Reactors
~
While recognising the need to analyze specific plants, I remain concerned that the plants selected may not include certain design features that can affect the source term. Two such features were discussed at the review meeting. The first is whether the concrete below the reactor vessel contains limestone since this would affect C09 release from reaction of core debris with the concrete. The second is whetHer water released early from a PWR primary system to the containment can flow to the cavity below the reactor vessel so that core debris might fall into a pool. This water would add to the water in the accumulators as water that could boil to produce a steam spike in a IMLB' accident that could th,reaten containment, and this steam spike might be larger than that for Zion or Surry. Cyb.ulski included an analysis of a Babcock and Wilcox plant where water can flood the cavity below the vessel and he calculated higher steam pressures for this case (-100 psi versus
-80 psi, I recall). We should be sure that there are no U or CE designs.out there where the steam spikes can get higher because I think we lire getting close to the threshold of containment failure. .
Since the effort in the source term review is focused on specific reactors, I I am left confused about the generality of the conclusions. I would hope that these reactors are typical enough that broad general conclusions e,an be l l reached. Recall that TID 14844 provided one set of numbers for all LUR'n. Then l ( WASH 1400 provided a set of numbers general enough but also limited enough j ' for people to assimilate, understand, and communicate. We have not yet talked about how general our conclusions will be from the present source term 9 l 1
-a- - . - - - , .s--v,,,-..- , - . , - - , , , . - - - - . - . , , - - - - +,-_--w,,,,.m. ,e_,-,-,n..,.- -, ,m - . , s,.,w + ,,n,
I i 4 I i i review, but I hope that the review of the five specific plants will lead to fairly general conclusions about all LWR source terms.
- 5. Gradual Slump Versus Coherent Slump Model in MARCH rhe answer provided by Denning for a gradual slump model was
{ persuasive, I thought. This should be written up as part of one of the elements supporting the computer code models. The data base to support the model is still weak, however, and I never got an answer as to how sensitive the source term is to variations in the model. I expect that there is relatively little sensitivity. The phenomenon that might have been affected most strongly by coherent versus gradual slump into a water pool in the lower head is a steam explosion. I am confident that a steam explosion even in a
' coherent slump cannot rupture the upper part of the vessel--much less the containment--but it might influence the later course of events in the primary system.
4
- 6. Condensation in the Auxiliarv Building in Sequence V I remain uneasy about the Battelle calculation that minimizes the effect
_ of condensation in the auxiliary building in the Sequence V scenario, but i I confess that I don't know what to do about it. Perhaps don't do anything i until we find out what the relative probability, hence risk, of the sequence is. l I understand what Battelle is saying, i.e. that the rain forest effect is over i by the time dry steam with fission products arrives. I expect that industry analyses will continue to disagree with this. Alternately there is the industry result hinted at the first meeting by Bob Burns that the pipes would plug up with aerosols before dry steam with fission products would get to the auxiliary building. If the probability for this accident remains the same order of magnitude as in WASH 1400, then we had better be pretty sure of j ourselves about the consequences since, the way Batalle is analyzing the
- sequence, it will continue to be the highest risk sequence, as it was in i WASH 1400.
! 7. Condensation and Fission Product Sweapout inside the Containment after Containment Failure l A remark by Cooper at the review meeting should be followed up. Battelle was discussing PWR containment failure from overpressure from a steam spike when Cooper pointed out that the resulting supersaturation would j lead to condensation which could sweep out much of the fission product-
.invento ry. I presume this part of the scenario would be calculated by the NAUA containment rode, but I doubt whether there is any provision in NAUA
! to analyze supersaturation, nucleation, and droplet growth. - i The phenomena may be important also' for MARK I drywell failure and ,~ MARK III containment failure if and when the f ailure results from steam pressure. t t l l l . l
- ._ . . _.-. . . - - - , , _ - _ . _ - - -- - . m .._,. - . - - . . _ . ._,_._.,._m _ - , _ , _ - _ . . _ - - , -
I envision the following scenario af ter containment failure by steam overpressure. 1 lipJ ! R.(..sy *- e~J:*.x -t W al--*d A;8 ^ a . - t... , v.t.T r>**a r;'- e9mm - (4:3L pes.n , t o gr; , s.A,na.J rin )
*-at" dar '- (e t=*g D at ff1W I**
s "'T*' e..t . sshn,%. s"**
~
6-. T The steam leaving the containment would flow through the break in choked flow at first-either isenthalpically or isentropically. The steam remaining inside the containment would expand isentropically to continue to fill up the space in the containment, as shown by the dashed curve on the P,T diagram. Hence the steam would be supersaturated and would want to condense; the
" supersaturation ratio", S, would be Py/P sat at the particular steam temperature.
Nucleation would occur throughout the containment at a rate that can be calculated with reasonable confidence from classical nucleation theory. The nuclei would then grow into droplets as condensatien proceeded. I would expect that a quite dense rain would occur that would sweep out most of the fission products and aerosols. On the other hand, the fog droplets might remain submicrometre in size, in which case they may not have time to settle out. These theories have been used for the formation of rain, for fog formation in cloud chambers and in steam turbines, and even by us at the University of Virginia for the successful analysis of droplet sizes in the condensation of UOg in ORNL's FAST tests (Nucl. Sci. & Eng., 83,, pp. 459-472, April 1983) . There are probably simpler methods than we used; but, regardless, some type of analysis should be applied if it is concluded that early failure from a steam spike is possible. In ORNL's aerosol experiments v a steam, have they ever looked at the effect of higher than atmospheric pressure saturated steam followed by a sudden decor.pression? G
-v -
INTER OFFICE .'.1EMO f$UN}lN [' .s M E7 SCIENCE APPLICATlONS,INC. l DATE: May 31, 1983 G, l'" l TO: I.B. Wall (EPRI) FROM: R.L. Ritzman (SAI) - v ;
SUBJECT:
MARCH 2.0 Cements Attached is a collection of comments which have accumulated over the past fcur months concerning the MARCH 2.0 Code and Draft User's Manual. The ccmments originate from both EPRI and SAI personnel who have been examining and using the code and its documentation. RLR/jas Enclosure
i COMMENTS ON THE MARCH 2.0 COMP'JTER CCDE AND ORAFT USER'S MANUAL The following comments pertain to the limited distribution version of the MARCH 2.0 computer code received by EPRI in January,1983 and the accompanying Draft User's Manual dated December 21, 1982. The comments are the result of examinations of the documentation by EPRI and SAI technical personnel and of efforts to use the code for calculating severe accident conditions in the PWR source term reanalysis work. General - The manual is certainly of use and it is somewhat better organized than the one for the 1.1 version. We were pleased to see the input and output parameters listed alphabetically for each subroutine. However, the descrip-tion of models and assumptions is too brief in some sections. Specific items of this type are identified in the following comments. It would have been helpful to have notified users which parameter definitions changed between the 1.1 and 2.0 versions. One example of this is the TGEX parameter in BOIL. In the 1.1 version this was defined as the average temperature of gases exiting the core while in the 2.0 version it is defined as the average gas temperature in the primary system. The 2.0 version could be made more user friendly by identifying in one place or perhaps cross-referencing all the parameters af fecting operation of j particular systems when such parameters exist in more than one subroutine in I the code. Examples of this are the IECC and ISPRA parameters which are in the j MARCH subroutine while other parameters controlling ECC operation are in the ECC subroutine and other parameters controlling containment spray operation are in the MACE subroutine. The comment is particul arly pertinent with respect to new users of the code who must learn such subtleties by trial and error. MARCH code capabilities in its present for n are very limited in handling the V-Sequence calculations. For this scenario, a series of calculations must be 4 performed in m1ica one s up :n t.ie ;ar ;n peu.ca :ne ::1.;emat1on neecea to i execute the next stap. T es c4i ai cluna invoive user inuoeling witnin MARCit capabilities which at times are not trivial. 1
Core Heat Transfer Modeling - We were pleased to note that the convective heat transfer between gas and rods in the core now contains a correlation for laminar flow conditions. i BOIL Subroutine Time Steo - We have discovered through running experience with the code that time step size during core uncovery can have a dramatic effect on predicted core heatup rates, degree of metal water reaction, and meltdown time. It would be helpful to alert users of the code about such critical areas in the code. 4 Primary' System Mixture Level Model - There are several comments pertaining to this model. (1) Because MARCH assumes a vertically uniform flow area, the user has to evaluate suitable elevations for breaks and safety / relief valves. It is not easy work. i (2) The mixture level in the hot leg and the mixture level in 1 the cold leg will be quite different due to the. difference in void fractions for the hot leg side and the cold leg side. MARCH has only one mixture level and it is for the j, hot leg because it considers the level swell in the core. t It is not adequate to use this mixture level for detemining whether the cold leg leak flow is liquid flow or gas flow. It may be better to use the collapsed. water level for the leak rate calculation in the cold leg. (3) MARCH has only one two-phase mul tiplier for leak flow modi fication. In the case where there are leaks both in the hot leg and cold leg, this model is not adequate. The i multiplier should not be applied for the cold leg leak rate calculation. j (4) Equation (3.105) is derived by assuming no void in the region above the top of the core. It's better to assume 2a
; void fraction, which is the core exit void fraction, to this i region. By assuming so, Eq. (3.105) will be written as:
[ Yg = (Yggg-Ha)/(1-2a) 4 2 l
Steam Generator Heat Transfer Model - Several comments regarding this modeling are: (1) It is not adequate to calculate a steam generator heat transfer coefficient based on the nomal condition. If the primary side is stagnant, the heat transfer coefficient through the SG tubes will be quite different fecm one at l normal condition. ' (2) Even when the heat transfer coefficient may not be so different from the one in steady state, Eq. (3.114) is not adequate for stagnant conditions. Under normal conditions the average temperature in the SG tubes is almost the same as the average temperature in the primary system. On the other hand, when the primary system is stagnant, the primary coolant temperature in the SG tubes will become very close to the secondary side temperature and will not represent the average primary temperature. In this situation, the effective heat transfer from the bulk of primary systc3 to the secondary side will be almost zero even though there may be a difference between T poo) and T 3g. (3) There is no reason why the SG heat transfer rate should depend on the core heat transfer model. (4) There is no description for why AY(79/Y(ggt is adequate to consider the degraded heat transfer due to primary water depl etion. Core Radiation Heat Transfer Model - Several comments concerning this modeling are as follows: - (1) BOIL considers many separated radiation models and there may be an overestimaticn of radiant heat. For example, it
. is questionable to consider the radiation from the fuel rod surface to the steam and again to consider the radiation ,
from the same rod surface to the adjacent nodes or structures unless gas absorption is taken into account. ! (2) The RHEAT model is just a fitting of a more detailed model. It is not clear whether such a fitting can be oossible or . not. Even when it's possible, it's very diffict.'t for the user to give adequate radiati,on interchange factors and emissivities as input. 3
i (3) The description for RHEAT radiation model is insufficient. (4) The value in Eq. (3.64), 0.173x10~3 , should be 0.173. Debris-Core Barrel Radiation Model - The c:mments concerning this model in the code are: (1) It is not adequate to consider radiation from a crust assuming its surface temperature is the same as the debris-head interfacial temperature. (2) Either of the two radiating areas,7R Fopen 3 in Eq. (3.134) 2 and TRg (1-Fopen) in Eq. (3.135), will be incorrect. (3) It's not proper to assume that all the radiated heat from debris is absorbed in the core barrel. (4) There is no description for the radiating interchange factor derivation. j (5) Debris melt radius R3 = R(FM)1/3 should be R g = R H(FM)1/3 ,
! Anyway, it is true only when debris is a perfect half sphere.
Miscellaneous Comments and Ouestions - The following comments or questions pertain to specific locations in the Draft User's Manual as indicated by page
~
numbers. j (P. 36) The total primary system volume V p is calculated by Eq. (3.1) using the input steam volume V3 . This is not convenient for the user. Yp should be given as input and V 3 should be calculated using V . 3 (P. 44)
"Because the new total system pressure, P', and the new saturation tempera-ture, TMT, are now known" should be "...are not known".
4
- - . , , - , - , , , ~ - - , - - - - - -
(P. 49)
" P3 ,7 = hydrogen partial pressure, psia" should be P3 ,7 = steam...".
(P. 49) The unit for the gas constant, ft /in3 F 2may be misleading. The unit, 3 2 (ibf/lbm) ft /in F is more understandable. (P. 51) PRIMP allows the maximum of 0.1% mass error in each time step. It seems too large to do numerous time steps of transient calculation, unless the mass in the primary system is conserved. (P. 53) There is no explanation for 5778 and 0.583 in Equation (3.42). (P. 61) IPp = 1 should be IPp (R)N(R)/NT = 1 in Eq. (3.58), where N(R) is the number of rods in radial zone R. It is not so easy for the user to give adequate P p 's. The code should have a program which normalizes P s using the inputted P 's. F F (P. 82) In the case of BWR's, channelbox oxidation can not be estimated only by in-
~
creasing the cladding thickness because the temperature of the channelbox is quite different from those of the fuel rod. Also, such an approach does not preserve Zircaloy surface area. There is no description of the model for zirconium-water reaction at a molten node. (P. 86) There is only partial description of parameters in Eq. (3.96). (P. 87) Missing. i i l 5 l
(P. 91) In MARCH 1.1, Eq. (3.107) P=P0/(2.0-M/MO) was written as Eq. (III.I.2) PO W P= 270 (1.0 + wMo ) . What is the difference of these two models? More discussion of the basis for the approach should be provided. lP. 93) There is, no description of how to calculate h in Eq. (3.109). ECC (P. 140) There is no description for how to calculate qi react. What is the order of , chemical reactions? (For example, when there are Fe, Zr, Cr and Ni, which reaction will take place first?) (P. 144)
! The assumption that " energy radiated from the top of the debris is used to decompose concrete" and "the fraction 1-FRW of the radiated energy is lost i from the system" are not adequate. 'More discussion of the reasoning behind this approach is needed.
(P. 178) The heat transfer correlation is for Oconee and the applicability to other reactors is of concern. f 6
. . ~ n . ~ . . - . _ . _. . . _ . - . - ..- - _ - -_- . - -_
1 ., Electnc Power l Aesearen institute i i June 10, 1983 2 Dr. Melvin Silberberg Division of Accident Evaluation U.S. Nuclear Regulatory Commission MS 1130 SS
! Washington, DC 20555 ,
Dear Dr. Silberberg:
This letter offers comments on the draf t of BMI-2104, j "Radionuclide Release Under LWR Specific Accident Condi-i tions (preliminary), Vol. II: A BWR Analysis." (1) We have been previously involved in peer reviews of NUREG-0772, Vol. I of NUREG-0956 in addition to the current review. After considering these three reviews, it seems that the resources being applied to the effort are not adequate to match the technical complexity of
~
the task as it has been defined. This is particularly
! true since the effort correctly requires high quality work in analyzing complicated physical phenomena and thorough visibility of all the steps in the analysis.
The latter factor especially requires special effort in writing and communication. T whether the significant effort required to address the problem as it has been defined is commensurate with the
, low probability of serious degraded core events is a !
matter which should be evaluated from time to time. } (2) The analyses are based on the results .of calculations j using a variety of codes. The descriptions of the codes up to this poi'nt have been generally qualitative, l and some of these descriptions are not even available so far. Therefore, the peer review on May 24 and 25 ". was very superficial. It is noted that Oak Ridge has 1 been assigned the task of preparing reports on " code validation and data base status" for ORIGEN, MARCH, MERGE, CORSOR, TRAP-MELT, OORCON, VENES A, SPARC, ICE CONDENSER MODEL and NAUA-4. The proposed scopes of 1 these ten reports is impressive; if the reports are ! i executed in a timely and thorough manner, they will be j j of great help in future peer reviews. It is further i gratifying to note that BCL is undertaking a sensitivity l analysis for key parameters. f hp/"(,4b) 3412 Minview Avenue. Po. ofrice som 10412. paio Aito. cA 94303 Teseonone M15' 855-2000 we,n.nton on,ce mo ueseeenu.erts on. nw. su,re m w.sn,nton. oc mmmi smm - l l
Dr. Melvin Silberberg Page Two June 10, 1983 Since these codes are often specialized, it would be helpful if independent peer reviews were conducted on each code by appropriate specialists so that their re-ports are available to'the current peer review group as soon as possible. It is of paramount importance for the codes to be pre-sented in such a fashion that the physical principles are visible and adequately tested so that the reactor safety community has confidence in the results. We feel that a latar peer review of BMI-2104 is neces-sary when the maths logy is better described. (3) As pointed out previ Lly, there is a paucity of infor-mation on the codes. Iowever, in the review of the part or BMI-2104 avai.able to ds, several particularly significant points deserve mention. These are the following. (a) The codes must be used, and if necessary modified, to predict the location, physical state, tempera-ture, etc. of the water in the primary coolant system, the drywell, the wetwell, and containment at various times in the accident sequence. Water is by far the constituent in greatest amount and can dissolve salts, wash down walls, etc. (b) Concern has been expressed concerning the applica-tion of MARCH to a BWR since BWRs contain channel boxes (shrouds). We understand that the BCL group certainly recognizes this problem, but the way in which the problem was approached has not been clearly described. (c) The core slumping model (page 5-5) directly af-facts the source-term since each part of the core that slumps is added to the source 'in the primary system.- Since the slumping depends on the size of the node, the results are user dependent. . Therefore, one part of the parametric ' study should' show how ' sensitive are. the results to the selec-tion of the nodes. In' addition, core slumping is described in the CORSOR description rather than MARCH. If we assume that slumping is determined by MARCH, it appears from Table 6.1 - (p. 6-37) that the BOIL model A was used for meltdown -- and model A of MARCH is for instantaneous core slumping when
Dr. Melvin Silberberg Page Three June 10, 1983 75% of the core is molten. Thus, there seems to be an inconsistency between the slumping model description (p. 5-5) and the actual input to the model (p. 6-37). Also (p. 6-37) " core slumping starts when lowest node in region is molten" --
- what if the upper nodes are molten first and
. freeze later. The slumping model also affects.
the timing of event. (d) We believe that SPARC may give low decontamination factors when the aerosol passes through a hot pool. A SPARC vs. SUPRA vs. experimental data comparison should be undertaken. (4) The Te releases proposed by BMI-2104 are higher in some cases than WASH-1400. The experimental evidence for the release of Te from core-concrete melt is essentially non-existent and should be promptly studied in an ex-perimental program. The rate of release of a vol'atile material fron a large molten mass will be very sensitive to scale. It is not explained how this matter was treated in VANESA. Additionally, the rate of release of Ta from the core-concrete melt will depend strongly on the rate of sparging. This in turn depends on type of concrete used. In this report, limestone which will release CO2 was_said to be assumed to be present. The type of concrete should be confirmed. Surry was found to not have limestone in its concrete and this may indeed be true for Peach Bottom. (5) Some of the statements (for example page 7-28) indicate that the recent BCL-EPRI scrubbing results in which it is shown the aerosol scrubbing by a hot pool is as ef- ! factive as by a cold pool have not been factored into ! BMI-2104. i (6) It area is not of 10 clegrft to wasusassumed, why a containaient which overwhelms dry well the break capa- ! city of the stand by gas treatment system (SGTS) . A smaller leak from the drywell could be handled by the SGTS, thereby not putting the reactor building under pressure and damaging its integrity. (7) Venting of wetwell through the SGTS at a deliberate rate may delay or preclude the failure of the drywell. We believe such a possibility is allowed in the emer-gency procedures for the Mark I BWRs.
Dr. Melvii. Silberberg Page Four June 10, 1983 (8) The temperatures shown in the report for the core and the BWR upper internals are much higher than those ob-tained by us. We have not been able to discover the source of the discrepancy so far. (9) The core slumping mode assumed in the analysis of the progression of the BWR core degradation appears to be different than that assumed for the Surry core degrada-tion, and it may be the reason for the "high" tempera-
, tures calculated in the BWR system.
Could this be explained and the effects of different assumptions on core slumping delineated? (10) We were troubled by the very large sensitivity (several hundred degrees) of the primary system temperatures to the nodalization of the Mark III primary and contain-ment system. This should be explored further (11) Hydrogen combustion loads are supposed to f ail the con-tainment in the Mark III TQUV sequency. This appears to neglect the information from recent experiments which show that the hydrogen combustion will be in the form of diffusion flames, which will not generate high pressures.
~
EPRI appreciates the opportunity to review BMI-2104. It is recognized that the evaluation of degraded core accidents as the methodology is now being developed is very complicated. BMI-2104 needs much more explanatory material before it can be subjected to a good peer review. We realize that the
; comments just presented are general,but in depth comments cannot be offered until the greater detail is presented.
EPRI also wishes to acknowledge the very significant effort on the part of dedicated people which has already been expended. , Sincerely yours,
, f f Dal Raj Sedgal Richard C. Vogel ,
i Senior Scienfitif Advisor
- Senior Program Manager l Nuclear Safety & Analysis Dept. Nuclear Safety & Analysis Dept. ;
i RCV:BRS :br cc: J.J. Taylor W.B. Loewenstein I I.B. Wall
C.A lt. . Tee: : e 5*' ?~: E*::
~.s *:25:.~t'9 1.'::l33 ATC..V!G l'iCTRIC C0iliPANY myg s
-= :- -. m -: g :.= ygggag. geng 3 7 .
s, .e -.' N K E 2
& Ob(WY June 6, 198 3 Mr. Jankowski LJ.S. Nuclear Regulatory Commission RES/ASTP0 Washington, DC 20555
Dear Mr. Jankowski:
At the last meeting of the Peer Review Group for the accident source term p rogran, it was proposed that tellurium be released at the same rate as iodine for high rates of Zircaloy oxidation. Partial justification for this thesis was data from TMI-2 as presented in a paper handed out at the meeting. As I stated for the record, it is my opinion that the tellurium data reported from TMI is inconsistant and inconclusive. If this data is to be used as a basis for tellurtum releases, the original counting , sheets should be reviewed to detennine the accuracy of the repo rted re sult s. As I understand, you presently are attempting to do this. I am including a brief discussion of the celuritan measurements that I an aware of and the questions raisec in my mind from the reported results.
~
I would appreciate b_eing kept infonned of your investigation. Sincerely yours, CCT% o C. D. Thomas, Jr. Radiological Engineering Group CDT/imm , Att achme nts k h *o w lkt e. 9Ad 1
- I u 2.W L C. con.
A wk u ves L. MS @ E t d l6wsi W
- o/04 6 SOM 'y
Some Comments on Reported Tellurium Measurements at TMI-II i Reactor Coolant S amp le #1 This sample was taken at 1600 on March 29, 1979__and analyzed by Bettis Atomic Power Laboratory (BAPL) and Babcox and Wilcox (B&W). B&W did not report Tellurium in the sample . BAPL reported only(132 Te at 0.086% of the total core inventory (compared to 8.41% for 131 I ). 11 It is suggested that this analysis be reviewed to determine if the presence of 132Te is in fact co rre ct . In particular, should other nuclides of Tellurium have been detected (probably a strong function of the time the sample was counted). In addition,
- again depending on the time the sample was counted, was a possible
- interf erence from 1 38Cs considered?
4 Reactor Coolant S amp le #2 This sample was taken at 0730 April 10, 1979 and was analyzed by BAPL, B&W, ORNL and Savannah River Laboratory (SRL).lZl With the exception of BAPL, the results are not generally available. Of interest in the BAPL results is that only 132Te and 103 Ru showed fractions of core inventory less than the i first RCS sample. The results of these sample analysis should be reviewed to
; subs tan tiate the 132 Te and examine the possibility of other Te's present.
'In particular, should 129Te and 129"Te be detected?
v Sump i The sump was sampled on August 20, 1979 and analyzed by ORNL.(3) The solution shows 129 Te present but no t 129mTe. The solids show 129"Te present but not 129Te. The reason for this and the absence of 1278Te and 127Te are not c le a r. f Another sump sample was taken by 3&W sometime later but never reported. A sample taken and analyzed by EG&G in May 1981 reported no Te in the sump j solution or solids. Should their AA technique have detected Te? Containment Air The containment atmosphere was sampled during the period April 29,19*0 to May 2,1980 and analyzed by EG&G.(4) The reported 129mTe concentration is guestionable as is evident from the footnote in Table 14. Should not 129Te,
- L27Te and 125mTe also have been , detected?
Penetration Seal Plate taken Augus t 28, 1979, was analyzed by ORNL.(3) 125"Te , This 127*re sample and $298Te were measured. The ratios do not appear correct, and the absence of 129Te and 127Te is not unde rs tood. ? J f
- . - - - . - , .- - - , . , - - - - - - .~ - _ - - . - . _ _ .--. ---- -
l Rafarsnc=s
- 1. C. A. Pelletier, C. D. Thomas, Jr., R. I. Ritzman, F. Tooper, Iodine 131 Behavior During the TMI-2 Accident, EPRI Nuclear Safety Analysis Center Report NSAC-30, November 1981.
- 2. W. N. Bishop, D. A. Nitti, N. P. Jacob, J. A. Daniel, Fission Product Release from the Fuel Following the TMI-2 Accident.
- 3. Letter R. E. Brookshank to Dr. Ben Rusche, Septem5er 12, 1979.
- 4. J. K. Hartwell, J. W. Mand lor, S. W. Duce, B. G. Motes, Characteri zation of the Three Mile Island Unit-2 Reactor Building Atmosphere prior to the Reactor Building Purge, GEND-005, May 1981.
i p r
{* l LETTERS ABOUT THE PEER REVIEW MEETINGS July 28, 29, 1983 l m_
[00ff$ On the Propagation of Error in Air Pollution Measurements John S. Evans Douglas W. Cooper, and Pate:ck Kinney Department of Environmental Science and Physiology Harvard School of Public Health Boston, MA 02115 Submitted for publication in Environmental Monitoring and Assessment. April 1983. M y 'J f .' > 0 Y
-- e
ABSTRACT Four methods for estimating the uncertainties in air pollution measurements are outlined. The approaches are: analytical solution - approxi-mation; application of distribution theory; experimentation: and simulation. The advantages and disadvantages of each method are illustrated using data from High-Volume air samplers, the instrument most commonly used for monitoring ambient concentrations of airborne particles. O 11
INTRODUCTION in air pollution studies, as elsewhere, important variables are actually derived from measurements of more fundamental quantities. Concentration, for example, is often determined by the change in mass of a filter, the volume dow rate, and the duration of sampling Uncertainties in the measurem2nts of each of these quantities thch combine to produce uncertainties in the inferred value of the variable of interest, concentration. This paper presents four approaches to estimating uncertainty in the derived quantity from uncertainties in the measurements of the fundamental quantities: (a) analytical solution - approximation (b) application of distribution theory (c) experimentation (d) simulation The application of each method is illustrated using the problem of determination of the total uncertainty in mass concentration measurements given by the High-Volume air sampler. We are addressing the use of a single set of measurements to use to calcu-late a single value of a derived quantity. Reasons for wanting to know the uncer-tainty in that derived quantity are partly determined by the use to which that quantity will be put, but also we might use the uncertainty estimate to decide how many replicate measurements to make (the uncertainty diminishing with the square root of the number of independent determination) or whether the equipment is suitable for the task, or for which component of the derived quan-tity it is rnost cost-oficctive to improve measurement precision.
" * = - **% e.* - - - .% .
Statistical uncertainties in a measured quantity, or " measurement errors." are often thought of as falling into two distinct classes, systematic and random. Systematic errors are those for which an assignable cause can be found and corrected, such as a faulty calibration, while random errors include all unassign-able and/or uncontrollable errors. This paper deals with so-called " random errors." those which increase uncertainty but do not create. systematic bias. (Of course, all errors have causes.) Surprisingly little attention has been paid in the recent air pollution litera-ture to these important aspects of data collection and interpretation. Most often, an analysis of random measurement errors (as opposed to systematic biases) begins and ends with an estimate of the variance of the overall distribu-tion of errors, often termed ' precision' (see e.g.. Fein and Bailey (1978)). The distribution itself is almost always assumed to be normal. A further rednement of great potential utility which occasionally appears is the attempt to identify and estimate each of several individual sources of error and to evaluate their relative induence on the total error (Jaklevic et al. (1901). ACCIH (1978) USEPA (1973)). The value of such an approach lies in its ability to indicate those com-ponents of the total measurement system upon which additional control efforts will be most effective in reducing overall measurement errors. The derived quantity we study here is the concentration as determined by a High-Volume Air sampler. The High-Volume air sampler draws air through a alter which is almost perfectly efficient (> 99.977.) at capturing particles. The mass concentration of particles in the sampled ai'r is estimated on the basis of the change in weight of the filter during the sampling interval are the volutne of air passed through the filter. Mathematically, the correct value f., is a func-tion of three variables: 2
_ _ m., ___ _ __-m. _ _ - . _ - - _ _ . _ . _ _ _ . _ _ _ . _ _ _ _ _ _ _ . . _ . I 1 4 i 4 { l f. =z izi'zi' (1) zi = net change in filter mass during sampling interval ( g) i zg = average flow rate during sampling interval (m3 / min) I z3 = duration of sampling interval (min) i i i j ) However, the observed TSP concentration, f depends not only upon zi. :: l j and z3, but also upon the errors in measurement of each, dzi, dza and dra: i I
- / =(z +dzi)(z e+d=2)-'(z s+d=s)-'
i (2) 1 i The error in measurement of the TSP concentration, df. is then given by: i j df =f -f, (3) I s j A complete investigation of measurement errors might have as its goals: i
- characterization of the distribution of errors, estimation of the mean and vari-ance of the distribution of errors, generation of confidence Intervals for indivi- ,
t
- dual values of f. based upon knowledge pf f and df and determination of the .
l i relative importance of each of several sources of error. I I ] METHODOLOGICAL ALTERNATIVES i 1 j Analytical Solution - Approximatian i j ! j A classical approach (Box et al.,1978) is to write the variance of the distri-l I l bution of errors in f as a function of the errors in the zj as: affas f(af / Ozj)8 aa,,lz [4) s= i l This formulation. referred to as Gauss's law of error propagation, assumes that the errors in the independent variables are uncorrelated and relles on the first ! I 3 i . sh==me = + - .
f ! I 4 term of a Taylor series expansion. thus limiting the accuracy of analysis to I j "small" errors in the zj. If f is: i-i ! / = f ajzy (5) { fa1 1 equation (4) gives: i 8
- a!f = f a 'a4,,
s (6) s-n i a weighted sum of the variances of the independent variables. If f is: '
/ =ri 'ze* 8 z ** (7)
! equation (4) can be re-arranged to give: a . j
- aff*f.saf(
p} i n,,is ' )8 (8) ! J f i a weighted sum of the squares of the coefficients of variation. These approxima-i tions become less accurate as the error contributions become larger, often the , i j cases of most interest, however. l The approach may be extended to situations in which errors are correlated. 4 See, for example, Bevington: i 3 1 l = j GIf
- 2= ((d/ / d2f )'a ssj '+2 (6/ / d2.j)(d/ / d2=)a ss,4, } (9) l anos j t
- Note that we will assume, for simplicity, that the errors are uncorrelated, even ;
I I though the variables themselves may wel! be correlated. In addition, although 1 i k 4
. i not widely realized. the analytical approach can be modified so that it will accommodate larger errors. Correction terms have been developed for large errors which are based on higher order derivatives (Seiler.1983).
Box et al (1978) recommend that the partial derivatives of the function be evaluated numerically rather than by the calculus. I.e., d- = [f (N*/+3a,,)-f ( ,j)]/ 33,, (10) a = [f (M*/ } ~/ (#'l-3a,,)]/ 3a,, (11) If these two estimates are approximately equal then their average may be used as a working estimate of af / d j. To illustrate the application of Gauss's law of error propagation, the errors in measurement of total suspended particulate matter with the High Volume air sampler were estimated by this approach. The data used in this analysis were drawn from the quality assurance documents of the Harvard School of Public Health Six City Study of Air Pollution and Lung Function (HSPH.1981). These data are summarized in Table I. The measurement error in net filter mass, dzi, was determined from a ran-dom sample of replicate analyses of filter mass. The measurement errors in average flow rate, d=a, and in sampling interval, dr a, were estimated by com-parison of measurements made using fleid procedures with measurements made using standard reference devices, e.g., a calibrated orifice and a calibrated timer. The measurements which served as the basis for estimation of these errors were drawn from a large data base generated by routine performance audits. Note that according to these records both the nel filter mass and the 5 --M m- -% _w,- + . . ,
3 i i 4 1 4-i i 1 Table 1. An Illustrative Data Set
- I Variable Mean '
i \
- xg net alter mass (yg) 5 1.34 x 10 1 .
1 { x2, average a w rate (m / min) 1.50 l x3sampling interval (min) 1.44 x 10 3 l Variable Variance dx , error in x t g 3.26 x 10 5 dx2 ff f I" *2 1.12 x 10-2 dx3' 'If f I" *3 0.90 x 10 l i i i 5 1 1 l I. - l 1 .
'! e f I
[ i f
'l i I i
e 9
\ . i j .
l i i i I 1 r [ j .- . - . . . . . . . . I 4 ,
sampling interval are known to within less than 0.5*;. In contrast, there is over ten times this uncertainty in the determination of average flow rate. The uncer-tainties in mass change and in time are almost negligible compared with those in now rate, which appears in the denominator. These data are useful in showing how to treat an error contribution that is not linearly related to the property sought. As recommended by Box, the partial derivatives were evaluated numeri-cally. However, as is illustrated below, in this example, the numerical solutions were not very different from those which would have been obtained analytically. As mentioned above, only za (the now rate) was subject to much uncertatnty. The analytic solution for the partial derivative of / with respect to ze in the vicinity of the means of zi, zg, and zs is: , Oz = -ziz-8z3-8 a = -41.36 (12) 1 3 (in units of yg/m per m / 3min flow rate change) and the estimates from the 3 numerical dit!erentiations are 40.45 and 42.31 having an average of 41.30, nearly identical to the analytic value. Using the datn in Table I and evaluating the partial derivatives and applica-tion of Causs's law of error propagation yields: afp2.14x 10-'(3.20x 10s)+ 1.71 x 103(1.12x 10-a), g,ggx to-3(0.00x 10)
= 0.08 x 10-8 + 1. 92x 10 & 1.67 x 10 -8 = 10.2 (13)
This is equivalent to a standard deviation due to mliasurement error, aq, of 4.30 3 g/m , Evaluating the true value of TSP at the means of z s, z e and23 gives f a g 3 62.0 pg/m , This corresponds to a coefficient of variation due to measurement l 6
error of 7.1%. It is also worth noting that virtually all of the measurement error i in f is due to measurement error in g, the riow rate. This 7* uncertainty esti- .
; mate is about twice that found in an inter laboratory investigation reported by McKee et al. (1972). The lower values found by McKee et al. are partially explain-able by the tendency of those involved in an inter laboratory comparison to be among the better practitioners and take greater pains with measurement tech-nique than is likely to be routine. Their lower estimate also includes the effect of eliminating one of the studied laboratories as ":!gnificantly different" from the others.
While this approach provides an estimate of the variance, a4f8 , of the distri-bution of measurement errors, it does not yield information about the nature of the distribution. Therefore, care is necessary in the formation of confidence intervals. A conservative approach involves forming canadence intervals based upon Chebyshev's it. equality: P(if, -f I >k au)< h (14) which indicates that 1/ km of the values of f, would lie outside the range l f ik a,f. Using the Chebyshev inequality it can be estimated that 00". of I observed TSP values in our data set lie within D.83 g/m3 of the true values, and that 95". of them lie within 19.0 pg/m3 of the true values. l' Although this inequality applies to any distribution with finite mean and variance, and thus is broadly applicable, much sharper confidence limits can be placed on / with some knowledge of the nature of the distribution of errors, df. Table !! illustrates this point. The 80 confidence interval for the normal distrt-i bution is 57". as wide as that given by the Chebyshev inequality. The difference = 7 4 m ----,r,-~v ww,----.--w- - - - - - - +
---9 m-%--- . s ,__ w- i----- e - + - - - - i--w ---- -- e----,- - - --E
. . _ . . - - . _. _ . . _ . - - . - - - - . _ - . . - . - . _ - ~ . _ - - _ - . - - . _ _ - . . . - - . - - _ - _ .
J l 1 . 4
; Table II. A comparison of confidence intervals from the normal
.I and Student's t distributions with the Chebyshev inequality (4). 1 i Number, k. of i Fraction of Values between x ks, and x+ks,
- Standard Deviations ,
i i normal Student's t Chebyshev ' I (3 d.f.) 5
! 1 0.682 0.608 > 0.000 I 2 0.954 0.851 > 0.750 3-0.997 0.939 >0.889 4 >0.999 0.968 l
>0.938 l
, 5 >0.999 0.984 >0.960 1 l
, Percent of Values Number, k. of Standard Deviations ;
Outside of % Interval x ks,<x<x+ks, normal Student's t Chebyshev i (3 d.f.) } 1 20 1.2S 1.64 2.24 10 1.64 2.35 3.16 i 5 1.R6 3.18 4.47 1 2.57 5.84 10.0 > l .l l ] 1 l i . i I I i i ,i . J i . i i i i ! 4 4
- e, i
4 1 9 l l l 1 t j adummew e w I-..._-_..-..,_,.___.-.-,,--,.>-----. . - . _ _ , . - .. - - - -,-, -_ - ,__,-,. .....-. ~ ,_..-, ,_ - _ ,,..~
I I i is even more pronounced for the 397. confidence intervals. Here the normat : intervalis only 267. as wide as that given by the Chebyshev. 4 Distributton Theon) In an attempt to tighten conadence intervals, and to avoid the problems inherent in use of the Taylor series expansion, one may study the theory of sta-Listical distributions for more exact analyses of propagation of error. t While not generally true, there are certain cases in which the parameters of the distribution of overall measurement error, df, may be deduced from the parameters of the distributions of the components of measurement error, i.e., d::,d: ,d:3 As a background for discussion, a few deanitions and remarks on the properties of two commonly encountered distributtons are in order. The normal or Caussian distribution is familiar to most air pollution special-ists due to the role it plays in air pollution modo!!!ng. It is a unimodal, sym-metric distribution with two parameters, p (its mean) and a (its standard devia-tion). The probability density function for the normal distribution is: l pd/ (z )=(#2Re)-lerp[-(:- )'/ 2a8] (15) fo r -== < x < =. The normal distribution has gained widespread use partly because of its convenience, partly because certain phenomena tend to fo!!ow normal distribu-tions. For errors to be distributed normally, positive and negative errors of a given magnitude must be equally likely. l It is well known that if z eand z are e independent random variables distri-8 1
buted normally with means g, yg and standard deviations ai,a2 that values of the function:
/ (zi. z) = at:i + agr a (16) are normally distributed with mean, yf = aggi + a 2. and standard deviation, af = [a f a' + af aa ]8 See, for example. Burington and May (1970).
{ The lognormal distribution also has widespread use in air pollution, as it is found that both airborne particle sizes and pollution concentration levels are often lognormally distributed. The lognormal is a unimodal, positively skewed 2 (i.e., long tail to right) distribution characterized by two parameters y and a. The parameter g is the mean of the distribution of the natural logarithms of z. The parsmeter a is the standard deviation of the distribution of the naturalloga-rithms of r. Its probabl!!ty density function is given by: Pd f( )=0 (17) for : s 0 Pdf(x)=(#E6az)-ta=p[-(inz - )/ 2a8] (18) for z > 0. The mean of the lognormal is exp(y + a s/2). The variance is exp[2g + a ][exp(a ) 1]. Often the lognormal is reparameterized using the geometric mean (or median). ,, and the geometric standard deviation, a,: 1 y, semp(#) (10) 9 m e e. - .
- ~.
7 - , .-. ,, ._ . - . _ - - - - . . , ,, , ,n--
1 1 . L 1 4 a, = exp(a) (20) ^ Variables may well be lognormally distributed when the factors caesing i deviation from the true values tend to give deviations proportional to the true 4 4 - value rather than deviations which are independent of the value being measured. f , If z is distributed lognormally with parameters y and a it can be shown ! (Altchison,1957) that values of the function: i i l i i [ <{ /(z)=ar" (21) i i i i are distributed lognormally with parameters yf = in(s) + ny and af = na. Similarly, if zi and ze are Independent random variables distributed tognormally j with parameters yi, ye, and ai, as it can be shown (Altchison,1957) that values of i the function: , J t I l f (zi.z s)=r *'z a*8 (22) i 6
- are lognormally distributed with . parameters and pf = ani+ayr ai = ta t al + alall.
1 J
)
Although there are many cases in which the results reviewed above will be exactly app!! cable, there are many other circumstances in which they may be approximately applicable. To extend the applicability of the approach it is use. fut to realize that in instances h which variables are distributed normally with j standard deviations much smaller than their means, the lognormal distribution
- i l
1 Is a close approximation to the normal. One simple method for obtaining an ! [ estimate of the geometric standard deviation Ef the lognormal distribution i I i t which well approximates the normal distribution is to use the square root of the ; ! ratio of the mean plus one standard deviation to the mean minus one standard I r [ 10 l a
deviation. To illustrate the application of this approach we return to our example involving the High Volume air sampler. It is clear that the measurement errors are smallin comparison to the values being measured for each of the three vari-ables. Extending the results from equation (21) to three variables and using the lognormal approximation to the normal for each of these variables we have: 8 s a o cuf=[otna,,+a,,d,,q,,d,,p/2
=[(4.2Sx 10-8)8 +(7.05x 10-8):+(2.07x 10-3) ]'/ := 0.0707 (23)
M8af EMins gY ansgM ns3
= 11. 8 -0.403 -7.27 = 4.13 (24)
The distribution theory approach g'<es an estimate of the coerncient of variation due to measurement error of about 7.1%. Here again [L is evident that most of the measurement error is contributed by uncertainty as to the now rate. Using
~
the approximation that the errors follow a lognormal distribution. :onfidence intervals may be generated. At the mean concentration of 02.2 pg/m3. 90*; of the true values would be expected to lie within the Interval 50.0 g/m3to 60.1 yg/m 3and 95% of the true values would be expected to lie within the interval 3 3 54.1 g/m to 71.5 pg/m . These confidence Intervals are assymetrical and smaller than those generated with Causs's law of error propagation and the Che-byshev inequality. Expression of confidence intervals for any asymmetric distri-button is not as straightforward as for a symmetric distribution. In that there is no natural choice for the percentiles over whictd the given percentage of the values are to be included; for example,90% of thu values could appear within the 1st to the Olst. 2nd to D2nd, etc. percentiles. If the percentiles are syminetric with respect to 50, the values w6ll not be, and vice versa, lleru we list the 95 11
, percent canadence limits for the tegnormal as those values representing the 2.Sth and 97.5 percentiles, respectively. Similarly, one has the choice to represent the confidence intervals with respect to the mean, the median, or some other measure of central tendency, depending on the distribution (the median is a natural choice for the lognormal). A potential pitfall with the distribution theory approach is related to uncer-tainty as to the distributions from which the underlying variables, e.g.' i, :g, z s, come. There are statistics such as the Kolmogorov Smirnov D statistic and the Shapiro-Wilk W-statistic which may be used to assess the likelihood that a particular sample arose from a specified distribution, e.g., normal. However, as Seiler (1982) has pointed out, without large samples it is often difficult to differentiate among several plausible candidate distributions. The most useful information for this purpose lies in the tails of the distributions. Thus, in most cases the underlying distributions are known only approxi-mately. And although the distribution theory approach is capable of producing exact results it is prudent to be cautious in making statements about the proba-bilities of very large errors, those in the tails of the distributton. Experimentation An alternative to analytical study of the propagation of error or appileation of statistical distribution theory is experimentation. In the analysis of overall measurement error in a monitoring system a com-mon experimental technique is so called " paired sampling. For example, to determine the overall measurement error in an air sampling and analysis tech-nique, two identical samplers might be co located and operated simultaneously 12 h ese gge =
for many sampling periods. The samples so collected would be analyzed in the same manner, and pollutant concentrations would be calculated using identical algorithms and computer code. The diderences between observations taken at the same time and place would be taken as evidence of overall measurement error. The most elementary form of analysis of the observed diderences. d. is to plot a histogram of th,eir values, compute the sample mean diderence, d. and compute the sample variance, sf. of the diderence values. If there is no evidence of a systematic diderence between the two paired sampling systems, then the sample variance, sf. of the dide.ences may be used to generate an estimate of the Ocmponent of the variance of the individual observations which is due to " noise" in the monitoring system. It can be shown that if the measurement errors of the two monitoring systems are independent of each other, and of the values being measured, and are equal. aq as [sf/ 3}:/a (25) This basic experimental approach may be litustrated with data available from two experiments conducted in conjunction with the Six Cities Study. In Ltertown. MA two lii Volume air samplers had been operated side by side each sixth day from 15 May 1979 to 33 December 1000. The data collected did act reflect any systematic diderence in the performance of the devices nor any didorence in their precision. The mean observed concentration from the 51 paired observations was 45.3 g/m3. The standarit deviation of the distribution of measurement errors was estimated to be 3.40 g/m 3. corresponding to a coedicient of variation due to measurement error of 7.50. A similar experiment involving 37 paired observations in Topeka, KS gave an estimate of the standard
~
13 emmam ,
deviation due to measurement error of 1.78 yg/m3at an observed mean con-centration of 32.4 g/m ,3thus a coet!icient of variation due to measurement error of 5.4*:. More complex forms of analysis, such as analysis of variance (ANOVA) or regression analysis, may be applicable to situations where the mean ditYerence is not zero, or where measurement errors appear to be related to (e.g., propor-tional to) the magnitude of the quantity being measured. These more complex forms of analysis may also be useful when estimates of the individual contribu-tions of severa1 components of measurement error are needed. While the experimental approach has the advantages of giving a relatively direct estimate of the total measurement error of a system and yielding some information on the distribution of measurement errors, it has some limitations and disadvantages. First, the procedure is likely to be relatively expensive. Second. as in any experiment, a fair amount of thought is necessary to its proper design and interpretation. Consideration must be given to the nature of " measurement errors." Caution must be exercised to ensure that the conditions that may influence " measurement error" are similar in the experiment to those found in the course of ordinary applications of the measurement system. To some degree costs and the potential for misinterpretation of results may be minim-tzed by following well established statistical methods for the design of experi-i monts. See, for example, Cochran and Cox (1957) or Box et al. (1970). Simulatton A fourth approach is simulation, l.a., mathematical experimentation. If one 14 men se - .
knows the distributions of several variables, e.g., zi , z a. n, and is interested in the distribution of values of f(zi,za. z ) an estimate may be obtained by:
- a. randomly drawing several sets of observations from the distributions of the x's; i b. evaluating / for each set of observations ca the x's; and
- c. summarizing the distribution of values of f which result.
Application of the simulation approach is illustrated with data from the Six Cities Study quality assurance program. Values of average now rate, sampling interval and the errors in determination of net alter mass, average now rate and sampling interval were assumed to be normally distributed. These distributional assumptions were based on the apparent normality of data on aow rates and sampling intervals, and in the case of the errors on the assumption that positive and negative errors of equal magnitude would be equally likely. Values of net n!ter mass were taken to be lognormal, since net alter trass is proportional to ambient TSP concentration which in these<1sta is distributed approximately log-normally. The parameter values derived from the Sir Cities data and used in the simulations are summarlzed in Table !!!. Simulated observations were generated using the SAS computer algorithms. For each simulated observation, the measurement error, df was calculated: df u(zg+dzi)(*a+dza )"'(z a&dz a)'l-riz '8a 2 3"' (26) From one simulation with 500 observations the standard deviation due to measurement error was estimated to be 0.93 g/ sit 0. The simulated mean true TSP level, f., in this run was 04.0 g/m .3 In this simulation 00". of the observa-tions were within an interval f,-Gyy/ m,8sf <f, &7 g/ m8 ; and 95". of the 15
Table !!!. Data Used in Simulations Variable Distribution Parameters xg, net alter mass ( g) legnormal yu, = 11.6 or , = 0.33 . x2, average a w rate (m3/ min) normal #,, = 1.5 a,,' = 0.075 x3, sampling interval (min) normal p,,a=,1440 a,, g7g dxg error in xg(yg) normal yo,e a
==3.3 0 x 10 5 m,
dr2, err r in x2 (* /*I") "Of * "I M*e=0 a a u, = 1.1 x 10 2 dx3. error in x3 (min)
+ normal pu, = 0 p u ,' = 9.0 I
l l 1 9
observations were within the interval f,-13gg/ m 8sf sf. + 16yg/ m 3 . Again, the simulated confidence intervals are tighter than those given by Gauss's law of error propagation and the Chebyshev inequality. They demonstrate the assymetry of the intervals given by the lognormal approximation, but are larger than those for the lognormal at its geometric mean of 62.2 g/m 3 i l Simulation has many advantages. It is relatively inexpensive. It produces an estimate of the distribution of the errors, even when the functional relation-ship is quite complex. The investigator may choose from a wide variety of distri-butions to model the input variables. If advantageous, non-parametric or discrete distributions may be used. Recently, computer codes, such as DEMOS, which facilitate simulation analysis of propagallon of uncertaintles, have become available (Henrion and Nair,1982). The utility of simulation may be extended by coupling it with statistical data analysis. Several features of the simulated TSP data are of interest in this regard. First, the mean of the distribution of simulated measurement errors d was not zero, even though the means of each of the components of measurement error was zero. The simulated mean error was 0.43 pg/m3 . Second, by plotting the simulated error, df, against the true TSP concentration /,, we could see that measurement errors increase as TSP level increases. This plot is shown as Figure 1. Third, by regressing the squared measurement error, df a, on the l squares of each of the components of measurement error, d=f, dzf, and d:] we l were able to determine the relative importance of each of the components. The results of the ordinary least squares regression were: df r._g,$4 x go-edzf + 3.01 x 103 def + 7.35 x 10*'dzj (27) L: 0.31 2G 0. t 1 16
.e e** .. amen. m, w
m._ _...-_m.. ~._-___.-_._----____.m - . . m.__.______.m _--~.._..______.&________..._m.-- . - _ . _ _ _ l
^
I t I, . 1 l 4 e I t 4 i d 40 ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' t' a : e' ] , 't !
- a
- j a a !
l
- I 1
T 20 - a a C - ! e a as I i ' a a ya a a a j as a a aa an a a a a a 3 j *
, a sa a a a a s asases as sa ; E na na a a a %A a ..aa. .aone. aa an a saa asmostCD6C ECCSCCaa aaC as a sa 3 $ O D&De a C O as e a e as sa a a
- m e C f - . C .cD a C t a ce Ccae a asoC as as an i a
-I E a s t s* E s ( 4 0 f ( # 8 8 6 assaab ata a as b a O e i r. C e asaae aaCE E a a a a i s a a t 1 I" ameaaCaC a aa a l E - Ca a La assa a b a a E a aa a as
) *- a a
-a a a -
'l a a
- a a a a a
u - i I
-20 - .a ;
a l l t 1 - t l - 1 -40 - - - - - - - - - - - - - O 80 160 240 ( , 320 ! i s TSPTRUE (pg/m') i i 1 i i i t 3 Figure 1. Sin.ulated liigh-Volusar sampling errors as a function of trias TSP levels. (A = l obse rva t ion , is = 2 obse rv.s t i on s . . . ) 1 J 4 4
-- - _ - . - - , _ . - - _ __. - - - _ _ _ - . . , - . - ~ . .- -_v,_ --.,----.--._.m._ , , , . , _ . . _ _. _. _. . . _ _ 4 -- , , . - . _ , . . - -
- . - _ . . - _ - - . .. ._ _ . - . - - - - . - . - - - . - = - - - . -
i f R2 = 0.144 F = 281 n = 5000 I l Here the only statistically signiacant variable is dz!, the square of the error in ] now rate. This result is consistent with both of the other analyses which permit-I ted investigation of the sources of error. It might seem that the coefficients 1 i from the regression equation should be approximately equal to the squares of 1 partial derivatives used in the application of Causs's Law. Comparison of equa- ] 1 . ] tions (12) and (27) indicates that although the numbers are qualitatively con- , I sistent, they are not equal - even allowing for the parameter uncertainty in j equation (27). For example, the coedicient of dzl in equation (27) is 3.91 t 0.15 l x 103. The squared partial derivative in equation (12) is 1,71 x 103. The reason ] j for the diderence is that the squared partial derivative from equation (12) is t evaluated, and therefore exactly correct, at thp mean values of zi, ze , and za . - i
- In contrast, the coedicient from equation (27) reflects a frequency weighted-J average of the squared partial derivative evaluated at each of the simulated sets i
- of values of zt,2 2. 2s-r j The low explanatory power of equation (27) (as indicated by the R2) is due !
i to this inadequacy of the linear model, i.e., the treatment of the partial deriva-l l tives as constants when in fact they are variables. This was confirmed by per-forming the regression illustrated below: i i ! 8 ' ~*'
- df = -17.1((s es z )8dz 8] + 1.23((ziz al8dzf )-41.5(( zgza, )8dzj] (2S)
I R2= 0.903 P = 15467 n=5000 i The quantitles in parentheses are the partial desivatives of / with respect to each of the components of error. The explanatory power of the model has i i
- increased dramatically. The significance of this result is in its implication con-i j cerning the accuracy of error estimates based on a Taylor series expansion in I i I
. 17 l -
4
4 l ) 1 circumstances where the partial derivatives vary substantially. 1 I Although the simulation approach has many advantages, there are certain I limitations. The accuracy of the results is dependent upon the accuracy of the t assumptions made about the distributions of the underlying variables. If the I functional relationship between the derived quantity and the underlying vari-l ables is complex and many variables are involved the procedure may become expensive. As with experimentation, costs may be reduced substantially 1
- through careful choice of the points at which the function is evaluated. For
; example, in some cases the efficiency of the simulation may be increased by j employing stratified sampling techniques, e.g., Latin Hypercube, rather than strictly random sampling (Iman et al' (1981)).
1 i
SUMMARY
l Using one or more of the approaches outlined above in many cases it may i 4
- j be possible to derive numerical estimates of the overall random error in air pol-I' :
lution measurement systems, and in some instances to determine the relative importance of the various components of measurement error. 1 f
! The error estimates made above are compared in Table IV. The first column I
- Indicates the method used. The next two columns show the lower and upper i
j bounds on the errors as calculated by the indicated method, the bounds being i ! for 80% of the errors and for 95% of the errors. The errors themselves are given ). I i here as percentages of the mean value of f appropriate to each method (recall t j that the paired sample data were for concentraltions somewhat smaller than j those of the data used in the other comparisons.) i ) Table IV shows that the analysis using distribution theory agreed quite well } l 18 I e
, essen e se -
l
=.-._.,,,-.,.,,m-. . , . . . - . ~ . . . , . . - , - _ _ , - , , . _ - , - .,_,,-,--m,._.,m,..-..,.-.,_,,,__-.,.- . _ _ . , _ , . - . _ - . . - , - , , , , ,
. . . .. . .. ._ . -- . . . . . . ~ . . ~ . . . . . . - .
3 s t i } Table IV. Error Estimates: Percentage of Errors Expected { Within Limits Shown, as Percentage of Mean i: s Method 80% of 95% of errors errors Analytic (Chebyshev)* t 15.97. e31.67. j Distribution theory (lognormal)* -8. 7. + 9. 5 P. -13.1. +15.0% 1 Expes:rnentation (paired samples) t9.7% t 14.7 F.
! t7.07. 210,67. ! S:mulalon** -9.3. + 10.8% -20.0. +24.77.
l d Assumed mean f = 62 pg/m .3
- l. " Assumed mean f = 64.8 yg/m .
i b 4 l t r 1 i I 1 1 1 i
with the results of the paired sample experiments. The simulation resuits indi-cated a broader error distribution. The Chebyshev inequality produced what seem to be artificially wide error bounds, due to its robust and conservative assumptions (that the distribution may have any form as long as the mean and variance are each finite). The analytical solution approximation approach is simple, and permits evaluation of the relative importance of the various components of error, but gives no estimate of the distribution of errors and is in some cases limited to analysis of sources of error which are small. Application of distribution theory is attractive in that it provides an estimate of the distribution of the errors and allows evaluation of their relative importance. However, there are only certain special cases in which the theory applies. Experimentation uses paired samples and provides a good estimate of the overall measurement error and the distribu-tion of these errors. However, experimentation is likely to be expensive. Simu-lation provides estimates of the overall measurement error, the distribution of errors, and may provide insight into the relative importance of various sources of error. However, it requires knowledge of the distributions of each of the vari-ables and error terms of interest. In many circumstances sequential application of two or more of these methods may prove to be advantageous. For example, the results of a prelim-inary analysis bsed on simulation might be useful in the design of an experi-ment. Or a quick and approximate estimate obtained by Causs's Law of Error Propagation or application of distribution theory:might Indicate whether more precise results would be worth pursuing. ) i 1 19
4 .) 4 REFERENCES l
- 1. Aitchison, J. and J.A.C. Brown. The Lognormal Distribution. Cambridge r University Press. Cambridge (1957).
- 2. American Conference of Governmental Industrial Hygienists. Air Sampling
- instruments for Evaluation of Atmospheric Contaminants (5th
- Ed). A.C.G.I.H., Cincinnati Ohio (1978).
j 3. Box, G.E.P., W.G. Hunter and J.S. Hunter. Statistics for Emperimenters. John Wiley and Sons. New York (1978).
- 4. Burington, R.S. and D.C. May. Handbook of Probability and Statistics tuith
! Tables (2nd Ed.). McGraw-Hill: New York (1970).
i i S. Cochran. W.G. and G.M. Cox. Experimental Designs, John Wiley and Sons, 1 Netu York,1957.
- 6. Fein. R.S. and B.S. Bailey. "Conductimetric and Pararososanaline Method SO2 M nit ring Uncertainties and Their Significance." Env. Sci. Tech. 12:4 463 46 (1978).
- 7. Harvard School of Public Health. Six Cities Air Pollution and Health Effects t
1 Study. Air Quality Group. Quality Assurance Manual. Boston,1981. I
- 8. Henrion. M. and I. Nair. " DEMOS User's Manual - Version !!." Department of f Engineering and Public Policy. Carnegic-Mellon University. Pittsburgh, PA.
4 1982. l l
- 9. Iman R.L., Davenport. J.M. and D.K. Ziegler. Latin Hypercube Sampling l l i (Program User's Guide). Sandia Laboratories Report SAND 791473 l 20
- _ -.- - = . . ._
i l . j Albuquerque,1980. 1
- 10. Jaklevic, J.M., R.C. Gatti, F.S. Goulding, and B.W. Loo, "A p-Gauge Method Applied to Aerosol Samples," Env. Sci. Tech. 15
- 6 (1981). ,
- 11. Leidel, N.A. et al., Occupational Exposure Sampilng Strategy Manual.
NationalInstitute for Occupational Safety and Health. Publication No. DHEW (NIOSH) 77-173, Cincinnati, OH (1977). 2 j 12. McKee, H.C., R.E. Childers. O. Saenz, Jr., T.W. Stanley, and J.H. Margeson, i 1 " Collaborative Testing of Methods to Measure Air Pollutants,1. The High- , Volume Method for Suspended Particulate Matter " /. Air Pollut. Control 1 ~ Assoc. 22:5. 342-351 (1972). I
- 13. Seiler, F.A., " Error Propagation for Large Errors Parts I and II," inhalation 3 Toxicology Research Institute, Albuquerque, N.M., (1982).
) 14. Shaeffer D.L., A Model Evaluation Methodology Applicable to Environmental J Assessment Models, Oak Ridge National Laboratory, Publication No. ORNL-
] 507, Oak Ridge, Tennessee (1979).
s t , 15. Shapiro, S.S. and M.B. Wilk. "An Analysis of Variance Test for Normality i (Complete Samples)," Biometrike. 52:91-611 (1965). 4
- 16. Stephens, M.A., "Use of the Kcimagorov-Smirnov, Cramer-Von Mises and
{ Related Staus.ics Without Extensive Tables," J. American Statistical Assoct-1 l ation. 69:730 (1974). . e - l 17. USEPA, Quality Control Practices in Processing Air Pollution Samples, ; i j USEPA Pub. No. APTD-1132 (1973). ; i ! 21
,n -,.-,,-.n . , . - - - , . .-.,, ,, -.,--, ,-,.,,,--.n,.- .
.,--_.-n.--, .,---,-._--.---.----,--~~,,..,,,-.----,n,, -
1 (CCffM _.,_.mn.._.. . . . - . i.~-..~ o i. .. . -
, CYCLONE DESIGN: SENSITIVITY. ELASTICITY AND ERROR ANALYSES Dotcus W. COOPER Enuronmental Heahn Sciences Department. Harsard S6hool of Pubbe Health. Boston m 02115. ll S A iheurd tar punwwm :9 Juh ino stasraes-The citeetneness fator. a = { -in epeneirationi) iprenure dror* is a usef ul ngure of merit for om g unione design oprimization. remaining the ume for a nstem of series parailes elemenn n it is for 'he indiudual elements, auuming thes act independentis The partui densaines of a thrwtion mien respect to its
$4riabies can be used ta determine its senutnetv. the change in the runction per unit change m the saruole .
~
elasticity. the tractional 6hange in the function per unit trxtional change in the sarubie; and error propasauon.thecontribution of the saruble's hoance to the function satunst An anaissn of the senutiutv. clatacity and error propaeation of exh of the mdependent sartables in the cMione effectneness tutor
- ndutes adsantages for dessens that use an esembis es high.e:Testnenen extor elements uprated as lower
' tom rates and or lo*er prewure drops than for those deugns that use a unele odone s
1%TRODL CTION - This artic!e presents an analysis ot' eyclone pert'orm- 1 ance equations for sensitnety, elasticity and error 3 propagation. Such analpes are useful in esploring optimal and near optimal designs- 2 One formulation of the cyclone design problem for a particular appiscation is to mmimrie penetration within certain constramts. such as total cost or system prenure drop Penetration n the tatto of the number of particles per unit time flowmg tro n the control :
)
deuce outlet to the number per umt t.me entering at its iniet. Similarly, penetration can be set at a deured lesel, p vi and the deugn adjusted to minimize tost or pressure drop Formulated in these terms. this is a constrained , optimization problem. typica!!) more dithcult to solse
* * " ~ "
than the unconstramed problem of masimitmg or ! 2 minir1 sting some ngure of mertt. A figure of merit that umphties the optimitation problem and has uwtui quahbes is the "e!Tectneness factor"ivt the negatne i logarithm of the penetration, multiphed by the re. - g ciprocal of the pressure drop JP acrou the cyclone ICeop-'.lWlt. ' i ' 4 = [-IniPnij JP i11 2 l in which the penetration n the penetration of a 2 a part Aular particle sue. Pit edg The ettectneness astor i i n an intenuve rather than estensne sariable ard h.as units whn.h are the interse of those of SP lt can be I 's l Elfestnenew txiors evi sor ,eoes paraitei ucwed as the solume of gas cleaned completely per ""'""- unit of energy in units such as m8 J ' ' il AP in in Nm J A sptem oi sencralidenocal dewes actmy m hiter's mherent ethca6y for part Autate remosal in series and or in parallel and acting mdependentiv comparison to its prenure drop t t ' 5 A K IHih The would hJse.: specm value ut ei. the sptem s *eflectne. effectneness fastor has recentiv been apphed to sw nos f actor." w hich would be lhe same as the salue f or lones 4 Cooper. lW14 and to the indnadual collecting
+:atn indnuiual deme. smte the demei pressure elements m a packed bed (Coope:r 1%Li and a nlter drops are additne and their penetrations multiplu.a. ICooper. 19Cbt Other thmus being equal.11 is toe. for each particle use t'>ee Feg la Thn ngure ut adsantageous in terms of eihtiency and or prewure mer:1 hJs been used in f:ltraison werk as a meJsure of a drop to use a wries of high..t dewes rather than a
, ; MS q.s h r: 1
.. _ . ~ =_ ____
-m m _m____ __m
,1 1
4%
- Dot iAAs % CooPsa i
single lower.q deuce (Cooper.19'6; Dirgo.1981, loganthms of the sanables- ! personal comma It is also adsantageous to use a set of such deuces m parallel. creatmg a system haung the 2 same q as each indnidual deuce. Maximizmg q means and perfor ms a multiple imear regression tot log tfion mmimizmg penetratton at rised pressure drop or togIxiand login The slopes of the regression b and c. minimizing pressure drop at fixed penetration. are the clasticities. the fractional changes m! due to i Identifying the maximal 4 is useful for systems for unit fractional changes m x or y: w hich pressure drop is an important cost factor or is a constramt. b = dlog(fhdlog(x) = idf f) idx xi ch Once the conditions which maximize q are found, c = dlogif) dlogiy) = tdf f) idt n 466 the penetration for a gnen particle size is simply: Error propaganon Pntd,) = exp{-qid, lip] Cl The vanance off. varifL can be obtamed from the m whach the particle size dependence has been noted vanance of x: the partial dernatne off with respect to explicitly. x.f.; the vanance of g and the partial derivative of f with respect to y. for small varunces with no ELASTICITY.sENsITl%IT)' OD ERROR AN AL)$ES correlation between the variances: Sensitn ity. elasticity and error analyses are outhned sarif) = ef,)harex&(f,)2 var 1 0 - H.O J.
~ 17) -
bnetly nest.after w hich they are apphed to a standard in which H.07. stands for higher order terms. The cyclone model usmg the etTectiseness factor. standard deviation is just the square root of the I i Gisen a function of several sanables, such asf (x. y L vanance. Rather than choosing a smgle pair of x and y one wants to know the erTect onfofchanges in x or y or around which to calculate the partial dernatnes. one perhaps the contnbution to errors m/due to errors in can calculate manyf.x,y triplets. then regress f on x - x or y. The answers can be determmed by sensitauty or and y. The regression coetTicients are the sensiteury elasticity analysis and error analysis. respectively. factors, but the regressions also gne the fractional The strengths and weaknesses of error and sensi. contnbutions of each of the variables to the sanance. l tiuty analysis hase been presented by Gardner er d which is the error analysis. 3 fl93IL who argued for error analysis as being more generally apphcable and more relevant. especially w hen the question is the allocation of scarce resources APPLICATION to impros e the estimates of parameters m models. This would apply to research on the improsement of O'84"#4"8" " cyclone designs. - The nenetration of a cyclone can be calculated using the fumulas presented by Leith and Licht I197:L and 4 N""""'T evaluated for many conditions by Leith and Mehta The sensitivity offix. y) to small changes in x and y 11973), who showed excellent agreement between this _, %w is the partial denvatne off with respect to x or y. theory and measured values for two well. defined Euluatmg such a derivative means choosmg an x,y expenmental insestigations. The use of the pair around which to vary x and y..The choice is not Leith-Licht model for optimization was presented in trnial and a ddTerent choice may produce substan. this journal (Leith and Mehta.1973), along with an i tially ditTerent partial denvatives. To get more rep. improved design derived from the model. There is ' resentatne salues of the partial denvatnes. one can some question. however.about the apphcability of the 6 7 choose many x.y pairs.evaluatefix.y) for those pairs model at high temperatures (Parker et at.199tt We and then regress f on x and y by multiple linear wdl perform our evaluations for typical atmosphene regression. The plane thus formed has as its temperatures. j i slopes the "best tit" values of the rate of change of tic formula for the penetration of particles et f lx. y) with respect to x and y over the range studied. If diameter 4, and density p, that are carned m a gas of i f(x 3p has substantial curvature, this estimate may be density 6 and uscosity p,is quite misleading, however. obscunng the difference in the relationships off and x and off and y in vanous g, gi a. 2,), ig, 4 domams. The sensitivity analysis does show shether in which T is .in mettial parameter f typically increases or decreases with x and with y. 9 , c. g g ; Ud'"C"V t C, is the Cunnmgham ali correction factor; n is the ' For correlatmg data which vary oser orders of sortes exponent: D is the d ameter for the cyclone (see magnitude. It is useful to use power functions. such as Fig. 2L and r is the inlet ve!ocity. calculated by dn iding
!IX. H = ax* f t .11 the solume flow rate by the height and width of the cyclone miet. : = Q ab. The sortes esponent was f
m3tead of linear functions. One transforms to the found empirica!!y to be proportional to the cyclone a d~s yfoY"Yhf l 1d .* 7 . 1
- - - , - - _ - - - _ _ . - - - ~ - - - - . ~ . -- ~ ~.
i Cgune acwn ,enutiun. einuem and error anaires v l i i
- c - ~ ce- shommg the strong effect of solume :fou rate and g cyclone duct eut diameter on pressure drop. along f
with the mtluence of the mlet area tam , e j; 3-- e s l
, Edurneness nctor .cnututtrws. ela<trarws
, The etfectneness factor can be put mto a power law
- 1. ! : =ea form as N../ q = 2iCPl' d"%bDf t 3Q2p. il4
- s. ,~~~"%. This form suggests that it wdl be most comement to i
J
' determme elasticity. f rom which the partul dernatises } .
y can be obtained easi!>.as can their squares The pow er. law form exponents for sescral sariables are shown
*s - e.w below for which the Cunningham shp correction has been neglected for particles of the sizes one typically j controls by cyclones: ,
saruble espotent
, , p, 1 12n + 2)
J, 2 m . 2,
~~-J p, - l 1:n + 2)
Fig. 2. Cyclone, with &mennons #* -I iLeith and Lu:ht.1972. Leith and 0 - 2
- I C"
- 23 Nienta.197h For a typical salue of n. n = 067,1 Cn + 2n = 03. The dumeter raised to the 0.14 power and is 067 for D
= I m at 233 K:II - niis proportional to the absolute esponents g.
show the percentage changes m y espected temperature raised to the 0 3 power i Alexander.194% rom a one-per cent change in the sarubles As the gas The factor C is a dimensionless combmation of solume rate of riow mcreases by I per cent. the cyclone dimension ratios: etTectoeness factor q wdl decrease by 1.7 per cent. The decreased penetration is more than offset by increased o I C = 4:tD' abl[2t1 -D;2 ifs'- a 21 - 6 + Z- - pressure drop. Thus, there is an mherent adsantage in ) - h l +J' .J'2t 3. (F- Z'D;2 - S1] aloi using several cyclones m a parallel arrangement. lettmg the tiow be dnided among them. This is the rationale in which the primes indicate the quantities hase been behind the "multicione'. The factors within C cannot , % =.w.m y,,.g made dimensionless by druding them by D:i = a D. has e their influences seen so readily, so we performed a
, F = h D. etc. Z' is an estimate of the distance the series of computer simulations for the conditions sortet extends below the gas cut duct IAlexander. shown in Table 1. hasmg chosen particles Spm in j
g l949t diamster hatng the density of water and penetrations i Z' = 2.3DHD2.obl82 111)
" 'I*'# "" * ' "I*
1 Table I shows the base conditions for three cyclonts: J and J is the cone diameter at the length Z l Leith.1979t one whach we " designed" usmg the optimization i These equations describmg cyclone penetration procedure of teith and Mehta l1973)and two standard made it ditlicult to mfer. without esaluating them. how designs,a high etTiciency design and a general purpose penetration will sary w hen one or more of the cyclone design taken from the monograph by Leith a 1979: The ' l i , dimensions is sarted.- rows hst the sariables studied: the dimensions of the , The equation for predicting pressare drop chosen by ~ eyelone, the particle and the gas variables. The last hs e
- Leith and Mehta (1973) was that deseloped by rows show the derned quantities C. In Pn. Pn. AP and Shepherd and Lapple il940t iv!The clasticities were calculated by talmg the base
~ ## "#* # "#*" E #'# '" '"'" D ' P" ##"
l AP=.Klah>Dbig,rp2 2 s ~. i l 2) - returnmg at to its ortgmal salue after the mcream. w hich Leith and Mehta found to gne almost as good Seseral concluuons can be drawn from the results agreement with esperimental results as did several presented in Table 1: ' other more comphcated etpressams.The salue for K is ' i Il The cyclone diameter is the saruble to w hich t he 16 for a cyclone with a standard tangential mlet etrectiseness factor is most sensitne. l iShepherd and Lapple. 1940t so this equation (2) The ratio D, D ts one to a hoch the deugn is also becomes: quite senutne, a characteristic esploited by the AP = %Q 2shDJ t136 Leith-Mehta design.
- i t
~
i .
- t r
y + ~mm, =.,- .-----.,--mm- p -s .~r-w- .
%-7m- m-m-----w-
.. . . .- - ._- - - ..~ _ .- _ _ . . c h
~~
I 443
, Doucs W Courrii TaNe 1, Base cases and elasucaties for three cyclone designs VaruNe Leith-\tctita ! or Stairmand Perry -L4 pie design migh etticiency ageneral deugni
- parameter Dimension Base Elast. Rase Elast Baw Elast.
D m 1 43 t 45 t uD - 0 43 14 0.42 05 0 31 0.! hD - at? 0 45 02 0.32 , D, D - 0.38 0 23 0.38 O bit 1.07 05
' SD -
t.81 05 4.83
# L.2 0.Os 0.5 4D 0.08 062$ 0 09 30 0.21 1.5 0 11 2.0 HD -
5.0 0 24 40 0.13 80 - 0.11 40 0 07 4 373 0.06 0 375 OM J, m 5=10** 425 act v, 0.60 5=10** 0 60 $ = 10' 0 60 kgm*8 1000 0 30 1000
._A 0.30 1000 0.30
% Nsm-2 1.8 = 10 O.30 1.8 = 10- 5 0,30 1.8 a 10-* Q30 6 kg m
- 8 1.2 - 0 99 1.2 - Q99 4
Q m' s *
- 1.2 - 0.99 1.78 -1.68 1.54 - 1.68 f.91 - l68
, C -
91 6 !$.1
)
.. -Ini PC 50 4 t39 1.04 1.01 h -
0 23 G3$ a36
.1P N m- 397 9:4 ggg$
4 m3 J l $.3 = 10 *
- 18 3 = 10-* 9 N = 10 ' *
- i i3) The ratios B,0 and 5,D do not hase much root of the relatne sanance is just the standard mtluence on the etTectneness factor. ,
devution divided by the mean. often referred to as the 146 The relatively large salues of H< D and hrD have coetitetent of sanation'. One can demonstrate that helnd the Leith4fehta design. as they noted. and the relatne vanance of a function which is of the hase mcreased the senntnity of the etTectn eness factor to these parameters. power law form will be' the sum of the products of' 4 the retatne vanances of each of the parameters and the
- 15) The gas solume t!ow rate and the gas density are 4
clasticities squared. For the example /u. p = ax*y'. s sanables to which the design is quite scrisative. ~ (6) The etTectneness factors for these I meter- **M' U " ** "WM,' diameter cyclones were such that the Leith-Lacht Thus. a 20 per cent uncertamty in solume flow rate design was supenor to the high.etriciency design. would cause y to have a coctTicient of vanation of which was supenor to the general-purpose deugn. about 20 x 1.7 = 34 per cent. The uncertainties of the This analyus suggests advantages for a senes/ parallel assembly of cyclones. each of which has a other vanables can be estimated and combmed as j indicated in Equations (?) and (166. smaller tiow through it than would a smgle device. lSee Appended to such error estimations should be our Fig.1.5 Recall that the q value for an assembly of
* * * "
- identical cyclones is the same as the indnidual g values. subjectne evalu.ation of the confidence w For example. if we replace the cyclone with twoaccurate the values from such thcones, even gnen completely mput data.
esciones identical to the nrst and arranged m parallel, j thus dividing the flow into two equal halves, the elasticity salue for flow of -1.7 means that the g cosCLt.510%5 ' effectiveness factor mcreases by 2 = 3.25. Since ' i the pressure drop decreases by 1,28 = 1,4. due to the The use of sensitmty, elasticity and error analyses decrease m flow rate through an indmdual umt. up to can aid in understandmg and desenbmg the behastor four of these cyclones could be used m senes in each of devices used in air pollution control, such as ,
~
- tiow path. If the ongmal pressure drop is not to be cyclones. These analyses can aid in optiminng and m 4 exceeded, an assembly of two parallel flow paths each estimating the impacts of changes or uncertamties m havmg two such cyclones in senes would have half the the vanables.
ongmal pressure drop (aP.) and allow less penet. The foregoing analysis of cyclones demonstrated t he ration than the smgle unit (Pao l: importance of cyclone diameter, gas tiow rate and I-it Pal /i-In Pn.1 = (q/q.llAP/dPo l density. and the cyclone dimennon ratio D,, D on the
! behavior of cyclones; it showed the relatne lack of
= 13 25Hid) = 1.62. 115) importance of some of the other cyclone dimension ratios.
Thus. Pn = Pn, i .a. If Pn = 0.5. then Pn = 0.33. A wt of cyclones m series, parallel combinations. rather than a ungle cyclone, theoretically can provide E//urirenm factor error anulna e ! perf ormance superior to the optimsied ungle. cyclone Denne the relative vanance off as var (f)/ga, the design and the clasticities or sensitnities can and in , variance divided by Ihe square of the mean. The square makmg design dectuons. Performmg such analyses J W'Wy?,-91
- i
.. .. x: l
( sene deugn senutnits, cia,ticity and error anaiy.es 439 should be preteratue to se:ecting only a standard Cooper D W il9s2bi Optimame niter nber dumcer design or to ca sulating an optimal Leith-\1ehta a r mo eni ni E m er-a,*rne 16. t ?29 i s.u doien for escry series paradei connguration con- C'ardner R H . O Neid R V. \tankin J 8, Carnes J H sidered. ilMll \ comparison et wnutiuts anaism and error
,ng,y, ,,.ised on a stream esoissiem model E. . .
\l..d 4hm,12.I't 190
- 4. k et.n...su. ment - Th:3 ork has berented trom ms dneus- Leith D 819795 Csdones in Hanah .9 Em in nm. ens,
. ions with soileagues Dr Daud Leith and John A Dirgo Ens.ne.nna iEdaed by Pereira N C and Wang L A i Humana Press. Chtten. N J Leith D and Licht w' il972:Cosiestion enheieno of cwtone R EF E R E M ES type particle sollectors. a new t heoretical approacn 4 I CA E. Ss mp kne. Air -19 11
\letander R \tsK. il949 Fundamentals of cyclone design Leith D and \lehri D ilv7h Cgione perform.ince and and operanon Pem. Au.arauw Inst Mm. \letou iNew design timmenen. Em un nment 7. ! ' ?49 Nrmt152-L 20t 223 Parker R., Tacn R , Cahert 5. Drenmel D and Aboott J C u per D % 194 Theoreti6a1 comparison ot erhetencs il9tle Particle collection in cwlones at hign temperature
-+*%*rG ^ -4 and power for ung:e-stage and muaiple stage particulate and hign pressure. Ent er S.: Tri amd 15. 451 45M stubbing. timmcor a Entironment 10.tont-lina e
shepnerd C B and Lappie C E. in9 ash Flow pattern and Cooper D W' i1941i \1memsung e> clone energy consump. pressure drop in cyclone dust co6iectors in.I Eriunu Ch.-m rion / iar Finli,e C.,ntn9 4u 31.1893 1104 32.124Al244 Cooper D W' t19*2a Filter Ndr energy-erhcient p.iding US Atomic E nergy Commnsson elv50i HundAmA . .n lumeter 1 or PWIus C.,*rned 4 32. 205 204 Irn.wl . w'ashington D C e
r 'df OM* g M e
t M * *
- e
-g;1.%
~
e
~ ~~
(QQ h y ( ,- l
;}
3 i sri sD) Ri rTILINE su sto r ON 49 ( ', o
- the weteht losumciem auennon to these considerations sometimes ! cads to erroneous l
sonclusmns. Thus to explain the shape of the path of a stream of tobacco smoke f the medium injected horizontally into a smoke chamber Prosad [84j. starting from the assumption
; ]
that smoke particles settle mdiudually, had to take the absurd value of 249 for the f '{ l i mean paruele radius. In fact the shape of the path was undoubtedly determined by the +
't.6) rate of settline of the stream as a whole because its density exceeded that of air on k account of the' carbon dioude contained in it. A striking e[ ample of rapid settling of a cloud is furnished by the " tire cloud" w hich descended with tremendous speed from the volcano Alont Pelee in 1902 and turned the town of Saint Pierre into cinders. p ,
I I .O Esidently the concentration of the disperse phase (volcanic ash etc.) was so great that j ! the density of the cloud. despite its high temperature, was much higher than that of y , air. lM ' If Stokeg shole etoud. A sery complicated system of mosement exists in cumulo-nimbus clouds contain-ing droplets of all sizes from r = 10p to r = 2-3 mm. In this case. under the 3 l j mrluence of the higher temperature of the cloud in comparison with the surrounding
}
i s air, a rapid rise of the whole cloud takes place at a rate up to 10 m sec-L while the ; y. l t13 4 drops of water in it are falling individually at speeds between 0-01 and 8-9 m sec-l. j 4 The resultant selocity of some droplets is therefore directed upwards while others t h ,
, gg, mose downwards. These phenomena play an essential role in the process of precipita- W uor. from c!ouds (see page 319).
i
; least in the ['I '
- r, when the la. THE NtOTION OF AN AEROSOL IN A CONFINED SPACE tional to the j acts Rn and For aerosols in an enclosure the motion of the particles includes that of the medium e
, . eases; caused by convection currents, artificial agitation, etc., as well as their own motion ! @
relatis e to the medium. Just now we are interested only in the latter, and shall examine ; C! I o ic$n. and
~
it f r particles settling under gravity. If the partic!cs of an aerosol occupying a space ] i
.h s inside and e nfmed by walls settle with a velocity V the medium moves in the reserse direction 5, and hence, with a mean velocity y V, where 5 is usually a very small fraction of the total volume of the disperse phase. Since the qiedium is entrained in the vicinity of the particles, .,
)
greater than i then in the spaces between them the selocity of the countertlow is greater then y V. der i gm m; fhus the rate of settling of partic!cs in the present case, unlike the monon of a free . doud, is less than that ofisolated particles in an infmite volume by the factor ! + x7 re ditTerence ! ate humidity b 1
- fore plays a' Acc rding to Cunningham [46] still another factor should be taken into account; g j
l ich is mainly in ihe derivation of Stokes' formula one of the boundary conditions is that the velocity ; j inside ami of the medium is zero at an infinitely great distance from a partic!c. When a cloud of a
- parueles settles in a contined space. however, the velocity of the medium is zero at a
- l istion of fuel ^tance e fr m the centre f a partic!c,where 2 e = n~' 'is the mean distance between
( I th the aid of "hacent particles. Thus each particle experiences the same resistance which it would , ; ip ! s. and so on. **nence at the eentre of a closed spherical vessel of radius e. According to Cunning-
" l
}
> ecol before 'im s calculations this resistance, on s Stokes approsimation, is equal to 6:Tr Vy i 6 l ration of the i! - 125r e). FollowingOscen,the cc.rection becomesless the greater the Reynolds j es the cloud
""" bE. All the authors occupied with this problem hase arrived by way of }
because the "'I) C mplicated, but not rigorous, considerations at a correction factor of I + xg- q k the 'Hh salues f x equ i t 5 5 [85], 7 0 [83l and 4 5 [86]. Rigorous solution of the ,) P'oblem is obviously extremely difficult. , I i l l l
ll
; t I
1 50 f o riti ur cu ssic s or sinosots fhe dnTerence 5ctueen correction f actors of the tspe 1 { - x9 til and I + x r n -
'C'" " "'
y w I - n tillis ofimportance because at the usual wiues of y in acrosok the factor ,
-t1 - v;i j a h n practically equal to I while the tactor til) may be a few per cent greater than I.b Wesdo j For the small values of , which are of interest thn problem has been insestigated .
t peer - { ; espenmentally only by Kermak (85) who measured the rate of settlement m monodis. De prm
" f , is consec-j perse suspensions of sartous ammal crythrocytes with radii 2-4, 3 0. 3 7 and 4-4p in water. It turned out that for y 004-008
- , the esperimental results agree well of r - b5!
i with a correction factor i + xy and x has values lying m the range 4 8-6-9 for ' des a . ! sanous ersthrocytes. Unfortunately tSe setthng rate of isolated particles was not eo ftietent c i ]
- measured in this work but was determs'ned by means of extrapolation. aerosob n t i j Thus from the rather scanty data ausble it may only be said that in the I settling tion. IJnfort j of acrosok in a contined space the resi'.tance of the medtum at low y n probably equal to 6.7r V., t 1 4-a hr;h enou
! x9 ) and x is eluse to 5 or 6. is necessar? ' { The rate of setthng of concentrated suspensions has become important recently therefore, il l j m connection with the fluidization of powders hee page 367). In the fluidized state { a concentration of particles for which the settling rate n equal to the llow velocity is { I automatically established. Experiments on fluidization hase led to the formula q15. \1( I 1 V/ = V,( 1 - y l', i (14.1) l where The mot i isoiatedV,' n the setthne rate of the entire system of particles and V, is that of an particle. motion in tl l tield is q E s' ; For sphencal partic!cs Lewn and Bowerman (87] and Richardson and Zaki(88] obtamed tlw same value 4 65 for the coefficient s. An approsimate theoretical calcula. of the partis { j tion of the settling rate was made by Richardson and Zaki who started from two ; models for the distnbution of spheres in space; they obtained two curves (V/, y) ! ! i one of which lies about 40 per cent higner and the other 20 per cent lower than the ne mos f 5 expenmental curse. Pf3CUC3I
'j '
{ In conclusion a phenomenon will be mentioned which is familiar to everyoneearth's gras i working with acrosols. When concentrated aerosols settle the upper boundary(90)and is Eh i l j
; usually flat and horizontal, a phenomenon which is exhibited both in the laboratory has played :
and in natural mists. The explanation is that, for an aerosol density exceeding that ACf05 of I the gas adjoining it, hydrostatic forces counteract any disturbance of the horizontal ser plates ar ; position of the upper boundary of the aerosol by consection.just as in liquids. Suchbservation 7 stabilization of the upper surface will be observed only when he particles move as aahonzonta l precedingwith whole the riedium, which necessitates a sufficiently high concentration (see where // is ( section). i The strengt ! l, The surface of acrosols dispersed in dense gases like chlorine or carbon dioxide of a panid etc., is particularly stable (89). ! ( sunultanea Many theoretical and experimental papers hase been devoted lately to the sedimen. dCPC"di"8 C t tation of particles in a limited space, or hmdered settling. Only equations which refer '"'C""t> b to scry small values of the volume fraction of the disperse phase, y,(the fraction of determmed the total volume which is filled by the disperse phase) will be given here. Following Cunmngham's idea (see p. 49), but allowing for backwards flow, IIsppel (610] and Kuwobara (611] obtamed at y - O the formula i V//V, = 1 - xp insome I wnh x = l 5 (610] and I 62 (6t t]. Brinkman (612] deduced the formula I , V//V both above i = 1 - 21 y and Hawksley (613] V//V, = 1 - 74 5 . Experiments (614l confirmed **"'C2I NCh the expression V//V, = 1 ! 9 . for almost isodisperse liquid suspensions The . followmg p 1 , 4 l
! ; I t
l I ' t 1
l I ,. { l i T I' s D Y R I t. r l L i N t. T R s OTION SI ('l g 3 g , , ,. ,, j results of all other msestigations can be esprewed by the formula I;' V, = 1 - x9 il , i crosoh the taetor = ( 1 + :<r r' with x = 4 0 [615. 616). 4 5 [617l and 5 4 (618l. The conclusion (see p. 50) j nt greater than I. that the selocity of hindered setthng at small concentrations depends on T raised to the l heen msestigated tirst power seems to be confirmed but no theoretical basis for this is apparent. rnent in monodn- The pnncipal difticuh> encountered in precision measurements of V// V, at small 10, 3-7 and 4 4 " l
> . is consection. Wikon [619] using scry dilute aqueous suspensions of glass spheres - <
results agree well lif r = l-5n. found that it was impossible to obtain stnctly scrtical trajectories of ! range 4 8-6 9 for particles at room temperature although they were reshzed at 4'C. when the thermal I l particles was not coefticient of water was equal to zero. Only in more concentrated suspensions and
""* acrosols is the downward gradient of concentration high enough to suppress convec.
hat in the setth.ng l
, tion. Unfortunately, this is of ten ignored. It is difficult to combine the two conditions,
)w 7 is probably l a hich enouch weight concentration and a low enough particle concentration. which n
j n'ecessary' for the neclect of coagulation in not serv coarse aerosols. It se q uportant recently therefore. that much sedimentation analysis of aerosols is erroneous. I he fluidized state
- e tiow velocity is j ;
- he formula 4 !!. MOTION OF PARTICLES IN VERTIC AL AND HORIZONTAL ELECTRIC FIELDS. PR ACTICAL APPLICATIONS t 14.11 ,
i i V, is that of an The motion of aerosol particles in an electnc field is no different in principle from ! motion in the earth's gravitational tield. The force acting on a particle in an electnc j j tic:d is qE where q is the charge on the particle and E the field strength. The velocity l on and Zaki[841 of the parti !c given by formula (8.2) is ;
! icoretical calcula-
/
I started from two Et"VEB = q I o curves (V/1) ! + A h 6:rrit. (15.1) , ! 7t lower than the The movement of particles in a vertical field is very interesting on account of the i I l 4 q , practical advantage obtained by the electric field being supenmposed upon the y
! iliar to everyone earth's gravitational field. The vertical electric tield method developed by Millikan !
I sper boundary is [90] and Ehrenhaft [91)is one of the most fruitful methods of studying aerosols and } in the laboratory has played a very large role in advancing knowledge in this field. b] I etceeding that of Aerosel partiefes are introduced into a chamber formed by two horizontal conden-i j of the honzontal j s in liquids. Such ser plates and having side wallt ofinsulating material provided with windows for the '}, observation, illumination and charging of the particles. Observations are made with I
.rticles move as a a honzontal microscope haVng an eyepiece graticule. The field strength E = Tilh. [
.t
)ncentration (see where /7 is the potential ditierence and h the distance between the condenser plates.
The strength and sense of the electric tield can be vaned as desired. The rate of fall
' r carbon dioxide of a particle V, is determined first with the field switched off and then under the '
simultaneous influence of the electric and gravitational i f elds. V, - V, or V, - V, ' ly to the sedimen, depending on the sense of the electric field. Hence V,is found. In addition, the field
.tions which refer i 1
I
. (the fraction of
'niensity E, w hich exactly balances the gravitationa! force on the particle is sometimes determined j.
I here. J backwards flow. E, = mg/q = y :rr ; giq. (15.21 t l '// V, = l - x 51 ,. in some chambers provision is made for vary ng the pressure between wide limits formula V/iV, Y ' both above and below atmosphenc preurc. The technique of working with the
. [614] confirmed Wnical licid method has been well set out in the literature [53. 92l: it permits the uspensions. The loHowing pr bicms to be solsed. %l i
i r i
?
- I i' l
_ _ _ . _ ( l
- - . - _ - _ _ _ - - ~ - ~ - . , _
.- .-_ = _ _ _ _ - - _ , - - . _ . _ _ -
( Ar n-o i. HERSERT S. ISBIN
- a. , . .o~ r. . . . ~ .
sr. Lows em, us ss4is !
- i. i ai ao... a I
i \ l September 5, 1983 ' To: Mel Silberberg ' , From Herb Isbin S.M Vithin h past month, a variety of accounts has been nade available with l the implication that the present NRC Source Terms will be shown to be much too 1 conservative. I am referring to h following publications as examples:
; ZPRIJ, July /Aug1983(ReactorAccidents: AGlobalRamaccassantcfConseqcw.oc) j ANS News, Aug 1983 (expectations of the ANS and IDCOR studies)
, Cham Engr Prog, Aug 1983 (R. C. Yogol's Robert E. Vilsen Award I4cture) , , Congressional Research Servios Review, May 1983 (Potentisl for Reducing ' W Prediated Consequences of Utzolear Pcwor Plant Aooidents) j NtielearSafety, pay / June,1963(LVRSafetyRccearchatEPHI: AnUpdr.tc) 1 AIF INF0, Aug 1983 (reference to and enolosure of EPHI J article; reference to the NRC draft report on h Prioe-Anderson Act) Additionally, the Sept 26-29, 1983, ANS Topical Moeting and h Cet 30-Nov 3,1983:. ANS Winter Meeting will include discussions of the source torne. I am planning to sitend the Topical Meeting. All inputs to your Office are to be received l' without bias and must be judged with the standards you are establishing. You have set goals for improving the data base and this is being aceospanied by a program for verification and validation of models. The need and i=portanoe i of sensitivity analyses have been rooognised. (However, the oral presentations ) at the July 28-29, 1983, Peer Review Group Meeting did not match my expectations.) ! The methods and criteria you have established for reosiving peer and advisory inpute are more than adequate; however the difficult assignment is what you do with the inputs to)chtsin prompt and effective resolutions. I nes the elements for the bookkeeping, but I do not know how you are handling the technical arrangements for h resolution and effective incorporation of any input on soonarios, data base, and modelling. You have established important and parallel efforts to sugreat h sotivitica through the use of ad hoo study groups, such as the one for evaluating moder end location of oontainment failures. Key elemente are the participants you han invited, the assignments being given, and the leaderchip being provided. Reenite from these groups bare yet to be made available. - ! I expoet the tempo and challenges of the inputs to inteneify, placing ! increasing demands on your Office to receive and evaluate improvements in the { data base, modelling, and in the evaluation of the course of portnisted nooident:. [ T(/ G { 3'WD l
2 HERBERT S. ISBIN
. asis uourune,... ,
ST. LOUIS P ARK. NN 554 4 6
.. i ai sao- .i 7 The concluding thought is that although you have a well conceived program, increased ittention must be given to the thorough, competent, and critical evaluation of the many inpute you have already received and will be receiving, the resulting impact on the BCL studies, the establishment of an improved date, base, and the assuranse that you are developing the proper analytical tools for evaluating oceseq2ences.
~
l e
VMitWMM
*if l'NIVEllSI'IT ()F gal.lFOltNIA. LOS ANGEL.ES UCLA p zk
. ........ ...... s.....,
....s. .... ......... , , , . . . . . . ..s...s...... ..s........, ,,s......
5.mb ( f 'If OYJ/ k
,- 6fG m_cnasms un sintctents norurmar sJ"Y 3 sonxn. .w esmstunna nu 4rrurn scies.rr i.o% ANr.RE$. cAUFOnm *.m2s V -
August 8, 1983 Dr. Melvin Silberberg Accident Source Term Program Office Office of Nuclear Regulatory Commission Washington, DC 20555 Cear Mel: I have reviewed the handouts and have read the material given to us at the Peer Review on July 28 and 29th. Before con .iting on the work of the Accident Source Term Program Office (ASTP0) and its contractors, I would like to make the following general comments. I was, and still am, particularly dismayed at the attitude taken by the IDCOR program, as voiced by Ed Fuller. ASTP0 has gone to great lengths in assuring a free and open dialogue during the course of its work. Progress, ' review, and criticism have been carried out in full public view. To announce that the IDCOR results will be completely different, without giving any details as to why, (except in some general vague manner), is inappropriate technically l and does not serve in the public interest. I expect that on some given day, IDCOR will deliver the results of two years work to NRC's doorstep and request instant evaluation and comparison with ASTP0 conclusions. If indeed their results are radically different, ample time and effort must be made available for review of IDCOR results, and recon-ciliation of any major differences. The Commission should be made aware of this situation, and every effort be made to move the IDCOR program into the public domain for appropriate review. 1
)['h
( e ()g- i
= - -
j -
- _ = _. .-- .- - _ _ . - _- . - - . _ - .
e t' ( - I Page 2. Letter to Dr. Melvin Silberberg (con't.) August 8, 1983 With respect to ASTP0's work, I comment as follows:
- 1. A number of my remarks made during the peer review were aimed at insuring that source-terms generated by ASTP0 were representative of a broad enough spectrum of accident sequences for each reactor type, so as to ensure some degree of " completeness". For example, there are no sequences for BWRs (both Mark I and Mark III containments) in which there is a B or B' (i.e.
1 failure of electric power to engineered safety features or failure to recover either onsite or offsite electric power within 1-3 hours following a transient initiated by loss of offsite AC power). Similiarly there are no loss of coolant accident (LOCA) sequences included for Mark III BWRs, and no small break LOCA's considered for Mark I, BWRs. I would like to explain why I am concerned about completeness, via two 1 examples. First, the results to date are based only on " internal initiators", i.e. external events such as earthquakes and fires are not considered. In the Zion PRA, seismic initiators increase core melt frequency but even more importantly they increase risk. While the SE sequence frequency (the earth-quake induced core-melt) by itselt, is small, it can be shown that by using mean values, the containment also fails structurally in about 3.5% of the cases, leading to an early release and hence the high risk. The remaining SE cases can fail containment by slow overpressurization. Without this external analysis, one might be led to only consider sequences leading to q slow overpressurization, when in fact the early release is risk dominant. '
~ Second the results to date are based on event trees which are " static",
and as such, miss a number of important considerations; operator intervention, partial success or failure of engineered safety features, use of non-safety grade features (e.g. use of CRD pumps for supplying coolant in a BWR LOCA) etc. The accident at TMI started as a transient and ended as a small break LOCA due to operator intervention, and.cannot be described by WASH-1400 type event trees. In view of this, I recommend very strongly that Battelle reconsider its 1 choice of sequences and include at least some LOCA's and B events for BWRs.
- 2. In making an effort to ensure some degree of " completeness",
l Battelle might attempt the following type exercise. For BWRs, there are two important combinations of core melt / containment failure: core melt before containment failure and core melt after containment failure. In addition, time sequencing is important, i.e. early core melts and late core melts. l h
, . - - - - - , , - - -- _,c.-- - -
r---- - - - , - - - - - . - - , - _ _ _
, - ~ , , - . ,
,p.----w
_ _- . . _ . _ - - __ . - - _. , ._ __~ . _ _ _ _ - . -. i I Page 3 Letter to Dr. Melvin Silberberg (con't.) August 8, 1983 In the Limerick PRA for example the accident sequences fall into four classes: three with early core melt (one with containment intact, but at low pressure; one with containment intact, but at high pressure, one with failed containment)
; and one with late core melt (containment failed prior to core melt).
Battelle might consider grouping the various sequences for Mark I and i III containments (Limerick is a Mark II), and check to see if the various
;' physically possible combinations of core melt /contaimnent failure are included.
A similar approach might be used for PWRs, although in this case the contain-ment is usually intact except for bypass (failure to isolate) and the V sequence.
- 3. In my last letter, I recommended that some parametrics on containment failure location and pressure should be undertaken. I fully appreciate Battelle's response (the problem with available time, funds, etc.), but at a minimum I would repeat the calculations for Surry and Peachbottom using the new codes (MARCH 2.0, NAUA etc.), but keeping the WASH 1400 containment failure pressures and location. In this way the source-term would be decoupled from containment
! questions, and the role of physical chemistry would be better understood, i
Furthermore I would like to see how the source term evolves with time for the l PWR long-time overpressurization sequences. i 4 As I centioned at the-Peer Review, some consideration should be given to sequences in which the ice compartment could be bypassed in the PWR-Ice condenser contaiments. There are return paths from the upper compart-ment to the lower compartment (the ventilation system represents one, although it has a baffle which is supposed to prevent this. However, these events are beyond the design basis so it would be an open question). Since the ice com- ' partment is an effective scrubber, the presence of the ice is important in determining the source term. Bypass would be an important consideration. So would the time that the ice melts.
- 5. I have had the opportunity to quickly review Volume 1 "PWR-Large, Dry Containment Draft Report" given out at the meeting. It appears that the treatment of the contaiment failure modes for each sequence (for the AB there are four, for the TMLB' there are two) are certainly along the lines I have i discussed above. I hope this serves as a model for the rest of the volumes.
- - , - - - - m - ., - ,,,, -~ - - - r =i--*rww- - -- - -- ' * - w e-"-- - - - - - - ' ' '" ' - -
4 1 . 4 j I Page 4 Letter to Dr. Melvin Stiberberg (con't.) August 8, 1983 ) - I i It is easy to be critical when one does not have to meet milestones, stay within budget, and respond to the peer review so I want to congratulate
, you constraints and Mike (time onand themoney meeting), under aswhich well as Battelle you on their efforts, are all working. A number given of the f
comments focused on the thermal hydraulic modeling; in particular nodalization j using MARCH. I agree strongly with these comments ard hope Battelle is re-sponsive to them. t i I hope these comments prove useful and I look forward to the next peer j review. Sincerely,
- h. S. M W.E. Kastenberg, j.
Professor of Engineering j WEK/shm i l cc: M. Jankowski - - R. Benero C. Ryder 1 1 I , i i s l : 1 1 j i
. [
i I l l e i ! ) -
- ~ ._ _
(
. . _ . __ _ . _ _ . _ , _ _ , _ _ _ _ . - . _ , _ . _ . . _ _ . _ __.,_ _ _._. , _ . - _ _ . , _ . _ _ , _ _ _ . , _ - . , _ _ _d
- 5. LEVY,INC.
4 0ff?/O Suite 725 g i 1999 South Boscom Avenue Compbell. CoWomio 95006-2233 LM/ USA 408/377 4870 August 1, 1983 Mr. Michael W. Jankowski Accident Source Term Program Office Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Commission Washington, D.C. 20555
Dear Mike:
The following are my comments about the BMI-2104 draft material submitted for and discussed at the third Peer Review meeting.
GENERAL COMMENT
S:
- 1. The implementation of a " state-of-the-art containment loads and capacity" program is a major step forward. The imoortance of this task cannot be over-emohasized and a continued analysis and (possibly) experimental program snould be olanned and implemented beyond 1983 and as long as the re-assessment of the source term is carried out.
- 2. It is suspected that reprioritizing of NRC tasks is con-tinously taking place as the source term program progresses. A more formal procedure might help put emphasis in the most important areas and such a pro-cess should be considered by NRC to especially avoid funds being diverted to areas which might have small impact on the source term. For example, it might be advisable to put more emphasis on CORSOR and fission products release (in addition to containment loads and failure modes). CORSOR is primarily empiri-cal. Furthermore, it fails to deal with fission products release when fuel geometry is not oreserved and in carticular when surface area to volume ratio decreases.
- 3. Some overall balance in modeling needs to be strived for. It is recognized thatit is very difficult to do but the models are sometimes overdetailed in areas of minimal importance and very approximate in critical areas. It is hoped that the validation and sensitivity / uncertainty programs _
will highlight sucn shortcomings. Botn of these programs are welcome additions. 4 Computer codes continue to be improved and modified. We now - have MARCH-2, NAUA modified TRAP-MELT-2, ... etc. It will be necessary soon to freeze some of the methods or the program may not close, it is particularly imoortant that the validation and sensitivity uncertainty programs be carried out with a fro 2en set of computer codes used to calculate the fission products release. C %'I! / j 6
o
- 5. LEVY, INC. .
August 1, 1983 COMMENTS ON BWR MARK I PEACH BOTTOM PLANT
- 1. This report fails to deal with the best estimate (i.e. realistic) conditions for the following reasons:
It always employs drywell containment failure which
, maximizes releases. It is suggested that releases be carried out for three modes of containment failure:
in drywell, in wetwell above suppression pool, and in wetwell below the water level in the suopression 001. Most recent PRA's recognize all three moces of tai ure and combine them probabilistically. The con-tainment loads and capacity program is expected to lead to a similar conclusion and should provide guidance on how to weigh the three modes of containment failure. Hydrogen generation rate accears to be excessive and produces early containment failure. This may be caused by inacequate modeling of the Zircaloy channels, failure to recognize steam starvation, and lack of eventual cutoff of hydrogen production. Another source of the excessive hydrogen may be the use of a single mode in the bottom plenum. Suostantial non-homogeneities will occur there and a multi mode model (similar to that being developed for PWR upper plenum)may be desirable. Extreme failures of safety systems are postulated. For instance, containment spray systems are not utilized when they are available (it makes sense not to use them when all ECCS systems are inoperable, but this is not valid in other sequendes). Here again, actuation of containment sprays will help redistribute non-condensible gases be-tween drywell and wetwell and delay containment failures. This would give more time to scrub fission products in the suppression pool. BWR Emergency coerating procedures are not taken full advantage of. Capability to vent the containment and activate the Automatic Depressurization System (ACS) will have great impact on the source term and need to be con-sidered. Excessive failure of mitigating systems is assumed when containment fails. This is not true depending upon mode and location of containment failure. Many, if not most. BWR's also do not lose ECCS pumps when containment is depressurized. .
- 5. LEVY, INC. . August 1, 1983
- 2. The present program does not olan recalculatino Peach Bottom. This should be reconsidered in view of the above comments. Another possibility is to deal with sucn snortcomings under the sensitivity / uncertainty program.
COMMENTS ON BWR MARK III GRAND GULF PLANT
- l. The sequence TW should be added or substituted for TPI be-cause the risks for TW are expectea to exceed those for TPI since the fission products would be released into a saturated pool for the TW sequence versus a subcooled pool for TPI.
- 2. The hydrogen burning model employs a sinole node. This leads to large cressure spikes which do not correctly describe the use of igniters at Grano Gulf. The pressure levels will decrease and become more realistic with a multinode burn model.
- 3. In all BWR Mark III sequences analyzed, retention of fission products in the outer containment is small. For that reason, consideration should be given to analyze failure of the outer containment from hydrogen burning and showing that it has a very small impact on the source term. If this is true indeed, there is no need to develop an improved hydrogen burning model for the Mark III, 4 It is interesting to compare primary system retention of fission products for Peach Bottom and Grand Gulf. They are very similar even through the sequences are different in timing and failures. This sugaests that primary system retention in BWR's may not be very sequence dependent. This could be important in terms of future efforts and improvements to TRAP-MELT.
Comments on PWR Ice Condenser Containment Design
- 1. The hydrogen burning employs a single node and thus fails the containment even though igniters are provided. As shown at the meeting, reduced pressures are much more likely and such cases should be added to Volume IV.
- 2. While maldistribution among ice compartments is not expected during a LOCA or when recirculating fans are operational, long into the sequence of events, the flow rates become small and involve high temperature natural con-vection currents. Maldistribution could occur and should be evaluated (a poten-tial flow model could bound sucn a case) especially for break flow locations close to an ice compartment. There is a synergistic effect, i.e., maldistribution gets worse and worse due to increased ice melting and increased trapping of fission products and heat production in the compartment with the highest flow.
- 3. In NUREG/CR-3248, impacti,on dominates when the ice is no longer present. Imoaction will change the flow area and its geometry. Equations (18) to (21) need to be modified to recognize the effect of imoaction uoan inter-c30 ting basket str1ps.
5.LEVV,INC. . August 1, 1983 1 COMMENTS ON VALIDATION / UNCERTAINTY PROGRAMS
- 1. These programs should recognize more than the equations used in the computer codes Number of nodes and types of nodino could have a substantial imoact. (See hydrogen burning and BWR lower plenum comments.)
- 2. The greatest sensitivity and uncertainty will come from pre-sumed safety system failures and the lack of application of emergency operating procedures. Because there mignt nave been some confusion aoout tnis comment at the third Peer Review meeting an amplification is provided herein. The variations that should be considered will be given only for BWR's for illustration purposes and should include the following:
a) use of containment spray systems where c~omplete loss of power does not take place, b) use of emergency core cooling systems where containment tallure does not hamper their comolete availability, c) use of CRD drive cooling and its possibility to quench the core melt within the vessel for some sequences, d) use of containment vent as permitted by operating procedures assuming and not assuming safety pumps can continue to pump water. In considering containment vent, it is important to recognize the long pipe and/or stack where considerable deposition of fission products will take place. e) use of Automatic Depressurization System (ADS) by operators could be timed to release fission products into a Saturated pool or a succooled pool before drywell or wetwell contain-ment failure takes place. In BWR's, the operators have the capability at any time with ADS to decide between core melt and containment failure. If drywell failure is to occur, it is best to have core melt proceed before the drywell fails. If wetwell failure is to occur, it might be preferable to delay core melt and the two possible options should be evaluated here. At the Third Peer Review meeting, I was attempting to convey the importance of considering all such limited variations because they may, in fact, identify interim source terms which might readily be achievable. I also like to add that the short discussion I had with Silberberg and Powers before catching my flight helped alleviate my potential concerns in this area. ,
- 3. The following are comments about some of the validation reports.
l o l SL LEUV, H4CL . August 1, 1983 i. MARCH-2 Reference to previous reviews of the MARCH code and its criticisms should be included. A comparison between MARCH-1 and -2 should be provided and the difference explained. Other significant shortcomings not mentioned at the
, meeting are:
I hydrogen burning single noding; single node in upper plenum for PWR and in bottom 4 plenum for BWR;
. hydrogen generation rate compared to other heatup codes especially for BWR geometries; inadequate containment modeling due to the provided series of containment volume. Parallel containment volumes should be provided.
. CORCON i The heat transfer model needs not be as complicated. A single mixed temperature with appropriate heat transfer coefficients at the bottom and top of the melt should be adequate. This is one area where efforts could be reduced. VANESSA I introduction of a liquid layer above the melt should be developed. 4 A verification program should be implemented in addition to the
- validation program. The number of inputs and options are so large that it would j be wise to show that the various codes are not user dependent. This would also 4
help increase the confidence in qua_lity assurance of the various codes.
- 5. Comoarison to other available and similar computer codes should j be orovided whenever possible, i
i
} Sincerely y,ours,
\' ( . /. .
j .4% y
- Salomon Levy 4
President 1 i 4 j SL:jm , j cc:M Silberberg-NRC - 3 A i y,- ,-n --
, __,-..v,. -- - , - . , - . , - , , e-.,--, w,m,,,_ _
,-m - , .-
. . . - - . . - - - _ - ~ _ _ - _ - - _ _ - - - _ .. _ - . _ . _ _ - . . - .
i R0WE 4 Myy/ ;= f.V z.wwu m ,a ,.. . 4 J w%
- 14400 Bellevue-Redmond Road. Suite 208. Believue. Wasnington 960J a206i 641-1620
{ i August 15, 1983 t Mel Silberberg i U.S. Nuclear Regulatory Comission 6915 Eastern Ave. 113033
- Silver Spring, MD 20910
Dear Mr. Silberberg:
I regret that I was unable to attend the peer review during July; however, I have reviewed many of the reports sent to me for that meeting. The results of my review are enclosed and it focuses on the themal-hydraulic aspects of the mathematical models used for the analysis. I would like to call your attention to the comment regarding the potential for fission product deoosition in the multiple primary system loops and steam generators of a PWR. The current approach in MERGE and TRAP-MELT does not seem to considen this when representing the primary system as a few control volumes connected in series. From a broader point of view ' the reports do not convey a strong technical basis for a best estimate analysis. There are numerous assumptions stated without justification and many qualifications of the analytical tools. I recognize the dif-ficulties here but feel that a better job could be done to tighten up the technical basis of the analysis. The results are giving a source term reduced somewhat from WASH-1400 but not by an order of magnitude. With all the uncertainties and tiualifications as currently stated, is it possible to make a case for a source term reduced from that of WASH-14007 I expect that you or others may wish to discuss these comments with me. Please feel free to call.
- Since ely, l
/ #Md/ 4 Donald S. Rowe I l
DSR:le Enclosure - 4 . g567/Jo/5' t
Comments by D. S. Rowe Draft Report: Description and State of Verification of the ORIGEN2 Code. I General Comments My comments on this report are quite general and editorial. The opening remarks on page 1 are very vague refering to the code as a "very flexible reactor physics code." The report needs a specific statement up front such as: ORIGEN2 calculates the radionuclide composition in the core, etc. Page 4 refers to Reference (2) and then states that rost of the calculations a61 trivial. I really don't think this is the case and i the writeup shpuld summarize the computational methods in a way to be informative to the reader who doesn't know anything about ORIGEN2. l l 1 l
l l ! Coucline to Other Codes The specific applications of MARCH 2 and its associated codes indi- ; l cate that MARCH 2 does "overall thennal hydraulics." MARCH 2 appears to
, do much more including calculations involving containment, core-concrete
( interaction and primary system such that there appears to be overlao
- with MERGE. TRAP-MELT, etc. A report section specifically devoted to i this topic would be useful.
Editorial Comment i Section !!! is nearly the entire report. Some breakup could help ; the overall organization of the report. l More illustrations are needed to show phenonena being modelled.
- l The text is heavy reading.
! SPECIFIC COMME!!TS i Pac _es 4 - 5 L , The core nodal treatment could be stated 'in somewhat more general ! terms. Each nadal volume contains volume fractions for fuel, cladding. l water (or stean), channel boxes, control rods, etc, based on the overall j dimensions of the core and its components. I Page 5 l Increasing the thickness of the cladding to include the Zircaloy J of the channel boxes needs further justification or model revision. The reaction of Zircaloy and steam depends on the exposed surface area. It would seem that putting the channel box and control blade mass into 6 a thicker cladding would underestimate the rate of reaction with steam. , t ! Page 6 . l j The modeling of the primary system as a single cylindrical volume l needs more explanation and justification. What happens to the steam f j generators, piping, pump, etc? How are the steam generations included j as heat sinks (or scurces)? 4 i s 1 i
- .. -- - - . . - - - - . . . _ _ _ _ . . . . . . - - - - - . - . - - = - - - - . - - -
l 1 } 4 j Comments by D. S. Rowe i i I I j Draft Report: Status of Validation of the MARCH 2 4 1 Computer Code, July ll, 1983. I j
GENERAL COMMENT
S ! Assumotions Many assumptions are made during the course of the discussion. While a few of them are discussed, most of them are not discussed. I realize that many assumptions are recuired to simolify the analytical model and the model must stand on that justification. Others can be i justified from the physics and order-of-nagnitude importance of the i pnenomena involved. Additional justification or statements regarding ) the assunctions would help strengthen and clarify the modeling discussion. ' I q Control Volumes ! r 4 , Mass and energy balances are stated but without reference to the control volumes that they represent. Illustrations of.the control I } volumes and the placement of. variables on those volumes would help l q clarify how the mass and energy balances are done. Each control volume i f would then also have a clear relationship with its neighboring volume i I by the mass and energy flows at their mutual boundary. The assembly of f all control volumes would represent the nodal description of the entire system. i Adjustable Inputs ! 1 i
- MARCH 2 contains a significant number of adjustable inputs. Thia can be useful for a parametric analysis but is not very satisfactory l for a "best estimate" calculation unless a set of default inputs are
(( selected. If this is not done it is. conceivable that a variety of
- answers could be obtained by " tuning' the inputs. Inputs to MARCH 2 need to be clearly identified if they are causing sensitivity in the results. ;
I I l
i I - I Pace 7 - The impact of the restriction to series connected containment volumes needs to be explained. Are there situations where convection loops j could exist between compartments that would affect the transport of fis-l sfon products? I believe that such loops could be included in MARCH 2 ' with a modest effort and without limiting the code's time step. Page 12 While mass and energy balances are discussed in this section, it is not entirely clear that a global mass and energy balance exists for the system. Does MARCH 2 do a global balance at each time step to
! assure mass and energy conservation?
An illustration of the primary system nodalization is needed to show the various features of what is modelled. Page 17 i l Equation (3.10) does not appear to be an energy balance. Is this supposed to be
, pro n W7pr CA "'~ = OreT
-4h j
where n and n+1 are old on new time respectively. Are there energy flow terms missing. I would prefer to see balance equations in the l above fom (with flow tems) because they have a closer analog to j the differential equations. They re also easily related to control volumes. Page 30
- The assumption of uniform radial temperature in fuel and cladding should be discussed and justified here, j Page 40 i
l None of the meltdown models consider the situation of the cladding I (or structure) melting before the UO2 . Is this more realistic than an 1 average core melt? What technical basis do we have to justify the approaches assumed? e**+
._---..m... - . y ..._ .,_, ,m. . , , , , , _ _-- , , . , . _ . , _ _ ~.--
N i ! Pace 43 4 See comment for page five regarding metal-water reactions. Pace 47 What is technical basis for forming spheres with U02 on the insiqe ] and zirconium on the outside? What is the impact of choosing the t various fraction options? Can we justify any of them? i Page 56 . It is not clear what is averaged and what is a single quantity. Is T 3gg an average? I J Pace 57 Is Equation (3.119) three equations for the unknowns T2 , Ti and Q? If so, then state as three euqations and three unknowns. Page 58
- Where are Equations (7.1) and 7.2)?
1 5 Pace 59 An illustration of the control volume (s) for the bottom head and ! debris energy balance would help clarify the heat transfer processes. 4 I Is radiation or convection from the top of the debris sufficient to i slow (or stop) the melting of the vessel bottom? How about convection 2 through the debris bed? 4 Pac _e 68 i Can the omission of debris entrainment, cavity sweep out, etc be ] justified? What velocities exist in the reactor cavity? I
- Pace 88 i Within the structure fiARCH2, the numerical solution for containment flows and pressures could be modified with modest effort to remove the
" series" restriction and allow " loops." This could be done without imposing time step limitations on the calculation.
Page 100 1 The variable g should be gc, the gravitational constant. i i I
. m
- Page 115 Reference for le Chatelier.
Page 118 Some typical time for flame Spread would be useful to the reader. Page 126 Why not inclu'de this core radiation heat transfer discussion with that starting on page 29. Page 128 The core barrel is heated by the outer two radial core nodes. How is this done in the axial direction where the core has many axial nodes and the core barrel is a single node. Pace 129 Is axial conduction in the fuel rods really necessary? Pace 143 I agree with many of the items brought out in III D. Some of these points could be brought out earlier in connection with the development of the nodels. Page 144 The potential for two-sided oxidation would appear to real as the core begins to slump and relocate. The ratio of the surface area to volume of zirconium would seem to be an important parameter. Page 147 Natural convection flows and associated tission product transport could be significant within containment compartments even if the results of MARCH 2 show long periods of stagnation. This would require a multi-dimensional containment analysis that can consider the coupled thermal response of the structure and hydraulics of the compartment. Page 148 Results of MARCH calculations compared to PBF tests should be included.
I I ] Comments by D. S. Rowe f' Partial Report: Radionuclide Release Under Specific LWR i Accident Conditions--Volume II, BWR, MARKI Design. I SPECIFIC COMME?tTS , Page 3-2 4 A brief statement pointing out that the letters refer to WASH-1400 accident sequences would help the first tine reader. Pace 5-1 The statement at the bottom of the page is not very positive. I suggest that statements be made regarding how MARCH 2 represents the i system and what it calculates. It can then be qualified regarding , f its uncertainties. It's the best we've got at this point inspite of its shortcomings. ! l j Pace 6-5 j Figure 6.3 is not very enlightening regarding the nadal treatment I in the core. The MARCH 2 representation contains many control volumes for purposes of core heat up. Thatpoint needs to be brought out. The 6 control volume treatment here may be satisfactory for cal-i culating pressure drops, flows and wall temperature but it is probably I too coarse for accuracy in the fission transport analysis because of numerical diffusion. Does the TRAP-t1ELT part of the analysis use this same nodal representation? If so, nodal sensitivity analysis should 4 be considered to determine the adequacy of the nodal treatment. This i need not be done on a full system calculation but it could be done as-separate effects calcuation for a standalone problem. , Pace 7-37 i I Reference to BWR1, BWR2, etc. in tables should be footnoted or j accompanied by reference to WASH-1400 so as not to be confused with the GE product line designation. l I l i
. , , - , - , - - . . . , - . . - , ~ - , - , . , , - , ,-n-. . , - - . , , , , . ~ -,
4 .f 4 ! Page 7-39 I i Reference to the hydrogen explosion destroying the reactor building is a rather dramatic statement that calls for more discussion or expla- !: nation. In the mind of the public this sounds like a reactor exploding. Pages 7-40, 7-41 l I This discussion is quite a let down. Are we any better off than WASH-1400? 1 i 4 Pace A-2 Can the operator be expected to shut off the two blowers? l i Page A-6 } It would be helpful to have some volume estimates for the aerosols
- in relationship to the capacity of the filters. The mass numbers are not very informative.
4 i 4 2 1 Y s e t 1 a a 8 i 4 1 1
. . _ . _ ,, . . _ - - . _ - _ = . . _ , _ . _ _ - ___ ,_ _ _ _ .. _ _ _ _ . - . - . . - _ _ _ _ , _ _ __.
Comments by D. S. Rowe Repert: Review of MERGE Code, P. Saha, June 1983.
GENERAL COMMENT
S I generally concur with the views expressed in this report. Fron a broader point of view, I don't understand why the calcula-tions of MERGE are not done in TRAP-MELT. If the mission of MERGE is to calculate primary system flows and structure temperatures for use in TRAP-PELT then it would seem appropriate to do it within the same ccde. Such an integrated approach would assure a consistent control volume apparoach and allow inclusion of relocating fission product heat sources along the primary system piping. If reevolution of fission products from surfaces is an incortant possibility that depends on an accurate calculation surface temperature, it would be necessary to replace the lumped parameter wall temperature model with a distributed parameter model. The extra work would be small with so few control volumes in the primary system. The primary system is represented in MERGE as a series of control volumes, however, PWRs have multiple loops. What justification is used to exclude transport through those loops and their stean generators? If the steam generators are effective heat sinks there could be signi-ficant steam flow and condensation in the steam generators. The steam flow could carry significant aerosals and fission products into the steam generators and associated primary system piping. I believe this could be an important phenomena that;has been omitted from the MERGE analysis.
Comments by D. S. Rowe Draft Report: Status of Validation of the TRAP-t1ELT Computer Code. T. S. Kress and A. L. Wright. 1 General Comments l This report is quite good as it deals with many of the issues 4 related to fission product transport.
- TRAP-MELT is primarily concerned with the transport of mass in the primary system. It does not, however, consider the volume of I
material being transported or deposited. This should be addressed be:ause it could have an important affect on the flow areas, pressure drops, surface heat transfer from pipes, etc. 1 The affect of heating by transported fisson products has been brought up several times and is worthy of mention again. This could be especially important regarding the reevolution of fission products caused by heating of a deposit on a structure surface. The nadal treatment for TRAP-MELT apparently is the sane as for MERGE. Those nodes (control volumes) are large and there are few of I them which can lead to substantial numerical diffusion of aerosals
; along the flo'w path. I suggest that a nodal sensitivity study be done i to explore the accuracy of the current nodal treatment. This could be i
done in a separate stand alone calculation rather than in a full system j calculation. Specific Comments Po. 12 From a numerical point of view, the well mixed volume assumption leads to numerical diffusion. This will cause a smearing of nonuni-form distributions of aerosals and a late calculation of last arrival of aerosals. The effect of numerical diffusion can be reduced by
- using more control volumes, Pg. 16 !
The equation M = CQ I is a definition for mass flow rate and not the incomoressible continuity equation. ! Po. 22 I would expect that the gravitational settling issue would be i related to aerosal size. If particles are large, they would settle by gravity at some velocity and not depend on the concentration. ) 1 i Pg. 29 I would like to specifically concur with the thoughts expressed ] concerning fission product heat sources, 2-D thermal hydraulics and t resuspension of deposited aerosals. 1
, i ir.~ w- ,4 +v e-e p., -
-w-w - w ,w - ,-
-r~--*- -----m-- -~-- y- --+---ws -er- ~~e ~- s-~-w- -+- w ,
LETTERS ABOUT THE PEER REVIEW MEETINGS October 12, 13, 1983 a
$ f,V', J' u
'.r (
- m ar ?
UNIVERSITY OF CALIFORNIA. LOS ANGELES .. . UCLA SEnsELET + D 4% is ta% IN E
- Lut4%CELES Al% E RitDE
- 14% DIEGO %4% f p4%CMo $ i ,$ $4%T484RS4A4
- S ANT 4 sRLZ
%O MECHANICS AND STRUCTURES DEPARTMENT SCHOOL OF ENGINEERING gND APPLIED SCIENCE LOS ANGELES. CALIFORNIA 90054 < October 24, 1983 j
i Mel Silberberg i Accident Source Tem Program Office , i Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Commission j Washington, DC 20555
Dear Mel:
I have reviewed the material handed out at the Peer Review Meeting of l October 12-13, 1983 as well as my notes and would like to transmit the following comments to you for consideration.
- 1. During the period beginrdng with core melt and ending with contain-ment lives: failu[gI:the 12 8.04 various days,isgjgpes 11 ': 2.28 of Iodine hours, deca 113 : with 20.8the hours, following 1139:hal 52.3f minutes, and 1135: 6.7 hours. All ;, ave comparable fission yields and activity.
Two questions come to mind: a) What happens to the cesium tied up as cesium-iodide, and is it important (as the iodide decays)? b) How significant (radiologically) are the daughter products and if they are, is their physical .1 chemistry important? For example 1135 decays to Xe135 which decays to i Cesium-135 with a 9.2 hour half-life, i
- 2. What can be said of the other important isoptopes. For example ,
! ruthenium is currently being lumped into the "other" category; i.e. in the l molten core. What can be said about its release. There are a number of other isotopes which would be important (e.g. Sr. Ba, Rb, etc.), and contribute
- to risk and hence the " risk perspective".
- 3. Most of the previous work conce ning the Zion plant assumed a flooded cavity for the TMLB' sequence. The Battelle work seems to imply a dry cavity.
1 The difference (wat vs. dry) affects the time, mode and location of containment failure, and hence the " source term". Why is there a difference?
Page 2 Letter to Mel Silberberg (con't.) October 24, 1983 4 In all of the cases shown for the Zion calculations, a point value is given for containment failure pressure. This value is treated as a " thresh-hold" with respect to the " source term". This viewpoint (as well as for other containments) leads to base mat penetration as the failure mode in the sequences shown for Zion. If containment failure pressure is more accurately portrayed by a distri-bution, there will be some probability (albeit small), that the containment would fail by slow-overpressure, followed by an airbourne release. This viewpoint is missing in the work presented during the meeting and should be assessed.
- 5. Adequate attention ~ to external events should be given where possible.
For example, we now have consideration of external events for Zion, Indian Point and Limerick. I believe we will see more for other PRAs such as GESSAR, Shoreham, etc. Seismic events may be initiators for suppression pool by-pass in some BWR suppression pool containments. At Zion they contributed to early containment failure. Such considerations should be factored into the Source Term Program. At present they don't appear to be.
- 6. I support your view that there is need for another meeting. I would recommend that the following be included, a) A clear statement of the changes made (in data as well as models),
during the course of the study is needed. Various changes were made (e.g. failure location for Peachbottom, concrete composition for Surry, etc.) which makes evaluation of the new codes and models difficult. These latter changes are often masked by the data changes, b) The Peer Reviewers need a clear statement of the various phenomena modelled (and not modelled) beginning with WASH-1400, and carried out thru this study, c) A clear understanding of the differences in the results obtained between the early Surry and Peachbottom (WASH-1400) analyses, the first Surry analysis using March 1.1 and the latest (March 2.0) Surry results, should be achieved. Battelle should have ample time! to digest their results, and describe and assess any changes in results, l
i i Page 3 Letter to Mel Silberberg (con't.) October 24, 1983 7 The job of putting the results of this study into perspective will be difficult. As I mentioned at the meeting, the source term takes on different significance when considered on different sites. Population density, meteorology, geology, etc. all play a role. It would be helpful to the Peer Review Group if some specifics were discussed at the next meeting. Lastly, I again congratulate your staff, and that of Battelle Laboratory for the effort they have expended. I look forward to the next Peer Review. Sincerely, Y. [. A W.E. Kastenberg, Professor Engineering and plied Science WEK/shm 1 cc: M. Jankowski C. Ryden I R. Benero _ 1 i
. . . - - - , - - . , . . . . - . _ - _ - , . . ~ - - , , . ~ . - - . , - , ,- . _ . . _ . . . - . . _ , . , , - , . . - . n-.e.,,..
- A
' ' ~
- 5. LEW INCORPORATED -
Suit. 72s 1999 South Boscom Aenus Campbell. Coldornio 95008 2233 USA 408/377-4870 Telex ITT 4994199 i October 17, 1983 M. N. Jankowski Office of Nuclear Fegulatory Research U.S. Nuclear Regulatory Commissi7n Washington, OC 20555 , l
Dear Mike:
The following are my comments about the material presented at the h Term Peer Review meetings of October 12 and 13.14R3. 1. The new gradual core slumping is in the right direction, but, in its present form, it raises considerable questions about the calculated results. Some of the concerns are: e As material moves f rom one node to the next, there is no check that it can be physically accommodated (i.e., is there any volume for it?). e During the slumping, no geometrical changes are assumed, even though surface and flow areas are expected to change con-siderably. The fuel and gas temperatures calculated from this r model appear too low,
- 2. The amount of radioactivity retained in the core is excessive for two reasons. First, the temperature in a melting core will force Also, because no changes in geometry are postu-some re-emission.
lated, substantial areas are employed for retention when, in fact, most will disappear with time during the melt process.
- 3. The use of homogeneous mixture in the upper plenum needs to be looked at, especially as we break it down in more and more axial nodes. There will not be complete temperature mixing and some high temperature streams may exist which will transport fission products without their deposition. Some parallet noding for sensitivity pur-poses may be worthwhile here. .
0 l] SO :l 30 ly y
- 5. L(VY INCORPORATED M. N. Jankowski October 17, 1983 Page 2 4 As the amounts of radioactivity retained in piping and containment increase, as in the recent calculations, there is an urgent need to take into account their location and heat production. The concern, here, is that they might lead to bypass of containment. For instance, during TMLR, enough fission products reach the steam generators to possibly produce steam generator tube f ailure and release to the secondary side and the environment. It is important to recognize that during this sequence of events, primary pressure is very high and, if we can agglomerate fission products at crev. ,
ices, dents, U-bends, we might bypass containment if the tubes fail. The same might be true for isolation valves, containment penetra. tions, etc. This needs a much closer look, especially as contain-ment overpressure failure probability decreases.
- 5. It was not clear what the impact was of the fog model addition in the containment. I have some serious reservations about its vali-dity. It might be preferable to just look at carryover of water with steam. Even with extremely well engineered separator-dryer systems, one can still carry 0.1 percent water with steam, and this should provide a more physical basis for including water drops in the containment.
- 6. The model developed by D. Powers for a water pool above the concrete may be optimistic. If there is no crust, as postulated in the model, there will be film boiling at the interface with the water pool. This pool boiling will be enhanced by the gas being generated in the corium-concrete reaction. Bubble sizes should be prescribed from Taylor instability, recognizing the impact of present impuri-ties upon surface tension. Heat generation in such bubbles may become important, especially as the water depth increases, and thus limit the scrubbing of fission products. Another note of caution is that the scrubbing efficiency proposed by Powers is considerably larger than that employed in the Peach Bottom or Grand Gulf suppression pool when, in fact, it should be the other way.
- 7. The quantitative uncertainty estimation will produce considerable debate. At this time, it should be viewed only as an experiment.
- 8. The UKAEA speaker referred to preliminary tests of molten fuel in boric acid which released iodine. It is worth noting that boric acid was present at TMI-2, and we .need to reconcile UKAEA results with the THI-2 accident.
! ' 5. LGVV NKOAPCAATED 4
- M. W. Jankowski f October 17, 1983 -
Page 3 ] Let me finally add that if another meeting is to be held, it is important i that all reports and results be in their final form and be provided to 3 the peer reviewers for advance review prior to the meeting. 1 Sincerely yours. (5 Salomon Levy l /rc j cc: M. Silberberg. US NRC i 1 l
\ ,
I i i e l - 4 l i l . a i i
College of Engineenng Nuclear Engineenng UNIVERSITY OF MISSOURI-COLUMBIA 1C26 Engineenng Columcea. ussoun 65211 Te=onone (314) 882 355o October 21, 1983 Mr. M. Silberberg Accident Source Term Program Office Nuclear Regulatory Research U.S. Nuclear Regulatory Comission Washington, D.C. 20555
Dear Mel:
SUBJECT:
BMI-2104 Review From the beginning of the peer review process on the subject work, I have been most impressed by the very substantial efforts undertaken by the principles (Batelle, Sandia Oak Ridge) and your staff. I believe to a very considerable extent the work is based on the state of the art, but I also believe first that the state of the art is not very gooc and second in some cases the computer codes involved d_o, o not use the best information that has been developed over the past several years. I , recognize that within the time constraints of the subject work it would have been both unwise and difficult to modify the computer programs. As such there was sufficient confusion in the review meeting of October 12,13 because of changes in some input data and use of MARCH 2 for SURREY-l. Now that the draft reports are nearing completion, it would be appropriate for me to bring my observations to your attention. To a large extent my observations are a result of my long standing research interests in aerosol mechanics, and my recent review of aerosol mechanics in nuclear reactor safety, Progress in Nuclear Energy ,12,,1 (1983). I had provided your office a preprint of this paper at the first peer review meeting, and also Chris Ryder has a copy of the just published paper. Chris has indicated that he has found the review paper quite pertinent and useful, and it is my purpose in this note only to sumarize some of my main observations. It appears to me that realistic (or conservative) description of a whole generally requires a correspondingly more realistic description of the parts, especially when the processes are highly dynamic and not well understood. Exceptions exist, but in general the local error must be much less than the global error. Thus if the o.verall release results for the release fractions are to be accurate within an order of magnitude, than the aerosol codes should be accurate at least to the first order. To achieve this goal, advances in several aspects of aerosol mechanics must be made. I have noted below the specific areas where advances are needed, continued....
. . - , , . , l
( () l 0IY '
i i . ! Mr. M. Silberberg
- ' October 21, 1983
- Page 2 I
4 and I have also commented upon the aspects where the release study aerosol
- computer codes (TRAP-MEl.T NAUA) do not use the state of the art infomation.
A ] (1) Sinale Particle Aerosol Mechanics: l i a) Condensation-Evaporation (Heat and Mass Transfer): i The computer programs use Mason's expression which is not valid
- for small aerosols. I have developed some better expressions, which i have been verified for isothermel conditions. An approximate ex-l pression for non-isothermal conditions can be readily constructed
- by using the results reported in nty review paper. I have presented
! such results in the past at NRC aerosol review group meetings. The !
- theoretical results, however, need to be improved (for cross-effects, j internal heat generation, strong condensation and evaporation, i surface effects), and experimental verification must be carried out.
- The latter particle can be accomplished electrodynamic balances Davis by(using)recently or the type of apparatus developed single used by Dr. Thomas for NRC sponsored accongnodation coefficients "
. measurement work at this University. !
b) Chemical Reactions and Chemisorption: i Relatively little infonnation for particles and gases of interest . .t is known. Again, I believe efforts of the type suggested in a) above { must be initiated. In respect to a question that was raised at the ! October 12,13 meeting, I believe the heat of vaporization (reaction) would be important if there is significant condensation (reaction) ; i and should not be neglected a priori. Recent. experimental evidence i ! strongly suggests that condensing / evaporating environments can have j strong synergestic effects on chemical reactions, and such evidence
- should be considered.
l c) Thermophores i s , Di ffus iophores i s , Photopho res t s : I The codes are using relatively crude expressions for thennophoresis , and diffusiophoresis. I have reported in my paper how better expressions '
. for these effects can be constructed by using results that have become available in recent years. I have also noted that much additional i j theoretical and experimental effort is needed - because recently, con-troversies regarding even the qualitative aspects have arisen.
j ' Photophoresis (or radio-phoresis), has not been accounted for in any of the codes. Either it should be established that it is not important , l T or it should be accounted for. ' j con ti nued. . . . . l )
- - . - _ ~ - - - _ - - . . - - _
.) ' i I 2 i 1 Mr. M. Silberberg i October 21, 1983 Page 3 1 d ) d) Deposition rates: 1 . l The models for deposition rates are entirely ad-hoc and serious questions can be raised regarding the expressions presently in use. , I believe first, consistent phenomenology should be developed (starting from general space dependent forms of the aerosol equation), and the
- deposition coefficients should be defined unambiguously both from i~
theoretical and experimental viewpoints. Also, new correlations for turbulent deposition have become available in the aerosol literature i 1 (for example, DeMota and Friedlander), and these should be considered in lieu of the expressions presently in use. Stefan flow and diffustophoresis must also be considered. < l
- e) Homogeneous Nucleation
r d I detected a certain tendency on the part of the code developers-Batelle - to reference old texts as these are readily available. Much I i has occured in the past several years, and better expressions have become available. Even these latter expressions are undergoing i revisions because of pmgress in kinetic theory and experimental j aspects, and while the inclination of the code developers can be understood in view of imense pressures to get results, I think there , J must be a continuing NRC effort to have available the latest, ' critically appraised information, in a useable form. i
- 2. Two Particle Aerosol Mechanics: -
i i The difficulties that we have with single particle mechanics, affect very strongly the two particle mechanics. In addition, new r { difficulties also arise. We all realize that the coagulation , i coefficients are poorly known, but progress has been made in the
! past years. Specifically the following should be considered:
a) BrownianCoagulation(8): g j All expressions that are available at this time have limitations. But TTe11 eve the expression I have given in my rwview paper perhaps ! is a I believe NAUA' good representation s expression should beofexamined the ideas advanced carefully. by Fuchs. I [ b) Gravitation Coagulation (8g ): : i Under NRC sponsorship, my studenp and I have conducted extensive f l work in this area. Our expressions are radically different from , j those being used in NAUA (TRAP-MELT). Based on some work we have j j done with our ASTD aerosol behavior program, ! believe that there i should be serious questions about the expressions being used in j BMI-2104 codes. [ l con ti nued. . . . . I l , I l
Mr. M. Silberberg October 21, 1983 Page 4 c) Turoulent Coagulation (ST) This is most difficult to quantify, but is likely to be important in primary coolant systems. T. K. Enomoto is looking at this question in detail in conjunction with his Ph.D. thesis under my supervision, and I should be able to comunicate a paper to your office within the next couple of months. I believe, many questions will require much further investigations as several assumptions have been made. Also, calculations corresponding to postulated PCS conditions must be made. d) Synergism: The assumption 8 = s + S n that the c des use, is unverified and has no scientific basis. g + 3 T All published work and our own investigations make it clear that synergism can be strong, and diffusiophoresis, thermophoresis, diffusion, kundsen drag forces, etc. can influence the calculated results markedly. . Efforts should be made to obtain better understanding of synergism, and improved 8 should be developed.
- 3. a) The Aerosol Behavior Ecuation:
The BMI-2104 codes use aerosol behavior equation that is based on strong "contrsction of description". There must be some examination of ' the manner in which effects of shape, velocities, structure, spatial
- dependence, densities etc. are " averaged" to arrive at the equation.
This will help define the coagulation coefficients and the deposition rates in some clear fashion. While resolution of all the questions will require much time and effort, some progress can be made in reasonably short time. I am not convinced that the aerosol equations are being solved in the best possible manners. With condensation, the equations become hyperbolic, and there is no recognition of this fact in the algorithms being used in the BMI-2104 codes. Based on my own experience with ASTO, I have little doubt that the BMI-2104 codes would yield useless results for strong condensation (evaporation) conditions. I believe with some effort the situation can be, and should be, improved. I would suggest in this conjunction that a class of problems must
! be used for computational benchmarking of the computer codes. Some
! analytical solutions are available, more are being developed, and the NRC must make sure that the codes use efficient numerical schenes j that yield accurate results. continued.....
! i ! Mr. M. Silberberg
- October 21, 1983 3
- Page 5 I
i Personally, I would recommend that efforts must be initiated for incorporating explicitly the spatial dependence of aerosol distribution especially in the PC5. I do not think that the TRAP-MELT approximations ) are consistent in this regard. ; 1 I recognize that NRC is a participant in several large scale { experimental efforts (MARV! KEN. T EAT, CSTF, NSPP) which will be crucial to understanding of aerosol behavior in nuclear reactor
- j accidents. I also think for data interpretation, guidance of experi- "
3 monts, and eventual development of the source tems that would meet { critical unbiased scientific appraisals, the BMI-2104 codes are not > adequate. I would like to urge, therefore, that a 1ggg ran program , j that would seek to advance the state of the art througn e er theoret- ' ical modeling efforts and small separate effects tests should also be developed. Such a program should be complimentary to the large scale experimental efforts now undemay, f No doubt every effort should be made to modify the present codes ! so that the known deficiencies are modified. I would recomend reactivation and much further strengthening of the efforts that had sought to prepare detailed and critical compilation of theoretical expressions and physical properties data as they pertain to aerosol i ' behavior in seve m accidents. - 1 i ! have appmciated the opportunities for particip'ation in the I ! peer review meetings, and the sustained NRC support of my aerosol ' j research work at this University. Some of the perspectives that ! i *have developed, are result of this work. If I can assist you further, j please do not hesitate to call upon me. t }; Sincerely, j /_m
._( t .J
-- _. 7
- Sudarshan K. Loyalka i
Professor of Nuclear Engineering i James C. Dowell Chair in Engineering j js t i CC: C. Ryder ! J. Larkins H. !sbin . ,1 W. K4stenberg 4 T. Kms: R. Adams i M. Tobias ! l I i I - i
R l l
\ ,t > ,
UNIVERSITY OF VIRGINIA l sCHoot. OF ENGINEERING AND APPt. LEO SCIENCE a - . CNAAbef?MWikka. 44940 y' OsPaatnesNt OF NuckaAm the4N64AINe AND ENe4NSGAING PNYSICS f fbSPNON 41 404 044 7936 M. ACTO 4 PACabafV December 9, 1983 I i l Michael W. Jankowski Accident Source Term Program Office Office of Nuclear Regulatory Research U. S. Nuclear Regulatory Comunission i Washington, D. C. 20555 l Dear Mike . i
. I an enclosing comments with regard to the last meeting of the Source i
Tern Review Committee. I apologize for sending in these remarks so late.
! I as also sending a copy of these comments directly to Mel Silberberg.
i Sincerely, W A. 5. Reynolds, ofessor A i Dept. of Nuclear Engineering l and Engineering Physics ABRiph
- t Encl. ~
i D , 'l g4 6 : p .) h V cpu.L. u) L . CLom i m i l (t) N. 9h kxno n) W , Pcuq pre - ; i C's) C . hoL%,i to) r2. %w i um te)a.Cyue<e.i M L. i u.) T . % s g o m, L7) 2. Wic.t.,. m t o% I Lf)O.Ocw g% 5 Sc7 / 3 0sy
o CO. 6 ON THE OCTOBER SOURCE TERM REVIEW MEETING AND ON ClAPTER V BMI-2104 A. B. Reynolds 8 December, 1983
- 1. For future development of CORSOR, J. L. Kelly and I suggest the use of Arrhenius plots of fission product release fractions, i.e.,
K versus 1/T, instead of the plots used in NUREC-0772 and BMI-2104.
~N , instead of the CORSOR (or Thus K would have the form, K,e NURIG-0772) form, AaBT, A detailed description of work by James L. Kelly and Michael E.
McGown and myself is appended to this report to support our recommenda-tion of the Arrhenius model.
- 2. I note that the wall and floor surface areas in the containment are tabulated in Table 6.6 of Volume V for the large PWR containment.
There is much more surface area from pipes and structures inside the containment. Are all of these areas included?
- 1 A common industry complaint is that neither all of the surface area I
nor all of the heat capacity of the pipes and structures inside the
> containments is included in calculating condensation and removal of fission products. It would be useful to list the internal heat , capacity used in the MARCI calculations and internal structure surface , areas in addition to the wall and floor uurface areas in the containment (and perhaps the upper planus, etc. in the primary system) so that industry people could compare with their analyses. (Not only would this be useful to industry people who often think Dattelle/NRC source term calculations are too conse'vative--it r would also be useful to explain to nuclear critics who think Battelle/NRC source term 4
calculations are too low why so much fission product depletion occurs in these volumes.) 1
I 1 4 ll 3. Another suggestion, similar to the #2 above, is to add sketches i that illustrate the large amount of structure (surface plus heat j capacity) and the tortuous pathways in the containment building (and perhaps in the primary system upper plenum). For example, i look at Fig. 4.4 in Volume 1 (or Fig. 4.7). This figure creates f the impression that there is a rather clear path between a pipe l break or the reactor cavity and the containment wall. Industry people stress that the pathways are actually complex, with many surfaces (pipes, concrete walls, etc.) along the way on which g material can condense or place out. Fig. 4.4 (and Fig. 4.7) is useful for its simplicity in illustrating the message of the figure.
- However, perhaps a companion figure should inmiediately follow each
; simplistic figure illustrating somehow the actual complexity of the geometry and structures that affect the pathways and attenuation of ,
the aerosols. (I am just sensitive to the people who tell me that . . it rained inside the TMI containment for 3 days due to all the surfaces and heat capacity, etc. I don't know whether this is true. And, if true, I don't know whether the Battelle codes (MARCI, I suppose) , would have predicted that.] f 4. Still another suggestion. Provide a detailed list (in tabular form, easy to see and reference) of all of the attenuation mechanisms being accounted for in each of the codes-especially TRAP-MELT and NAUA. These are listed to some extent in Volume I, Section 5.3, but maybe they could be collected and highlighted in a table. I recently saw l a list produced by Milt Levensch (of Bechtel, and current ANS president) .l of all the natural mechanisms that would take out aerosols, and he talked as if many of these were not considered in codes. I expect that they are being considered in codes as effective as those being 4 2 i i
used by Battelle, but a list would allow people to see this.
- 5. Since little detail was provided by Westinghouse for the upper plenum, the suggestion I made in items 2-4 may not be very relevant for this region. I note that Battelle discusses this problem in I
j6.1ofvolumeV. I understood Baccelle people to say at the last meeting that this lack of information created a problem for them, I but they seem to say in { 6.1 that it's not really a problem because flow through detailed geometry could not be analyzed anyway. Despite industry's admonitions to take all surfaces and geometries into account, I suggest that NRC/Battelle can only take the most conserva-y tive approach in the absence of details from industry. Hence, to the extent that industry fails to provide details of the geometry, they J' damage their own case. i
- 6. I reiterate what many have said: The report needs more schematics j
showing the control volumes in MARCH. Figures like Fig. 6.3 in . Volume V are woefully inadequate. In Fig. 6.2 (Vol V), what are structures 1, 2, and 3 for example? They are not illustrated anywhere. The gas outlet (Fig. 6.2) 6 approaches the structure 1 temperature; is this coincidence or would the reason be obvious if I had any concept of where the gas outlet was relative to the position of struccure 17 , In Fig. 6.la, Vol V, where are ROD (__, __)?
, I thought I had figured out the formula for TRO (,__, __) in Volume I, but I can't figure out 4
the numbering sequence for ROD (_ , J . It seems crucial to me that l once Battelle goes to the trouble I of plotting temperatures for positions, or rods, that they should define (by another picture) what I ] the numbers mean. Moreover, ROD (12,1) appears twice on Fig. 6.la. 3
- , - , , . ,er,+m ,----r v. .-- . - - - - .g,--~.. -, , , , ,. . .y ,,.--e . , . - - , m--- ,e---- r.-,, -,,-+---egw--eyeg-
. - __ . _ _ _ _ _ __ . . _ . . _ ._ _ _ _ . _~ _ _ _ __ _ . _ . _ _ _ _ _ __ _ _ __
M Also ROD (12,1) in Fig. 6.lb looks like neither of the j ROD's (12,1) in Fig. 6.la. Things like this really discourage the i reader from going on. What are Volumes 2 and 3 on Figures 6.11 and 6.12 of Vol V7 Again these volumes were not explained, but their temperatures are plotted.
- 7. I would like to know the correlation for h used for film boiling in the model presented by Powers at the October meeting. In fact. I-would like to have a writeup of all of the calculations presented by
] Powers at that meeting in order to resember what he was doing and its I significance.
- 8. I do not recall a satisfactory response to Kastenburg's question about possible bypass of the suppression pool for BWR accidents. Are ,
scenarios possible for dry well well failure or blowout of penetrations through the dry well so that the suppression pool is bypassed? Is it Batte11e's (and NRC's) position that bypass of the suppression pool.is
- I impossible for MARK III but perhaps possible for MARK I? I would like
~
l to have the position restated at the January meeting.
< 9. I see that Battelle is trying to answer the of ten repeated question i
"Where is all the water?" by showing water inventories, for example, in Tables 6.3 and 6.4 in Volume V. This is a useful addition. I suggest
! that inventories at time zero should also be included in those tables i
- for quick reference to allow the reader to assure himself easily and j
I quickly that a complete water mass balance exists at all stages of ! the calculation. i l 10. Bravo for Zumwalt's plots. Tho' importance of early containment i failure on I and Cs release was recognized by most of us early in the 1 y i review, which prompted our concern about understanding contaiment 4 l 7_ ,- c --- mm.--_.,y-- ., ,...m73 ..-.__ _ __-m__ m_n.,n ym-_- -.,,y,.. me-3_p,m-,_.y,~,, _ , ,m.. , , ,,r ,_, ;y. -_-, - - - - .~., , , -
, failure better before definitive answers about source term could 1 be obtained. The Zumwalt-type plot displays information in a useful way. One note I might add to Zumwalt's plots. He shows the fractional releases for the AB-8 as occurring at essentially zero time (since containment is bypassed). However I looked at Figure 7.3 of Volume 1
- V and noted that releases (from the auxiliary building) occur at a i
j later time. I have plotted the results for iodine on Zumwelt's plot (see next page). 1 It would be useful to add the type of containment failure to Zumwalt's points, i.e. list y, 6, c, etc. beside each point.
, 11. I suggest adding containment failure times to the final release 1
fractions in Section 7. (They currently appear in Volume V on a table with times for core melting and vessel failure). I found myself searching for these times when I was checking Zumwalt's curves, and
- trying to see which kind of containment failures (8, y, 5, c)
~
corresponded to his data points. Since Zumwalt's plot is useful, it ) would also be useful to put the containment failure times directly on the final tables in Section 7. (The only confusion arises for the t 8 sequence since " containment failure" exists from the start while the ! final release occurs such later.)
- 12. I recommend that review committee members be supplied with the papers directly related to source term presented by the national lab researchers at the NRC safety research meeting at Caithersburg in October. These were discussed!in the "Inside NRC" in the October 31 issue in which the impression is that so many problems have been identified that the source term can hardly be specified. Relevant 5
._ .-.- -- _ . - . - - - - - - ~ - - - _ _ . - - _ - - _ -
4 t o I N 5- S C LE A 5 E V 5. C : cw M i.N -
- :"' .U RC T: MI i t i i eT ,
T~A YI i
- '. ')
t 2 Gr ' 12 M -> Y fgj k 7 -. 59 Qay h?VRR YTLM(I $ J:w S. - y b w ~pW x. ; ~.,,
'Eh 7
.- a l'
p6tf p 1*p SECvovAu TLM3
'l
- J.
.,3 w' A ' S EhJ oYAH TLM h3
\N 1 -
- 10 '" T (9 np 5 gL
. .. O .
- ' g%
t \
- 1 OI I
=$-v u
I
- [ O- .-g /
E LL .
- MW <
t.o .. d --- Il 3 N TT._,vI G x -
~
C n ,s of
- j'),13 ,2eg y , '. :
po u, w .:. 2. ca cG. lm) UM suP=vd I U ., b ,, f, C (-% ...A L . .- j,1-a fp.,1<.1 lN .ub< n,a l w/ v. 2 F. p- 71} # %) .
- i.0 - a -
l0 9..-:
.:: = , :..
* .^:::=.'
C 7 .- ' 10 30 10 0 300 1000 3000
- . __ _ _.- l " . . - .- l l l l l l FA1LU R E tim E (MINUTES)
I
' I papers by Gieseke. Kress. Wichener. !.ipinski, Powers. Niemczyk !,
and Owczarski were discussed in "Inside NRC." l 5 i I t I i 4 9 . .I
' i 1
n
. 8 r
b b e i !
, t, I
i e I 3 .e l 4 f I i t 4 4 f i 6 1, i i
d 8 l@ r' Ei.ctnc Power
- p. .ren miitui.
November 1, 1983 Dr. Molvin Silberborg Division of Accident Evaluation U.S. Nuclear Regulatory Commission MS 1130 SS j Washington, DC 20555 l Dear Dr. Silborborg l This letter formalizon EPRI's comments on draft sections j of report BMI-2104 presented at the Poor Review Mooting i of October 12 and 13, 1983. Unfortunately, I was unablo to personally attend, but Frank Rahn represented EPRI. The following comments are his, but I agroo with them. 1 At the mooting the ranults for the PWR plants (Sequoyah, l Surry and :lon) were discuaned as well as the sonnitivity I study work being performed at Sandia. While the DMI-2104 calculations appear to be much improved over previous efforts, we find that it still is not possible to trace the calculations be the annumptionn used. This loavon un uncomfortable as to whether we really understand the information presented. Of particulse importanco, in our opinion, is the location of the water during the course of the accident, a question rained nevoral timos in our previous comments. The latent calculations give a bottor answer to this point, but a completo inventory as a function of timo is not yet available. The prononco of oven small amounts of moisture, in the PCD atmosphore or piping systema in quite critical in ovalue'ing fission product reloano, onpocially when dealing with .ydroncopic finnion product spacios. This has boon otton shown in experimonts, but does not appear to be reflected in the calculations. One other gonoral comment regards the probability and modo of containment failure. Wo undoratand that quantions related to those anpoets and that of any 4ssociated finnion product attenuation in being separately addresnod, and will bo included in the final report prior to issuanco. Wo considor this part of tho analysis to be critical to the conclusions drawn, parti-
/
(ynylAl' m: soie. 4.~. r%u cev. no. w r.io Ann <:4 04m 4.i,non.. m m un Withingt0A CHmp thoQ Mts gg husetts A.. , NM Quore 100. Weghongton, DC .NJ)M I:0!I t!! 91:2 i
Dr. M lvin Silb rborg
- , Novemb r 1, 1983 Page Two cularly with regards to the timing and magnitude of postulated releases. We, therefore, feel that the Peer Review Panel should have an opportunity to review these data prior to the report's issuance.
Specific comments related to the latest Peer Mooting follows e We wish to underscore the information relative to the PWR check valve "V" accident scenarios presented by Stono and Webster Engineering Company at the mooting. The NRC defi-nition of this important accident scenario seems to be grossly at variance with what might really happen, especially in regard to the flooding of compartments, the draining of the RWST, and the timing of core uncovery, e The assumptions of containment failure timing appear to be completely arbitrary. Hopefully, the still ongoing work assessing containment questions will renolve this question. e Although it was stated at the Peer Revihw Meeting that the adiabatic assumption was not used in calculating hydrogen combustion response (and that even if it woro used, only a few psi dif ference in peak pressures would result), we do not fool confident of the analyson and would like to see further confirmation of the calculations. e The behavior of tellurium modelled in the report is still questionable. The German papors presented at the recent Cambridge Safety Topical Mooting did not support the Oak Ridge data, their extrapolations and subsequent conclusions. Finally, and perhaps most importantly, is that no " stopping critoria" has boon presented, i.e., guidelines as to when the hypothetical scenarios becomo no improbable that they no longer deservo serious consideration. We urge the NRC and its con-tractors to more clearly clucidata those critoria in order to put the current work in perspectivo and to provido a point at which this work can be terminated. EPRI apprecistos the opportunity to comment on this work. We recognize the largo effort to dato, and the obvious improvement over past drafts of the report. Wo encourage you to continuo your efforts in this regard. Sincoroly yours, 144 d Richard C. Vogel Sonior Scientific Advisor Nuclear Safety & Analysis Dept. RCViss cci J, Taylor, W. Loewenstoin, F. Rahn, R. Sohgal
.~ _. . _ . _ _ - . . - . . _ - - _ _ _ - . - . _ _ _ - .
y e
&h-! !;- , b. -
e .p x October 47, 1983 , 7~. I
~p 1007 Birmingham Drive ' i!* M >
~
p \ Cardiff.by.the Sea, California 92007 , A Fear Review Report To: Accident Source Term Program Office, RES, NRC. ' Attention: M.W. Jankowski and M. Silberberg / i
Subject:
Foot Meeting (of Oct. 14 13) Cossents and Suggestions Regarding BMI.2104
\
By: L.R. Zumwelt, Consultant and Profenor Emeritus of ! Nuclear Engineering,N.C.S.U. i . l The reviewer was greatly impressed by the qvantity and quality of the j work on BM12104 presented by the personnel of Battelle Columbus and also , by the studies and proposed work presented by personnel of the Sandia and i j the Oak Ridge laboratories. On the other hand the reviewer had expected more l complete versions of Volumes IV, V and VI than were actually presented but understands there was not time enough by the October 12 13 Peer Review Meet. ( ing to complete the required accident sequence calculations and to fully re. flect on the results that were obtained. It is understood that a substantial editing effort will be made before draft versions of the report volumes are completed. Specific comuments follow. 1
- 1. The sero point of time for any figure or table relative to the zero point
] , j (start) of given accident sequences should be clear. ! 2. For a review of the chemical interactions of fission product elements or compounds with surfaces and aerosols in the RCS Lt is highly desireable to 1-haves ! (a) Specification of the meterial of the surface (alloy comyosition), ! mass transfer equivalent area (perhaps, equivalent heat transfer area will do), and surface tesgeratures of the walls of the control volumes ] (taking decay heat into account, if possible) as a function of time. 1
- 1 i
i
._,_. __ _ . . _ _ _ _ _ _ _ _ _ _ . _ _ , _ _ ~ _ . . _ . _ . , _ _ _ _ _ _ . _ _ _ . . _ . _ . . , _ _ _ _ . , . _ _ _ _ , _ . _ _ . _ _ - . , - . _ , . , _ _
i i L.R.Zusaci t 10/27/83 p (b) The composition ( H.g/M.0 2 ratio ) and temperature of the gas j passing through the control volume as a function of time. * (c) The aerosol composition and mass per unit gas volume in the con-j trol volumes as a. function of time. The composition should include
- the elements B, Ag, Cd and In from control material in the core or ..
i primary coolant. This detail may not all need to be included in the i ! Report Vol mes but should be available. 1 l l 3. Volume IV " bottom line" data (i.e. the total release of cesium, iodine, 1 i tellurium to the environment) did not include a ecgarison with WASH-1400 l i , j estimates. Should not this be done? 1 J j Rather than make more comuments on the details of the report volumes, the reviewer would like to present a look at the " bottom line" data avail. ., able as of the Oct. 12-13 Meeting. This is done by plotting, on a log-log scale total release to the en-vironment (release fraction) vs. failure time of containment of a given ! reactor plant accident sequence. 'In the graph a containment isolation } } failure release fraction is taken as occuring at a failure time of one i !; minute (rather than zero time, at minus infinity on log tiec scale). .
. t l Similarily the release fraction point for a no failure esse is' p1 Iced at i
j 10,000 minutes (167 hours) rather than at plus infinity. These plotting l choices are somewhat arbitrary but seem reasonable, j The plotting of total release fraction (RF) versus containment failure
! time (TCF) uns done separately for cesium, iodine and tellurium. The graphs i enable the viewer to readily see all the release results for a given fission i
Product. These are ue11 separated due to the sensitivity of RF to type of i
- reactor plant, accident sequence and TCF.
1 4 l [ i
l ' L.R.Zuanselt 10/27/83 The plot also enables the viewer to get an idea of the reasonableness l or uncertainty of RF values. The reviewer believes this is possible because, t intuitively from the nature of the physical system and calculational models involved, it would seem that for a given reactor plant and a given accident sequence the functional relationship between RF and TCF should be a smooth curve wherein RF decreases with increasing TCF and approaches an asynytote i corresponding to the no containment failure case for the given reactor plant i and accident sequence. Reference is now made to the three attached graphs corresponding to cesium, iodine and tellurismi release. A brief study of the graphs indicates i the 'following:
- 1. The pattern of points and curves for cesium, iodine and tellurium are similar. The differences are due principally to differences in the j chemistry of these fission product elements. The similarity of pattern in.
1 dicates other phenomena independent of chemistry, such as aerosol behavior, are quite important in determining release. }
' 2. Only calculations on SURRY AB cover a sufficient variety of con-
! tainment failure cases to permit drawing an approximately complete RF vs. TCF curve. For the cesium and iodine graphs the solid curve is drawn as in-between the results of SURRY AB Volumel and Volume V calculations while , the dashed curve is an approximate fic of the SURRY AB Voltme 7 results alone. These latter curves have a shape similar to that given in the tellurium graph which fits both Volume I and Volume V results fairly well. There appears to
- be an anomaly in Volume V results in that the AB-T (TCF = 152 min) fractional 1
l release (0.16) is higher than ABf fractional release (7.1 x 10 2),
- 3. Two reactor plant, accident sequences had only two points for drawing a segment of the curve. The others have only one. This is inanediately I 3- l
\
4 \ l
._ . _ -_ _ _ _ _. - _ _ _ . ~ _ _ _ . _ . . . . _ . _ . _ _ . __. _. - . . _ _ . _ , . _ _ . . _ . _ _ . _ _ _ _ _
. L.R.Zumwalt l 10/27/83
, i seen by inspection of the graphs. '
- It is recommended that this type of graph be used to present BMI.2104 results. Also, it is a useful graph for program management
- to follow re.
sults, check for anomalies and indicate possible need for additional calcu. lations to be made on the . variation of total fractional release with contain. - ment failure time. l e l i ) I J , 4 1
** e 4
4 4 4
.G.stu M RELEA S E VS CONTAINriENT FAILURE T1M E 4 l ) ! l l
. I .L
& 3. . _ .
7-- . Z'
' ~ %.6... ~-----. ._
- S U R RY ~4 e ......i r10-' , . __.
ss'u.fepp x (. .... ...su k RY T LM S 11 . . .
.r=._ _=. ,
i . ,,, p
. _ _. 2 . .._
. . . . . . . =. ..
,c . .- .
Z. e.
- 10. ._3. ..O$ E O,U nv4 M TLM ee e a e
,Y . . _.I _ _ --- -
5 EQUOYA H TLMB, - x----- Q.. gz G.. I . . . . _ . . . _ . . . . _-3 . C
.._y i. . . -"10 CC _. ..
e .i.ug.a . p. _
. p e -+->...e-4.*ee. en
_ ,0 <. g , W. f LE GE N D rI
.l.O~c b.1d
. A ccioENT S EOU i B' CE 210 N 'TLMB -E G- SURRY AB
~
IV . A SURRY T t.M.B, -
- 10 W SURRY $s D ,
y
, O ZIO N TLMB
__ q o- e
$- S' E Q U o 'f A H TL.M b .. _ .
Y5 _ fpSEa.UOY'AH TLM .
. Ml-7 / o
%. .I,. ., ,.
,., 9 . .. .
. . _ .. o. .. . .. .
l . .
.5.5h 10 .3 O aoo 300 10h 0 3Od'b) 5 ! . ..__..._,...l _-
l .. l [ [ l .........l._. FAtLU R E. T'lM E (MIN U T ES ) ,
TUDIN E R E LE A S E V s CONTAlN ME.N i FAILURE IME
- 1. - .
h . a ... l .] .r l h .. .. D. ., .. ,
. 8 .I
- m. .E
'L -. , . . . , . - - .
L -
' ! .a...,-. . . .L. . . . . _
. ~ _.
' I . .' .. .; .4.J . . . . . _ .
__.T a. _ . p. * .,..;..;...
. .. l
,..- .1 . . . s. . . . . . . . . . . .
i ' .4 . 4.4 __ ;- . ___ . . . '. , .. .o. .
.4 i J. ..g y. '
j . 7 ,a N. , , i 1
.L .j._..
-. q..
.i
- m. ,,,. ...
9' .; , , . . .. , . .
.l .. - ...,.
, i I f ' '
L.J._ .
'.4.
* *. * .s 4.L4. L.J.. .....:..
- m . . . . . .
, , l ._ .. . . . . . .. ..
.4. ,W . .
_ m..m .
' 4.;._. *. ..a. . . . . , . . _ .
J..... . .. . 4{ .. . . . - _ { .
~ . . ~ . . . . . . . . . -
' ' ' ' .e ; 4.. . . . . . . . . . -
........,q'
' I ' ' * *
j ,' d.... ' ..E'.' -
+ .4 - '. . , -.
4 I .- . . . . . .. ,
'- "- .. _- SEQ u o f.AH.TtM S.
' ' ' I __ - ' . '
.. ., . . . _ . . . '-_ . . . . . . . . . _ . - .4-. . . . . . . ..
' ' L.... ....6- . .. .. . . . . . . - - - -
' i '
n. 4_ . ... . . l -
.. . . ] . .
M',, ' y_
-....a.-.
._e.
1 --- . . .\. . .. . I'l .. f ' ' _... . J. . .
,,,ll ',' '
{
.1
}
- . _ . .,.J...-. . . - . .
{'Q={ \- , y.,
=
f1 . - . . . . __
- ..j ... _ .. . . . . . . - . . . . . . .
. 5.
.. ___ L . .. . . . . < . . . ..
. ..k I'I _ . . . . . . _ - .
p._ _ _ .,, lQ. .
.. e .
'i LMB-F
. . . _ _ _ . n
- u. ..
- - z,,,.1 O N _ . . . ao h g ... . -. ... . . . . - . -10 . . . . . .
[.f
~ . . . .
~
S U RRYSQ.E
- i. .- .
L. . _ _ . - . _ . . . . _
.1..g:.....
L ._;
.. 6 . . .. ':. . . ..
. j. . ; .. . ...i..._.. . a 6. .. . _
' ~
. . . l . . ..t . ; ; . ._ .
. . . . . .... . .. . . .. . ._: l'. -.. ...
. . 4., .--- .
L. . .; _4.J
. ~ ;. . . . , . . . . . . _ . . , .
. . . . . . . . . .. .. . .._ .L 1'_
. , _..-..J,.'...:....... '-...J.._. I g -9 ..._'i .
i . ,
.p.. . . . . . . . . .
....7
.....-_,L. ._
.V' i
.1_ 1
- ; . ,. . ,_.. ...s -. .-- .
. . '. :. r ;
. . ,' .'. ; L.. - L4 a ' I
'1 ,:
_.J. s ; . '. .
..;.J . .L, . ; . -
..i., .
_. .& : L . .. , .... ...... . . . . , i.........i...
.....;.......a..*
.. . ......-.-{1--.+--.'-~.
. . .7....... .. ... ,.f.
i 1. o._.;.., .... i - 300 1000 i: +s3 0. - ... 10 0 -. . p . ..., y __ _.. > . [. . . .
...l....
_ . . p. . .. . . . ..q.
.. . .. .....:. . . . _ . .3.0.. }00 EARVRE I.1LE_(MIN.UTES) -
TELLURIUM RELEASE VS. CONTAlNMENT FAlLUR E TIM E
._...4 .
s.
. . . . ..b. .t_ _ . . . _ _ _ _
-I. m _ . .... S U. R R M _A. B _u. - . r__ .
.a 1
_= - _.._1.__.._.....
.__.____-S M M, l
l .l ;1;.t _ . +' i - ' ' 2 --i ..*_ _- *
*s.
He ,y'
- S U R. R L. _. 45. 24.Lg.
i- l .
* .s -v . .
.- i
.x
~
s s
\
i
.. 1 . . . . . X. N ,,,, ._ ,, _
'_ .. 4
' .j!:_ , _ .. . _ . .:..__
- -vr
- r. . . . .
g... g _ _. ;
- TO_g_ .
.___m.,_
_a. ,. YZ l
..?_seouo.M. .A- TER$x._
. . . =
m 1 8 ' __-A- - 4_ 3 --._.. i . O .__. . . _ -. .
.. __ g -
F, .-i . _._ . . dS EQU.OYA H-TLM_. _ _.. ___ _...... ~ ,- - - __ . .. _ +. .
%,,/ . . ._ . . . . _ . . _.,.
A .: ; ... .._ . .._ , . __L _s._ yQ ..
. _ _ _ _ _ . .r- .,.--
, . . . _ _ . _ . . . - _ . . . . .- ---- . -- Z i o N -T L M B -
.._..._ _ __ _ _ . _ . _.)
_.. ........_.._...p _Q . . . . .._.._.__._ _, C_
-{-Q. S__._<(/)
L_
- _ _ . . . ._..._... g ,
,,_.f Q.. O. _ . . .
-\07 -
SURRX_$,D -
. ... . . . . a. - {
_ _ _ .sr 6 . q_ _.._p..______.
.- . j.
. ... -. _.. . . . . . . . ... . , 3 ,
; . ~ . _ ,.; . , _ ;._... .. _ . . .i.. ..
-+ . i-*~---- j ,
,,. .. --.t------*- .. , ,
,- . ... . . n- e- . . . . . . , - - = : . .
.;--"- T- t-* ,
i i , ,...i . - - - . - : ;- . j ,
* - _..- . - * . .
- _ _ -q
. ;- - r-* :
, g. _ .
-_T*
_n i . -
,,q_ . .
_i - -*.- _- ,
= --
,,--1
; m--d.. -.---rT-. . r~" +. .
6 ., , ,
._._,.....,,.c,- .a n
. . ,. . 4
.__. . . . , ... c. . .
*. _ . . _ _ a . y . .,_. .
._ J. !0 ._ +:
- .J
. ._3'_.__
. . . ___ i .._., _.b.3 . . . J. 0 "._ . - ._...d, ...._O i ._. - ._1. ..O. , _.=__ .. ' 3.0. _, .,p0. . r.t_.0.. .....0_.. -..p. - 0. 00w
- 1. .
. 3. 0. 1r. - . FA l LU R.E_ TI M E IMI NUT ES)
- - - - - , , , e - - - - -
I O
+-
e i l I LETTERS ABOUT THE PEER REVIEW MEETINGS January 26, 27, 1984 C 4 t' \_' e
' ~ ~ ~ ~ ' ' - - - - - _ _ _ . . , __ _
l
I bbl'hhbilC r.htibCitiG LiG:3Gi,Giat6 ' i
; ' Wigshaw Lane Cutcheth
, Warriagton WA3 4NE D u Telex 629301
., Telephone Warrington (0925) 31244
; Extension 72o5 2nd rebruary, 1984.
Mr. M. Jankowski, USMRC, t Silver Spring,
-* Nashington 20555,
} USA.
8 i . t l
Dear Mike,
I BMI 2104 Peer Review Meeting, January, 1984 The discussion at this meeting covered most of the issues I wished to raise
! 'j more than adequately, however there are one or two maaller points which it 9
may be worth drawing your attention to at this stage: 1. In vblume
# V there seems to be some confusion in the nt==nclature for the TMLB delayed containment failure accident. For example, in tables 6.2 and 6.3 it is referred to as E and elsewhere as E. I am not ,
sure whether the above or below grdu,nd options for con *mia-at failure
' make any difference to the source term in practice, but in any event the
,r situation needs to be clarified.
- 2. Volume 7 leaves me in some doubt as to whether it is a hot or cold-leg
- V sequence which is analysed. Mike Kuhlman assures me that it is a cold-leg sequence, and the inclusion of' the steam generator in figs. 6-20 and
] 7-9 to 12 suggests that this is, in fact, the case. However, that being H so, I find it difficult to understand why the steam generator (module 6) 1 9 l appears to retain so little of the fission products and aerosol. This is in contrast to the S2D-t case, for example, in figs. 7-14 to 16, where the i steam generator is a good site for retention.
- , 3. Volume VI says that it does not analyse the V sequence for Zion because it is much less likely than for Surry. From the table on page 4-2 of Volume g VI it seems that the V sequence might contribute quite significantly to
.i : risk if the associated source term approximates to that for Surry. However,
.! this is not necessarily an argument for analysing the V sequence for Zion.
E. 4. I note that EMI 2104 does not analyse any severe accidents involving staan
.f$; generator tube rupture. Is this perhaps because the USNRC feels that "U.t there would be no new phenomena involved which are not being covered in 4
g'.. the existing <=1<n1=*4ans? . 4, 5. In 1ERGE/ TRAP-4ELT the corwinnsation of fission product vapours in the upper T plenum is assmed to be controlled by mass transfer factors only. In q - fact, heat transfer factors may well be important, (cf the trea* ment of
';- stama condensation in Miutcs, the results of which are then carried over into MlWA) . I cannot see this point acknowledged anywhere in the documentation ,
}*.{:A 6 f55W' ' 06 6United Kingdom 4 Atomic Energy Authonty L : .. /
\ .. .
for the meeting. It has been suggested that if heat transfer factors were properly accounted for the amount of CsI and CaCH condensation on aerosol might be substantially reduced, with a corresponding increase in condensation at walls. 6. We have been unable to discover from the documentation how NAUA obtains its steam supersaturation values from the FMRCH output. We cannot cosament therefore on the validity of the condensation modelling in NAUA. l I hope these comments are helpful. May I say that I found the meeting very interesting and stimulating at the end of a year's abstinencel It was also a pleasure to renew old acquaintances. Once again thank you for an excellent dinner. . i
~l Best Wishes, i
I Yours sincerely, f A* P f* ' g; V,* %
- t
; F. ABBEY O .
, c.c. _ _
i
;, N. R. Hayns, NST
~I P. N. Clough, NST
,l I. H. Dunbar, NST Files (2) e P
e
- b;
*i
- i'.
in..<
. rs =2
~,b
.: 4 N
- 5
_ y . - , _ .
(asflenten
~
THE PENNSYLVANIA STATE UNIVERSITY l fb 152 DAVEY LAsORATORY UNIVERStrY PARK. PENNSYLVANIA 16802 College of Science Depeammes of chemmay l 4 February 10, 1984 ) Dr. Mal Silberberg i Assistant Director j R====rch and Techaf e=1 Support U.S. Nuclear Regulatory Commission i 7915 Eastern Avenue Silver Springs, ND 20910
Dear Mal:
J 1 As a follow-up to our recent meeting and your request for general comments concerning issues in reactor safety research which still need to be resolved, I offer the following comments. These pertain in particular to the BMI 2104 analysis and I have made most of them verbally at various meetings. However, some of them are worth reiterating in light of some of the discussions which surfaced at the IDCOR review.
- 1. Perhaps the single most important issue which deserves attention concerns the chemistry of fission products following surface adsorption and possible desorption. In.my opinion, there is a great uncertainty in this area. First, one knows very little about the chemistry of adsorbed fission product species. In particular, the most stable condensed state is often not the same as the most stable vapor species, especially in the case of high-temperature vapors. Interactions with the substrata surface may alter the chemical form, and reactions with impurities in the or on the surfaces may i
play an important role. One would expect that the fission products once adsorbed would be less volatile and that the more stable condensed phase species, once formed, might also be less subject to revaporization, but this clearly needs to be investigated on a case-by-case basis. Second, a related ' point is that during reevolution, as the surface temperature rises the partial pressure of the reevolved species will lead to a just saturated vapor and it would, be intersating to follow the time history of this gas-phase parcel to ascertain where sufficient supersaturation would reexist for particle formation to be re-initiated. That is a case where nucleacion expressions may have to be invoked to do a proper treatment. 1 The various codes require suitable deposition velocities for the irreversible adsorption of such species as CsI, Cs0E, add tellurium species; a stronger data base is required for the adequate treatment of these, i gg y. ~ . c0 g(')GCET . ;.Q: s 3 -
*-?'- ., .a >. AN EQUAL OPPORTUNITY UNIVERSITY
) .
... _.--.--. . .... . - . - - . . . ~ . - - . . . - - - - - - . . -- .
1 2
%,, Dr. Mal Silberberg February 10, 1984 4
Pge 1 I I
- 2. In cases where fission products are adsorbing on wetted surfaces i or into liquid pools, an ideal solution assumption is often not justified '
in creating the condensation / solution removal phenomenon. Consideration needs to be given to issues such as the potential contact between such
! :' compounds as cesium iodide and borates (a point raised by Lloyd Zumwalt),
leading to iodine evolution as well as the potential alteration of tellurium
, and cesium hydroxide chemiseries, for example, following the adsorption /
desorption process. I
- 3. I still have a great concern over the accuracy with which the free energies of formation of various high temperature fission product compounds are known. In many cases these are extrapolated over wide ranges of tempera-ture and a related matter is uncertainties in the heats of vaporization. Not only would the free energy influence the concentration of thra sygies in the gas phase, but would also, together with the uncertainty in the vaporization thermochemistry, lead to a potentially big error in the creatment of con-densation and/or nucleation. A detailed assessment of the thermochemical properties of important species should be made and the ranges of uncertainties i
O h 14 6 t 7 4 i= icivier e te=1 =1 = ca che predictions of the overall model. It should be realized that an
=a 1 1*c =1=== i=
uncertainty in the thermody===ics of one species, can also influence predicted concentrations and hence reactions with others.
- 4. Another thermochemical property to which very little attention has been given is the activity coefficients of various fission products in either the molten fuel / control rod / structural material or in the melt along with concrete. The ideal solution approximation is certainly better than simply ignoring the influence of solution on partial pressures, but any sort of i
tam 14stic model will ultimately require a handle on chemical activity co-efficients.
< 5. A related matter is the development of a better model for the vapot-ization of control rod materials. According to Mike Kohlman, the Battelle model predicts silver to be up to fifty percent of the inert aerosol. Such a prediction could be in error if the partial pressure of silver, as an example, is suitably reduced by its interaction (hence its activity coefficient) with tHe alloy. Related statements could be made about other materials such as ein and zirconium.
- l 6. There are several uncertainties with regard to the chemistry of I
aerosols. First, experiments need to be designed for comparison with model I a=le=1=tions which take account of coagulation of various chemically unlike constituents. Related to this, is the necessity for handling the problem
- O e.~
- w. , .
! g...... M% S,4 . i 42 -. 5. -l ~ l
- m. ,,m .m
- , . ~ ~ . - ~ . . . . .-
'Os. Dr. Mel Silberberg February 10, 1984 Page 3 of potential cheatcal interactions between species as they co-cluster into an aerosol particle possibly 1==<thg to some chemical reaction and potential reevolution of certain vapor species, e.g., free iodine. As mentioned by otheres as well, a related problem worthy of some attention is consideration of the salt effect of hydroscopic aerosols as this night influence the con-j densation of water vapor on them. This would only be important where the , salt-to-water ratio is very large.
Another question is whether coagglomeration is an important process to crest theoretically. Consideration should be given to developing suitable experi-ments for testing model predictions on this matter.
- 7. There are still a number of questions concerning nucleacion in the
] primary system in the PAesence of a high radiation field; ion induced nucleation may assist in the process. Although not very accurate, there are models for creating the influence of ions and some calculations should be done to ascertain whether or not this could be an important process. h 8. A related question of nucleation comes up in considering the competitive process of water condensation on aerosols. Attempts should be made to ascertain whether, under very high supersaturation conditions and low aerosol concentrations, nucleacion of water vapor can occur instead of only water condensation onto parti _cles.,
- 9. In the context of water condensation on aerosols, one wonders whether or not appropriate creatments are available for concurrently changing the collision and settling shape factors in order to account for the changing geometry of both hydroscopic and hydrophilic particles. Comparison should be made between theoretical predictions and experiments made in NSPP. .
10 Some carefully designed experiments need to be made to address the entire issue of particia re.-suspension. This af c,ht be especially important in a case whera there is a. pressure pulse and a :.arge change in hydrodynamic flow due to vessel failure. II. Although it doesn't look as though the decontamination factors will ! maka a tr====nt== difference, the question. of the hydrodynamic properties ! of bubbles in suppression pools arises when one gives consideration to l surfactants. Attempts should be made to adcartain whether these changes could I be important in the case of " natural surfactants" that might be present as impurities in suppression pool sumps, s A./ - h is: .
; *}N-y m.
- 7. .: - -&
m e a. ,, n . -
~ T " **
O y Dr. Mel Silberberg February 10, 1984 Page 4 i I i i I hope this stammary of some of my comments and suggestions is of value. i Best regards. Sincerely, Y A. Walford Castleman, Jr. Professor of Chemistry AUC:bai O O M e e f 1
-Q , - . %. . u , * -
_ _ .e<- ,. --
~
__ s4- - - .
~
~
H;//W 1MMCE33Hailford ====.=m.=== f' 8450458 February 6, 1984 ' i i M. W. Jankowski Mail Stop 113055 Accident Source Term Program 0,ffice Office of Nuclear Regulatory Research
- U. S. Nuclear Regulatory Commission
! Washington, DC 20555 C39 TENTS ON BMI-2104, VOLUMES I THROUGH VI, RADIONUCLIDE RELEASE UNDER SPECIFIC LWR ACCIDENT CONDITIONS (DRAFTS FOR PEER REVIEW)
At your request I have reviewed the six draft volumes of BMI-2104 and have attended four of the five peer review meetings. I have also reviewed the supporting document on the status of computer code validation, ORNL/TM-8842. Many of my comments were expressed orally during the review sessions, but I an attaching a list of consents for your consideration. I b; I' wish to thank NRC for the opportunity of serving as a peer reviewer of the source tem reassessment program. I appreciate your position of actively seeking comments and your demonstrated willingness to incorporate them when
' possible. The NRC contractors (SCL,.0RNL, SNL, PNL) are to be commended for their good work. ) If you have questions concerning my comments, please call me on FTS 440-1584.
i i R. K. 1111ard l i Fellow Engineer I I dht t I
Attachment:
Peer Review Comments DOE /RL-AMAR - RJ Myjak (w/o attachment) KR Absher i i 00E/HQ-8TP - Sr. Division Director, Safety and Physics Asst. Director, Safety and Physics USMtc - M__c " L. -. T FRS/TMC-ANL - DR Ferguson jh l L Baker, Jr.
..,f N ~ M
, mustue Conseaws umeelemy of Wuessnesse Doest. Cup / Opemens the Messes tasseenes essespuest t,aamesary fu the useet 4
_--* ~ 3~ _
I 1
. s ATTACHMENT TU 8450458 CON 4ENTS ON BMI-2104, ORAFT VOLUMES I THROUGH VI, l
"RADIONUCLID?. RELEASE UNDER SPECIFIC LWR ACCIDENT CONDITIONS" R. K. Hilliard February 2, 1984 i
- 1. ' The . methodology appears logical, although interfacing 8 differ.ent computer codes seems cumbersame and increases the chance of input errors.
i
- 2. The present approach uses newly developed computer codes that are by and large unvalidated. The ORNL review of the status of validation is very hel;ful in describing the code bases and state of validation.
However, until adequate validation has been achieved, the calculated source terms should be considered tentative.
- 3. The documen'ts have very little interpretation of the data, comparison between volumes, or other insights obtained by the authors during the performance of the work.
conclusions.would be yatuable. An additional volume providing overall t ('
- 4. The QUEST uncertainty analysis appears to be little more than a i
sesitivity study. Thagh interating, its valu is Hatted unten the
?
true uncertainty on the end product can be quantified. The oral l presentation, although given by excellent speakers, was much too rapid
! and the material too new for me to evaluate it adequately.
- 5. Only Volume I gives the reIease' to environs of FP species other than I, i
Cs and Te. A similar treatment for Ba-Sr, Ru and La in the other 4 volumes would be useful.
; 6.
i More comparisons of calculated source terms with WASH-1400 would be { informative.
- 7. An important removal mechanism, impaction, has been omitted from the TRAP-MELT code. This mechanism may be dominant for some conditions (flow normal to surfaces, bends in pipe, etc).
- 8. The effect. of FP decay heat should be accounted for, not only for the effect arr thermal hydraulic conditions, but for possible melting of steel pipe or components and revolatilization of deposited material.
This topic. was discussed at length during the last peer review meeting, and there was general agreement that it could have an important effect on the release to the environs.. It should be studied carefully, i i
. s.
g ( I"
,w~ -- - ---
-c -
me . m g - -- "
- .m,,-r- -- , _ - -. , _y, _ _ ~ , - . , - - --__ . -
w- --_ ._._.,. ,_,,_-,.-,, . - . ,
.l
. l 1
ATTACHMENT TO 8450458 Page 2 h 9. Properties of tne gas mixture (H , steam) should be used, rather than
- ' those of pure steam. Agoodcomp$1ationofmanypropertiesofmixtures is given in ANL/ENG/TH-01.
j
- 10. The presence of liquid water in the flow path is likely in several of i ,
the sequences analyzed as dry systems (TML8' , V). A more realistic analysis should be made for these cases. ' 11. Aodine is assumed to exist only as Cs!. There appear to be suitable conditions for some formation of I and organic forms. High ! tamperature exposure in t!)e presence of kxygen and reaction with boric acid are two known processes for I g fonsation. The effect shoulW be evaluated.
- 12. Hydrogen may burn as a standing flame at the point of release to the containment atmosphere. The effect on aerosol properties and behavior i
4 should be evaluated.
- 13. The deposition in containment leakage paths has been ignored. Although difficult to quantify, some assessment of this potentially important effect should be made.
- 14. The effect of ionizing radiation on FP behavior has been completely ignored in the codes and in the ORNL assessment of code validation j
status. The effect may as negligible, but should be addressed in the ORNL/TM-8842 document. O
- 15. For the SURAY V-Sequence, it appears that the break point in the low 4 pressure interface piping would be under water at the end of the blowdown. This would have a major effect on retention of radionuclides. When the grimary reactor vessel melts through, water l
j would be. sucked back through the piping and PRV to equalize pressure in the initially sub-atmospheric containment building. What is the effect on release of fps to the environs? t ' 16. Based on my experience, I believe that the initial particle size used l in TRAP-MELT is too small. A mechanistic code, such as the ANL RAFT code,may be useful in predicting initial particle sizes. i 1 l
- 17. Uniform chemical composition of all aerosol particles, independent of {
~ size, is assumed in both TRAP-MELT and NAUA codes. This may not be l correct and l i
the effect of nonuniform agglomeration should be I evaluated. t ! I i ! 18. The assumption is made in both TRAP-MELT and NAUA that settling is due to the difference between vertical convective flow and Stokes settling l velocities. This is not. a valid asstamption for flow in boundary layers near horizontal surfaces, and the uhcorrected Stokes velocity should be used (with Cunningham and Klyachko corrections, as appropriate). 19 Resuspension of deposited aerosol in the RCS and containment has been ignored. The importance of this phenomenon should be assessed. O i , r. a ww.;r m .m;.:na-ms:- ' ? .i .- - - ~ - J.L.%..,u,., a ,. =,w_ -- -
- ~!
- - - - -, *** 3*-- "~' .
}
. . l ATTACHMENT TO 8450458
, Page 3 p .
'v
- 20. Volume V, p.
6-68. Table 6.2. TMLB-4, should probably be changed to
- 21. All cases analyzed asstamed no operator intervention in the course of the accident. How is operator action addressed?
- 22. the A good editing effort would greatly help the reader in understandinI contents of the six volumes.
23.' source Good progress teries. has been made during the past few years in assessing the for the accuracyMuch of themore work magnitudes. remains before confidence can be claimed predicted f e 5 4 6 6 I I I l . 4
)
- 1 i
i : j E
\C ;
i : 1 4 l - - l t l i I 4 - 1
's i
i b
- O O 79.. '
~;
~~.m.'. - L a w.n.. .,.o - _ - . -
- - - - - - - - - - ~ .
.m*
j
# f
.~ UNIVERSITY OF CALIFORNIA, LOS ANGELES . . UCLi 1
'.%f BEREELET
- 04Wi$
- IRvlNE
- LOS ANGELES
- RIVER $1DE 14M DIECU $4M FRANCISCO f m : e.;
$44T4 84RB4AA * $4NT4 CRL'2 mye MECHANICS AND STRUCTUllES DEPARTMENT sCHOOt. OF ENGINEERENC AND AFPUED SCIENCE LOS ANGELES. CAUTORNTA $1024 i
; February 20, 1984 I
Mal Mi1F M - y Arr 4delt Scurce TorIn Wam Mfim
- Mffm of } air 1==r Regulatrzty Research l W-h4=Jtzm, DC 20555 i
j
Dear Mal:
i I have revimund the ma*=einta handed out at the Peer Review Meeting j h., cm January 26 and 27th,1984, as well as my notes and wish to ccament i I as follone: t
-! 1. Battella coli=*=is fezzk. In listaning to the Battella presentaticns, l I still get the feeLux; that they are so busy ruming < =1m1=tisms and +
r ,---imy reports that they have not had a chance to reflect cm their
- - - -- .1 4 ' d s. For example, it is d4**4m1t to tall 54-G c changes-in the source ter n are due to diff==. s in nrvinia, in data or due to plant changen (e.g. the ccncrete m ---:+1*4m problen) .
I also had the feeling that nony of the r r===sts nada at and folicwing the October peer review were not censidered. W4-11y, the report cm the Ice Ccndenser (Sequoyah) appears to be
=4=1==d4=y with respect to the p i=1 for the ignitor system <-an=4M failure. 'Ihm report gives the 4 ---s=im that the ignitors will cause cen+minnent failure when tryac is generated. It is not clear frcan the " report that this ney m1y occur for rapid hydrogen generaticn and a parti-cular range of hp -,--dair ratics. In short, it sounds like the ignitors do unre haza than good.
With respect to the BIGH9ERK III ochtaimeent, I was pleased to see that pool-bypass is being ccmsidered, and I await the results. a r-V
~
i {5o }/L] / 6 0
. , . n ,w.u., . ..w.,~.
- a. .. r . -*. ... .
. . ~ . .- _ .-- . - - - - . - - - - - - - . .
e- " - - - "
--. . - . - . - - . - - ~- - - - . - _- . _ - - - ---- - - - - . - - .- -
f Page 2.
)
O Intter to Mr. Mal S41P- :-w (ccn't.) - February 21, 1984 i I Iastly, Di Miaman of Stma and Webster inada a rather cmvincing argunant w----mim the presence of water during a PWR "V" seque1ce. Battella should address this issue and d4=m== its inpact cn their surry and zicn work. ' 2. Sandia Work. Sa Sandia progrant (QUEST) has dev=1rTad into a j
;-;---- '=11y %=d. meer**4= since the October meeting. I still belima study and not an w however, that it is a very well defined and eascuted sensitivity A analysis. Although it is d474"mit to assess the implicatims of the study without the final report, I have the Fn11 =4 5 =v- --*4= with respect to the code igut uncertainty.
As I understood the presentatim, this part of the study has as its focus the effect as the source tem frca "different, but resscmable, inputs ; to the codes". It would be helpful if the Sandia team could determina the "h=~4= y" between reasonable and unreastmable igut. For =='Ti m,
,' are there physical or chandcal limits to the igut data? And if so how I
do the source terms change at these limits. Alturnatively, are there sets of igue data that make the " source tems" i @ysical, or sana j of the system varimhlam (e.g. pressures, Ptures, etc.) non-physical, His ----- --+ would give sea well defined 1tnits to the range of outp.tts due to igut "i.z J.nty". i i
!{ Mith respect to ?- - f ar ical s W.nty, the ranges ccmsidered are not as clear cut. I amt ecnomened with both negimetod phenomena and uncertain phanczuna. Phancaena Amnefa as a result of both i@ut condities
!' and r w @ . A case in point is the transition fram 1==inar to m'1=rst flow. If con has a code 1ddch caly ~1~1=+== laminar flow, you- can change 1
the input data, the ocmditions7 vary systna paramtars etc., but still get 4 the wrceg results h=r mi== scusa F =- -- is =i ==4 m, and this , , +- { won't ecover it. One has tz know =ar4a*4 that fbr a given set of conditicms, the flow regina cban,yme. I don't see this type of +4nir45 refiscted, as { I. listened to the r : a :==. ! <anarmi I =~ai'= the various -A placed en the prmemn, ine tivi 43 tian and acmay. I urge you to allow your ccmtractors scans tima for j reflectim cm the work they have generated. i 2mnk you again for the ei mity to be a part of this effect. i SADCsEElya 1 wu!wa. -
- 2. m -
Profesace of F*r4a=== 47 ,,14=d Science i i NsM
- i F oc: R. Benero, 20C (AS'IPO)
I D M. .Taniernamiri, IGC (ASIPO) i ! C. Ryder, NIC (ASIPO) ; O
- t l
i I
= . . . . - - . _ -
l __
Kastonberg
, UNIVERSITY OF CALIFORNIA, LOS ANGELES UCLA l BEREELEY a DAvts = IRVINE = Los ANCELES
- Riv ERSIDE
- SAN DtECO
- 54N FRANCISCO
- SANT4 SASB4AA S ANTA CRUZ ntECHANICS AND STEL*CTURES DEFARTMENT SCHOOL OF ENGINEERING AND AFFLIED SCIENCE LOS ANCELES. CALIFORNIA 90014 i
! i Petruary 21, 1984 1 1 - l f Melvin R41 6 8- 3
- Accident Scurce Term Fr-w Office nerim of pv-in c wit =*niy Research U s. Mv-1 =mc pm,=i1 =+,wy ec==n ..i c.s 4
==*4 = ton, DC 20555 i
Dear Mal:
Ih11r=ri7 my ataandance at the NRC/IDO'R meetirry at Fissial Product Balease and h= met (Fisi uny 7-8, 1984), I have the fnlieweim em===1ts tahich may be relevant to the NRC Source Term FD y lo i
- 1. As I mattimarf in my 1stter fn11rneim the October peer review, I have sana**11"'*4'="' concern over the ie-L F4==4= rAi., beymd mainm,
; iodh and 121 par *4N1='" ZuthW1i1Ea, strattilmt, harium and lantharn=- I still <===*4m "othor= category. Daare is alib the r- .li - of:5ti-im theen into the the =w w.
qcmtim of the treatment of silver, as well as a heet of other elements that ultimately go into the aerosol (e.g. Ng, Cd, Ni) . I believe thans cmsideratials are i ! iw-L for both the Battelle facek, as well as the NRC Staff reqx3rt tdt.tch will h*= a ate the results, and provide a risk g-- 2 -+1iive.
- 2. Die ===*4= of the re-evnh*4m and resuspensim of 74-4m p *- L-plated out in the prianzy systuus is still open. At present, Battelle ,
does ret a:nsider this phenamuna. For same scenarios, this can be an iwAL Mh*4m to the source term. Die relative timiny of vassal faiinre with respect to cuenh==t failure is iwi.mut. I would re-cannend a limited, but well 4---b- out, set of <-=1-1=*4mm by Battelle i which could put a r-- ,+ dve en this.
- 3. 2:a degram to idiich the vessel is r A -i at 7=41"= is an iw6L mnmideratim in aerosol and pressure generatim in the ccm*=4-t.
Die influmnr= of the Sandia work (under, D. Powers) at the StI u.yu,.i is not clear. Itare ie-i.ly, it will be ici-t in providing 3. gut to
- the staffs' risk g-- - _iive.
1 Jgynys e k ~
'~T- ~~
~
\ . - - - - -- -- -- - - - - - - - -
Oc Page 2. February 21, 1984
. Istter to Mr. Mel Aih-M+g (ccn't.)
- 4. One of tM nest iw-d. uncera4*4a= is in describing the core malt-dows and vessel perittration processes, since this drives tM rest of the c=1m1 *4==. mis uncertainty should be a major focus of the Sandia QUEST prmJmn-
- 5. mere is still nucts irmi inty ra;=M4m ccmtairmant failiiru.
Se BCL work crm=4 dam structural failitre. Others place f.J.s third behind failisre of seals and p--Laticns, and eqni t =rit hatches. It is still not clear hcw the two ccxra4==nt wodcing groups (failure and r ~ = - +) wil2. feed infonnatias into the source term project. In i I additicn, hir wodt should feed into the QUEST gw. Battalle still has not made an effort tu e- % the sensitivity of cort =4==nt failure as the " source terms". In ==mry, there are several areas of urmL2inty Msich should be dealt with in greater detai.1 by the Soum Program. O S y,
-ta r . m ' W.E. Kastenberg, Professor of h74MM and M 14ad Scis ce ! wp:/sta i
[ cx:: R. Benero, NIC (AS'IPO) 1 M. Jankowski, NIC (ASIPO) C. Ryder, NRC (ASTPO) O _ g ,_ . ,wy, ,
---m - - ' , - , - -- =-r a 14 * "
w :,w m,,- -
- - . - e
Kelly UNIVERSITY OF VIRGINIA sCHCol. OF ENGINEERING AND APPUED SCIENCE v u CxAnunrEeviu.E. EE . DEPAftTMENT Ofr NUCLEAR ENGINEERfMS AND ENGINEERING PHYSICS YgLEPMONE SO4 354 73 3g REACTOft FAC3UTY February 2,1984 Mr. W. M. Jankowski Accident Source Term Program Office Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Conaission Was5ington, D. C. 20555
Dear Mike:
Here are a few connents that I have about the material presented at the last peer review meeting, as well as at previous meetings. (1) The methyl iodide problem has still not been addressed, nor has the effect of a strong radiation field on the chemistry of the various species. (2) In treatment of flow in the upper plenum at Surry, it was assumed that the domed region above the upper plenum was essentially isolated from the upper plenum. Is this treatment justified in all PWR's? O (2) The intact RCS ioops can serve as effective heat sinks (and fission pro-duct sinks) as degradation of the PWR core proceeds. Is this effect significant in reducing the release of fission products from the RCS? (4) How sensitive are the releases _ in an AB sequence to the size of the assumed hole in the piping? I understand BCL assumed a hole of 8-in. diameter. (5) Wannan's point about the crucial nature of timing of the resuspension of fission products and the failure of the containment was interesting. This point needs to be studied at length. (6) In Volume IV, how are the DF values in Table 7.5 calculated? The foot-note is ambiguous and the values for 145 and 160 minutes seem anomalous. For Table - 7.7, the DF should be defined in a foot-note. Why the discrepancy between the values of DF for the TMLB' and TML accidents? (7) I would like to see a sunnary of the releases to the environment, as estimated by WASH-1400, NUREG-0772, and BMI-2104 for the spectrum of accidents examined. I look fonvard to learning what ID,COR will come out with. I have enjoyed 3 participating in the review meetings. - Yours truly, h / Kelly, Professor J. L. Dept. of Nuclear Engineering 1 l and Engineering Physics ! Ld:ph W
- e, . .
\
#el/y i l
TEMPERATURE DEPENDENCE OF FISSION PRODUCT RELEASE RATES James L. Kelly, A. B. Reynolds, and Michael E. McGown, Department of Nuclear Engineering and Engineering Physics University of Virginia Charlottesville, Virginia 22901 December, 1983 In NUREG-0772 [1] and s'ubsequent " source term" reports (e.g., NUREG/ CR-2629. [2_] and BMI-2104 Q]), the release rates of the fission products from the fuel prior to core slumping were estimated by the Albrecht.-Wild.model [4_] . . In this model, a fractional release rate,'K, is defined for a fission product such that the amount, M, of'the fission product remaining in the fuel after a time, t, is determined by h=-KM. (1) Defining F as the fractional release of the fission product, so that F is. C eRea1 to 1 - wM , here M is the initia1 amount of the fission product pre-o o sent, Eq. (1) gives
~
F=1-e . (2) In the Albrecht-Wild model, the fractional release rate, K, is a function of temperature, namely, ' K = Ae . (3) L Values of A and B have been determined for a number of fission products by fitting Eqs. (2) and (3) to experimental data. The model was further developed in Ref. 1 in which the overall temperature tinge was divided into three parts (800 C < T $ 1400 C,1400 C < T $ 2200 ', T > 2200 C) with values of A and B defined for each part. More recent values for A and B are presented in Ref. 2. ,
. 1 C
b -
. \
In reviewing the fission product release data reported by Albrecht et al. (Kfk) [4-7] and Lorenz et al. (ORNL) [7 _15,], it seems to us more reasonable that the fractional release rate should exhibit the usual Arrhenius temperature dependence of the form K = K,e 4 RT , (4) BT instead of the Ae form used by Albrecht and Wild. Q may be interpreted to represent the activation energy for the rate-controlling release mechanism. R is the universal gas constant, 8.317 J/mol K (1.986 cal /mol K) . To test this idea, and to determine values for K, and Q, published release data for each fission product were plotted as in K vs 1/T. For the most part, data from experiments conducted in an air or inert atmosphere were used, although a few of the data were obtained in a steam environment. (l Albrecht!s data were obtained at the SASCHA facility [6]; Lorenz's HI tests were based on fission product release from fuel irradiated in the H. R. Rotiinson PWR. . The results are presented in Figures 1-13. A single straight line was selected to represen*. the data for each plot, and in most cases one line correlates the data adequately over the entire temperature range. The corresponding values of K, and Q for each fission product are reported in Table I. Based on the indicated Q values, the fission products may be divided into two major categories: Kr, Sr, Mo, Ag, Sb Te, I, Cs, Ba Q = . 200-300 kJ/mol Zr, Ru, Ce, Nd Q = . 700-1100 kJ/mol For purposes of comparing the Arrhenius release rate model with the , Albrecht-Wild model, curves corresponding to the latter model are also shown (as das,hed lines) in the figures, using values for A and B reported in Ref. 2.
+
I' - - - _ - - . ..____}}