ML20059M020

Summary of 930825 Meeting W/Representatives of Aeclt,Aecl & Aecb Re Staff Calculations Using Aecl Code Suite to Evaluate Large Break LOCA W/O Scram for Candu 3 Design

ML20059M020
Person / Time
Issue date:	11/05/1993
From:	Scaletti D Office of Nuclear Reactor Regulation
To:	Office of Nuclear Reactor Regulation
References
PROJECT-679A NUDOCS 9311180107
Download: ML20059M020 (3)

Text

i-lpa meog'o, e

r-E 1W E

UNITED STATES

("

'"l NUCLEAR REGULATORY COMMISSION j

e gm j wassiwoTon, o.c. 2osnaooi November 5,1993 Project No. 679 APPLICANT: Atomic Energy of Canada Limited / Technologies (AECLT)

PROJECT:

Preapplication Review of the CANDU 3 Reactor Design

SUMMARY

OF MEETING HELD WITH AECLT TO DISCUSS TOPICS RELATED TO THE STAFF'S USE OF THE AECL CODE SUITE TO RUN LARGE BREAK LOCA CALCULATIONS On August 25, 1993, the staff met with representatives of AECLT, AECL, and ALCB to discuss staff calculations using an AECL code suite to evaluate a large break LOCA without scram for the CANDU 3 design. The purpose of this meeting was to allow peer review of the staff's calculations from the stand-point of code limitations and boundary conditions. Also, discussions were held regarding the staff's effort in identifying and assessing the database (raattor physics, thermal-hydraulics, A00s, DBAs, and severe accidents) for the CANDU 3 design.

AECLT indicated they felt that the INEL results as they related to the timing of the event were generally what they would expect if they had run a LBLOCA without shutdown event using the CATHENA code; however, they disagreed with the implied results. AECLT expressed a concern that analyzing a LBLOCA without shutdown may be an inappropriate use of the CATHENA code, and stated that they do not run the event for the CANDU reactors. AECLT stated that CATHENA was not designed for fuel melt projections and that they have a LOCA i

code to be used for this purpose. AECLT felt it was inappropriate to assume no shutdown and also break all four inlet headers. The LBLOCA without scram had been previously evaluated by Ontario Hydro for their Pickering Plant using Ontario Hydro codes.

AECLT indicated that they would provide additional information relative to the CANDU 3 database and additional comments on the staff's effort to run the AECL r

CATHENA code to model a large break LOCA with out shut down. Additional l

information was provided on September 10, 1993 (Document Control Desk from Victor Snell, AECLT).

mm y[Ume q -*

7~ WI f { {'[3 N% 2 ' ^ C b > *

• OdN '

002062 i

\\

f tu t/7?

9311180107 931105 i

PDR PROJ 679A PDR U3 g

150033 4

l i

d J

r-I AECL Technologies

-3_

November 5, 1993 CANDU Project No. 679 cc:

Louis N. Rib, Licensing Consultant AECL Technologies 9210 Corporate Boulevard, Suite 410 Rockville, Maryland 20850 Bernie Ewing, Manager Studies and Codification Division Atomic Energy Control Board P.O. Box 1046, Station B 270 Albert Street Ottawa, Ontario, Canada K1P SS9 A.M. Mortada Aly, Senior Project Officer Advanced Projects Licensing Group Studies and Codification Division Atomic Energy Control Board P.O. Box 1046, Station B 270 Albert Street Ottawa, Ontario, Canada KlP SS9 Project Director - CANDU-3 AECL CANDU 2251 Speakman Drive Mississaugua, Ontario, Canada L5K IB2 L. Manning Muntzing Newman & Holtzinger, P.C.

1615 L Street, N.W., Suite 1000 Washington, DC 20036 Steve Goldberg, Budget Examiner Office of Management and Budget l

725 17th Street, NW.

Washington, DC 20503 A.D. Hink Vice President / General Manager AECL Technologies 9210 Corporate Boulevard, Suite 410 Rockville, Maryland 20850 i

i

1. p%g 4

UNITED STATES fe g

NUCLEAR REGULATORY COMMISSION 5

E W ASHINGTON, D. C. 20555 o

4

,o 9 *****

SEP 1 1993 MEMORANDUM FOR:

Louis M. Shotkin, Chief, RPSB/DSR/RES 3 <

THRU:

Ralph 0. Meyer, Section Leader, RPSB/DSR/RES gf[

FROM:

David D. Ebert, RPSB, DSR/RES

SUBJECT:

MINUTES OF THE MEETING WITH AECL ON AUGUST 25, 1993 A meeting with AECL was arranged by Dino Scaletti of NRR/PDAR to discuss the results of transient calculations, using the AECL suite of codes, for the CANDU 3 design subjected to a large-break LOCA with failure to shutdown.

These calculations were performed by RES staff and contractors. Also discussed during the meeting were contractors' preliminary data-base assessments for reactor physics, thermal-hydraulics, and fuel behavior.

Copies of all the overheads and handouts presented at the meeting are included as attachments.

Rex Shumway and Jerry Judd of INEL presented results of transient analyses performed by INEL and RES (David Ebert) using the AECL code suite. The following sensitivity studies were perforined on a base-case LBLOCA with failure to shutdown:

(a) D,0 purity, (b) critical heat flux, (c) heat transfer coefficient, (d) initial steady-state power level, and (e) number of broken inlet headers.

In addition, a LBLOCA with delayed scram using the shutdown rods and a rod withdrawal transient with failure to shutdown (an ATWS) were investigated.

AECL indicated general agreement with the transient analysis results but had several comments. AECL stated that they had not performed LBLOCA with failure to shutdown for any of the CANDU designs, but that Ontario Hydro had undertaken such a study of the Pickering reactor design. The AECL code suite was not used in that particular study, however. Ontario Hydro used their own set of codes.

i AECL indicated that the fuel temperature calculational model in CATHENA is somewhat crude.

A more accurate value for maximum fuel centerline temperature under nominal operating conditions is about 2000 *K, no+ 2700 *K as obtained from the CATHENA calculations. There was some question about how the ranges i

of parameter variations were selected. The response to this question was that they were arbitrarily selected to provide a spread, and not based on an uncertainty analysis. AECL commented that a LBLOCA with failure to shutdown is an extremely unlikely accident. This accident scenario has been classified as EC-IV (Residual Risk) in NUREG/CR-6065, Systems Analysis of the CANDU 3 Reactor, and would be reviewed in that context.

The status of the data base assessment tasks were presented by:

(a) Jerry Judd of INEL (reactor physics for EC-I and EC-II), (b) Rex Shumway of INEL (thermal-hydraulics for EC-I and EC-II), (c) Don Hagrman of INEL (fuel Co M a y 7vu m

l l

SEP 1 1993 l

Louis M. Shotkin 2

behavior for EC-I & EC-II), and (d) Anthony Wright of ORNL (severe accident phenomena for EC-III and EC-IV) l l

AECL stated that some additional reactor physics experiments have been undertaken at Chalk River Nuclear Laboratories, and more are planned in

)

cooperation with the CANDU Owners Group (COG). They indicated that reactivity as a function of void fraction has been measured, and measurements are being I

considered using mixed-oxide fuel, using hot fuel with and without hot coolant, with moderator poison, and with stratified void. They also indicated that there will not be much depleted fuel in CANDU 3, and that they believe measurements on such fuel would not be warranted. AECL thought that the effects of fission products on the void reactivity worth would not be very large and therefore would not be very important.

In response to a request by AECL, a draft report entitled CANDU PHYSICS CODE DATABASE ASSESSMENT by Jerry L. Judd dated May 1993 was distributed at the meeting. This preliminary report is also included in the attachments.

AECL indicated that more experimental information concerning two-phase flow in inlet and outlet headers is asailable and would be sent to INEL by the Whiteshell Nuclear Research Establishment in the near future.

The phenomenon of pellet / cladding interaction (PCI) behavior in CANDU fuel rods was discussed.

It was noted by AECL that there is extensive experience with power ramping in fuel rods when they are inserted at power during fueling operations.

AECL disagreed with a conclusion concerning the adequacy of the moderator heat transfer data base during a severe accident situation.

They indicated that the experimental data base is quite extensive. AECL mentioned that in reactivity induced accidents, such as a LOCA with failure to shutdown, the fuel would progress through a melt situation rather than breakup as would happen in a light water reactor. The primary reason they gave for this behavior i.s that the prompt neutron lifetime is about ten times. longer in a heavy-water reactor than in a light-water reactor, resulting in a more gradual energy deposition rate.

In an effort to verify the general behavior of the AECL code suite, David i

Ebert presented a simplified transient simulation model of a LBLOCA with failure to shutdown. His results showed reasonable agreement for power evoluticn with the AECL code suite. He also obtained, through integration of' the power evolution, the energy deposited (in cal /grm) during the base case i

LOCA with failure to shutdown and with a scram using only control rods.

AECL agreed with the calculated energy density deposition rates by mentioning that they had calculated similar results. David Ebert also briefly mentioned i

energy density deposition rates for several other reactor types for major i

reactivity-induced accidents.

Donald Carlson presented his plan for an independent in-house analysis of i

coolant void reactivity worth in the CANDU 3 design using a suite of recently acquired reactor physics computer codes, COMBINE /ANISN and MCNP.

He presented a preliminary calculation of neutron flux spectra for the nominal and coolant i

SEP 11993 Louis M. Shotkin 3

voided cases. This work is being undertaken in an effort to better understand the physical basis of the positive void coefficient in the CANDU 3 design.

AECL expressed some. concern about the adequacy of the ID transport-theory calculations of ANISN in which the fuel pins are modelled as ring shaped regions.

Don Carlson indicated that spot checks to be provided by the 2D continuous-energy Monte Carlo code, MCMP, would address these concerns. He also mentioned that he had obtained reasonably good agreement for the void worth calculations between POWDERPUFS and ANISN calculations. AECL wondered-why the NRC would be investigating design changes needed to reduce the value of the positive void coefficient. The NRC response was that the physical mechanism of the void coefficient would be better understood by going through this exercise and that the NRC would not try to modify the CANDU 3 design.

u

- l C. 5L !

David D. Ebert, RPSB/DSR/RES Attachments:

As stated cc:

Dino Scaletti (w/ Enclosure)

Ed Throm (w/o Enclosure)

J

i Review of the CANDU Reactor Physics Experimental Database J. L. Judd August 25,1993 Idaho National l

Engineering Laboratory e

g C!

[

c r-l

DEFICIENCIES

. Only one measurement of the void reactivity for 37-element bundles has been made. Additional measurements should be performed.

. Limited depleted fuel void reactivity measurements have been performed. These were performed with simulated depleted fuel with no assessment of the effect of fission products on void reactivity.

. No measurement of full power reactivity coefficients (i.e.

power, fuel, coolant) appear to have been made.

No uncertainty analysis of experimental / code prediction comparisons has been made. Many comparisons have been performed in a qualitative manner.

..Use of an inconsistent code suite for comparisons. Many of the experimental / code prediction comparisons have been made with codes other than CERBERUS (i.e.

WHIRLAWAY and EXTERMINATOR).

........_,...-..,_.........m.

.......,.i

.~..--....

C l

RECOMMENDATIONS i

Perform additional void reactivity measurements for 37-element bundles.

Perform an assessment of the effects of using simulated depleted fuel for void reactivity measurements.

Assess the viability of performing void reactivity measurements with depleted fuel from existing reactors.

i Perform Full power temperature coefficient measurements.

i Use a consistent code suite for all comparisons to experimental data.

Perform a uncertainty analysis.of all experimental /

prediction comparisons.

r

• , p

...m

..m.-....

,; _.,.. -. _.,... _.... _,.~

a, g

Preliminary Assessment of CANDU 3 Experimental Thermal-hydraulic Databases Rex Shumway Cal Slater

/NEL

^-

Idal10 National EngIncering Laboratory

Assessment Approach o identify the limiting thermci hydraulic physical phenomena in se!ncted limiting transients for licensing a CANDU 3.

o Review status of experimental dctabases for the limiting phenomena.

o Assess the databases with regarc' to their completeness for benchmarking code features important to limiting phenomena and transients in CANDU reactor,

systems.

o identify deficiencies in the databases.

o identify areas where calculated.results are so important that multiple, i

independent databases should be used for validation.

i l

~

o Assess need for additional integral tests.

i l

e 'y a

...~.-

.......~.i...... -...-~

.._.....,_-...._.--~...,._._,-..~...-.-...s

. ~............ -.

. ~.

.,.g i

F S, elected Representative Transients for Thermal-hydraulic Phenomena o Break in feeder tube headers (Large Break) o Pressurizer relief valve failure (Small Break) o Feedwater piping break (Undercooling) o Loss of Class IV power ( Loss of essential Power) 4 These transients have been selected as those that best represent the expected important thermal-hydraulic phenomena. This selection has been based on existing analysis and engineering judgement. As i

more."inhouse" simulation analysis becomes ava,lable, we may i

revise or add to the list.

._m

..mm_-.....

a Preliminary Conclusions 4

o in general, the Canadian reactor safety program has fccused on the important elements of the transient response of CANDU reactor systems and produced a very broad database.

@ We have identified some General areas that may require additional experimental data:

(1) Two-phase flow phenomena in the inlet and outlet headers and feeders I

(2) Steam-generator performance during small-breaks and natural circulation situations (3) Full scale two-phase pump performance in other than 1st quadrant

.,._.-__._.m.

. ~... -

, _... ~ - _ - -

..... ~. - - -

IJA 4/a r/o in EGG-NRE-?"

CANDU PHYSICS CODE DATABASE ASSESSMENT Jerry L. Judd Published May 1993 Idaho National Engineering Laboratory EG&G Idaho Idaho Falls, Idaho 83415 h

Prepared for the U.S. Nuclear Regulatory Commission Washington, D.C. 20555 Under DOE Idaho Field Office Contract DE-AC07-76ID01570 FIN L2445 1

1

.P 5

CONTENTS t

1.

INTR O D UCTIO N............................................................................ 5 t

.~

2.

CAND U-3 C O RE D E S IG N................................................................. 5 3.

AE C L-C AND U D ATAB AS E.................................................................... 5 3.1 Cross Section Generation Methods............................................. 6

~

~

3.1.1 POWDERPUFS.V...

.... 6 3.1.2 MULTICELL 6

1 3.2 Co re Sim ulati on Me thods.............................................................. 7 3.2.1 FMDP.. -

7 i

h '3 3.2.2 CERBERUS

....-......7 i

3.3 Coolant Reactivity Coefficient Comparisons................................ 7 g 3 }s 3.3.1 Coolant Void Reactivity Coefficient--

... 7 3.3.2 Heat Transport System Temperature Reactivity Coefficient 9

l 4.[

3.4 Moderator Temperature Reactivity Coefficient Comparisons...10 l

c-

/

3.5 Flux Distribution Comparisons.................................................10 g 'l 3.6 Reactivity Device Worth Comparisons......................................11 j 1--Q 3.7 Time Dependent Core Response Comparisons......................11 y'

3.7.1 SRL Process Development Pile Experimental Companson 11 3.7.2 Shutoff Rod Actuation Experimental Comparisons..

12 l

3.7.3 Liquid Poison injection System Expenmental Compansons 12 4.

U. S. D O E D ATAB ASE........................................................................... 13 5.

D ATABAS E DE FICIE NCIES............................................................... 13 6.

CONCLUSIONS AND RECOMMENDATIONS.................................14 7.

RE F E RE N C E S...................................................................................... 14 i

L 6

e F

i i

l 1

I J

l 11 Ancra$rymaker/canduS4hysics.databaseTOC. doc April 26,1993 1

4 6 -

TABLES Table 1.

Void Reactivity Measure ment S ummary................................................ 9 Table 2.

HTS Temperature Reactivity Measurement Summary........................... 9 Table 3.

Moderator Temperature Co:f6cient Measurement Summary..................10 Table 4.

Elux Distribution Measurement Summary.............................................. I1 Table 5.

Reactivity Device Measurement Summary........................................ I1 Table 6.

Time Dependent Core Response Measurement Summary......................13 i

.l i

i i

iv AuersM'znaler/candu3/ physics.databaseLOT. doc Apru 26,1993

[.

1. INTRODUCTION AECL Technologies of the United States is in the process of submitting to the USNRC an advanced reactor design, CANDU-3. CANDU-3 is an advanced heavy-water reactor (HWR) design which has mechanical and hydraulic characteristics significantly different from light-water reactors (LWRs).

One of the objectives of this project was to review and assess the adequacy and identify any deficiencies in the existing AECL experimental and analytical physics database with respect to the CANDU-3 reactor transient response. This assessment was performed by examining AECL reports and documents relating to experimental physics measurements and computer codes and researching information in the U.S for heavy water systems.

2. CANDU-3 CORE LESIGN The CANDU-3 is the latest version of the CANDU pressurized heavy water reactor designs developed by AECL. The design consists of 232 horizontal pressure tubes arranged in a square lattice in a calandria vessel containing low pressure, low temperature heavy water. The pressure tubes are made of zirconium-niobium and are surrounded by Zircaloy calandria tubes. Each pressure tube contams twelve fuel bundles containing 37 fuel pins and the coolant used is heavy water at high pressure. The fuel pins are natural enriched uranium dioxide pellets clad with Zircaloy.

The CANDU-3 reactor core utilizes on-power refueling which reduces the excess reactivity requirements and permits continuous flux shaping to yield an optimum power distribution. Control and shutdown devices and in-core instrumentation devices are located in tubes that perpendicular to the fuel channels and are in the low pressure, low temperature moderator.

The CANDU-3 core design resuits in a well thermalized neutron spectrum and flexible operating characteristics.

3. AECL-CANDU DATABASE f

An assessment of an experimental database includes a review of the cal [c

,. w ~k l

methods used, the experimental measurementcand the-experimental facilities. This report covers-only the-first, two items andthtfassumption is made that e

v

' experimental facilities used are appropriate for validating CANDU-3 physics data.

s Section 3.1 gives a brief description oTthe cross sectionisrTeration methods or CANDU lattices. Section 3.2 gives a brief description of the core static and transient simulation methods. Section 3.3 presents the AECL experimental data on coolant void and heat transport system temperature reactivity coefficients and a comparison to the predictions of the AECL computer codes. Section 3.4 presents the AECL experimental data on moderator void and temperature reactivity coefficients and a comparison of the predictions of the AECL computer codes. Section 3.5 presents a AuerQ2/ maker /candu3/ physics. database. doc Apr0 26,1993 5

comparison of the AECL measured and predicted flux distributions. Section 3.6 presents a comparison of measured and predicted rod worth. Finally, section 3.7 presents a comparison of measured and predicted transient core response to shutoff rod actuation and liquid poison injection.

3.1 Cross Section Generation Methods I

Cross sections for CANDU lattices are calculated with the Powderpufs-V and 2

Multicell computer codes. Powderpufs-V calculates the basic two-group cross l

sections for a given fuel bundle design and lattice spacing and Multicell calculates the effects on the cross sections for the presence of a reactivity device. The reactivity device may be a shutoff rod, an adjuster rod, a zone controller, or a liquid poison plume.

3.L1 POWDERPUFS-V The Powderpurs-V code is used to calculate two-group cross sections for use in l

three-dimensional diffusion codes for reactor analyses. The Powderpufs-V code uses simple empirical relationships to evaluate the four-factor formula. A simple three-region lattice cellis used which models homogenized fuel, an annulus region vdth coolant, pressure and calandria tubes, and a moderator region. Micro-scopic cross sections are evaluated using the Westcott convention. Resonance absorption and fast -

238 fission is only accounted for in U

. Two-group cross sections are then inferred from the four factor data and leakage estimates.

The coefficients used in the empirical relationships used by Powderpurs-V have been checked against critical experiments,65, 3.L2 MULTICELL Multicell is used to calculate the effect of reactivity device or incore instrument in the lattice on the cross sections calculated by Powderpufs-V. The results from a Multicell calculation are the incremental cross sections to be added to the Powderpufs-V base cross sections to account for the presence of the reactivity device or incore instrument. These incremental cross sections are calculated once at

~

nominal conditions and used for all coiElitions. This is aieasonable assumption in that the reactivity devices are in the low pressure moderator separated from the fuel and that the n'eutron spectrum at thoseIlocatiissToesnTEhange apFr~ei:iaByWith

~

~

~

coolant or fuel changes.

The three-dimensional diffusion equations are solved by Multicell. Internal boundary conditions, which are based on integral transport theory, are used at the reactivity device and fuel channel boundaries. The converged flux distribution is then used to obtain neutron reaction rates in each materialin the supercell and these reaction rates are then used to calculate cell-averaged cross sections. The incremental cross sections are then calculated as the difference between the Multicell cell average cross sections and the Powderpufs-V cell average cross sections without the reactivity device present Auer42/maler/candu3' physics. database.dx April 26.1993 6

P b

V oA

.y-c r__

3.2 Core Simulation Methods Three-dimensional calculations are performed with coarse mesh diffusion methods. This technique is valid for the CANDU-3 design because the migration is 2

2 much larger than that of LWRs (350 cm versus 55 cm ). Two codes are primarily 3

used to perform CANDU-3 calculations. FMDP is used for the static design 4

calculations and CERBERUS is used for the transient calculations. Both FMDP and CERBERUS use the same techniques to solve the spatial diffusion equations, but CERBERUS uses the Improved Quasi-Static Method for the time dependent solution.

3.2.1 F M D P FMDP is a computer program for core design and fuel management calculations.

It calculates the neutron flux and power distributions by solving tie two-group, three-dimensional neutron diffusion equations. The diffusion equations are sc17ed with a Gauss-Seidel iterative procedure. The code is capable of modeling reactivity device worth, temperature and void reactivity coefficients, time averaged and core follow depletion calculation, xenon transients, and other types of calculations.

3.2.2 CERBERUS CERBERUS solves the space-dependent neutron kinetics equations. The f

i diffusion equation solver used in FMDP is used in CERBERUS with additions to implement the Improved Quasi-Static (IQS) method to solve the time-dependent portion of the equations. The reader is referred to Reference 4 for a more detailed description of the IQS implementation in CERBERUS. The code is capable of modeling control and shutoff rod movement during a trip, poison injection, and' response due to other changes in fuel, coolant, or moderator changes.

Several other computer codes are used in conjunction with CERBERUS for H

transient analyses. MATMAP is used to set up the geometric model, assign cross 12 sections to the various regions, and describe the reactivity devices. CERBSPOW is used to convert the CERBERUS calculated power distribution into the format 18 needed by the CATHENA thermal-hydraulic code.

3.3 Coolant Reactivity Coefficient Comparisons This section discusses the comparisons of the measured and calculated reactivity coefEcients that result from changes in the conditions in the fuel channels. These include the coolant void reactivity and heat transport system (HTS) temperature coefficient. These reactivity coefficients are important for all transients the result in the heatup of the coolant. The void reactivity is always positive and the HTS temperature coefficient can be positive or negative, l

3.3.1 Coolant Void Reactivity Coefficient References 5 and 6 have been reviewed to assess the adequacy of the experimental database for predicting the void reactivity coefEcient. Only one series j

/ users $tymder!ca.ndurphysics. database. doc April 2S,1993 7

i i

~

of expedments were performed with 37-element fuel bundles, D 0 coolant, and a 2

square lattice pitch as used in the CANDU-3 design. This series of experiments was performed in the ZED-2 facility with fresh fuel bundles for the Bruce reactors. In this experiment, the critical material bucklings for the lattice with D 0 as coolant 2

and air as coolant were measured and the difference in material bucklings was used as an estimate of the void reactivity. The difference in material bucklings was overestimated by POWERPUFS-V.

#~

I Many other similar experiments have been performed at ZED-2 with 19 and 28 element bundles in a hexagonallattice arrangement. The comparisons of Powderpufs-V to experimental data for these expedments also exhibits that Powderpufs-V overpredicts the void reactivity for lattice pitches in the range of that expectidTn~tFCXNDU-3 design. Additionally, exp'eddents have been performed in

~

~

ZED-2 with H O as a coolant and D 0 as moderator for 28 element fuel bundles in a 2

2 hexagonal lattice which also exhibit the Powderpufs-V overprediction of void reactivity.

8

~A series of experiments was performed at SRL in the Process Development Pile (PDP) with simulated burned fuel. The fuel bundles contained 31 aluminum clad fuel pins on a triangular pitch instead of arranged in concentric rings as in CANDU fuel bundles. The fuel bundle pitch was square and varied from 19.00 cm to 28.65 cm. The experiments were performed with air and D 0 as coolant and D 0 as the ~

2 2

moderator. Once again Powderpufs-V overpredicted the void reactivity for all An experiment was performed during the start-up of Gentilly-1 which measured the critical moderator heights with voided fuel channels and with H O coolant in the 2

fuel channels. The void reactivity was underpredicted by 2 milli k which was stated to be within the experimental uncertainty. This measurement is not consistent with other void reactivity measurements, but that may be due to the use of H O as 2

coolant instead of D 0, and to the fact that WHIRLAWAYM was used as the core 2

~

~~

~ ~ ~ ~ ~

modeling code instead of FMDP or CEIniEliOS~ ~ ~~ ~

In summary, there have been many measurements of void reactivity at ZED-2 and other facilities. Table 1 summarizes the void reactivity comparisons performed j

f and gives the ranges of parameters over which the measurements were mad _e Unfortunately, only one measurement is applicable tithe CANDU-3 design. It is,N recommended to perform other void reactivity measurements with 37 element fuel i

bundles and perform a rigorous statistical analysis of the comparisons with predicted values. The predictions must be performed with the computer codes,i.e.

j Powderpufs-V and CERBERUS, that will be used for the safety analyses.

x

,,-Gh

.'s i

a.

J i

1

/userMrftmaker/ca.ndu3' physics. database. doc Aprd 26,1993 8

Table 1. Void Reactivity Measurement Summary Facility Fuel Type Range of Parameters ZED-2 UO 19.28.37-element, D O Fresh fuel 2

2 Coolant SRL PDP 31-element Mixed Oxide, Simulated irradiated fuel D O Coolant 2

Gentilly-1 Fresh UO, H O Coolant Fresh fuel 2

2 3.3.2 Heat Transport System Temperature Reactivity Coefficient The isothermal temperature coefficient of the heat transport was measured during the commissioning of the Douglas Point, Pickenng-A units 3 and 4, and Bruce A unit 2 reactors. This parameter was also measured with burned fuel at i

Douglas Point and Pickering-A unit 2. The commissioning tests had relatively good agreement with the exception of some discrepancies above 150 C that were explained as unknown systematic measurement error, inadequacy of the Powderpufs-V bare pile model, or the neglect of depleted fuel bundles in the initial core. The Pickering-A unit 2 burned fuel comparison is adequate given that a single burnup was assumed for the Powderpurs-V calculations and the core had a burnup distribution. The Dougl_as Point _ comparison-was-performed with-Powderp

-V and D

EXTERMINATOR. The Douglas Point measurements were more negative than Powderpurs-V/ EXTERMINATOR calculation.

r.,fy I ~

utteuc M j

Tible Eiumnidriies Die HTSlemp rature reactivityt6etticient compansons e

performed an_d.gives_the ranges of parameters over-which the measurgere made.It is recommended that detailed calculations of the meastuements be performed with Powderpurs-V/FMDP or Powderpufs-V/CERBERUS to validate thej

-modeling techniquerrusedirrthe salety analyses. All the atsove mennoned calculations were performed with Powderpurs-V inly usihg a bare pile model.TDh1 y

doesn't adequately validate the complete code sequence for calculation of the HTS s

temperature reactivity coefficient. Once again, it does not appear that a rigorous

~

statistical analysis of the comparisons has been performed.

~~~

Table 2. HTS Temperature Reactivity Measurement Summary Facility Fuel Type Range of Parameters Douglas Point Fresh and irradiated UO2 40 C to 250 C Pickering-A Units 3 and 4 Fresh UO2 50 C to 270 C Pickering A Unit 2 Irradiated UO2 80 C to 250 C Bruce-A Unit 2 Fresh UO,37-element 50 C to 238 C 2

Arser@ymaker/candu3'phynes. database. doc Aprd 26,1993 9

3.4 Moderator Temperature Reactivity CoefBeient Comparisons Moderator temperature reactivity coefficients have been measured during commissioning tests at Douglas Point and Bruce-A unit 2. Additionally, experiments were performed with irradiated fuel at Douglas Point. Powderpufs-V alone was used.

to calculate the moderator temperature coefficient using a bare pile mode for the commissioning tests and it predicts a more positive coefficient than that measured.

The Douglas Point measured coefficient with irradiated fuelis more positive thanN what Powderpufs-V predicts, but the measurement uncertainty is quite large and 1

the Powderpufs V prediction is within one sigma of the measured value.

l Table 3 summarizes the moderator temperature reactivity coefficient comparisons performed and gives the ranges of parameters over which the measurements were made. It is recommended that detailed calculations of the measurements be performed with Powderpufs-V/FMDP or Powderpufs-V/

CERBERUS to validate the modeling techniques used in_the safety analyses. All the above mentioned calculations were performed with Powderpufs-V only using a bare pile model. This does not adequately validate the complete code sequence for calculation of the moderator temperature reactivity coefficient. No rigorous i

statistical analysis of the comparisons has been performed.

i Table 3. Moderator Temperature Coefficient Measurement Summary i

Facility Fuel Type Range of Parameters Douglas Point Fresh and irradiated UOy 25 C to 35 C Bruce-A Unit 2 Fresh UO,37-element 30 C to 79 C i

2 3.5 Flux Distribution Comparisons.

Comparisons of predicted and measured flux distributions have been performed,7 The flux distributions were measured in the Pickering and Bruce 5

7 reactors during commissioning. In most experiments, copper wire activ tio,n was' used to measure the flux distribution and a traveling fission chambe ts:a few cases.

}

The radial and axial flux distributions are well predicted for a variety of reactivity 5

device positioning. Highly skewed flux distributions were measured by inserting reactivity devices asymmetrically and these skewed flux distributions were well predicted also.

l Table 4 summarizes the power distribution comparisons performed and gives the ranges of parameters over which the measurements were made. These comparisons were performed with Powderpurs-V, Multicell, and CHEBY-based codes (i.e. FMDP or CERBERUS) so the comparisons are iTalid-for validating the safety analysis. code package. No rigorous statistical analysis of the comparisons has been performedN l

x i

NsersitytTakericandu35hysics database doc Apnl 26.1993 10 l

l

.r 4

Table 4. Flux Distribution Measurement Summary Facility Fu-lType Range of Parameters Pickering-A Fresh UO, Range of rod pattems result-ing is skewed distributions Bruce-A Fresh UO,,37-element Range of rod patterns result-ing is skewed distributions 3.6 Reactivity Device Worth Comparisons Reactivity device reactivity has been measurad at ZED-2, Pickering-A, and Bruce-A. Reference 5 only reports calculations performed for the Pickering-A and Bruce-A measurements, so these are the only ones discussed. In general, the comparisons are good and there does not appear to be any consistent error. The difference between measured and calculated values does not exceed 7%.

Table 5 summarizes the reactivity device worth comparisons performed and gives the ranges of parameters over which the measurements were made. The calculations were performed with Powderpurs-V, Multicell, and FMDP. These 3 calculations validate the safety analysis code system, but no statistical analysi,s hp been performed to calculate uncertamties.

t

• /

p..

Table 5. Reactivity Device MeasurementImmary Facility Fuel Type Range of Parameters Pickering-A Fresh UO2 Cobalt Adjusters, Shutoff Rod Banks, individual Shut-off rods, and Zone Control-lers Bruce-A Fresh UO,37-element Shutoff Rod Banks and indi-2 vidual Shutoff rods 3.7 Time Dependent Core Response Comparisons A variety of transient simulations have been performed by AECL. Three of these are discussed below. All were performed with Powderpurs-V, Multicell, and CERBERUS. Several reports mentioned a series of transients to be performed at I

ZED-2, but no reference c Efil be'foundToc~um-'enting dese exper&iienTs ~ i any

~~

o comparisons to the experiments: Table 6 siinm~arizesWeb dependent core response measurbme6G. ~ ~

3.7.1 SRL Process Development Pile Experimental Comparison Reference 8 presents a comparison of CERBERUS results with a transient at the Process Development Pile at the Savannah River Laboratory in the U.S. The bervirymaker/candu3/phystes database. doc Apnl 26,1993 u

i

experiment consisted of a free-fall insertion of fuel bearing rods into off-center locations of the core. The transient is super-critical, i.e. inserted reactivity greater than zero but less than the delayed neutron fraction. A comparison of excore detector readings and predictions is shown over the duration of the transient and CERBERUS accurately predicts the excore response. This comparison shows that t

CERBERUS would adequately model the.regonse of the core for pos.i_tNeTelsctivity insertion events such as coolant voiding or reactivity device ejectiortas long as the reactivity-doer nnexceed one dolla(CERBERUS is probably accurate for p critical eve % but this has not been validated.

~~

t.

3.7.2 Shutoff Rod Actuation Experimental Comparisons Reference 9 discusses shutoff rod drop experiments performef at the Bruce-A' s

units 1 and 2. Thirty rod tests were performed at units 1 and 2, and a 28 rod test was also performed at unit 2. CERBERUS calculated powers are comp'aredTEion chamber response. The CERBERUS powers are normalized to 100 at the start of the transient as are the ion chamber readings. The transient reen~e k -

redicted as is the asymptotic power level after full insertion of the rods. The dynamic i

reactivity of the shutoff rods is greater than the static value which validates the flux uhane_retardatiortdue_to delayed. neutron holdup.

,j i

The comparisons demonstrates the validity of CERBERUS for rod drop analysi.f.

It should yield conservative results for these types of events. "

3 7.3 Liquid Poison Injection System Experimental Comparisons Reference 10 discusses Liquid Poison Shutdown System (LISS) experiments and the comparison to calculations with CERBERUS. Additional computer codes, ALITRIG, INCON, and PJET, were used to model poison movement into the modergtbr for LTSS ehnts.ALITRIG is used to model the poison movement into the core, i.e. the jet or plume. INCON uses the ALITRIC.mL m' parates jet lengths versus time. PJET calculates the average concentration of a jet and Multicell then produces increment.al cross sections for a given jet length and concentration.

The Gentilly 2 test was performed with 5 of the 6 poison injection nozzles operating. CERBERUS predicts a slightly slower power decrease and reaches a higher asymptotic power level than that measured. The measurements were taken j

from both an ion chamber and three incere fission chambers. QERBERUS predicts the instrument response on the injection side very well, but the two ins _truments on I

the side. opposite the i.niection are conservativ.e.

The Pt. Lepreau test was performed with all 6 poison injection nozzles operating.

This test was performed from 60% of full power after a xenon transient following a down. power maneuver from 100% power. The down-power maneuver and xenon transient were modeled with CHEBEMAX, a module of FMDP. Instrument responses were obtained from 15 detectors, both ion and fission chambers. Similar resul:s were obtained from Pt. Lepreau as from the Gentilly-2 tests with the exception of uncompensated detectors. With the exception of the uncompensated j

detectors, all predicted responses were conservative with respect to the measured i

i taermir),tas tr candu3/physus. database. doc Apnl 26.1993 12 l

4

,-g s

u.a.s a

--,..ans

+

c

.~.

a

.r s

I 1

j data. The uncompensated detectors were all spare detectors of unknown history and therefore were d,eeme_d suspect by the AECL analysts, Tlie compt-ison of the CERBERUS methodology to the two experiments is quit good and is adequate for validation of the method for LISS actuation.

1 s

Table 6. Time Dependent Core Response Measurement Summary Facility Fuel Type Range of Parameters

- i SRL PDP High enriched U, triangular Positive reactivity insenion lattice of fuel pins Bruce-A units I and 2 Fresh UO,37-element 28 and 30 shutoff rod drop-2 sub-critical transient mea-j surements Gentilly-2 Fresh UO2 5 out of 6 poison injection nozzles operating.

Pt. Lepreau Irradiated UO2 6 out of 6 poison injection nozzles operating, at 60%

j power following a xenon i

transent

4. U.S. DOE DATABASE The U.S. DOE database deals with high enriched heavy water systems and therefore is not appropriate for validating Powderpurs-V. As discussed above,' -

]

transient data from PDP at SRL has already been used to validate CERBERUS..No i

detailed review of the DOE database has been performed, so there may be other l

useful transient experiments for validating CERBERUS.-

- p

].d/ %.b

5. DATABASE DEFICIENCIES

1. 7n" R e,q One fundamental deficiency in the database is the lack of adequate experiments i

to determine the void reactivity for 37-element fuel bundles in a square lattice l

typical of the CANDU-3. Only one experiment has been performed and that was with fresh fuel. It is highly recommended to perfonn additional 37-element void

- experiments with 37-element fuel with fresh and irradiated (or simulated irradiated) fuel for the CANDU-3 pitch._,

Comparisons of void rea<+:vity for simulated irradiated fuel have been

{

perfonned. An assessment o. :he validity of the burnup simulation, the use of fuel pins in a triangular pitch, and the use aluminum cladding need to be performed and documented for these comparisons.

.nervppmaker/ca.ndu3/phynes. database doc Apnl 26.1993 13

l 1

No measurements of the HTS temperature coefficient at full power were identified. It is recommended that these measurements be made at operating i

plants and comparisons made with Powderpufs-V/CERBERUS.

1

~

A good portion of the analyses discussed in this report were performed over many years and with core simulators other than FMDP or CERBERUS.

It is recommended that all comparisons be redone with the latest versions of j

Powderpufs-V/CERBERUS and documented in a single report. In addition, no statistical analyses of the calculated to measured differences has been performed. It is recommended to perform such an analysis to assign j

quantifiable uncer:ainties to parameters calculated by the Powderpufs-V/

l CERBERUS code system.

.* w' -

i r..,. I

e. n.- A

'W*-

l

6. CONCLUSIONS AND RECOMMENDATIONS

-.-; 4 y..

i

/j{[{['

l The review of the physics databases associated with transient analysis has uncovered only one major database deficiency, void reactivity for the 37 -

7, g7 j

element fuel bundle at depleted or equilibrium burnup conditions. Previous'

. _j h_

void reactivity measurements have been performed on 19 and 28 element fuell

,., )

l bundles with either fresh fuel or mixed oxide fuel to simulate depleted fuel.

.ge.

There has been no assessment on the validity of using mixed oxide fuel -

6 instead of depleted fuel and the net effect on void reactivity uncertainty and [h-.!

there have been no void reactivity measurements with mixed oxide fuel for 37 element fuel bundles.

p.

In general, there is a good deal of experimental data available. The comparisons should be redone using the latest versions of the codes to used in j

the safety analyses and documented in a single report. A statistical analysis of the comparisons should also be performed to calculate uncertainties for the l

various parameters.

7. REFERENCES i E.S.Y. Tin and D.B. Miller,"Powderpufs-V Physics Manual."TDAI-31 Part I of 3,

[

July 1979.

.i 2 A.R. Dastur and D.B. Buss," Multicell-A 3-D Program for the Simulation of Reactiv-ity Devices in CANDU Reactors." AECL-7544, February 1983.

p 3 A.L Wight and R. Sibley," Fuel Management Design Program - FMDP Part 1 Program I

Description."TDAI-105, August 1977.

4 A.R. Dastur. B. Rouben and D.B. Buss. "CERBERUS - A Code for Solving the Space-t Dependent Neutron Kinetics Equations in Bree Dimensions." TTR-392, February l

1992.

5 A.R. Dastur. "The Neutronic Basis of the Numerical Models Used in CANDU Simula-tion." TTR-333. February 1991.

i 14

• uservjtymnkerreandu3' physics. database. doc Apnl 26.1993 m.

i 6 E.S.Y. Tin and D.B. Miller. "Powderpuis-V Comparison with Experiment." TDAI-31 Part 3 of 3. July 1979.

7 V.K.Mohindra and G. Kugler. " Measurements and Simulations of Thermal Neutron Flux Distributions in Bruce A," AECL-5973. April 1978.

8 G. Kugler and A.R. Dastur," Accuracy of the Improved Quasistatic Space-Time Method Checked with Experiment." AECL-5553. October 1976.

9 A.R. Dastur, et. al. " Confirmation of CANDU Shutdown System Design and Perfor-mance During Commissioning," AECL-5914, October 1977.

10 R. D. McArthur, et.al., " Simulation of CANDU 6 (CANDU 600) LISS Tests," 16th Annual Nuclear Simulation Symposium, St. John, New Brunswick, August 1991.

11 B. Rouben, "MATMAP Program Description," TDAI-318. October 1984 12 B. Rouben and R.D. McAdiur,"CERBSPOW - Program Desenption and User's Man-ual." TTR-334, May 1991.

13 D.J. Richards, B.N. Hanna. -t al, "CATHENA Theoretical Manual." THB-CD-002, November 1989.

14 T.B. Fowler and M.L. Tobias. "WHIRLAWAY: A Three Dimensional Two Group Neu-tron Diffusion Code for IBM 7090 Computer," ORNL-3150.

15 T.B. Fowler. M.L. Tobias, and D.R. Vondy, " EXTERMINATOR-2: A Fortran IV Code for Solving Multigroup Neutron Diffusion Equations in Two Dimensions," ORNL-4078. Oak Ridge National Laboratory,1967.

.I I

1

)

users /JrJunker/candu3/physws database doc Apnl 26,1993 15

g CANDU-3 Fuel Behavior Data Base for EC-I and EC-Il 2

Idaho Presented by Engine r ng La b o r a tory

Fuel Behaviors Considered for Class i Events Integrity after power ramp (pellet - cladding mechanical interaction)

Hydriding i

Vibration and potential for fretting a

• : Sheath deformation

• i

• Sheath and pressure. tube oxidation *

-

• Fuel temperatures and stored energy

• t Boundary conditions for class ll and 111 event analysis 84964 DLH 0893 002 t

P s

A

,.,....---..,....----m-m.,.~.-m.-

~, -.. - -. * ~ ~ - -.~.- - --.

~

-- ~-----i

l Data Typep:

a Class i Fuel Integrity Data 1

Integrity after power ramp I

- CANDU operating.. experience

- CANDU loss-of-integrity PIE and CANDU and. LWR hourglassing analysis Hydriding (LWR and CMDU problems e

attributed to sources inside sheath).

1

- CANLUB graphite hydrogen remaining

- - Residual H 0 in UO 2

2

- PIE of fuel without significent internal H or D (end support plates)

- Separate effects hydriding.with.Zry-4 and Zry-2.5 Nb

.. m-.- -

mm v.

e e.-

--ew -

w e.

w,-.ev-.---,=,v,--*=,..-.+=-*ew=-wa--

• ww.,-+,-w-++.==e.'

.-:r

--+

w++=m-.--.-'e-4...

r e se w-e e -,- - = =

m-e-~4 vv

=e w r -e--

-h-,erw evaee'

<*e+

w e-r-awn 4 r r*- a w -w w

9 Data Types:

Class I Fuel Integrity Data - 2

• Vibration and potential for fretting PIE of fuel, fuel dimensions, and coolant flow (Needed only if fretting is observed):

- Fuel element bundle normal mode tests

- Vibration analysis

- Zry-4 / Zry-2.5 Nb elastic moduli at 300 C and versus fluence

- UO elastic moduli (compressionJ' at 2

300-2700. C 64964 DLH 0693 004

,e p

Data Types:

Boundary Conditions for Class > I

= Sheath deformation

- UO densification, swelling, and creep versus fast 2

neutron flux, 300-2700 C, UO density 2

- Zry-4 / Zry-2.5 Nb compressive creep versus stress and neutron flux at 300 C ? (Creep rate faster than necessary time resolution?)

- Axial growth not considered (Design renders length changes unimportant)

• Sheath and pressure tube oxidation i

- Rate constants at 300 C in D 0.with CANDU 2

oxygen potential M964 DLH 0893 005

- - - -,.. ~,.. _..

,.~..........,.

- -.. ~ ~

Data Types:

Boundary Conditions for Class >

I 1

1

• Fuel temperatures and stored energy q

- UO heat capacity, thermal conductivity 2

at.300 - 2700 C, initial UO density, vs. burnup 2

- Zry-4/Zry-2.5 Nb heat capacity, thermal conductivity 1

at 300 C

- Zry oxide thermal conductivity at 300 C

- UQ/Zry interface heat-conductance with CANLUB liner

- UQ/Zry gap conductance.(creep rate faster than necessary. time resolution?)

1

- mn...

4

.'s

s s

Summary of Existence of Data For Class I Events (References in Handouts)

Extensive fuel integrity data available from operating experience LWR data bases can.be applied for most c

EC > 1 B.C. s Exceptions:

UO /Zry interface heat conductance-with 2

CANLUB liner Indications that Zry-2.5 Nbfl uptake 2

differs from Zry-4 H uptake s

2

-Coolant oxygen potential effects on Zry oxidation M984 OLH 0893 00F 4

s

...-r.,

-w-

.,e

.-...-.,--,m_.

m

,,,,-c

-.-w-

+, - -, - -

-e.-,v

.w.-

w.*e e,.-.-i**=.-=--

-. v w E,- % -a,,. 5 _..,. -

Fuel Behaviors Consideied for i

~

Class II Events Sheath / pressure tube expansion and/or sag Sheath rupture. potential i

Sheath and pressure tube. oxidation Fuel bundle disassembly Thermal response *

• Boundary conditions for Class lli event analysis M964-OLH oc33 00s 8

- e -.

...,,_.--.,L.,....,.-,

t F

Data Types Needed to Desc' ribe Class II Events Sheath / pressure tube expansion and/or sag Integral deformation tests with local internal heat source contact and horizontal fuel Zry-4 / Zry-2.5 Nb stress / strain / strain-rate equation (versus temperature, fluence, initial coldwork, annealing, oxide layer thickness)

- Anisotropy parameters? (depends on integral test results)

Sheath rupture potential

- Fuel rod tests with local internal heat source contact of sheath:and horizontal fuel

- Sheath rupture criterion ht964 Dd VJS3 009 '

___.____.__.__._____._..___________.-_...._.___.___m_____.m__-._______.,________._______m_

1 --m m m-aww 2--

e-r e

e-,e ra e_e.,n.a-we e >

=ee

=-r=-

4e w

v w-m w e-

=r-m t-ae-

Data Types Needed to Describe Class II Events Fuel bundle disassembly

- Integral (full assembly) tests l

Sheath and pressure tube oxidation i

- Rate constants at 300-1800 C Thermal response

- UO heat capacity, thermal conductivity:at 2

300 - 2700 C, initial UO density, vs. burnup 2

- Zry-4/Zry-2.5.Nb heat capacity, thermal i

conductivity at 300 - 1800 C

- Zry oxide heat capacity, thermal. conductivity at 300 - 1800 C

- UO /Zry interface heat conductance with 2

CANLUB' liner

- UQ/Zry gap conductance

. M964 DLH 0093 Oto i.

,e e

.....,,-,...-,-..,.-.-.-.~.-,..-.m_

-.-m.-.----.-..._,,mu.-----.,

..... ~,.

-._.--,.-...-....,,m,

=_..,_....

Summary of Existence of Data For Class II Events (References in handouls)

Extensive " bottom-up" data for sheath / pressure tube expansion and/or sag Full assembly " top down" data fpr horizontal deformation and sheath rupture with internal heat source not located M964 DLH 0893 011 w-grw - me miyrr-..

=T-wr=-*

rr er

• --T'e--*w

=*e--

ewwwe "m

es--<s-.

e-

' we h -

• ee-w'-ge

-e--

3u w wm e-M4--'.r---.--o-+we-

==-se

-w es w = --

e.-e---e.--zu

,s*ur ww a

w.

-- e a-mw.

uy se i,e.-..

v-.

-w-4,-4 mm--

m.4

l L

v l

4 Prioritized Areas for CANDU Database Enhancements -

EC-1 dnd EC-Il Fuel Beha\\rior

• No Class I enhancement needs identified

• Full assembly deformation and burst tests (Horizontal axis with internal heat source &

CANLUB)

• Ft/H O fluid flow pressure differential 2

with deformed test assembly M964-OLH 0093 012

/=-

Computer Models Developed for CANDU EC-1 and EC-Il Analysis Parallel LWR Analysis Models CANDU LWR Description ELESIM

• FRAPCON - Single fuel rod. Normal operation ELESTRES FRAP-S GAPCON ELOCA

• FRAP-T

- Single fuel rod. Up to melt FROM

• COXIDE

- Full range Zircaloy oxidation model-CAFE

• FRAll

- Probabilistic failure model (Ramp SCC)

BEAM

• AXISM

- Detailed stresses, collapse pressure, lateral vibration frequencies, and buckling load BOW

• AXISM

- Endplate bending and bowing M964 DLH 0093 013 L

Modifications Required to Use LWR Codes for CANDUs FRAPCON - Add D O properties, Zry - 2.5 Nb properties, 2

CANLUB, CANDU power ramp integrity i

correlations, heat transfer. package for horizontal flow,and modify sheath oxidation modal for CANDU coolant FRAP-T

- Modify sheath deformation model to consider l

sag and gravity-driven pellet-sheath contact COXIDE

- None

~

FRAll

- Total rewrite. FRAIL has not been supported AXISYM Add consideration of lateral vibration frequencies

_-,--,--------._._.,,_,_.-.__;.,.-.~__..-............-._.,...~.-..._-...-_.,__,i..

... _.. >.,..... _ - -, - -.. _ _. -.. _. - ~. - ~ - - - -..... - _.,

.. ~

1 Conclusion' LWR code activity is so low that use of either LWR or CANDU Codes would be a cold start for United States' staff LWR Code Authors FRAPCON Gary Berna (208) 526-1985 Don Lanning (509).376-2212 Carl Beyer (509) 376-2382 FRAP-T Larry Siefken (208) 526-9319 AXISYM Gail Price (208) 526-9685 M964 DLH 0893 Ott 4

6

-e

.- e

,...e.---n

+--v-er -

o-r

...--m -

1=e v wwmew

-v re

+-e ar ies -;.---o,..:.r--r

.-wam=m*s*me, e

w

..-w,w--i

,yr--w e 1mvr =~ r n v-+-,

s-e-y, w*w w =tS'-r

Table 2.3.3.1 Data Types Useful for Confirming Key Phenomena in the Area of f uel Behavior for Representative CAN003 Sequences of Event Category I KEY CANDU3 DATA TYPES REPRESENTATIVE EC-1 SEQUENCES COMMENTS FUEL BEllAVIOR USEFUL FOR PilEN0MENA CONflRMING SB-A LOP-A SCRAM-A PilENOMENA PARAMETER RANGES Of DATA Integrity After Power CANDU Operating Ramp (PCHI)

Experience PIE and Stress Analysis flydriding CANLUB Residual flydrogen U0, Residual 11/)

PIE of fuel Without hi internal H.

(e.g.

I End Support Plates)

Separate Effects i

flydriding with Zry-4 and Zry-2.5 Nb s

y

=

e Table 2.3.3.1 Data Types Useful for Confirming Key Phenomena in the Area of fuel Behavior for Representative CANDU3 Sequences of Event Category I (Continued).

KEY CANDU3 DATA TYPES REPRESENTATIVE EC-1 SEQUENCES COMMENTS FUEL BEllAVIOR USEFUL FOR PilEN0HENA CONF 1RHING SB-A LOP-A SCRAM-A PNENOMENA PARAMETER RANGES OF DATA Vibration and Potential PIE of Fuel, Fuel for Fretting Dimensions, and Coolant Flow if Fretting Observed:

Bundle Normal Mode Tests Vibration Analysis Results Zry-4 / Zry-2.5 Hb Elastic Modul1 00, Elastic Modul1 (Compression)

..-......-.-,---.,-....r,.,--.----+.+.r

-+r---

--e.-.,-w*

Table 2.3.3.1 Data Types Useful for Confirming Key Phenomena in the Area of f uel Behavior for Representative CANDU3 Sequentes of Event Category I (Continued).

KEY CANDU3 DATA TYPES REPRESENTATIVE EC-1 SEQUENCES COMMENTS FUEL BEHAVIOR USEFUL FOR PilEN0MENA CONFIRMING SB-A l0P-A SCRAM-A PilENOMENA PARAMETER RANGES OF DATA Sheath Deformation

• U0, Densification 00, Swelling 00 Creep 2

Zry-4 / Zry-2.5 Nb Compressive Creep Versus Stress (Is the Creep-down Rate Faster Than Necessary Time Resolution?)

Sheath and Prepsure Rate Constants in Tube Oxidation CANDU Coolant Oxygen Potential Boundary Conditions-for Class 11 and til Event Analysis m,,

_. _..., ~ -........... _ -., _ -, - -.

.,._.,, ~. -.,. _

s Table 2.3.3.1 Data Types Useful for Confirming Key Phenomena in the Area of fuel Behavior for Representative CANDU3 Sequentes of Event Category I

{ Continued).

KEY CANDU3 DATA TYPES REPRESENTATIVE EC-1 SEQUENCES COMMENTS FUEL BEllAVIOR USEFUL FOR PilEN0HENA CONFIRMING SB-A 10P-A SCRAM-A PilEN0HENA PARAMETER RANGES OF DA1A fuel Temperatyres and UO,lleat Capacity Stored Energy UO Thermal Co$ductivity Zry-4 / Zry-2.5 Nb lleat Capacity Zry-4 / Zry-2.5 Nb Thermal Conductivity Zry 0xide Thermal Conductivity

/ Zry Interface UO,t Conductance With llea CANLUB Liner UO,ductance With/ Zry Gap lleat Con CANLUB Liner (Is the Creep-down Rate faster Than Necessary Time Resolution?)

Boundary Conditions for Class 11 and !!! Event Analysis l

Table 2.3.3.2 Data types Useful for Confirming Key Phenomena in the Area of fuel Behavior for Representative CANDU3 Sequences-of Event Category 11 KEY CANDU3 DATA TYPES REPRESENTATIVE EC-Il SEQUENCES COMMENTS FUEL BEllAVIOR USEFUL FOR PHENOMENA CONFIRMING LB-K L0fWTB B End-fitting Failure PHENOMENA PARAMETER RANGES Of DATA Sheath / Pressure Tube Integral Deformation Expansion and/or Sag Tests With Local Internal lleat Source Contact Zry-4 / Zry-2.5 Nb Stress-Strain-Strain i

Rate Equation Anisotropy Parameters? (Depends on Integral Test Results) i i

a Y

/

e

..-,_--,..m

-,,.. ~.

...-.,_..-...e-.

.,... _ - -......, - ~. - - _. - - - - - -.

s Table 2.3.3.2 Data Types Useful for Confirming Key Phenomena in the Area of fuel Schavior for Representative CAN003 SequePres of Event Category 11 (Continued).

KEY CANDU3 DATA TYPES REPRESENTATIVE EC-II SEQUENCES COMMENTS FUEL BEllAVIOR USEFUL FOR PilEN0MENA CONFIRMING LB-K 10fWTB-B End-fitting failure PilEN0MENA PARAMETER RANGES Of DATA Sheath Rupture fuel Rod Tests With Putential Local Internal lleat Source Contact of Sheath F

Sheath Rupture Criterion fuel Bundle Disassembly Integral (Full Assembly) Tests Sheath and Pressure Rate Constants Tube Oxidation 5

A

..... ~ -.

...c 1

Table 2.3.3.2 Data Types Useful for Confirming Key Phenomena in the Area of fuel Behavior for Representative CANDU3 Sequences of Event Category II (Continued).

KEY CANDU3 DATA TYPES REPRESENTATIVE EC-Il SEQUENCES COMMENTS FUEL BEllAVIOR USEFUL FOR

'-~

Pil[NOMENA CONflRMING LB-K 10fWTB-B End-fitting failure PilENOMENA

--~~~ - ~~

PARAMETER RANGES Of DATA Thermal Response

• UO,lleat Capacity UO lhermal Conductivity l

Zry-4 / Zry-2.5 Nb l

Ileat Capacity Zry-4 / Zry-2.5 Nb Thermal Conductivity Zry Oxide lleat Capacity l

Zry 0xide Thermal Conductivity UO,/ Zry Interface lleat Conductance With CANLUB Liner U0,ductance With/ Zry Gap lleat Con CANLUB liner Boundary Conditions for. Class III Event Analysis e

/

EVENT CLASS I FUEL BEHAVIOR DATA AS OF AUGUST I

23, 1993 i

INTEGRITY AFTER POWER RAMP CANDU Operating Experience M. R. Floyd, D. A. Leach, R. E. Moeller, R. R. Elder, R. J. Chenier and D. O'Brien, "Behaviour of Bruce NGS-A Fuel Irradiated to a Fuel. Chalk River. Canada. OctoberBurnup of - 500 MWh/kgU, 4-8 1992, pp. 2 2-60

[

Fuel Cladding - A Review," Journal of Nuclear Mate pp. 249-292 172, 1990, W. J.

Penn, R.

K.

Lo, J.

C.

Wood, "CANDU Fuel - Power Ramp Performance Criteria," Hyclear Technoloav 34, July 1977 pp. 249-268 T.

J.

Carter,

" Experimental Investigation of Various Pellet Technolocy 45, September 1379, pp. 166-176. Geometrie PIE and Stress Analysis M.

R.

Floyd, R. J. Chenter, D.

A. Leach, and R. R. Elder, "An Overview of the Examination of Fuel as Follow-Vo to the 1 November Overpower Transient in Pickering NGS-A Unit 1, " Thi rd International Conference on CANDU Fuel Chalk River Ca 4-8 1992, pp. 2 2-29.

M. J. F. Notley, *PRAMP - A Model for Calculating the Effect of Variable Power Histories on the Occurence of Fuel Defects," Third International Conference on CANDU Fuel. Chalk River. Pi Canada. October 4-8 1992_, pp. 5 5-20.

M. Tayal, E. Millen, R. Seinoha, G. Valli, "A Semi-Mechanistic Aoproach to Calculate the Probability of Fuel Defects," Third international Conference on CANDU Fuel. Chalk River. Ca 4-8 1992, pp. 5 5-37.

1 r,. Hallgrimson, M. Tayal, B. Wong, R. Aboud, *Recent Validations of the ELESTRES Code," Third International Conference on CANDil C_halk River. Canada. October Fuel.

4-8 1902, pp. 5 5-65.

P. K. Chan, K. G. Irving, and J. R. Mitchell, "The Role of Zr I C in Minimizing Stress Corrosion Cracking in Fuel Cl adding, ", Thi rd Jnternational Conference on CANDU Fuel. Chalk River. Canada, Octobe 4-8 1992, pp. 7 7-40.

HYDRIDING CANLUB Residual Hydrocen re Fuel Defect Investigation at Point Lepreau "

i n,

]

Detober 4-e ]p92, pp. 2-3O CANDU Fuel. Chalk River. Canada.

24

t 00 Residual H O

~

2 2

LWR Data Base PIE of Fuel Without Internal H.

(e.g. End Support Plates)

M. R. Floyd, D. A. Leach, R. E. Moeller, R. R. Elder, R. J. Chenier and D. O'Brien, "Behaviour of Bruce NGS-A Fuel Irradiated to a Burnup of - 500 MWh/kgU," Third International Conference on CANDU Fuel. Chalk River. Canada. October 4-8 1992, pp. 2 2-60.

Vincent F. Urbanic, Brian D. Warr, Angela Hanolescu, C. K. Chow, and Michael W. Shanahan, " Oxidation and Deuterium Uptake of Zr-2.5Nb l

Pressure Tubes in CANDU-PHW Reactors, Zirconium in the Nuclear Industry:

Eiohth International Synoosium. ASTM STP 10?3, L. F. P.

Van Swam and C. M. Eucken, Eds., American Society for Testing and Materials, Philadelphia, 1989, pp. 20-34.

V.

F.

Urbanic, B.

Cox, G.

J.

Field, "Long-Term Corrosion and Deuterium Uptake in CANDU-PHW Pressure Tubes, Zirconium in the Wuclear Industry:

Seventh international Symposium.

ASTM bTP 939, R. B. Anderson and L. F. P. Van Swam, Eds., American Society for Testing and Material,, Philadelphia,1987, pp.189-205.

Separate Effects Hydriding with Zry-2 and Zry-4 LWR Data Base Separate Effects Hydriding with Zry-2.5 Nb See references under " PIE of Fuel Without Internal H".

VIBRATION AND POTENTI AL FOR FRETTING PIE of Fuel, Fuel Dimensions, and Coolant Flow J. Judah, " Overview of Fuel Inspections at the Darlington Nuclear Generating Station," Third International Conference on CANDU Fuel.

Chalk River. Canada. October 4-8 1992, pp. 3 3-22.

T. J. Carter, K. M. Wasywich, M. R. Floyd, R. R. Hosbons, and M. G.

Maguire, "An Overview of Fuel Ex.minations at Chalk River and Whiteshell Laboratories in Supp;rt of the Darlington Fuel Examination," Third International Conference on CANDU Fuel Chalk River. Canada. October 4-8 1992, pp. 3 3-36.

M. P. Paidoussis, "An Experimental Study of Vibration of Flexible Cylinders Induced by Nominally Axial Flow," Nuclear Science and Enoineerino 35, 1969, pp. 127-138.

Bundle Normal Mode Tests M. Tayal, B. J. Wong, J. H. K. Lau, A. M. Nicholson, " Assessing the Mechanical Performance of a Fuel Bundle:

BEAM Code Description,"

Third International Conference on CANDU Fuel. Chalk River. Canad11 Octcber 4-8 1992, pp. 3 3-63.

I Vibration Analysis Results M.

J.

Pettigrev.

Flow-induced Vibration Analysis of Nuclear i

Components, AECL-6219, August 1978.

M. Tayal, B. J. Wong, J. H. K. Lau, A. M Nicholson, " Assessing the Mechanical Performance of a Fuel Bundle:

BEAM Code Description,"

Third International Conference on CANDU Fuel. Chalk River. Canada.

October 4-8 1992, pp. 3 3-63.

M. Cojan, " Fuel Vibration and Fretting Wear," Third international Conference on CANDU Fuel. Chalk River. Canada. October 4-81992, pp.

7 7 - f>5.

Zry-2 and Zry-4 Elastic Moduli LWR Data Bzse Zry-2.5 Nb Elastic Moduli M. E. Rosinger, "The Elastic Properties of Zirconium Alloy Fuel l

Cladding and Pressure Tubing Materials," Journal.of Nuclear Materials 79, 1979, pp. 170-179.

i D. O. Northwood, I. M. London, L. E. Bahen, " Elastic Constants of Zirconium Alloys," Journal of Nuclear Materials 55,1979, pp. 299 -

310.

U0 Elastic Moduli (Compression) 2 LWR Data Base SHEATV nCFORMATION 00 Densification 2

LWR Data Base 00 Swelling 3

LWR Data Base UO Creep 2

LWR Data Base Zry-2 and Zry-4 Compressive Creep Versus Stress M. J. F. Notley, "PRAMP - A Model for Calculating the Effect of Variable Power Histories on the Occurence of Fuel Poer Ramp Defects," Third International Conference on CANDU Fuel. Chalk River.

Canada. October.4-8 1992, pp. 5 5-20.

Zry-2.5 Nb Compressive Creep Versus Stress M. J. F. Notley, *PRAMP - A Model for Calculating the Effect of Variable Power Histories on the Occurence of Fuel Po e Ramp Defects," Third International Conference on CANDU Fuel. Chalk kiver.

Canada. October 4-8 1992, pp. 5 5-20.

Allan R.Causey, Richard A. Holt, and Stuart R. Macewen, "In-Reactor Creep of Zr-2.5 Nb,"

Zirconium in the Nuclear industry:

Sixth International Syroosium. ASTM 824. D. G. Franklin and R. B. Adamson.

Eds. A,nerican Society for Testino and Materials 1984, pp.269-288.

F

I s

i SHEATH AND PRESSURE TUBE OXIDATION

'f Zry-2 and Zry-4 Rate Law / Constants in CANDU Coolant V.

F.

Urbanic, B.

Cox, G.

J.

Field, "Long-Term Corrosion ' and i

Deuterium Uptake in CANDU-PHW Pressure Tubes, Zirconium in the j

Nuclear industry:

Seventh International Svmoosium.

ASTM STP 939, i

R. B. Anderson and L. F. P. Van Swam, Eds., American Society for Testing and Materials, Philadelphia,1987, pp.189-205.

Zry-2.5 Nb Rate Law / Constants in CANDU Coolant M. R. Floyd, D. A. Leach, R. E. Hoeller, R. R. Elder, R. J. Chenier l

and D. O'Brien, "Behaviour of Bruce NGS-A Fuel Irradiated to a i

Burnup of - 500 MWh/kgU," Third International Conference on CANDU Fuel. Chalk River. Canada. October 4-8 1992, pp. 2 2 50.

1 i

Vincent F. Urbanic, Brian D. Warr, Angela Manolescu, C. K. Chow, and Michael W. Shanahan, " Oxidation and Deuterium Uptake of Zr-2.5Nb i

Pressure Tubes in CANDU-PHW Reactors, Zirconium in the Nuclear Industry:

Eichth International Synoosium. ASTM STP 1023, L. F. P.

Van Swam and C. M. Eucken, Eds., American Society for Testing and

(

Materials, Philadelphia, 1989, pp. 20-34.

j V.

F.

Urbanic, B.

Cox, G.

J.

Field, "Long-Term Corrosion and j

Deuterium Uptake in CANDU-PHW Pressure Tubes, Zi'rconium in the Nuclear Industry:

Seventh International Symoosium.

ASTM STP 939, R. B. Anderson and L. F. P. Van Swam, Eds., American Society for Testing and Materials, Philadelphia,1987, pp.189-205.

FUEL TEMPERATURES AND STORED ENERGY UO Heat Capacity i

2 LWR Data Base UO; Thermal Conductivity I

LWR Data Base j

s Zry-2 and Zry-4 Heat Capacity i

LWR Data Base j

1 Zry-2.S Nb Heat Capacity Zry-2 and Zry-4 Thermal Conductivity

[

LWR Data-Base Zry-2.S Nb Thermal Conductivity I

+

Zry 0xide Thermal Conductivity LWR Data Base UO / Zry Interface Heat Conductance With CANLUB Liner l

g l

i

[

U0 / Zry Gap Heat Conductance With CANLUB Liner 2

l I

P EVENT CLASS 11 FUEL BEHAVIOR DATA AS OF AUGUST 23, 1993 SHEATH / PRESSURE TUBE EXPANSION AND/OR SAG Integral Deformation Tests With Local Internal Heat Source Contact Note:

The PBF tests were not horizontal and the NRX X2 loop tests may not have been.

i J. A. Walsworth, P. J. Fehrenbach, J. H. K. Lau, and E. Kohn, Calculated and Measured Behaviour of Zircalov Fuel Sheaths durino in-Reactor LOCA Tests, AECL-9268, October 1986.

Zry-2 and Zry-4 Stress-Strain-$ train Rate Equation S. Sagat, H. E. Sills, J. A. Walsworth, D. E. Foote, and D. F.

Shields, Deformation and Failure of Zircalov Fuel Sheaths Under LOCA Conditions, AECL-7754, October 1982 2ry-2.5 Nb Stress-Strain-Strain Rate Equation

Amin, S. Rizkalla, Rameshwar Choubey, and John J. Jonas, "Fl ow Softening and Anneal Hardening in Two-Phase Zr-2.5 Nb," Zirconium in the huclear Industry: Sixth International Symoosium. ASTM 824. D.

G. Franklin and R. B. Adamson. Eds.. American Society for Testino and Materials. 1984, pp. 176-199.

Zry-2 and Zry-4 Anisotropy Parametei; LWR Data Base Zry-2.5 Nb Anisotropy Parameters Allan R.Causey, Richard A. Holt, and Stuart R. Macewen, "In-Reactor Creep of 2r-2.5 Nb,"

Zirconium in the Nuclear Industry:

Sixth international Symoosium. ASTM 824. D. G. Franklin and R. B. Adamson.

Eds.. American Societ y for Testino and Materials.1984, pp.269-288.

SHEATH RUPTURE POTENTIAL Fuel ~ Rod Tests With Local Internal Heat Source Contact of Sheath Note: The PBF tests were not horizontal and the NRX X2 loop tests may not have been.

James Adams, Robert McCormick, Richard K.

McCardell, Zoel R.

Martinson, Peter Kalish, PBF.CANDU Fuel Element loss-of-Coolant i

Accident Exceriment Test Results Recort, EGG-2384, May 1985.

J.

A.

Walsworth, P. J. Fehrenbach, J. H.

K.

Lau, and E. Kohn, Calculated and Measured Benaviour of Zircalov Fuel Sheaths durino in-Reactor LOCA Tests, AECL-9268, October 1986.

'i Zry-2 and Zry-4 Rupture Criterion S. Sagat. H. E. Sills, '.

A.

Walsworth, D. E. Foote, and D. F.

Shields. Deformation and Failure of Zircalov Fuel Sheaths Under LOCA Concitions, AECL-7754, October 1982.

Iry-2.5 Nb Rupture Criterion

i FUEL BUNDLE DISASSEEEMBLY Integral (Full Assembly) Tests P. M. Mathew, D. G. Evans, S. Wadsworth, E. Kohn, M Notley, G. I.

Hadaller, " Disassembly' Behaviour of CANDU Bundles under impact,'

Third International Conference on CANDU Fuel. Chalk River. Canada.

October 4-8 1992, pp. 4 4-10.

t SHEATH AND PRESSURE TUBE OXIDATION Zry-2 and Zry-4 Rate Law / Constants l

LWR Data Base Zry-2.5 Rate Law / Constants THERMAL RESPONSE 00 Heat Capacity 2

LWR Data Base UO Thermal Conductivity 2

LWR Data Base Zry-2 and Zry 4 Heat Capacity e

LWR Data Base r

Zry-2.5 Nb Heat Capacity Zry-2 and Zry-4 Thermal Conductivity LWR Data Base Zry-2.5 Nb Thermal Conductivity Zry 0xide Heat Capacity LWR Data Base Zry Oxide Thermal Conductivity LWR Data Base 00 / Zry Interface Heat Conductance With CANLUB Liner 2

U0 / Zry Gap Heat Conductance With CANLUB Liner 2

P 1

i I

t

Anthony L. Wright OAK RIDGE NATIONAL I.ABORATORY i

SEVERE ACCIDENT DATA BASE ASSESSMENT FOR THE CANDU 3 IMACTOR DESIGN Presented at the NRC/AECL Meeting August 25,1993

...,-,,.-,.,.e..

...-+,

= - + - - -

N

^- * --

~ " ' " * * * - * ~ ' " " ' ' ' ' * ' ' * " ' " * * ' ' * ~ * " * ' ' ' ' ' ' ' ' '

~

OUTLINE FOR THIS PRESENTATION:

Data Base Assessment Methodology Description of Event Sequences Used as the Basis for the Assessment Discussion'of Relevant Severe Accident Phenomena

~

Discussion of Preliminary Results of Severe Accident Data Base Assessment Discussion of NRC Code Modification Needs Summary and Conclusions

=..

= - -.-.-

r THE NRC REQUESTED ORNL TO PERFORM A SEVERE ACCIDENT DATA BASE ASSESSMENT FOR THE CANDU 3 REACTOR ~ DESIGN i

ASSESSMENT METIIODOLOGY:

Identify. Representative Severe Accident Sequences Based on Information in the ORNI.

CANDU 3 Systems Analysis Study (NUREG/CR-6065).

1 Compile Lists of the Relevant Severe Accident Phenomena Expected in These Severe Accident Sequences.

Assess the Existing Severe Accident Data Base for its Coverage of These Phenomena.

GROUND RULE:

Experimental Data Requirements Should Not Be More Stringent Than Those Currently Employed for Existing LWR Designs Identify Phenomenological Areas Where Experimental: Data Base Enhancements Are Needed Identify Modifications Needed So' That MELCOR Could Be sed For CANDU 3 Severe Accident Analyses www

.--e,,

.. +,~,-w--.,,.,.,.--..-,-.r-.

w--,-*r.

r-w--

,,.--ei xy, v,,,.

.rr-w,~.,%-e,~%

v,

.-. + ~u+

-+,.-.,,. - -,,,.,.-.-

5-m.,.

~,,ev,.--.-.--w,

I t

" EVENT SEQUENCE" AND " CONSEQUENCE CLASS" DEFINITIONS IlSED IN TIIE CANDU 3 SYSTEMS ANALYSIS STUDY EVENT CATEGORY DEFINITIONS:

EC-1:.

~ Anticipated Operational Occurrences EC-II:

~ Design Basis Accidents EC-III:

~ Severe Accidents EC-IV:

- ~ Residual Risk i -

CONSEQUENCE CLASS DEFINITIONS:

CC-1:

No significant fuel damage, no FP release from fuel.

CC-2:

. Moderate fuel damage, release of fuel rod gap inventory, coolable-geometry maintained.

CC-3:

Extensive core damage, large FP release from fuel and potential relocation of core debris.

THE MAIN FOCUS OF OUR ASSESSMENT IS ON EVENT SEQUENCES wrf1I CC-3 CONSEQUENCES.

CANDU 3 EVENT SEQUENCES USED AS BASIS FOR THE SEVERE ACCIDENT DATA BASE ASSESSMENT NUMBER ALTERNATE EVENT DESCRIPTION' IDENTIFIER

• EC-IV-1 LB-AC A large-break loss-of-coolant sequence where the reactivity systems SDSl and SDS2 do not work, resulting in a positive reactivity insertion to the fuel and subsequent major core damage.

EC-IV-2 LB-G A large-break loss-of-coolant sequence where the high-pressure injection (llPI) portion of the emergency core-cooling system (ECCS) fails, and the moderator cooling system (MCS) does not cool the fuel.

EC-Ill-1 SBLFB-II A small-break loss-of-coolant sequence, initiated by a feeder-tube break, where steam-generator crash cooling (CC) fails so that the high-pressure ECCS injection cannot be initiated. llowever, in this sequence the MCS does not fail and can cool the afrected fuel channels.

Source: NUREG/CR-0665 (ORNL/FM-12396), Systems Analysis of the CANDU 3 Reactor (to be published). The " alternate identifier" noted above is the event sequence identifier from the NUREG report.

1 In NUREG-6065, EC-IV-1 AND EC-lV-2 had CC-3 Consequences, and EC-Ill-1 had CC-2 Consequences.

a e

+ +, -

a

1 CANQU3 c

. E..

._A g

m,-.. a:.o.=

3

%<v p#

.ZH

., _m e d-1

'W

~

PvIdF,,,-

i

,m....s, L

i p

== =o.n s

)

.or sgw naziottsows eines I

s-,uo :.

/

5

_, c f

.t a

.i w* ~.g.*

,E;

.rs "d

l A

a

)

oc t "'*

• fl

$ *W T

tr y~

y

{

fi a

s 1

C

.,,c

1. g

c

r,

"n s m

[

Ii ! I t

i a hiv 2

~

h

^

px v7 6

3.

ca usema of CK**EA tNo p* ELD SatLL M

p.au mes pygt c>usenet AssemLY

& alls M

i S" ".," \\

Asso ENO FTfTodB

" J"S::;'

J F10URE 2.1 REACTOR ASSMT 2-6 a

D k

2 Appendix E j

$b 3

gO Z

s\\

5 4

\\

5 U

i s

t i

k g

h E

E

\\

s l

• x

\\

\\.

\\\\

i

\\

h j

s 1

r 3

\\

3

\\

8 I

\\

\\,

l i

a E

\\,-

h 4

Ey

\\

4 i

'\\\\

5

=

i

\\

g

'. - \\

w B.

s s1 i

5 1

a SN'CR-6065

!!6

l' Appendix A CRNL-DWG $3-3204 ETD 1CANDU3 r

1 ww\\x N

'N FUEL ELEMENT

\\WW\\\\

D$ PRIMARY COOLANT 0 0 MODERATOR 2

PRESSURE TUBE CALANDRIA TUBE J

Figure A 2 Unit cell for 37-element fuel 117 NUREG/CR-6065

4 POSSIBLE ACCIDENT PROGRESSION DESCRIPTION FOR CANDU-3 EC-IV-1 Rapid Fuel Heatup and Fuel Melting Rupture of Pressure and Calandria Tubes Rupture of Calandria Vessel, Loss of Moderator in Calandria Vessel, Shutdown of Reactivity Insertion Pressurization of Containment Due to Calandria Vessel Rupture Melt Progression for Disrupted Fuel within the Calandria Vessel:

Debris / Melt Behavior in Calandria Vessel; a

Possible FCIs?

Calandria Vessel Melt-through?

Fuel Behavior in Shield Tank?

Shield Tank Melt-through, Core-Concrete i

Interactions?

l

i.

\\

J POSSIBLE ACCIDENT PROGRESSION FOR CANDU-3 EC-IV-2 l

Fuel Heatup by Decay Heat, Zircaloy Oxidation, Fuel /Zircaloy Eutectic Formation Fuel Channel Ballooning and/or Sag; Pressure Tube Contact with Calandria Tube Moderator Heat Transfer due to PT/CT Contact (Assumed Unsuccessful for EC-IV-2")

Continued Fuel Channel Heatup, PT/CT Rupture Fuel Failure Propagation in Calandria Vessel, Potential for FCIs Debris / Melt Behavior on Calandria Floor, Calandria Vessel Failure?

Debris / Melt Behavior in Shield Tank, Potential for FCIs, q

Shield Tank Failure, Core-Concrete Interactions 9 a

Challenge to Containment due to Fuel Failure Propagation EC-III-l would end with successful PT/CT contact moderator heat transfer.

t i

MA.IOR SEVERE ACCIDENT PIIENOMENOLOGICAL AREAS EVALIIATED IN TIIIS STUDY:

PilENOMENOLOGICAL AllEA GENERAL ASSESSMENT Fuci n havior and Core Melt Progression Experimental Data Needs Exist for Some e

Phenomena Fuel-Coolant Interactions Existing LWR Phenomena Data flase (Steam Explosions)

Adequate, But Integral Tests May Be Needed For CANDU 3 Fuel Melt Configurations Reactor Vessel Failure:

Canadian Data Base Adequate (Calandria Tubes, Calandria Vessel and Shield Tank) liigh-Pressure Melt Ejection and DCll Not a Consideration for CANDU 3 Core-Concrete Interactions Existing LWR Data Base Adequate Ilydrogen Combustion Existing Canadian and LWR Data llase Adequate Containment Failure Data Base Existing Canadian and LWR Data 13ase Adequate Integral Test Needs -

Integral Test Needs identified

,n.

+

y

-. ~,

.-.--n

,r.

th

'O

'q 2.5 lixistence of Severe Accident Data Bases liir CANI)ll 3 Tabic 25.2 lisistence of 1 uel llehavior and Core Melt Progression Data llases liir Representative CANDD 3 I!C-Ill and liC-IV Sequences CANDU 3 FUEL DEllAVIOR EXISTING AND COVER AGE OF COMMENTS AND CORE MELT PLANNED DATA DASES OVER AI.I.

PROGRESSION PilENOMENA DATA II ASI!

Initial Fuel lleatup Existing 1.WR Data flase, Same as existing LWR Canadian Data Base coverage Clad Sheath Failure Existing LWR Data Base, Adequate Canadian Data Base Zircaloy Metal Oxidation Existing I,WR Data Base, Same as existing LWR Canadian Data Base coverage Fuel /Zircaloy Eutectic Formation Existing LWR Data Base, Same as existing LWR Canadian Data Base coverage Fuel Channel Ballooning and/or Sag Canadian Data Base Adequate Moderator llcal Transfer During Canadian Data Base May De inadequate Canadian Data Base based on Pressure /Calandria Tube Contact experiments performed in early 1980's with fuel pin simulators; not clear that results represent actual fuel bundle perfonnance.

Pressure Tube Rupture Canadian Data Base Adequate Calandria Tube Rupture Canadian Data Base Adequate Fuel Failure Propagation within the None Identified Inadequate Overall Nature of Failure Calandria Vessel Propagation Process Not Known.

l Debris / Melt Behavior on the -

Existing LWR Data Base May De inadequate Nature / Geometry of Fuel Debris l

Calandria Vessel Floor Ded Not Known.

l

m-Fucl Failure Propagation within the None identified inadequate Overall Nature of Failme Shield Tank Propagation Process Not Knimo.

Debris /htelt lichavior on the Shield lixisting 1.WR Data llase hiay lle inadequate Nature / Geometry of Fuel Debris Tank I loor lied Not Known.

Transport of Alcit/ Debris through fixisting 1.WR Data Base Same as existing 1.WR Failed Shield Tank to Concrete coverage liasemat Rapid Energy insertion to Fuel lixisting 1.WR/RI A' Data hiay lie inadequate I.WR/EIA* data may not be Channels llase applicaole to CANDil fuels Fuel Channel Breakup Existing LWR /RIA* Data hf ay Be inadequate I WR/RIA* data may not be Base applicaSle to CANDtl linels Pressure Tube and Calandria Tube Canadian Data Base Adequate Failure Criteria for Pressure Pulses Failed Fuel Propagation within None inadequate Overall Nature of Failure Calandria Vessel Propagation Process Not Known

• RIA: Reactivity insertion Accident 4

P g

4 rm-

- - + -,--~

-,...-a

-nv

-,---.-,,v-,,

+

---

r a

,v-a

}

THE LWR /RIA DATA BASE MAY BE INADEQUATE FOR j

CANDU 3 LOCA-WITHOUT-SCRAM SEVERE ACCIDENT ASSESb4ENTS The following comes from an AECB report titled "CANDU FUEL BEHAVIOUR I

DURING LARGE BREAK LOSS-OF-COOLANT POWER TRANSIENTS" (V. J.

Langman, March 1985)

"The LWR test fuel design and RIA power transient conditions vary considerably from-CANDU fuel and LOCA power transients."

l "For example, the LWR test fuel design used in the RIA tests and the CANDU fuel.

design have the following differences which may affect fuel behaviour in rapid power l

excursions;"

I 1.

U-235 enrichments of the LWR test fuels are in the mnge 5 to 20 per cent where:

CANDU fuel is 0.7 per cent; i

2.

LWR test fuel is smaller in diameter than CANDU fuel; i

3.

LWR test fuel has a thicker fuel sheath than CANDU fuel; l

4.

CANDU fuel sheath has bearing pads and/or spacers and associated heat-treated zones; LWR test fuel does not; 5.

LWR test fuel vertical; CANDU fuel horizontal; 6.

CANDU fuel elements have a complex radial flux distribution; LWR test fuel generally has a simple single-element flux distribution; i

7.

LWR test fuel free-standing with fuel / sheath gap; CANDU fuel collapses under normal operating conditions and has no sheath gap; and

'l l

8.

Void volume of LWR test fuel much larger than in CANDU fuel.

l f

i

t i

LWR /RIA AND CANDU 3 LOCA-WITHOUT-SCRAM DIFFERENCES (cont.)

"Also, RIA power transients are different from those in CANDU LOCA's in the following respects, which may effect fuel behaviour in rapid power excursions:"

j 1.

The initial fuel power in RIA tests is zero; CANDU fuel is at full power at the beginning of a LOCA; 2.

Initial fuel temperature distribution is relatively flat in RIA tests; parabolic in -

CANDU fuel at power; i

3.

Initial volume-averaged fuel temperature is much lower in RIA tests than for CANDU fuel at power; 4.

Rate of power rise in RIA tests is more rapid than for CANDU LOCA power pulses; and 5.

Duration of power excursion in RIA tests is shorter than for CANDU LOCA power transients.

i THE AECB REPORT CONCLUDED THAT:

the results of the LWR RIA and OPTRAN tests cannot be directl applied to the behaviour of CANDU fuel during large break LOCA power transients."

,.--s

PRELIMINARY

SUMMARY

OF PRIORITIZED AREAS FOR CANDU 3 SEVERE ACCIDENT DATA BASE NEEDS PitIORITY PIIENOMENA COMMENTS I

Rapid Energy insertion to CANDU Fuel Integral tests with CANDU horizontal fuel geometries may be Channels, and Fuel Channel Dreakup needed.

I Moderator IIcal Transfer During Adequacy of moderator heat transfer cooling of fuel for Pressure /Calandria Tube Contact LOCA without ECCS is critical to AECL claim that no melt progression occurs in CANDU 3 accidents. Integral confirmatory tests may be needed with prototypic CANDU 3 fuel channel geometries.

2 Fuel Failure Propagation and Nature of fuel failure propagation in calandria vessel not Debris / Melt Behavior within the known.1his would influence the character of the debris / melt Calandria Vessel that could attack the calandria vessel floor. Separate effects phenomena tests may be needed.

3 Potential for Fuel-Coolant Interactions There is an extensive LWR data base related to 1. Wit fuel-with CANDU 3 Molten Fuel and coolant interaction phenomena. Ilowever, integral tests may Coolant Configurations be needed to identify FCI potentials for CANDU 3 fuel melt and coolant geometries.

3 Fuel Failure Propagation and Fuel behavior in shield tank for severe accidents dependent Debris / Melt Behavior within the on fuel behavior in calandria vessel. Separate efTects Shield Tank phenomena tests may still be necessary.

I i

NRC CODE MODIFICATION NEEDS FOR CANDU 3 SEVERE ACCIDENT-MODELING The MELCOR Code Will Be Used By The NRC To Perform CANDU 3 Severe Accident Assessments For EC-IV-2 Type Accidents Much Of MELCOR Would Not Have To Be Modified For CANDU 3 EC-IV-2 Modeling (Hydrogen Combustion, FCIs, Core-Concrete Interactions, etc.)

A New MELCOR Core Package Would Have To Be Written For EC-IV-2 Modeling:

Pressure /Calandria Tube Fuel Channel Model Fuel Failure Propagation Model Possible Modifications to Debris / Melt Progression Models Calandria and Shield Tank Failure Models CANDU 3 Specific Geometry Models MELCOR Is Not Capable Of Modeling Rapid Energy insertion And Fuel Channel Breakup Phase Of EC-IV-1 (LOCA-Without-Scram) Accidents i

O O y0 6

SUMMARY

AND CONCLUSIONS A Structured Severe Accident Data Base Assessment is Being Performed For The CANDU 3 Reactor Design The Assessment Is Based On CANDU 3 Event Sequences That Are Expected To llave Severe Accident Consequences (Extensive Core Damage)

Summary of Preliminary Experimental Data Base Assessment Results:

Much of the LWR and Canadian Data Base is Applicable to CANDU 3 Severe Accidents Ilowever, Experimental Needs IIave Been Identified Associated With:

1.

Rapid Energy Insertion to CANDU 3 Fuel, and Fuel Channel Breakup, 2.

Moderator Heat Transfer During PT/CT Contact, 3.

Potential for FCIs for CANDU 3 Fuel-Melt / Coolant Configurations, and 4.

Fuel Failure Propagation and Debris / Melt Behavior in the Calandria Vessel and Shield Tank.

A New MELCOR Core Package Will llave To Be Developed To Pennit Use Of MELCOR For CANDU 3 Severe Accident Modeling MELCOR Is Not Capable Of Modeling Rapid Energy Insertion And Fuel Channel Breakup In LOCA-Without-Scram Accidents

. 6 PLAN FOR INDEPENDENT IN-HOUSE ANALYSIS OF COOLANT VOID REACTIVITY-IN CANDU 3 Donald E. Carlson i

NRC/RES August 25,1993 t

i f

meartv 5 T w' crW W wV W 1* tg'MWNL'N trs1T99B 7' wd we dt"6-r 'h'W'WemN+4D4"

'*4-hv't' 48FV'iqa4'M

"W4*N"Tr&*6 % et-'T 5mwiW*-

T 9'*4-*eWDehJf'

• E**

swe-Ta WP P4944 4--

it-w 'e 9-

+-epP-e=+9e-es=~e-e,

&ei e &

www-'>t

-h+v w-ewe

+de

-e me t-ereeowe*=

e

INDEPENDENT ANALYSIS OF CANDU-3 COOLANT VOID REACTIVITY Objectives 1.

Check the accuracy of PPV-calculated void reactivities by comparing them to those calculated with COMBINE /ANISN and with MCNP.

2.

Develop a clear understanding of the underlying mechanisms of coolant void reactivity.

3.

Test our understanding of void reactivity mechanisms by hypothesizing design changes that would make the coolant void reactivity negative.

6 6

5 A $$

3

-.,,,,__.-...,e.,,,...,

,.,m,~,..,,,--,.,,,m,g._.e.,m...,-,....w.,

,,,,,,y.,~,-

,--..<..,,-.4 n,......

...m.

..y,.

...m,.

.,__m..

s,

,.m

c...

INDEPENDENT ANALYSIS OF CANDU-3 COOLANT VOID REACTIVITY Phase-I Analvsis Plan Step 1. Perform and compare PPV and COMBINE /ANISN calculations of infinite lattice void reactivity for selected cases with varying:

fuel burnup fuel temperature moderator boron level moderator temperature coolant purity coolant temperature (constant density) coolant void fraction

I INDEPENDENT ANALYSIS OF CANDU-3 COOLANT VOID REACTIVITY Phase-I Analysis Plan (Cont.)

Step 2. Perform COMBINE /ANISN calculations to examine the infinite lattice void reactivity reductions that would result from hypothetical changes in:

fuel enrichment lattice spacing l

.other parameters l

I l

~...

! ~

a...

INDEPENDENT ANALYSIS OF CANDU-3 COOLANT VOID REACTIVITY Phase-l Analysis Plan (Cont.)

Step 3. Use the COMBINE /ANISN results from Steps 1 and 2 to derive preliminary insights into the void reactivity mechanisms and compare these insights with explanations found in the CANDU literature.

Step 4. Perform additional COMBINE /ANISN calculations as needed to resolve outstanding questions.

.m

- -. ~ ~. -

m

INDEPENDENT ANALYSIS OF CANDU-3 COOLANT VOID REACTIVITY Phase-Il Analysis Plan For selected cases from Phase I, perform spot-checking calculations with the continuous-energy Monte Carlo code, MCNP.

Run additional PPV and COMBINE /ANISN cases to resolve differences as needed.

4 l

l l

l l

m e

---e-

-,. = - -. -.. =,

w,

i I

i Flux per unit lethargy

~'

G

.=A a

O m

.O

.O o

o ao O

. o

'o O

n L

k O.

O L

4

~%

me i

a O

r O

6 A

o L

m 3

t R

e.A Q

J

< Q t

O l

i A

I o

6 l

u

< T1 i

9. o l

-.4 O

L Q. o CD Q.

l Q. CD l

C.

l O

4 O

O.

i o

3 3

s

~

Q E

l o

6 a

w 9

M I

O tlD i

)

I

. ENCLOSURE 3

,t :

i 4

MEETING CANDU 3 PROJECT NO. 679 NRC - AECL/T AUGUST 25,_1993 NAME ORGANIZATION PHONE 1.

I d

S(4LG~iTI fl/LRlAn M,U.CR.t3t?.

(20 *).<0i,t - lI 0 4

// ir i

>>'rf T; l.

2 rs

,,-m'5 2.

-=

3. A f. hhotb <AJGtti AECL car >bb (t w n E M o y o

^

~,c r o n_ <s. S

• M L4-AECL-thcLA ACCL.

0 4.

5. ALy Ha r:(1 A k AFCA

/4 t t199T-1793 s.Ldw eD A. rmis

• O isci-as=ssn.us

%o 5 N 2 i-OI y2ns-4

,. MAG 8Y Eti-NB~/ARY A6c8 (O2) 991- 0(Il

%ded Y OuAp<

AFCt-Tc~il

.dat 4/7 oo </

8.

9.

$$o^'s $. UnbAb ONS-4 /E* EVY~~ b8 8 hlAn Ra cil NCClCES (bilif9z-3559 10.

%8'c 7. M (f h A6c L - LAdd

/4//d ru-9040 11.

12.

OEM So u B6^/

A CC G Ch w h.!

4th) i k3 -40 %

13..k.c u l__3 0 N

_T.t_/_C_/ _

6 oEl, 52 6-70) m 14.

Eowa nev'T*4Em NEL/ADAA/90AA 2.08-fo 4 -1Itf Ten > lA k (' r^$6cn hlRN-lAf.NlRf5h hI-492-OOIA 15.

Ar!o dle ve"h Ah7b/FF3//2PSE

.3C1-492-3E3L 16.

~* " ? :> %N>

3 at - Y Y3 ~SfL' 17.

18.

Den %rmn Ef;W: rd Ju, /zust.) {act) SM mo 19.

$6L CYJm m Y$ b 01 in

$YEo l3d$ $$$' YI h

DM d Eber[

WRC / EEG[RPTG 3el-Pf2-si3&[

20.

21.

22.

23.

24.

25.

R s' ~*

28.

29.

30.

.i*

AECL Technologies November 5, 1993

! is a meeting summary by David Ebert, RES.

Presentation view graphs are provided as Enclosure 2.

The meeting attendees are identified in.

N "-;r2,M Uy Dino Scaletti, Sr. Project Manager Advanced Reactors Project Directorate Division of Advanced Reactors and License Renewal Office of Nuclear Reactor Regulation

Enclosures:

1.

Meeting Summary 2.

Presentation Viewgraphs 3.

List of Attendees cc w/ enclosures:

See next page Distribution:

Distribution:

(w/all encls.)

(w/o Enclosure 3)

WUpshaw, DIP 3/H/5 Central File TEMurley 12G18 RMeyer NLS 169 PDR FMiraglia 12G18 PShea PDAR R/F DMCrutchfield ACRS (10)

P 315 DCScaletti WDTravers OGC 15B18 EDThrom EJordan MNBB 3701 DCScaletti CANDU R/F l'

0FC LA:PDST __ %

SP Ah APD:PDAR NAME PShe'aibsk DScaletti EThrom Y DATE 11/ \\ /93 11/6/93 11/8/93 OfflCIAL P.ECORD COPY DOCUMENT NAME: CANDSUMB.25

,