ML20207A334
| ML20207A334 | |
| Person / Time | |
|---|---|
| Issue date: | 07/19/1985 |
| From: | Hopenfeld J NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| To: | Shotkin L NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| Shared Package | |
| ML20207A169 | List:
|
| References | |
| FOIA-86-678, RTR-NUREG-1116 NUDOCS 8611100219 | |
| Download: ML20207A334 (8) | |
Text
- _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _
v.
~
JUL 191935 MEMORANDUM FOR: Louis M. Shotkin, Chief Reactor Systems Research Branch Divisions of Accident Evaluation FROM: Jorani Hopenfeld Reactor Systems Research Branch Division of Accident Evaluation
SUBJECT:
STEAM EXPLOSIONS
References:
(1) Memo fm Dircks to Asseltine dtd 06/18/85, " Steam Explosions" (2) Memo fm Asseltine to Dircks dtd 04/25/85, " Steam Explosions" ,
(3) Memo fm Kelber to Distribution dtd 03/14/85,
" Additional Information on Steam Explosions" (4) Memo fm Telford to Curtis dtd 12/18/85,
" Highlighting of the Steam Explosion Review Group Meeting" (5) Memo fm Hopenfeld to Bassett dtd 01/15/85 " Steam Explosions" Reference (1), which was prepared in response to Commissioner Asseltine's request (Reference (2)), states that, "the staff believes the review in NUREG-1116 provides an adequate basis for the current position;" i.e., the probability of containment failure from steam explosions is very low. This is to bring to your attention that not all the members of the NRC staff appear to agree with this conclusion.
In Reference (3), Kelber points out that the SERG report may not have included considerations of all relevant factors. Additional information was developed during the review of the draft of the document, but was omitted from NUREG-1116 because the latter was limited to reporting only on events at the SERG meeting.
Kelber recommended that the additional informaticn be given careful consideration before taking any resolving actions.
In Reference (4), reporting on the SERG meeting, Telford points out that the SERG members were not certain at the meeting about the " upper bounds" and "best estimates"ofcgntainme failure. The upper bounds of the best estimates ranged from 10- to 10-1 In Reference (5), I pointed out that none of the SERG members provided an l estimate of the probability of containment failure, which is technically l defensible. It was also indicated in the reference that when model predictions are compared with large steam explosions, there is no agreement between theory and experiments.
w
> t9 MA-af.-Qs Q 8611100219 861030 PDR CURRAN 86-678 FOIA PDR o
.c
. Louis M. Shotkin - 2-Reference (1) points out that the author of the SNL report disagreed with many of the statements of the SERG experts; the fact that some members of the NPC technical staff.also disagreed with the SERG is conspicuously missing in that document. -
- /
P ~'
,?,
.' ' g , .
o ?^. an Joram Hopenfeld Reactor Systems Research Branch Division of Accident Evaluation cc: R. B. Minogue, RES D. F. Ross, RES
- 0. E. Bassett, DAE W. M. Morrison, DAE .
R. Curtis, CSRB J. Telford, CSRB M. Silberberg, FSRB C. Kelber, RES C. Allen, CRBRP0 -
J'.'Rosenthal, RSB ,
f P
Distribution '
Subject 'e Chron, '
f Cir ' '
Branch R/F ,
RBMinogue DFRoss '
OEBassett WMMorrison WDBeckner JHopenfeld -
JHopenfeld R/F
- JTelford < ~~- -
RCurtis .
CKel ber , ,,
DAE:RSRB
_> ? h DAE:RSRB DAEi CSRB fflY DAE:CSRB wi}>
DAE:RSRB$ti- DA:SP, Q JHo'penfeld:kw:kd:bsd -
JTefford~ RCurtis LMShgtkin CKilber 07/;g/85 WDB[eckner 07/h85 07/ /35 07/ f/85 07/1985 07/.;/85
+
=
f i
n
- r l 9 m/* '
L ?... . c:' . . - __.
w fl+;< V *f .
Fd M. n e I W3^
f-4 1.4 Molten Fuel-Coolant Interactions Procram (B. W. Marshall Jr., 6427 M. Berman, 6427) , ,
n .w, The objective of this program is to develop an understanding of the nature of Fuel-Coolant Interactions (FCIs) .during hypothetical accidents in Light Water Reactors (LWRs). FCIs can occur in the core region, in the reactor lotter plenu's, or in the cavity below the vessel. They can occur when melt .1 falls into water or when water falls into melt. The under-st,anding of FCIA achieved - in this program should be suffi- s cient to resolve the key reactor safety issu'es for both terminated and unterminated accidents. Models are being developed to determine: ,
- 1. The rates and magnitudes of steam and hydrogenhen-eration due to FCIs; - ,
.s
~
- 2. The probability and cons'equences of direct contain-ment failure by steam explosions; 4 .
- 3. The influence of FCIs on accident progression and the _ nature of the source term (including fission-product- chemistry, release rate, particle size, and dispersal); ,
- 4. The consequences of pouring water on the melt in X order to terminate an accident; and
- 5. The characteristics of the debris produced by FCIs, including particle size distributions, porosity, and coolability.
e 1.4.1 Current Progress and Technical Highlights 1.4.1.1 EXO-FITS Experiments >
(B. W. Marshall, Jr., 6427: M. S. Krein, 6427)
No experiments were conducted at the EXO-FITS fability. The evaluation of the pressure transducers'and mounting hardware used during an FCI experiment began and is..
'. described in Section 1.4.1.2.
~!
1.4.1.2 High-Explosive Simulation of a Steam Explosion (B. W. Marshall Jr., 6427; M. Krein, 6427: J. Fisk, 6427)
The experimental . evaluation of the pressure transducers and mounting hardware began. Preliminary tests were conducted in an effort to-distinguish the appropriate charge size for the experimental matrix. Evaluation of these first- few tests (in a 56-cm rigid chamber) revealed significant variations in peak pressures radially at an axial position.
The times at which these measured peaks occurred, however, appeared consistent within a few tens o f; microseconds, s
y W
s q
i
. i-implying that the charge was most likely in the center of the enamber as desired. This leads us to believe that either gauge mountings, gauge types, or the experimental configuration of the high explosive were yielding these anomalous results.
Two ongoing efforts- to help diagnose these problems are (1) CSQ will model a top x-y view of the water chamber and prima-cord mandrel to determine the effect, at the gauge locations, of burning one wrap of prima-cord around the ma'ndrel, *and the ignition location will also be changed relative to the gauge location to determine if this influences the pressure magnitudes and response time; and (2) Experiments will be conducted using one strand of prima-cord (PETN) down the center of the water chamber.
These experiments will allow CSQ to model the experimental setup precisely since this is truly a two-dimensional problem and the equations-of-state of the PETN are known.
Results from this work will be reported as they warrant.
1.4.1.3 Iron / Alumina Thermite Evaluation (B. W. Marshall Jr., 6427; G. B. St. Clair, 6427)
During this reporting period, we received results from x-ray diffraction as to the oxidation state of the iron powder used in the iron / alumina thermite. Qualitatively, the results indicate that the powder as received from the manufacturer is primarily FeO with lower levels of Fe304 and very small amounts of Fe2O3 When the powder is baked at 800*C, the oxidation states of the iron powder change to essentially all Fe2 O3 with Fe3 04 also being present in much lower amounts. We are currently having the same samples analyzed using M5ssbauer to verify these results.
These results are very preliminary and may change depending upon the results of other analysis and observations.
, 1.4.1.4 Analysis of the FITS Experiments (O. Seebold, 6427; B. W. Marshall Jr., 6427; M. S.
Krein, 6427)
Work continued on the FITS-5D data reduction report. The experimental gas phase pressures recorded during the FITS-SD experiment were compared to the Hicks-Menzies theoretical maximum final state pressure. This comparison indicates
! that the maximum possible coolant-to-fuel mass ratio was about 9.6. The maximum mass ratio from the experiment would be 19.9 if all_the water and fuel participated in the FCI.
Therefore, based upon the theoretical calculations, we can bound the mass of coolant that participates in an FCI for a I ~
given fuel mass (i.e., the mass which was delivered from the crucible to the water chamber). In the FITS-5D experiment, we have justifiably reduced the maximum possible mass ratio from 19.9 to approximately 9.6, a factor of approximately two.
r M
yt Five of the liigh KXplosive (HEX) calibration tests have been transferred to the VAX and the data have been reduced and plotted. Evaluation of these data is currently underway and will continue into the next reporting period.
The data from the 50-kg MDX-2 experiment have also been transferred from the FITS site computer to the VAX in Area V. Work on this test will continue as time becomes available.
1.'4 .1. 5 Modeling of Explosion Propagation and Structural Loading Using CSQ-II (K. L. Schoenefeld, 6425: M. F. Young, 6425)
During this bimonthly period, the basic analysis of the ,
FITS-3B experiment was completed. Using the run which produced the closest fit of the water-phase pressure data as '
the base case, input parameters to CSO-II were varied to evaluate their effects. The parameters considered included the deposited specific energy, the density of the mixture region, the detonation velocity,- and the width of the detonation front. These data are currently being digitized and plotted using TAMGRAF and will be reported as they warrant.
CSQ-11 was also used extensively during this period to help analyze the high-explosive tests being conducted at the FITS facility. In'the high-explosive portion of CSQ-II, the JWL equation of state is used for the explosive PETN and the analytical equation of state is used for the water.
Problems were also encountered with the tabular equations of state for water (i.e., the pressure in the region of the high explosive exceeded the maximum value in the table of 2000 bars).
1.4.1.6 Upgrade to TEXAS: TEXAS- 2 Step (M. F. Young 6425)
Debugging of T.EXAS-2 step continued, and the second test problem was successfully completed. This benchmark problem simulated a boiling water column. Initiation of boiling reduced the timestep from 0.1 s to 12.5 as: after the initial boiling disturbance, the timestep increased to the input maximum of 0.1 s. Several corrections were made to the heat transfer coefficients and interphase friction coefficients in the bubbly flow regime. A subroutine to calculate the wall friction was also added. Also, the basic SETS method of coupling the conservation variables (density and energy) more closely to the pressure iteration variables (pressure, volume fractions, and temperature) was added to
. TEXAS.
All of the work currently being put into TEXAS-2 step is expected to be transportable to the MELPROG Integrated
t-
' Fuel-Coolant- Interaction (IFCI) module, except for the equations of- s t a te ~. This will allow rapid transfer of routines developed for TEXAS to IFCI as well as support to the modeling of the coarse fragmentation and explosion
-Phases.
1.4.1.7 'Modeling of Coarse Fragmentation and Hydrogen and Steam Generation
. (M. L. Corradini. UW)
We' continued to develop benchmark problems to test the TEXAS computer program. We plan to use these benchmark problems to . help us determine if the models and constitutive rela-tions used in TEXAS are reasonable. We also prepared for the meeting at Sandia and presented our latest results in modeling the FITS experiments for hydrogen and steam gener-ation.
1.4.1.8 Modeling of the Steam Explosion Phase (M. L. Corradini,.UW)
We continued our efforts in modeling the FITS experiments.
In particular, we began to model a group of 10 experiments with the nonequilibrium ' parametric steam explosion model.
Using the consistent methodology we have developed in the past f ew months for calculating the initial conditions for the computer analysis, we predicted the results of these FITS tests (MD-19. MDC-2, FITS-3A, SA, 2B, 3B, 6B, 7B. 9B.
and RC-2). In all cases we consistently predicted the conversion ratio to be higher than what was measured. We are now varying the two important - input parameters that control the rate of energy exchange in the explosion (mass of fuel fragmented and the explosion propagation velocity).
to match the test results. One interesting result is that we predict the conversion ratio for RC-2 to be approximately three times as large as in the other tests, primarily because of the rigid wall constraint in contrast to weak wall constraints. We presented some of these results in the meeting at Sandia in March.
1.4.1.9 Modeling of Film Collapse and Fuel Fragmentation (M. L. Corradini, UW) l Discussions were held with researchers at Sandia about our recent results concerning the possible mechanisms for rapid 4
fuel fragmentation. As a result- of these discussions, we-i began to look at ' the multi-dimensional behavior of Taylor
}
j instabilities and the details of liquid jet impingement on another liquid surface.
j' 1.4.2 Presentations, Visits, and Meetings Attended 1
i On February 13, M. Berman gave a presentation to 0. Bassett and R. Curtis of the USNRC. Topics of discussion included i
j i
-- -. - ._ . .=- _ - - . - - -
t
-f current program status and future research plans for the FCI Program. Also, on February 14. M. Berman gave a presenta-tion and tour of the FITS facility to USNRC Commissioner 4 F. Bernthal, O. Ba s s e t t ~, and R. Curtis. Topics discussed during these meetings included Sandia's current views about a-mode failure due to steam explosions and the importance of large-scale experiments to resolve the large-scale mixing uncertainties. i
- Ort March 12 .and 13, the " Fuel-Coolant Interaction Working Group" meeting was held at Sandia. K. Wolhetz and R. McQueen from Los Alamos National Laboratories attended the meeting and discussed the volcanism experiments being conducted.
These experiments involve up to 100 kg of iron / alumina ther-mite and display explosive behavior similar to the FITS experiments. Other topics discussed included the FCI modeling efforts here at Sandia, efforts at the University of Wisconsin, the current experimental efforts at FITS, large scale (SEALS) experiments, and future program planning.
1.4.3 Documentation M. Berman prepared a response to G. Bankoff's letter on February 4, 1985 to the SERG (Steam Explosion Review Group). This memo discussed our views related to some of the assumptions used to model steam explosion phenomena with the PHOENIX code. .
On February 28, 1985. M. Berman wrote a memorandum to J. Telford summarizing our reservations concerning the deliberations and conclusions of the Steam Explosion Review Group.
1.5 Hydrocen Behavior (J. T. Hitchcock, 6427: M. Berman, 6427) l The major concerns regarding hydrogen in LWRs are that the static or dynamic pressure loads from combustion may breach containment or that important, safety-related equipment may j be damaged due to either pressure loads or high tempera-l tures. In order to assess the possible threats, it is necessary to understand how hydrogen is produced, how it is transported and mixed within containment, and how it combusts.
The objectives of this program are (1) to quantify the threat to nuclear power plants (containment structure, safety equipment, and the primary system) posed by hydrogen combustion: (2) to disseminate information on hydrogen behavior and control; and (3) to provide programmatic and technical assistance to the, NRC on hydrogen-related matters.
1 I
, u 1.5.1 Current Progress and Technical Highlights 1.5.1.1 HECTR Analysis and Code Development (C. C. Wong, 6427)
The HECTR (Hydrogen Event: Containment Transient Response) code is a reactor accident analysis tool designed to cal-culate the transport and combustion of hydrogen and the transient response of the containment. It was developed to meet urgent NRC licensing needs and continues to be a major tool for predicting both local and global conditions during combustion sequences. It has been successfully applied to BWR Mark III, PWR ice condenser, and PWR large, dry contain-ments.
We continued to work on the development of a diffusian flame model for HECTR. Two approaches are being evaluated. The first is the empirical approach, which utilizes empirical correlations to characterize the diffusion flame process and the flow induced by the flame. The other is the lumped parameter approach, which involves defining one or more
- flame compartments and having HECTR calculate the gas properties within these compartments, such as the average pressure and temperature rises assuming complete hydrogen combustion. When completed, these two models will be assessed against the Nevada Test Site (NTS) continuous injection experiment--Test C3.
The assessment of HECTR against the large-scale hydrogen combustion experiment performed at the Nevada Test Site continued. The latest tests that we analyzed included standard (dry) and steam-laden premixed tests. The standard
+
tests had concentrations of hydrogen which varied between 5.3 percent and 9.9 percent with low concentrations of steam
(~4.4 percent). The steam-laden tests had concentrations i of hydrogen which varied between 6.9 percent to 12.9 percent l with relatively higher concentration of steam (-28.5 per-cent). The results (thow that the calculated peak pressure ratios compare reasonably well with the experimental data. l However, HECTR overpredicts the peak gas pressure and tem-perature. We suspect that HECTR underpredicts the radiative i and convective heat transfer during combustion. More detailed assessment of these two heat transfer models is underway. I We continued to provide technical assistance to the NRC by modifying the HECTR and ACHILES codes according to the needs of their computer facility located at the Brookhaven National Laboratory.
. We have released Version 1.0 of the HECTR and ACHILES codes to the Nuclear Energy Software Center (NESC) at the Argonne National Laboratory. The release of these codes is limited i
_. - _ , _ _ _ _ _ . - _ - - _ _ _ . . _ _ - - _ - . _ , . . - _ _ _ _ . _ . . _ - - - - - - - - - - - . -- -