ML20028B499
| ML20028B499 | |
| Person / Time | |
|---|---|
| Site: | Dresden  |
| Issue date: | 10/31/1982 |
| From: | Chon Davis EG&G, INC. |
| To: | Odan F NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| References | |
| CON-FIN-A-6047 EGG-NTAP-6090, NUDOCS 8212020110 | |
| Download: ML20028B499 (243) | |
Text
{{#Wiki_filter:h$ $W O $Y Mff0k $Vi eef$ $ EGG-NTAP-6090 October 1982 V AN ANALYSIS OF A SPECTRlDi 0F LARGE-BREAK LOCAs IN A BWR/3 USING TRAC-BD1 Volume 1: Overview and Summary i
- C. B. Davis Idaho Nation,al Engineering Laboratory Operated by the U.S. Department of Energy ,
- . . _ . 3, h * @y '
N ' N mur- s; % 4W[ ,
]1 f yy- .; A'.? ,. 4_ -
4 s
;_, pry p:&.da :c -
MkEWMC5-IN c ?' 1
%-- ari n gs 6 :
F A--. v w v 2 T * * " , , q .i- p p ^ -
- --% . m - -
-vg-gey w[ 7 ' ,ssa:ami W - - ,T7~
l
.; ul.1 Tw? y,.
# - Mi.1 ;,, , , _,y . ,; ,. . . .
'O ;-
t
- s. ,-
A
.g>N% " n .
This is an informal report intended for use as a preliminary or working document Prepared for the
\ U.S. NUCLEAR REGULATORY COMMISSION Under Contract No..DE-AC07-761001570 FIN No. A6047 g g g g ,d,h, 8212020110 821031 PDR RES 82120201JO PDR
i
)4 EGcG ,o.u.. . .
OORM 4 usa (GSG 398 INTERIM REPORT I Accession No. _ Report No. EGG-NIAM090 Contract Program or Project
Title:
NRC Technical Assistance Program Division Subject of this Document: AN ANALYSIS OF A SPECTRUM 0F LARGE-BREAK LOCAs IN A BWR/3 USING TRAC-BD1 Volume 1: Overview and Summary Type of Document: Technical Report Author (s): C. B. Davis
' ' .]
Date of Document: October 1982 7~ 4
\ !
N_/ Responsible NRC Individual and NRC Office or Division: F. Odar, RES This document was prepared primarily for preliminary or internal use. it has not received full review and approval. Since there may be substantive changes, this document should not be considered final. l t EG&G Idaho, Inc. l . . Idaho Falls, Idaho 83415 Prepared for the U.S. Nuclear Regulatory Commission Washington, D.C. l Under DOE Contract No. DE-AC07 761D01570 NRC FIN No. A6047
) INTERIM REPORT
O ABSTRACT A spectrum of hypothetical large-break loss-of-coolant accidents in the Dresden Unit 3 Boiling Water Reactor was analyzed with the TRAC-BD1 Version 12 thermal-hydraulic computer code. This report describes TRAC-BD1 models of Dresden Unit 3 and the results of the loss-of-coolant analyses. V J lh O FIN Ho. A6047 Code Assessment and Applications 11
j J
SUMMARY
./
Analyses of hypothetical l loss-of-coolant accidents in the Dresden
! Unit 3 Boiling Water Reactor (BWR) have been performed using the TRAC-BD1 Version 12 computer code. Dresden Unit 3 is a BWR/3 and is owned and , operated by the Commonwealth Edison Company of Chicago, Illinois. TRAC-BD1 Version 12-is an advanced best-estimate computer code for the thermal-hydraulic analysis of a Loss-of-Coolant Accident (LOCA) ir, a BWR.
TRAC-BD1 calculations were performed at the request of the Nuclear
" Regulatory Commission (NRC) to audit LOCA analyses made by the Exxon Nuclear Company. Exxon's LOCA calculations were required by the NRC prior to granting a license to reload Dresden Unit 3.with Exxon fuel. The results of the TRAC-BD1 LOCA calculations and the conclusions derived therefrom are the primary products of this' task. The actual audit of the Exxon calculations will be performed by the NRC.
Three different TRAC-BD1 models of_Dresden 3 were developed in order
. /O . to perform the audit calculations. The 'three models varied.significantly in the amount of geometric detail represented and the computer time
' required to calculate a given transient. A three-dimensional detailed model, a:two-dimensional detailed model, and a one-dimensional scoping model of Dresden 3 were developed. All three models are documented in this report. The TRAC-BD1 models can be used to represent transients other than the large-break-LOCAs for which they were developed although some modifications may be required. The transient to be analyzed will determine which of the three models is most cost-effective for a given application.
A spectrum of large-break LOCAs was analyzed with TRAC-BD1 in order to audit the Exxon calculations. Each LOCA was assumed to be initiated by a large, double-ended, guillotine break in the suction piping of a recirculation pump. Four. LOCA transients, with total break areas of 200*4,
~160%, 120%, and 80% of the flow area of the recirculation pump suction piping, were analyzed. The two-dimensional detailed model, which represented axial and radial, but not azimuthal, gradients within the vessel was used to perform the LOCA calculations because it was thought to (q.) ) 'be the most cost-effective of the models.
111
For each break size, a calculation was run past the time of peak cladding temperature. The peak cladding temperature for each calculation was 450 K or more below the licensing limit of 1478 K. The LOCA initiated by a 200% break resulted in the highest peak cladding temperature (1021 K). The peak cladding temperature occurred during reflood in the LOCAs. initiated by 200% and 160% breaks. A core rewet prior to the start
- of reflood in the 120% and 80% break transients kept the peak cladding temperatures below 700 K. '
Sensitivity studies which primarily investigated the effect of nodalization on the results, were performed. The refill of the lower plenum was sensitive to the nodalization of the lower plenum and jet pumps. Relatively detailed nodalizations were required in order to get realistic refill rates. The low powered fuel channels were also important to lower plenum refill as they were the major liquid flow path between the upper plenum and the lover plenum. The following major conclusions were derived from the results of the analysis presented herein
- 1. A best-estimate analysis of a spectrum of large-break LOCAs in Dresden 3 yielded peak cladding temperatures far below licensing limits.
- 2. A 200% break in Dresden 3 appears to be the most limiting transient of those LOCAs initiated by a large double-ended offset shear break in the suction piping of a recirculation pump.
- 3. A detailed assessment of TRAC-BD1 is required for the refill and reflood portions of a LOCA. .
O iv
m
, m}
( CONTENTS ABSTRACT-..................... ........................................ 11
SUMMARY
............. ............................................... iii
-e _
' 1.'
. INTRODUCTION ................................................... . 1
- 2. DEFINITION OF' TRANSIENTS ......................................... 3
- 3. rCODE' DESCRIPTION ................................................. 4
- 4 .'
MODEL DESCRIPTIONS ............................................... 6
- 5. RESULTS .......................................................... 13 5.11 Break Spectrum Results'.....................................
1.3 5.2 M'ultidimensional Effects ........................ .......... 18
' 5.3 Sensitivity Studies ........................................ 24
- 6. CON C LU S I ON S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 36 s 7. REFERENCES ....................................................... 39 FIGURES 1 1. TRAC-B01 two-dimensional detailed model of Dresden 3 for a large-break LOCA ................................................. 7 a
' 2. -TRAC-BD1 two-dimensional detailed vessel model of Dresden 3 ...... 8
- 3. TRAC-BD1 one-dimensional scoping vessel model of Dresden 3 for a large-break LOCA ........................................... 9
- 4. TRAC-BD1 one-dimensional scoping vessel model of-Dresden 3 .. . 10
- 5. The'effect of break size on upper plenum pressure history ..... . . 14
, 6. .The effect of break size on cladding temperature history at the maximum power zone ......................................... . 14 n
- 7. Radial variation in cladding temperature--200% break ............. 20
- 8. Radial variation in normalized channel outlet mass flow--
200% break ............................................ .......... 20 y i I-i e s- r.,..w .,w~.ce-,+ ..v-.,*.-,-. eva,,.-i . - - - , . - . - , . , , - , , . - - . , - , . . _ , , - - , . -- .,.-.. ,,- ,,.- . , - . -%-, .- - - , ,nm- pg~. --r-
- 9. Radial variation in liquid velocity at the top of the bypass--200% break ....... . .......... ....... ...... .. ..... 21
- 10. Radial variation in bypass liquid level--200% break . ..... . . . 21
- 11. Radial variation in upper plenum void fraction--200% break .. .. 23
- 12. Radial variation in lower plenum liquid level--200% break . .. ... 23 e
- 13. Lower plenum and jet pump nodalizatons . .. .... ..... .. . ... 26
- 14. The effect of nodalization on lower plenum fluid mass--
200%a'reak .. . .... ..... .... .. ...... ..... . .. 27
- 15. The effect of nodalization on flow through the intact loop jet pump suction--200% break ... . . . . ..... ... .... . .... 27
- 16. The effect of LPCS temperature on lower plenum refill--
200% break .... . .... ... ... ... . .... .... .. ..... .. 29
- 17. The effect of jet pump nodalization on lower plenum mass during blowdown--200% break ..... .. .. ......... .. .. .... ... 29 b 18. The effect of model nodalization on upper plenum pressure history--200% break ...... . ....... ......... .............. . 31
- 19. The effect of model nodalization on vessel-side breal flow--
200% break ................... . .. . ......... .... . ... . 31
- 20. The effect of model nodalization on lower plenum fluid mass--
200% break ..... ...... ........ .......... .... ..... ..... . .. 32
- 21. The effect of model nodalization on average powered fuel rod temperature--200% break ...... .. .. ............ .... ...... . . 32
- 22. The effect of the fine mesh model on high powered fuel rod temperature--200% break ......... . ..... ... .. . .. ... 34 TABLES
- 1. Initial conditions ... . .. . . . ... . ....... . .. 12
- 2. Sequence of events . .. . .. .. .... ... ... .. .. 15
- 3. Peak cladding temperature and time to peak cladding temperature vs. break size . ..... ..... . . .... . . .. . .. . .
17 vi
f v [ p AN- ANALYSIS OF A SPECTRUM OF LARGE-BREAK LOCA's IN A'BWR/3 USING TRAC-BD1
- 1. INTRODUCTION y Analyses of hypothetical _ loss-of-coolant. accidents in.the Dresden Unit 3 Boiling Water Reactor (BWR) have been performed using the TRAC-BD1
.. computer code. Dresden Unit 3 is a BWR/3 and is owned and operated by the Commonwealth Edison Company of Chicago, Illinois. TRAC-BD1 1 is an a'dvanced; be'st-estimate computer code for the thermal-hydraulic analysis of a Loss-of-Coolant Accident (LOCA) in a BWR. TRAC-BD1 calculations were performed to audit LOCA analyses made by the'-Exxon Nuclear _ Company.
Exxon's LOCA calculations were required 'by the Nuclear Regulatory Commission (NRC) prior to granting a license to reload Dresden Unit 3 with-Exxon fuel. The results of the TRAC-BD1 LOCA calculations and the
. conclusions derived therefrom-are the primary products'of this task. The actual audit of the Exxon calculations will be per_ formed by the NRC.
(j' A spectrum of large-break LOCAs was analyzed with TRAC-BD1 in order to audit the Exxon calculations. Each LOCA was assumed to be initiated by a large double-ended, guillotine break in the piping on the suction side of a recirculation pump. Four LOCA transients, with total break areas of 200%, 160%, 120% and 80% of the flow area of a recirculation pump suction pipe, were analyzed.
-Three TRAC-BD1 models of Dresden Unit 3 were developed to perform the audit calculations. The models vary significantly in the geometric detail represented and the computer time required to calculate a given transient.
The most detailed model represented the vessel as a three-dimensional thermal-hydraulic component. This model is subsequently referred to as the three-dimensional detailed model. A two-dimensional. detailed model, which does not represent azimuthal gradients within the vessel, was also developed. A one-dimensional TRAC-BD1 model, hereafter called the scoping model, was also developed. The scoping model was designed to be a relatively fast running model which could calculate the general trends of a ('"j 1 transient but would not provide the detail or accuracy of the 4 multidimensional models. 1 L _ _ - - -
The two-dimensional detailed and one-dimensional scoping models were used to analyze the LOCAs. Comparison of the results from the two models provided information on the importance of multidimensional effects during a large-break LOCA. The two-dimensional detailed model was thought to be more cost effective than the three-dimensional model for performing large break LOCA calculations because azimuthal effects probably do not dominate e a LOCA. Conseqeently, the three-dimensional detailed model was not used in the audit calculath.ns. However, the three-dimensional model has been ~ developed and could be used to analyze transients in which azimuthal effects are significant. The remainder of this volume provides an overview and summary of the LOCA analyses. A definition of the LOCA transients is presented in Section 2. A description of TRAC-BD1 is presented in Section 3. A brief description of the two-dimensional detailed and the one-dimensional scoping models is presented in Section 4. An overview of the results of the analyses of the large-break LOCA calculations is presented in Section 5. Results of sensitivity studies are also presented in Section 5. Conclusions derived from the analyses of the LOCA transients are presented in Section 6. References are presented in Section 7. Appendices have been assembled which contain details of the TRAC-BD1 models and calculations which may not be of interest to the casual reader. These appendices constitute Volume 2 of this report. Appendix A contains a detailed description of the TRAC-BD1 models of Dresden 3. Results of the four LOCA calculations are documented in Appendices B through E. e O 2 j
- g. -
t
,, E 2. DEFINITION OF TRANSIENTS A _ /1 s
- A: spectrum of large-break LOCA transients, with total: break areas
; corresponding to 200%, =160%,- 120%, or 80% of the flow area of the recirculation pump. suction piping, were analyzed during this' study.
3 . Differences between transients-were due to the assumption of a different total-break area. Each LOCA was assumed to be initiated by a large,
' double-ended, guillotine break. The break was assumed to occur at the lowest point:in .ths pump suction piping. ?The total-break ~ area was divided equally between the vessel-side'and' pump-sido breaks. The spectrum of ;f break sizes corresponded'to a spectrum of. effective discharge coefficients. The 200% break transient corresponded to an effective discharge coefficient of 1.0 based on the area of the recirculation pump suction piping. The 160%, 120%, and 80% break transients corresponded to effective discharge coefficients of 0.8, 0.6, and 0.4, respectively.
A. loss-of-offsite power was1 assumed to occur simultaneously with the
<[ 'y r.pening of the break. The loss-of-offsite power caused a trip of the #
\d r! circulation pumps and caused the_ isolation of the main steam, main feedwater, and control rod drive hydraulic systems.
The safety system _ single failure criteria was applied to the admission valve vi the Low Pressure Coolant Injection (LPCI) line in the unbroken recirculati_on loop. The LPCI admission valve was assumed to fail closed _w hich prevented any LPCI from reaching the-reactor vessel. This single
~
_ failure thus-effectively failed the entire LPCI system. The core was assumed to be filled with; fresh (beginning-of-life) Exxon
~
fuel. . The core was assumed to be operating at 102% rated thermal power at the start of the transient. The maximum average planar itnear heat generation rate was assumed to be at the proposed Technical Specification limit.2 The assumed initial and boundary conditions for the trans'ients were generally " conservative" rather than "best-estimate". Conservatisms
'i ncluded a 20% increase in decay heat and. degraded injection rates of emergency core coolant. A more thorough description of the assumed d, ) boundary conditions is presented in Appendix A.
-v 3
1
- l l
l l
- 3. CODE DESCRIPTION TRAC-BD1 is an advanced, best-estimate computer code designed for the thermal-hydraulic analysis of a LDCA in a BWR. TRAC-BD1 is being developed at the Idaho National Engineering Laboratory (INEL) and is a version of the transient reactor analysis code (TRAC) being developed at Los Alamos e National Laboratory. TRAC-BD1 provides a unified treatment of a LOCA in a BWR including the blowdown, refill, and reflood phases. The code contains '
a full two-fluid thermal-hydraulic model of two phase flow. The two-fluid model allows both phase slip and thermal nonequilibrium within a V hydrodynamic cell. The code provides a three-dimensional thermal-hydraulic treatment of the reactor pressure vessel; other BWR components are treated one-dimensionally. TRAC-BD1 has models to represent nonhomogeneous critical flow and countercurrent flow limiting (CCFL) at the upper tie plate and side entry orifice in a BWR. The code can calculate reactor power with a point 4inetics model which includes the effects of reactivity feedback and scram. The code has models to represent the control system of a BWR. TRAC-BD1 has a detailed heat-transfer model which includes radiation effects within the fuel bundle. Several different types of hydrodynamic components, including PIPE, VALVE, CHAN, VESSEL, PUMP, JETP, TEE, BREAK, and FILL components, are used in TRAC-BDI. The BREAK and FILL components are used to impose thermal-hydraulic boundary conditions. The other components are used to represent different types of hardware such as a pipe, valve, fuel channel, recirculation pump, jet pump, or reactor pressure vessel. ' The reactor vessel can be represented as a three-dimensional . thermal-hydraulic component. A cylindrical coordinate system is employed in which the user defines the radial, angular, and axial coordinates of mesh cell boundaries. The user-input radial coordinates define rings in the vessel which TRAC-BD1 uses to calculate radial gradients in thermal-hydraulic parameters. The angular coordinates define azimuthal segments which TRAC-BD1 uses to calculate thermal-hydraulic gradients in 4
N the azimuthal (theta) direction. The axial coordinates define levels which TRAC-BD1 uses to calculate axial gradients in thermal-hydraulic parameters. The LOCA calculations were performed with TRAC-BD1 Version B002 which is the current working copy of Version 12. TRAC-BD1 Version 8002 has been , stored under Configuration Control Number F00874 at the INEL Computer Science Laboratory. e o O 5
~~ -
i
- 4. MODEL DESCRIPTIONS A two-dimensional detailed model of Dresden 3 was used to perform the spectrum of large-break LOCA calculations. A sensitivity calculation was performed with a one-dimensional scoping model. Each model represented the reactor vessel, the recirculation loops, and portions of the feedwater, 4 steam, control rod hydraulic, and safety systems. The modeled safety systems included high pressure coolant injecton (HPCI), low pressure core '
spray (LPCS), and automatic depressurization system (ADS). Figure I shows the identification number and the relative location of each componert in the two-dimensional detailed model. The two recirculation loops in Dresden 3 were each modeled as a separate flow path. The recirculation loops were modeled identically except for differences caused by representing a large, double-ended break in the broken loop. The vessel-side break was represented by Junction 14. The pump-side break was represented by Junction 15. The nodalization of the two-dimensional detailed vessel model is illustrated by Figure 2. The reactor vessel was represented with one azimuthal segment, four radial rings, and thirteen axial levels. The core was modeled with three radial rings to allow explicit representation of high powered, average powered, and low powered fuel channels. The two-dimensional detailed model contained 155 hydrodynamic cells. The VESSEL component was modeled with 52 cells. The one-dimensional components, including BREAKS and FILLS, were noded with 103 cells. The model contained 206 heat structures including 20 double slabs and 38 lumped narameter slabs in the VESSEL component. The remaining 148 heat structures were associated with one-dimensional components.
- Figure 3 shows the number and relative location of each component in the one-dimensional scoping model. The nodalization of the vessel model is illustrated by Figure 4. The reactor vessel was represented with one azimuthal segment, two radial rings, and eight axial levels. The core was modeled with one radial ring which allowed the model to represent only an average powered fuel assembly. The scoping model contained 78 hydrodynamic cells. The VESSEL component was modeled with 16 cells. The one-dimensional components were noded with 62 cells. The model contained 6
_ _ . . _ = _ . _ ._ _ . . .
*' M
^y N-i ( ' '
l Break (Main steam) I 75 b81-116 X PE as Fill - Itil (AUS) F"~%~ ] EC N. (I 4# fill (IN*ct) l~I']SC2 Itil (Hale feed) r7r pen. ii i3, lgy , 4 41 4) I I 3 /'80 of i 4 i Broken Recirculation 6 41 IJ 4 15 Loop I 15 ') /I II Intact Recirculation >- , -et Loop , 4 jf} ,{ j 8 33 s) :I 19 I I l y g vpo pa -- l 1 1 g 65 63 68 f f6 sit *5
- i
,, ,_I ,JuncLton I g ,, ,,,
I cougesent 1 ' 15 51 _ _ _ 1'
,a R ut.uun mma rill (cRD) Contairmnent
{' -- 14 3 1 L t.GEtiD-ADS - Autonatic depressurization systein NI 19 I 120 M 37 2 4 CRD - Control rod drive .
, lil'CI - liigli pressure coolant injection l Li'CS - Low pressure core spriy 4
Figure 1. TRAC-BD1 two-dimensional detailed model of Dresden 3 for 3 a large-b, eak LOCA. . t
- . ~ -- -- , - . - . - . . . . -
CE?iTERLINE 3ECTICN A-A PRESSURE VESSEL WALL RI!G 3 f :CWNCCMER N RI G2 \ \ 3 6
\
AIG 1 -> \ g
-\ s -
l8> l >
/
j LCW-70WERE3 CMANNEL
' AVERAGE PCERE3 CHANNEL SYPASS > -
,' HIGH.7CWERED CHANNEL LEGEND:
CCRE ACS - Automatic decressuri:stien system 3HRCUh '!PC: - 4tgn ;ressure coolant injecticn LPCS - Low pressure care scray
;3.320 I
S[ LEVEL 13 { : "AIN STE.4 I l 17.564 --------m---------J l i
, STEAM LEVEL 12 , CRYERS '
I 2C5 13.500
' S UA N CR3 l
s?C: LEVEL 11 'T
-~~ ~~~}
LEVEL 10 h io.m -. _ - UIPEf?Tb - - _ - ,_ _~ _~ ~_ __-.
- tAIN FEE 3 LEVEL 93 9.503 i LPCs LEVEL 8 c:
a A la
- i. - I i_l ' l .
y 5 i G ' vA I'333
$: [~hSd ~N:
~ ~
LEVEL 7 6.500 _ g .gg --.E ' -
~h q -
~E TU LEVEL b 5.267 f UU - ...
JET :UPP LEVEL 5 4.431 -- -----8----- , -- e : 70 :,E::RC'.UT:CN et:'9 LEVEL 4 i---- ', ' ' 3m - -- r--- LEVEL 3 f. o 91 1 - - - - - - - LEVEL 2 1.295 - r - - r - - ' - - - - LEVEL 1 0.0 O.3 0.7881 Z.0123 2.6C51 3.1377 e uCIus (m) e Figure 2. TRAC-BD1 two-dimensional detailed vessel model of Dresden 3. 8
. + *
- O b \
i N , Y. i 1 t Break (Main steani) I ys kE l&X @ h49 fill fill (ADS) i sn M as si d i fill (llPCI) l 38 I 4WY l2= 4# ' fill (Main feed) i ___ u d 4 it i-.al,, 43 l t' lfor 6]l y . , 18
.l' !
J
$ g Broken liecirculation loop 1- '3 13
-- _.-.J
, e intact Recirculation 8' 33 19 .
toop l l
,3 q rhites:
t,) } .1 '"a uen l g, i coai.oac=t i II
,3 Containment 8 " 3
- H a" 8"*5
u 15 ) i
, \1 ._
W D1 17 a m) fill (CRD) (2' Lf.GEND: i ADS - Auttnatic depressurizationi systcui LRD - Control rod drive llPCI - liigli pressure t.colant injection
, LPCS - Low pressure core spray >
figure 3. TRAC-BD1 one- dimensional scoping model of Dresden 3 for a large-break LOCA. i
CENTERLINE SE;i!ON A 1. RI?;G 2 s PRESSURE /ESSEL WALL
'N N
'N, 00WNCOMER
\ s/
\ , e RING 1 -
'43 BYPASS
- \
AVERAGE-PC'.;ERED CMANNEL LEGEND: i CCRE ADS - Automatic decressurization sy.te SHROUD HPCI . Hign pressure coolant injection LPCS - Low ::ressure core spray 20.920 , 9 STEAM
- LEVEL 8 ----> MAIN STEAM DOME l 17.564 STEAM DRYER LEVEL 7 p ADS 13.036 - - - - - - - - - - - - - - -
LEVEL 6 _ SEPARTORS 4 HPCI Ji!, 11.036 . _- _ - _ _ . _ _ _ . . LEVEL 5 g UPPER PLENUM j 4- 14AIN FEED 3 9.509 - - - -_-. . - _ . 4 LPCS
- LEVEL 4 $ A 7.913 A( _ _ _ , , ,, ,
j , LEVEL 3 5.267
$N
< o. u
[4 -JET PUMP l LEVEL 2 l 3.023 . . . . . . . . . . . . . > TO RECIRCULATION PUMP LEVEL 1 . 0.0 . 0.0 2.6051 3.1877
-> RADIUS (m)
Figure 4. TRAC-BD1 one-dimensional scoping vessel model of Dresden 3. 10
y ( )- 76 heat structures including 12 double slabs and 11 lumped paramet'er slabs in.the VESSEL component. The remaining 53 heat structures-were associated with one-dimensional components. The TRAC-B01.models of Dresden 3 are described in detail in g 'l Appendix A. The only differences between the models used for the LOCA
-l transients'and the models described in Appendix A were related to representing a large break in the recirculation loop piping. The models described in Appendix' A represent Dresden 3 for_ normal operation; ruptures =
in the piping were not. represented. The models described in Appendix A were modified for.the LOCA calculations by renoding the pump suction piping inithe broken recirculation loop and inserting two BREAK components to ! represent containment pressure. In addition, changes were made to.the models to decrease!the computer time required to perform the LOCA calculations. The_ broken loop jet pump was renoded to allow the code to execute at larger time steps. Junction 11, which connects the broken recirculation loop and the vessel, was attached to an axial rather than
-[D radial face of a vessel cell to allow the code to execute at larger time steps. .The fluid. volume, hydraulic resistance, and elevation of ~the broken i recirculation-loop.were preserved.
~
The initial conditions for the four LOCA trans'ients are shown in Table 1. .The initial conditions were obtained in the following manner. The models described in Appendix A were run in a steady state mode until sufficiently steady results were obtained. The TRAC-BD1 control system was used to control reactor pressure, downcomer liquid level, and total jet pump discharge flow rate. The pressure, temperature, void, and. velocity a-distributions were extracted from the steady state runs and incorporated in
~the models used to calculate the LOCAs. The results of the steady state runs were similar for both models as shown in Table 1. Differences in initial conditions.of the two models were generally less than 1%. The
~
' largest difference in' initial conditions of the two models was in the
.le'akage.from the fuel channels to the bypass. The radial resolution provided in-.the detailed model allowed a more accurate epresentation of the radial power profile and the side entry orifice distribution which slightly affected the leakage from the fuel channels.
-(%. J) ll
TABLE 1. INITIAL CONDITIONS Two-Dimensional One-Dimensional Parameter Detailed Model Scoping Model Reactor power * [Mh'(t)] 2577.54 2577.54 e Steam dome pressure (MPa) 7.033 7.033 Feedwater flow rate (kg/s) 1251. 1247. Feedwater temperature (K) 441.0 441.0 Total recirculation loop flow (kg/s) 4560. 4522. i Recirculation pump speed (rad /s) 170.7 169.1 Total jet pump discharge flow (kg/s) 12348, 12348. Bypass flowb (%) 10.67 10.25 Channel leakageb (%) 5.01 4.59
. Guide tube leakageb (%) 5.66 5.66 a.; =
Control rod drive flow (kg/s) 2.268 2.268 Control rod drive temperature (K) 322.0 322.0 Lower plenum temperature (K) 548.9 548.9 Downcomer liquid levelc (m) 10.479 10.485
- a. Corresponds to 102% rated thermal power,
- b. Expressed as a percentage of total jet pump discharge flow.
- c. Corresponds to the level above the bottom of the baffle plate. The '
baffle plate separates the downcomer and lower plenum and is 3.023 m above the bottom of the lower plenum. . O 12
r_ t-
,n
/ i
- V 5. RESULTS- ;
. An overview of' the results of the break' spectrum analysis is presented in-Section.5.1. Calculated multidtmensional effects in the vessel are described in Section 5.2. _ Sensitivity calculations are described in d :Section 5.3. 'A comparison between LOCA calculations with the two-dimensional: detailed model and the'one-dimensional scoping model is
&lso presented in Section'5.3.
5.1 Break Spectrum Results Four LOCA calculations, with total break areas of 200%, 160%, 120%,
.and 80% of-the flow area of the recirculation pump suction piping, were performed with the'f.wo-dimensional detailed model shown in Figures 1
- and.2. Upper plenum pressure versus' time is shown in Figure 5 for each of the four.LOCA calculations. As expected, a decrease in break size resulted p in a slower depletion of' coolant,.a slower blow'down rate, and a higher e
(' pressure. Corresponding events generally occurred later in time as the break area decreased,~as illustrated by Table 2, due'to the slower blowdown rate. The calculated peak cladding temperature and the time at which the _ peak cladding temperature occurred are shown versus break size in Table 3. The largest peak cladding temperature..was 1021 K and'was calculated to occur.in the.200% break transient. Hence, the 200% break was the most limiting transient. The peak cladding temperature for the 160% break was only 15.K lower than for the 200% break. Thus, the calculated behavior of th'e 160% break was very close to that observed in the 200% break. In both-calculations, the turnaround in cladding temperature was caused by reflooding the high powered channel. The peak cladding temperatures for the 120% and 80% breaks were at least 300 K lower than for the other two calculations. The lower peak cladding temperatures were caused by calculated rewets which occurred prior to the start of reflood. The peak
= cladding temperature for each transient was far below the licensing limit of 1478 K~ set by Appendix K.3 The cladding temperatures remained low ~'Q enough that the metal water reaction model was never actuated.
Consequently, oxidation of the cladding was not calculated. 13
1 i 8000 - ' e I 200% BREAK O '60% BREJ.K -1000 g a 120% BREAK y
$ 6000 _ O 807. BREAK , g
-800 O #
8 8 3 m x '
~
3 - o 4000 ~ _-600 $ L 0 4 I Q.
-400 - 2000 -
e E 3 0 -
-200 $
0 -
' ' - - " "- n .,
0 0 0 50 10 0 15 0 200 Figure 5. Time (s) The effect of break size on upper plenum pressure history. 1200 , , i 200% BREAK 0 160% BREAK -1500 1000 - A 120% BREAK O 80% BREAK 2 v C 0 c 800 --
-1000 e U t ~ 3
{ -500 400 -
/ m M -
Dryout
-0 200 ' '
O 50 10 0 15 0 200 Figure 6. Time (s) l The e ff ect of break size on cladding temperature i history of the maximum power zone. ' 14
,s _
y 5 m ji
, TABLE 2_. SEQUENCE OF EVENTS
?: ~ vs
~,
' ' Time (s) _.
200%; 160% 1' 20% - 80% Event- -
' Break Break Break- Break lOJ -
Break opened, loss o* offsite power. . 0.0- :0.0 0.0 0.0~
'- * ~
t
'. Scram signal ~ generated;'feedwater terminated - 0.5 0.5 0.5 0.5
. Low'waterilevel' signal"-
i: 1.3-' 1.5 1.5' f 2.0 i Control rod drive flow terminated 3.0 3.0. . 3.0 3.0 ! Low-low water.-level signal * - 3.2 3.2 3.4 4.3
- Control rods fully' inserted 4.5 4.5 4.5 4.5 MSIVs closed 5.0 5.0 5.0 5.0 .
- q. - Jet: pump' suction uncovered b 6.7 7.1 7.6 9.1
)
d Recirculation'line uncoveredD 11.0 11.9 13.6 17.1
, Lower plenum flashiiig- 12 13 15 20 s HPCI. initiated. 23.2 23.2 23.4 24.3 i
. Intact loop isolation valve closed.
32.5- 32.5 32.5 32.5 [ Dryout'at the peak power zone c 37 40 59 78
' LPCS' initiated. 44.8- 48.1- 54.6 68.8 d
;- ; . Rated LPCS~ delivered 54.4 59.6 70.1 94.3 HPCI terminated 72.0 77.5 90.8 128.5 Jet, pumps renoded' 72.5 77.5 90.8 .135.0
; Lower plenum refill started. 75 76- 92 98 I
LBackflow from containment to vessel -110 130 --* --* Reflood: initiated 160 160 --* --*
\
l
, 15 1
I I TABLE 2. (continued) Time (s) 200% 160% 120% 80% Event Break Break Break Break , Peak cladding temperature obtained 175 174 120 140 Calculation terminated' 190 196 132 160
- a. The ' low and low-low water level signals correspond to collapsed downcomer liquid levels of 12.80 and 11.28 m, respectively, above the bottom of the lower plenum.
- b. The times at which the jet pump suction and recirculation line uncovered correspond to the times the collapsed downcomer liquid levels dropped to the elevation of the jet pump suction and the top of the reactor vessel outlet nozzle, respectively.
- c. Dryout corresponds to the time the void fraction in the center of the
' heat;ed length in the high powered channel approached unity.
- d. Rated LPCS corresponds to a flow of 4500 gpm per pump which is the flow required by the Dresden Unit 3 Technic'al Specifications.
~
- e. The event did not occur in the calculation.
l 16
. . m h/x TABLEI3.' ' PEAK CLADDING TEMPERATijRE AND < TIME TO PEAK CLADDING TEMPERATURE
'NS. BREAK-: SIZE
- . Total L Break - Peak C1 adding I- Time 'to Peak
, Area * > Temperatureb Cladding Temperature 4 .(%) -(K) (s) *
'C
=200- 1021 175
.160 1006 174
~*- . . . ~120 637- ~120
- , 80- ~ 699 140
.a ,
.a. The total break' area was the combined area of' the vessel-side b'reak-and D the pump-side. break normalized ~ to the cross-sectional flow area of - the j- -
recirculation pump suction piping.
- b. The peak cladding temperature occurred at- the center of the heatea length, which corresponds'to the maximum power zone, in the 200%'and_160% break-transients. Tne peak cladding temperdture for the other transients occurred
- j. '
in thefcell'above:th6 maximum power zone. e'
- r f
sl - .
~
y a ., a 4 fp F I' ') L a J .e-A 5 1
\
p p 17 1 gyy yyw yvw---
~ ,
i
,\ <
Cladding temperature at. the maximum power zone in the core versus time is shown in Figure 6 for each of the four LOCA calculations. The dryout - time, which corresponds to the time of significant rod heatup, increasea as - the break size decreased due to a slower depletion of the liquid inventory $ associated with a slower blowdown. lfhe effect of break size on dryout time was the principal cause of the variatiUn in cladding temperature between
, j the calculations. In the 120% and B0% _ break transients, dryout occurred later and leakage of LPCS liquid into the bottom of the channels was '
h sufficient to rewet the fuel rods before their surface tewperature exceeded 3 the Leidenfrost temperature. A lensittvity ca7culation in which a rewet did not occur in the 120% break transient when a small change was made to the model is described in Appendix I/. The calculations indicate that the s 1 potential for a rewet exists prior to the start of reflood for the 120% arh 80% break transients. However, the actual occurrence of a rewet cannot be } ' assured even though one was caYculated because of the difficulty involved _ a for a computer code to accurately predict rewet phenomena. Consequently, relatively large uncertainty exists in the calculated peak cladding
~
temperatures for the 120% and 80% break transients because the peak ' cladding temperatures would have been significantly higher if rewets had s not been calculated. However, even if '.he rewets had not been calculated, the peak cladding temperatures would hade been less than the 1021 K ' i reported for the 200% bruk transient. i Furthermore, the fact ,that a higher j peak cladding temperature Ns chiculatid for the 80% break thin for the h 120% break is probably not significant'considening the uncerta'nty in the i I rewet calculation. y
- , r 4
5.2 '. Multidimensional Effec'ts > ' c - Since TRAC-BD1 has a sique capability for multidimension3 1 i thermal-hydraulic analysis within the vessel, a diicussion nf the , calculated multidimensionai effects is presented to t'ocument some of the ~ i effects which would be absent in a one-dimensiona} analyy,s. The TRAC-BD1 , 3 3 calculations were performed with a two-dimensional model which represeated j radial and axial effects. i Calculated axial effects are 'destdibed in 3 Appendices B through E. A\ discussion of the most significant radial effects is presented below. 3 i 2 3 18 E
*} .
]; }[ {j[
~ - - - -
1 4
+
jw
\" , .
):
~
' t, LThe thermal' respense of. the core contained the most significant radial feffects cbserved in the LOCA calculations. Cladding' temperatures at the
. maximum power zone in the high powered, average powered, and low powered fuel' channels are shown in Figure 7 for the 200% break transient. The calculated temperatures strongly show-the influence of the radial power
; ,- profile-with the highest temperatures occurring in the high powered channels and the lowest temperatures in the low powered channels. A
;similar radial variation'in cladding temperature occurred in the 160% break transient. Less pronounced radial variations occurred in the 120% and 80% break calculations because the lower peak cladding-temperatures for
>1 these transients' allowed less room for variation.
, iA pronounced' radial profile in the flow of LPCS from the upper plenum to the . lower plenum was calculated during the. period of lower plenum refill. Figure ~8 shows the integrated mass flow rate at the upper tie d . plate in the high powered, average powered, and low powered channels, u_ normalized on a per channel basis, during refill in the 200% break transient. In'the high powered and average powered channels, the gM integrated mass flow was positive and increasing which was. indicative ofa m net mass flow up through the upper tie plate. The CCFL model allowed a D- i small amount of liquid flow down through the upper tie plate but the downward liquid flow was less than the upward steam flow. CCFL breakdown I._ at the upper tie plate in the low powered channels began at 75 s which
'k lallowed liquid to drain from the upper plenum down through the low powered
' channels into the lower plenum. The primary flow path for LPCS to the l
lower plenum was through th'e low powered channels; 60% of the LPCS reaching s the -lower plenum between 75 s and 150 s flowed from the upper plenum down f{ i through the low powered channels. Multiple channel effects are expected to 4 be'important for an accurate' representation of the refill process. 3 A pronounced radial profile in the flow from the bypass to the upper plenum was also calculated. The calculated liquid velocity from the bypass to the upper plenum for the first and third vessel rings in the 200% break transient'is shown in Figure 9. The third vessel ring is the outermost c
, ring inside the core shroud and contains the low powered channels. The . ,3 !
first. vessel ring-is the innermost ring and contains the high powered N._ / - a- .n k~ _
.f
- f 19 1
$gf
1300 HIGH-POWERED CHANNEL 0 O AVERAGE-POWERED CHANNEL 110 0 _- A ~ m ~-1500
- 900 -
~
g 2 - o U $ .
-1000 -
700 - E E H f g ' E o 500 -
- -500
" 0 0 m a m 3 300 ' i ,
50 10 0 15 0 200 Figure 7. Time (s) Radial variation in cladding t empe r a t u r e--2007. b r eak. m 50_ , i -
- 10 0 O
on ^ 5 E 3 0 -- cc o c c c-o
~
f 0 G _u- '
--0 3
m y Start of CCFL " E -50
- breakdown
- 10 0 g
i g s V 6 -
@ -10 0
- -200 0 y HIGH-POWERED CHANNEL 5 ,
O AVERAGE-POWERED CHANilEL 2 A C LOW-POWERED CHANNEL
-15 0 ' _ _3co O 50 10 0 15 0 200 Figure 8. Time (s)
Radial variation in normalized channel ou tlet mass fl ow--2007. b r eak. 20
f^y
- 10 , , ,
7 j {, l( - 20
^
N 1 , i N
- OC O*O h ,
k , !--0 .b hk x t i
! i 3
=- .pd
; ( ,, =
i
- I
. -20
- O 4 ;
o -10 -~ - y b -
- -40 5 RING 1 0 ' RING 3
- -60
-20 ' ' ' O O 50 10 0 15 0 200 Time (s)
Figure 9. Radici vorlation in liquid velocity of the top of the bypass--2007. break. G 5 . , ,
. RING 1 15
.(
- o -
v E
--10 v 3--
b $ 8 ! 8
; 2 -
oh m - E ' E 4 1-I p u , d[I i 0 O O 50 10 0 15 0 200 Time (s) Figure 10. Radial vorlation in bypass liquid level--2007. break. U 21
channels. The liquid velocity in the second vessel ring, which contains the average powered channels, is not shown for reasons of clarity but was generally between the two velbcities which are shown. The liquid velocity in the outer ring was generally less than the velocity in the inner ring. CCFL breakdown at the upper core grid, which separates the bypass and upper plenum, began at 80 s in the third vessel ring. The CCFL breakdown allowed , LPCS to drain from the upper plenum into the bypass. The third vessel ring was the preferred LPCS flow path between the upper plenum and the bypass. . Only a relatively small fraction of the LPCS flowed down through the inner two rings. The radial velocity profile was primarily caused by the distribution of axial flow area in the bypass. The inner two rings had a relatively small area due to blockage by the fuel assemblies and control rods. The third vessel ring had a relatively large flow area due to a 0.3 m space between the outermost fuel assembly and the core shroud. The preferential flow of liquid from the upper plenum down into the outer ring of the bypass caused a radial variation in liquid distribution within the bypass. Figure 10 shows collapsed liquid levels in the first and third vessel rings in the bypass for the 200*.' break transient. The liquid level in the second vessel ring is not shown for reasons of clarity but was generally close to the level in the first ring. The collapsed liquid levels were computed from the void fractions and heights of the cells in the corresponding bypass rings. The liquid level was generally higher in the outer ring of the bypass due to the preferential flow of liquid into the outer ring from the upper plenum. The gradient in levels then caused the liquid to flow radially towards the center of the vessel. The radial distribution of void fraction in the upper plenum for the , 200*4 break calculation is shown in Figure 11. The void fractions for the three vessel rings were similar until shortly after LPCS injection began. - The CCFL model then limited the amount of LPCS draining into the bypass and fuel channels which resulted in a decrease in upper plenum void fractions. A pronounced radial gradient in void fraction was calculated after LPCS began. LPCS was injected into the third vessel ring which caused the void fraction to be lowest in that ring. The void fraction was much higher in the inner two rings with a relatively small gradient between those two 22 i
[ b'
'wJ '
, "dr f
/ ^
' 0.8 j J f , i n
3 aj
) -- 0.8 I
%' M
-* 8 c;
- I
- 8 j 0.6 n- f -- 0.6
, O
]O h I b
L egan 3 0.4 -- ._,g4 2 o jlt
> o RING 1 {
. 0.2 -- O RING 2 g -- 0.2 A RING 3 '
)
ah l Start of CCFL breakdown 0 ' ' ' O.0 0 50 10 0 15 0 200 Fi gur e 11. Time (s) Radial variation in upper. plenum void fraction---200% break.
\ j).
[N i i RING 1 5 O RING 3
~
-15
^4 -
O E - O . i 1 O es E 3-- _ -10 y'
% 2 L- 52 -
~
.5 Q
, 5 1 - _
N. Start of refill O O O 50 10 0 15 0 200 Figure 17 Time (s) Radial vorlation in lower plenum IIquid level--200%
- g. br eak.
>. a>
23 l 9 , ,, , - -. ~ - - . . - - , . . ,e
rings. The radial void gradient between the second and third rings became less pronounced after CCFL breakdown began at 75 s. The void gradient was relatively small after 110 s when most of the liquid had drained from the upper plenum. TRAC-BD1 Version 12 dces not have a spray distribution model v and cannot account for the size and orientation of the spray nozzles. A spray distribution model would probably affect the void gradient shown in ' Figure 11. Data comparisons will have to be performed before the accuracy of the calculated void gradient in the upper plenum is known. A small radial variation in lower plenom liquid inventory was calculated. Figure 12 shows collapsed liquid levels in the first and third vessel rings in the lower plenum for the 200% break transient. The liquid level in the second vessel ring is not shown for reasons of clarity but was generally close to the level in the first ring. An insignificant radial gradient in liquid level was calculated until 75 s when lower plenum refill began. The preferential flow of liquid down through the low powered channels during refill caused the level in the third ring to be slightly higher than in the first ring. Comparisons between the results of the two-dimensional and one-dimensional models are presented in Section 5.3 to document the influence of radial effects on the overall system response. 5.3 Sensitivity Studies Studies were performed to investigate the sensitivity of the calculated results to nodalization, LPCS temperature, and heat-transfer options. The results of these sensitivity studies are described below.
- Sensitivity studies were performed which showed that the calculated refill behavior was significantly affected by the nodalization of the jet pumps and lower plenum. The effect of jet pump and lower plenum nodalization on the calculated results for the 200% break transient was determined with base and sensitivity calculations. The base calculation was performed with the two-dimensional detailed model described in Section 4; results of the base calculation were presented in Section 5.1 24
MT 1 I
,-s
( '# [ 'and in Appendix B. ~ Figure 13 shows the jet pump and lower plenum
~ nodalizations used-in the. base and sensitivity calculations. The lower plenum below the jet pump discharge was modeled with three axial levels in ithe base calculation and one;1evel in the sensitivity calculation. The jet pumps were modeled with the TRAC-801 JETP component in the sensitivity calculation. The JET'P component uses three total cells to represent the mixing throat, diffusor, and discharge regions of a jet pump. In the base
=
fealculation, the jet pumps were modeled with JETP components until 72.5 s; TEE components, which modeled the mixing throat, diffusor and discharge regions with nine total cells, represented the jet pumps after 72.5 s. 1 The total fluid mass in the-lower plenum versus time for the base and sensitivity calculations is shown in Figure 14. The figure will be discussed in three time periods: blowdown, refill, and reflood. Relatively small differences were observed between the base and sensitivity calculations during blowdown (0.0 to 75 s). In fact, the difference in
-minimum lower plenum inventory between the two calculations was only 3's.
. During refill (75 to 160 s),- the lower plenum filled at a significantly V faster rate in the base calculation. Based.on an extrapolation of the refill rates, the lower plenum would be completely liquid full at 195 s in the base case but not until 300 s in the sensitivity calculation. The difference in refill rates was due to a difference in the calculated
. behavior ~of the jet pumps. The integ ated mass flow rate through the
, intact loop jet pump suction to the downcomer during refill is shown in Figure 15. Although not shown, the1 integrated mass flow rate from the broken loop jet pump-is similar. In the base calculation, only a small amountiof mass flowed from the lower plenum through the jet pumps to the 1
downcomer during refill. A relatively fast refill rate was obtained because virtually none of the LPCS reaching the lower plenum was lost through the jet pumps to the downcomer. In the sensitivity calculation, over-half of the LPCS reaching the lower plenum flowed through the jet pumps to the downcomer which resulted in a relatively slow fill rate. The relatively coarse jet pump nodalization used in the sensitivity calculation tended to promote liquid flow through the jet pumps due to numerical A. diffusion. The more detailed jet pump nodalization used in the base
/ \'
( ) calculation resulted in a better representation of the mixture level, and a q,.? 25
VESSEL VESSEL t ' CENTERLINE l _ CEffTERLINE _ i i ! C DOWNCOMER l i l l l l '
, , (--MIXING THROAT l l ' I I
i j _ l l l I I 1 _ I _
~
I CORE PLATE ! ._ l _ l ! _ - DIFFUSER l_ _ _1__ , _ _l _ _ _ _l _ I --- i l l l l 3 m I l i
~~~ i I I l ~-'
l <H ISCHARGE _ _ _ _ _ _ l_ _ oBAFFLE PLATE I I l N I
--L__ __,____
I l I I l l l l I l
, 4- LOWER PLATE l l l l
I I ! l l
-b 1- 79- i l l l l l l l l 1 (fiote: Jet pump nodalization l g / , l l shown applicable only af ter I L //
s) GASE CALCULATIOil SENTIVITY CALCULATION Figure 13. Lower plenum and jet pump nodalizations.
, , , j
.. . _ _- m . _ .
.A s
\
i 2'd 60000 , , . i i BASE CALCULATION
.O SENSITIVITY CALCULATION - 1.2 0 '10' 50000 -
Start Of reflood
~
- ' 00*'O 4 40000 -
m
..^
V y .
- 8.0 0 '10' E
.c
.. ( C 30000 -
- 6.0 0'10' O E 2 O
'20000 - -
2
- 4.0 0'10 10000,-
~- 2.0 0*10' atart.of refill O ' ' ' ' '
O.00 0 50 10 0 15 0 200 250 300 Time (s) Fi gu r e 14. The eff ect of nodalization on lower plenum fluid q mass--200% break. I ) 40000 , , , , , BASE CALCULATION
^
. O) x O SENSITIVITY CALCULATION
-80000 $a
- 30000 -' -
y w -60000 0 0 $ E 20000 -
~ o
-40000 E
'O
# V
+ W O
' O u
7 10000
~
20000 f C
- C 0 O. O O 50 10 0 15 0 200 250 300 Time (s)
Fi gu r e 15. The eff ect of nodalization on fl ow through the intact loop le t pump suction--200% break. m i . 27
,,mme- , e,--,,--.wa,- -
_ , - . -- . , , , , - , - , , , . - - , - , . , , , , - . . - - . . , - _ , . , _ , _ - , , , , . . . , ._--_,n.,. .-----w-.,-,,.,_..,,--,--w-
~ more pronounced separation of the phases within the jet pumps, which prevented liquid flow through the jet pumps. The more accurate representation of the area change within the diffusor was probably responsible for most of the effect of jet pump nodalization on refill rate. The more detailed lower plenum nodalization used in the base calculation allowed a mare accurate representation of the pressure and void
- fraction at the exit of the jet pumps which promoted the formation of a mixture level within the jet pumps and reduced the liquid flow through the jet pumps. About half of the effect of nodalization on refill rate shown in Figure 14 was due to lower plenum nodalization; the other half of the effect was due to jet pump nodalization. The refill rate in the base calculation was thought to be more realistic than the rate in the sensitivity calculation. The results shown indicate that the current JETP component in TRAC-BD1, which uses a fixed, coarse nodalization, is not adequate for some transients. Code development personnel will incorporate a new JETP component, which will allow a flexible, user-defined nodalization, in a future version of TRAC-BD1.
Figure 14 shows that the refill of the lower plenum was halted at O 160 s in the base calculation when reflood started and lower plenum liquid started flowing into the core. The start of core reflood occurred before lower plenum refill was completed. The halting of the lower plenum refill and the premature start of reflood was not thought to be realistic. The behavior perceived as unrealistic was probably caused by LPCS-induced condensation in the upper plenum. A sensitivity calculation was restarted from the base calculation in which warm (375 K) LPCS was used rather than the cold (308 K) LPCS used in the base calculation. The effect of LPCS temperature on the calculated fluid mass in the lower plenum is shown in Figure 16. In the sensitivity calculation, there was an initial redistribution of mass between the core and lower plenum but then the refill rate returned to a rate consistent with that of the base calculation prior to 160 s. The sensitivity study demonstrated that the calculated refill /reflood hydraulics are sensitive to LPCS-induced condensation. The calculated hydraulics of the TRAC-BD1 code should be assessed against appropriate refill and reflood data. Only a few such assessments have been performed to date. Consequo tly, the uncertainties in the code's calculation of refill and reflood phenomena have not yet been quantified. 28
w V
' y"-~5
{ d_, i _ v/ =60000 ,
, i
, WARM LPCS O COLD-LPCS - 1.2 0*10' 50000- -
C
- 1.0 0*10'
.40000 -
.p. ~^
f - l' E v, - 8.0 0*10' . o ._
.; - , ,30000
.$i -
- 6.0 0*10' $
2- 0 20000 - 2
- 4.0 0*10' 10000 -
-- 2.0 0*10'
.0 O.00 10- 50 10 0 150 200
- Time .(s)
Figu r e 16. The effect of LPCS. temperature on lower plenum r e fill--200% b r ock.~
..n\
I ..
\ l" k/. 60000 , , _,
DETAILED NODALIZATION O COARSE NODAL!ZATION 1.2 0*10*
~ 50000
-c
- 1.0 0*10' 40000 -
^
.v j -
- 8.0 0*10' o E
,. 30000~ -
- V y -
- 8.0 0*10' $
'2 O
- 20000 -
4,00 1o* 10000_ -- C O
- 2.0 0*10' 0 ' ' '
O.00 0 20 - 40 60 80 Time (s) F igu r e 17. The eff ect of Jet pump nodalization on lower plenum mass during blowdown--200% breck. O b , 29
>~ -,, , - - - , - .---.--,i--,J .r--, .._----_--,-,~~,,---,,n.,---,. .. ,-~ . - . - +, - . - - - -
The results presentea in Figure 14 showed that the calculated lower plenum mass during refill was significantly affected by the jet pump nodalization. A sensitivity calculation was performed to determine if the jet pump nodalization significantly affected the results during blowdown. The ansitivity calculation was restarted from the base calculation of the 200% break at 15 s with the detailed jet pump nodalization shown in
- Figure 14. The lower plenum mass in the base and sensitivity calculations is shown in Figure 17. The lower plenum mass was moderately sensitive to '
the jet pump nodalization during blowdown. A 20% increase in minimum lower plenum inventory was calculated with the detailed jet pump nodalization. A 20% increase in minimum lower plenum inventory would correspond to a 5 s earlier refill of the lower plenum. The results shown in Figure 17 further illustrate the need for a new JETP component which allows a flexible, user-defined nodalization. An analysis was performed to determine the overall effect of nodalization on the calculated results of a large-break LOCA. Calculations of the 200% break transient were performed with the detailed and scoping models described in Section 4. The models generally calculated similar results during blowdown. Figure 18 shows that the two models calculated nearly identical traces of upper plenum pressure. The calculated break flows were also nearly identical as illustrated by Figure 19 which shows mass flow rate out the vessel-side break. However, during refill the two models calculated significantly different S havior. Figure 20 shows lower plenum fluid mass for the two calculations. With the detailed model, lower plenum refill started at 75 s when CCFL breakdown began at the upper tie plate in the low powered channels. CCFL breakdown allowed liquid to drain from the upper plenum through the low powered channels into the lower plenar causing a relatively fast refill. A relatively slow refill was . calcu ated with the scoping model because of coarse nodalization and lack of multiple fuel channels. As discussed previously, a coarse nodalization of the lower plenum and jet pumps can result in a slow refill rate. Hewever, the major reason for the different refill behavior between calculations was that the scoping model represented only a single, average powered fuel channel. Since the low powered channels were not modeled explicitly, the major refill flow path between the lower plenum and 30
,s 7x '
.t )
.w/
s .. 8000 , , , DETAILED MODEL 5 O SCOPING MODEL 1000 ^ n o 0
$ 6000 -
'n 6 v -
-800 O
. u. e
- 3 '
- m. 3 g 4000-_ _-600 $
u e o.- ' g -
-400 2 2000 -
- 3 0
> o
-200 >
0 O O 50 10 0 15 0 200 Ti'me (s) Figure 18. The e f f ec t o f model nodalization on upper plenum O J'
. pressure history--200% break.
15000 , , , DETAILED MODEL - 30000 O SCOPING MODEL' - n ~ 10000 - 7
- 20000 N
-- Q ;
m E d$ e v
$ 5000_- -
10000 3
. C 2 m
m. () m O- M 2 o._ - 10 90O j
.. e
__o 4
-5000' ' ' ' - -10000 0 50 10 0 15 0 200 Time (s)
Figure 19. The effect of model nodalization on vessel-side break
~
[ flow--200% break.
\v 31
60000 , DETAILED MODEL
- O SCOPING MODEL - 1.2 0 '10' 50000 -
C
- - 1.0 0'10' 40000 -
3 E 9a - - 8.0 0*10 ,o v - g 30000 -
- - 6.0 0'10' E n 0 2 2 20000 -
\'
- % / - 4.0 0 *10' 10000 _- .
""Q m-
- 2.0 0*10' 0 O.00 0 50 10 0 150 200 Time (s)
Figure 20. The eff ecf of model nodalization on lower plenum fluid mass--200% b reak. 600 i , , O DETAILED MODEL { q O SCOPING MODEL 550 - g I -500 C v L o
- 500 l -
- s j
3 l 3
+- - -400 O
g
' L
[450 - { E E -
-300 0 d -
400 -
- -200 350 O SO 10 0 15 0 200 Time (s)
Figure 21 The eff ect of model nodaliza tion on avoroge-powered f uel rod t empera tur e--2007. break. 32
f _ ( j upper plenum was not present in.the scoping model. A slow refill rate was
'~
a consequence of not explicitly modeling the low powered channels. The
' thermal response of the core was also significantly different between the
.two. calculations. Calculated cladding temperatures at the center of the heated length in.the average powered channel are shown in Figure 21. The t scoping model calculated a lower maximum cladding temperature, an earlier temperature turnaround, and an earlier quench. Although the maximum V
. cladding temperatures were less than 80 K different, a different type of behavior was calculated with the two models. With the detailed model, leakage of liquid from the bypass into the average powered channel was responsible for a bottom-up quench. The CCFL model allowed only a small amount of liquid down through the: upper tie plate. With the scoping modal, a top-down quench was calculated due to a h_igher, but still CCFL limited,
- ! ' liquid flow'down through ~ the upper tie plate. The detailed model would be
. preferred over the scoping model for additional large-break LOCA calculations because of'its more realistic refill behavior and the extra n
-radialdetailprovidedin~thecordthermalrespons'e. The scoping model ran
-about'2.5 times faster than the detailed model.
(by) A final sensitivity calculation was performed to investigate the effect of _reflood heat-transfer. options on the thermal response of the core
-in the 200i break. In the base calculation, the fine mesh model was turned on near the start of reflood. The fine mesh model was designed to represent enhanced heat transfer during reflood due to axial heat conduction in the cladding and precursory cooling near the quench front.
. In the sensitivity calculation, the fine mesh model was not used; reflood heat transfer was calculated with the regular TRAC-BD1 heat-transfer package. Figure 22 shows the effect of the fine mesh model on the cladding temperature at the center of the heated length in the high powered
-channel. The cladding temperature turned around and began decreasing about 10 s after the fine mesh model was turned on in the base calculation. The cladding temperature continued to increase in the sensitivity calculation i in which the fine mesh model was not used. Based on the discussion presented in Appendix B, the heat transfer with the fine mesh model was thought to be more realistic considering the calculated hydraulic
() conditions. A code update, which improved the quench front propagation 33
I 110 0 - I ' WITH FINE MESH -1500 O WITHouT FINE MEsg m 1050 -
^
M v _ l'- 4
-1400 v o
u o 3 ' 3 ~ 5 1000
- ~
u o 8 a ( -1300 1 E
- E s o 9s0 ~
- H
~1200 900 15 0 16 0 170 18 0 19 0 200 Figure 22.
IIme (S) The eff ect of the fine mesh model on high-powered fuel rod t emperature--2007. br eak. O 4 e O 34
_~ . . - . - - -_.-._. i s criterion, was required to1obtain reasonable results with the fine mesh t
- model. The TRAC-BD1 code developers plan to incorporate this update in a ;
! . future version of.the code. -However,'it is~ generally recognized that
~
j TRAC-BD1.has4a major weakness inithe calculation of reflood heat transfer.
- The' code developers plan to correct this. weakness by eventually replacing. l g- the fine-mesh-model~with.a more sophisticated moving mesh model.
- v. ,
_h.. ' 4 l a t 1 4 k i 1 6 i e- ! ? 6 1- i Y 1 W' s
)
4 e f i 1 4 t 1 6 l 35 i r
n-
- 6. CONCLUSIONS Conclusions derived as part of this study relative to the large-break LOCAs are presented below.
- 1. A best-estimate analysis of a spectrum of large-break LOCAs in 1L Dresden Unit 3 yielded peak cladding temperatures far below licensing limits.
The peak cladding temperature for each large-break LOCA was more than 450 K below the licensing limit of 1478 K as shown by Table 3.
- 2. A 200% break in Dresden Unit 3 appears to be the most limiting transient of those LOCAs initiated by a large double-ended offset shear break in the suction piping of a recirculation pump.
The analysis of a spectrum of large break LOCAs showed that the maximum peak cladding temperature was obtained for the 200% break. The 160% break was very similar to the 200% break but resulted in a 15 K lower peak cladding temperature.
- 3. The potential for a core rewet prior to the start of reflood exists in transients initiated by " smaller" large-break LOCAs.
A core rewet, which resulted in peak cladding temperatures below 700 K, was calculated for the 120% and 80% breaks. However, the actual occurrence of the predicted rewets cannot be assured
- because of the sensitivity of the calculations to relatively ,
minor changes in the model. The peak cladding temperatures for the 120% and 80% breaks would have been less than that obtained for the 200% break even without the core rewet. O 36 l
~ ~ -
( ) 4. The refill process is significantly affected by multiple channel effects.
'The primary liquid flow path between the upper plenum and lower plenum during refill was calculated to be through the low powered D' channels. The low powered channels should be modeled explicitly as a separate. flow' path in order to accurately represent lower
-= plenum refill.
' Conclusions relative to TRAC-BD1 as a tool for performing the audit
. calculations of large-break LOCAs in Dresden 3 are presented below.
- 1. The TRAC-BD1 calculated results were generally thought to be reasonable except as detailed below.
- 2. A detailed assessment of TRAC-BD1 is required for the refill and reflood portions of a large-break 'LOCA.
/m
( )
' 'k / Analysis indicated that' the weakest areas of the audit calculations were in refill and reflood. The calculated reflood hydraulics and heat transfer need particular attention. The reflood heat-transfer calculation was more realistic'with the fine mesh model than without it. However, additional code
' development and assessment are required in order to improve
-modeling techniques and quantify code uncertainty.
- 3. The calculated refill of the lower plenum is sensitive to the e
nodalization of the lower plenum and jet pumps. The nodalization of the jet pumps and lower plenum can affect the time of lower plenum refill by as much as 100 s as described in Section 5.3. More than one level in the lower plenum below the jet pump discharge and more than three cells in the jet pump below the suction were required to obtain a realistic refill 7] calculation. A nonstandard jet pump model was required to obtain Q) a' realistic refill calculation. 37
- 4. A detailed model should be used rather than a one-dimensional scoping model for any future large-break LOCA calculations.
A one-dimensional scoping model does not contain the radial detail required to represent the thermal response of the high powered channels. In addition, the lack of an explicit i representation of the low powered channels caused the calculated lower plenum refill rate to be too low as described in . Section 5.3. O O I O1 38
_ . w. - 4 J *O
- 7. REFERENCES 11.- JJ. W.-Spore et-a'l., TRAC-BD1: An Advanced Best Estimate Computer
. Program for Boiling Water Reactor. Loss-of-Coolant Accident Analysis,
- Volume 1: Model Description, October 1981, NUREG/CR-2178, EGG-2109.
l2. LJ. E; Krajicek, .Dresden Unit 3 LOCA Analysis Using the ENC EXEM , Evaluation Model MAPLHGR Results, November 1981, Exxon Nuclear ' Oy Company, Inc. , XN-NF-81-75(P).
, 3 .- NCFR-Part-50,." Acceptance Criteria for Emergency Core Cooling System
.for Light-Water-Cooled Nuclear. Power Plants," Federal Register,
; Volume ~39, No. 3, (January 4, 1974).
w
"~2 _
+
, t 3 b t d d 39
EGG-NTAP-6090 October 1982 AN ANALYSIS OF A SPECTRUM 0F LARGE-BREAK LOCAs IN A BWR/3 USING TRAC-BD1 Volume 2: : Appendices l l l' C. B. Davis ! 6' 1 l
-Idaho National Engineering Laboratory Operated by the U.S. Department of Energy
( %Qq ~- . _ _ ,,, fx
. p .;:, ; ;6 . (M M mumur, q _ ..
md, &g g .. . -, -
-# 4,g
% 6.- 7.... m =>====
WTwMh&5~L^3Z^p~L g*. '
'& % Q cure W b5 E 25'm: n -++A
_ _,--we e u-
- R*r r-w ,, s
' -- W'.ml
.j
. M-$ -
,anszuulW" T]l . ,%. , -
;;m===* e][mamarma==={d(.'/
h,y ,, ~ .
"Q& '%
N; -
, % , e J :
~ wA 4 .
This is an informal report intended for use as a preliminary or working document Prepared for the U.S. NUCLEAR REGULATORY COMMISSION Under DOE ' Contract No. DE-AC07-761001570 Q b 6 E O Idaho FIN No. A6047 Q er
EGzG. . . ,-
. FORM EG4G 398 (Rev 03 82)
INTERIM REPORT Accession No. Report No. Contract Program or Project
Title:
NRC Technical Assistance Program Division e . Subject of this Document: AN ANALYSIS OF A SPECTRUM 0F LARGE-BREAK LOCAs IN A BWR/3 USING TRAC-BD1 Volume 2: Appendices Type of Document: ' Technical Report Author (s): C. B. Davis Date of Document: October 1982 Responsible NRC Individual and NRC Office or Division: F..Odar, RES This document was prepared primarily for preliminary or internat use. it has not received fu!I review and approval. Since there may be substantive changes, this document should not be considered final. l EG&G Idaho, Inc. Idaho Falls, Idaho 83415 v Prepared for the U.S. Nuclear Regulatory Commission Washington, D.C.
. Under DOE Contract No. DE-AC07 761D01570 NRC FIN No. A6047 INTERIM REPORT
A A
- s ,
w Y
- n
, b .,
e CONTENTS
#n. .": ,'. . .
l APPENDIX'A--MODEL' DESCRIPTIONS ........................................ A-1
. APPENDIX B--200% LARGE-BREAK LOCA ..................................... B-1 4
- . ' APPENDIX-C--160% LARGE-BREAK'LOCA .....................................
.- C-1
-- w:
.APPENDIXLD--120% LARGE-BREAK LOCA ..................................... D-1 e, -APPENDIX E- 80% LARGE-BREAK.LOCA ..................................... E-1
\
./-
w ' e &V
,w a
i 9
- D_
t 4 4 11 i
(. '[. b-i b e . b APPENDIX A l MODEL DESCRIPTIONS h r i o a ? 8 e ? 3- 0 5 O f. A-1 I \. . - _ . - - - - . _ . _ . _ _ - - --_ ,. - - ,. - - -.- -. - , _ ----..-
g - ,. , iw l C b' g e . - , I ' m,.,s - C P - '
' APPENDIX 2A' n "
MODEL DESCRIPTIONS
~
k .
. CONTENTS m.. -
- I;- O' i A c
m.^
' >1; :; I N T RODUCT I ON . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A m s ., ,.. .. ,, 9
- 2. tTHREE-DIMENSIONALfDETAILED M0 DEL' DESCRIPTION'................... A-5 m
s [ _g. . <
;2.1[HydrodynamicNodalization;................................. A-5 -
'2.1.1 : Recirculation Loops ............................. ;A-6
'2.1.2 ~
Reactor Vessel .................................. A-11
- 4 2.1.3 Fuel Channels ................................... A-19
-2.1.4 Guide Tubes ...........'.......................... A-21 2.1;5 Balance:of P1 ant................................. A-22 ? 12.2 Heat Transfer Nodalization ................................ A-24 2.'2.11 ; Recirculation Loops ............................. A-24 12.2.2: Reactor Vessel .................................. A-25 2.2 3 . Fuel Channels:...................................- A-26
- 2.2.4' Guide Tubes ..................................... A-28
<2.2;E Balance of P1 ant'................................ A-28'
[ w 2 . 3 D C ode : 0 p t i o n s ' . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-28 p_ L2.3.1' Main Program-Control ............................ 'A-29
, . 2.3.9.. Component ....................................... A-29 2.3.3. . Power .=.......................................... A-30
~
p =2.4 Boundary Conditions'....................................... A-30 2.4.1 ADS ............................................. A-30 2.4;2- HPCI ............................................ A-32 *
. 2.' 4. 3 LPCS ............................................ A-34 e , 2;4.4. LPCI ............................................ A-36 s 2;4.5 Steam, Feed, CRD,- and Recirculation Pump Systems .................................... A-36 -
'a-o LM y 2.5. Initial..' Conditions ........................................ A-36 v 7 s
?3. 0THER MODELS ................................................... A-39 3.1' Two-Dimensional Detailed Model ............................ A-39 c-F -3.2 One-Dimensional Scoping Model ............................. A-39 y
f'
- 4. -MODEL LIMITATIONS .............................................. A-45
[ i y A-2
s FIGURES-
-A-1. LTRAC-BD1three-dimensionaldetailedmodelofDresden3........ A-7 A-2. . Detailed nodalization of the intact recirculation ~ loop ........ A-8 A-3. TRAC-BD1 three-dimensional vessel model ....................... A-12 s i ~A-4. ' Radial and azimuthal nodalization of the
; three-dimensional' detailed' vessel model ....................... A-13 A-5. 'Axialinodalization of the three-dimensional detailed vessel model ............................... ......... A-17
- ~A-6. .-Axial nodalization of the CHAN' component ...................... A-20 A-7. TRAC-BD1.two-dimensional detailed model of Dresden 3 .......... A-40
[A-8. TRAC-BD1 two-dimensional vessel-model ......................... A-41 EA-9. Radial and azimuthal nodalization of the two-dimensional detailed vessel'model ......................... A-42 A-10. TRAC-BD1 one-dimensional scoping model of Dresden 3 . . . . . . . . . . . -A-43
- A-11. TRAC-BD1 one-dimensional vessel.model ......................... A-44
- A-12. Scop'ing nodalization of the intact ecirculation loop ......... A-46 A-13. Axial nodalization of the scoping. vessel model ................ A-47 A-14. Radial nodalization of the scoping ~ vessel model ............... A-48 TABLES A-1. Correspondence between the physical and TRAC-B01 components in the intact recirculation loop ................... A-9 L o" -A-2. Cell volumes and flow areas in'the vessel component ........... A-14 A-3. Lumped parameter heat slabs ................................... A-27 A-4. ADS flow ...................................................... A-31 i
1A-5. HPCI. flow ..................................................... A-33 A-6. LPCS flow ...................................... .............. A-35 A-7. Initial conditions ............................................ A-38 l A-3 w .,
<- Y
.,. APPENDIX A '
' " MODEL DESCRIPTIONS 3
1..INTRODUCT.j0N 1 Mathematical models of tpe Dresden Unit 3 Boiling Water Reactor (BWR)' q* , have been developed. Dresden Unit 3 is a BWR/3..and is owned and operated:
- by-the Commonwealth Ed'sen-Company of Chicago, Illinois. The mathematical
~
'models of Dresden 3 are input. decks for the TRAC-BD1 computer code.A-I
- TRAC-BD1 is an advanced best estimate computer code for the thermal '
k' ,
?
hydraulic analysis of a less-of-coolantf accident (LOCA) in a BWR. . , s The TRAC-BD1 models of Dresden 3 were.specifically developed toi perform audit calculations'of LOCA analyses made by Exxon Nuclear Company. The Exxon LOCA calculations'were required by the Nuclear Regulatory l' <
~
Commission prior.to granting la license to load Exxon fuel into Dresden ' Unit'3. The. TRAC-BD1' mode,1s~~r~e'generalandcanbeusedtoanaiyze$hermal a '
^
- hydraulic transients other than thos'e' required for .the audit calculations- N
~
.although'some modifications may be required. 3 4
g
'4 .
.The TRAC-BD1 models were developed primarily from'information obtaihed '
i from the Exxon. Nuclear Company which Exxon used to develop a RELAX model of Dresden 3. RELAXisthethermal-h)drauliccomputercodewhichExgonused -( t to calculate the blowdown portion' cf the LOCA. Someofthedata$ ^
~
- incorporated.in the' TRAC-BD1~ model'sfrepresent information which 1.E y proprietary to Exxon Nuclear Company and/or Commonwealth Edison Co.ipany.
Proprietary information is not prestnted in this appendix. ; c . 6 Three TRAC-BD1 models of Dresden Unit 3 were developed and will be described in this appendix. The models vary significantly in the geometric '
. detail. represented and the computer time required to calculate a given transient. The most detailed model represented the vessel as a
, - - three-dimensional. thermal-hydraulic component. This model is subsequently
. referred to as the three-dimensional detailed model. cA,two-dimensional detailed model which does not represent azimuthal gradients within the t
t. _ A-4 - _ - _ _ _ _ _ _ _ = _ . _ .
,my
- q. . .
i Q; Q1 ws { vessel was also developed. The two-dimensional model was thought to be more cost-effective than the three-dimensional model for performing the f' l ;, audit calculations because azimuthal effects are probably not dominant during a LOCA. A one-dimensional TRAC-BD1 model, hereafter called the
>~
- scoping model, was also developed. The scoping model was designed to be a W" relatively fast running model which could calculate the general trends of a transient but would not provide the detail or accuracy of the v
'~
x multidimensional models, w k The purpose of this appendix is to document the TRAC-BD1 models of
;. Dresden 3 so that potential users will understand the modeling philosophy, 4 assumptions, and limitations inherent within the models. The remainder of
- this appendix documents the TRAC-BD1 models. The three-dimensional T detailed model is described in Section 2. Descriptions of the nodalization, selection of code options, boundary conditions, and initial conditions are presented. The two-dimensional detailed model and the scoping model are described in Section 3. Limitations of the models are listed in Section 4.
- 2. THREE-DIMENSIONAL DETAILED MODEL DESCRIPTION The three-dimensional detailed TRAC-BD1 model of Dresden 3 is described in the following section. Descriptions of the nodalalizations used in the hydrodynamic and heat transfer calculations are presented in i Sections 2.1 and 2.2, respectively. The user-selected code input options employed in the model are described in Section 2.3. The boundary and initial conditions for the model are described in Sections 2.4 and 2.5,
. .respectively. .The model wu developed such that it could be easily modified to perform a large break LOCA calculation.
o 2.1 Hydrodynamic Nodalization Several different types of hydrodynamic components, including PIPE, VALVE, CHAN, VESSEL, PUMP, TEE, JETP, BREAK, and FILL components, are used in TRAC-BD1. The BREAK and FILL components are used to impose thermal-hydraulic boundary conditions. The other components can be used to A-5
t b represent different types of hardware such as a pipe, valve, fuel channel,
- reactor pressure' vessel, jet pump, or pump. The user can node all the-components, except BREAK and FILL components, with as many hydrodynamic cells as desired. - All the different types of TRAC-BD1 components were used in the three-dimensional detailed model of Dresden 3. -The reactor vessel, both recirculation loops, or portions of.the feed, steam, and safety systems were represented in the TRAC-BD1 model. Figure A-1 shows the
- identification number and the relative _ location of each component within sthe TRAC model. The model illustrated by Figure A-1 represents Dresden 3
- (for~ normal operation only. ' No broken or failed components are modeled.
'2.1.1- Recirculation Loops The Dresden 3 BWR contains two recirculation loops; each loop was represented as a separate flow path in the TRAC-BD1 model. The first loop,
= consisting of components 1 through 5, will hereafter be called the intact
-loop. The second loop, consisting of components 11 through 15, is the loop in which. recirculation line breaks'will eventually be modeled and hence is
. called the broken loop.~ Although the recirculation loops in Dresden 3 are
, lightly different from each other, the recirculation loops in the TRAC-BD1 model.are identical; each loop in the model represents an average of the two. loops in-the plant.
A schematic of a recirculation loop is shown in Figure A-2. Divisions between TRAC-801 components and between cells within components are noted in the figure. The correspondence between the TRAC-BD1 components and the physical components in the intact recirculation loop is -illustrated in Table A-1. The table also presents the volume, length, and elevation of . each cell. Tt.a.modeling philosophy used in the development of the TRAC-B01 model of the recirculation. loops is described below.
- Only the major flow path in each recirculation loop was modeled; none of the loop penetrations were represented. Dresden 3 has isolation valves near-the suction and discharge of each recirculation pump. The model ;
. frepresente'd th'se e valves as pieces of straight pipe of the same schedule as A-6
a .. . .
~
Urcak (Main steam) i is pas.(ys-)c _ p_91. y 4, rijis y ,, as..[y,1) al.l 77 l Break (Hain steam) Fill (AUS) <m l E *C--]i? IM , f .'CI33 1. 59 ;- 3 9 i Fill (ADS) l 38 alm T 47 i4a ,38 , 3, g 3o i 39 Fill (ilPCI) ) !
- i Fi i (siPCI)
Fill,(Main feed) i it i- 82 .,, j; g 8_l, ,01_p 73 ),8$- M 4 i Fill (Main feed) 45 43 41 42 44 46 1-
- f. t 1 3 1 1 1 6 }I 7 '" 3 0 f17
.$ Y j 5 4! 4] 4l 42 4 4 tij 15 1 /L '3 11 72 '4 /6 11
.o .
L _' _ __ l 3" Intact Recirculation " ~ ~ N loop a 35 33 II 32 3i 16 19 Loop 5? i3 5; ,4 51 g St , 4 i i l , l 65[63}61}{62jH{66
. g g os 3 r21 rn ,4 ra Notes:
I i i 1 Junction I l4 {6 1 Component 1
.c3-).3 g 2 Fills.(CRD) 12 gg x yahes Legend- J Oaccircaiatioa""es ADS - Automatic depressurization system :
I CRD - Control rod drive ' ilPCI - liigh pressure coolant injection LPCS - Low pressure core spray Figure A-1. TRAC-BD1 three-dimensional detailed model of Dresden 3.
DRIVELIllE Cr . N0ZZLE 1# JET PUMP REACTOR OUTLET. > N0ZIM .. 4 d TYPICAL OF 10) REACTOR INLET ?. THROAT N021LE ~
~j..
, r >
' .i .. . r DIFFUSER r '4-VESSEL WALL
~~' 5 .
- EZ72 - __ M -- DISCHARGE NORMAL FLOW N AFFLE B PLATE DIRECTION 4 JET PUMP RISER (TYPICAL 0F_5)
L -) y -- g MANIFOLD pump SUCTI0il 1 RECIRCULATION PUMP 4 PIPING. m PUMP DISCHARGE r ... 4-- PIPING
---*S \ 1: / >
~
Zi ! 3 %: g LEGEtID L'
- 1 COMP 0tlEllT 1
--- CELL BOUNDARY ISOLATI0tl VALVES- " "
Figure A-2. Detailed nodalization of the intact recirculation loop. A-8
o' . j ,
. TABLE A-1. CORRESPONDENCE BETWEEN THE PHYSICAL AND TRAC-BD1 COMPONENTS IN THE INTACT RECIRCULATION. LOOP YOI"** Elevation
- Physical TRAC Cell 3 Component Component Number (m ) Lenfth (m (m) Comments Pump suction 1 1 1.097 3.035 4.102 Cell 1 includes the reactor. vessel piping 2 0.9193 2.731 1.670 outlet nozzle. Cell 4 contains the 3 1.205 3.581 -1.4 86 isolation valve.
4 1.205 3.581 -5.067 5 1.234 3.666 -7.922 Rec irculation 2 1 0.7079 1.27 -6.730 pump 2 0.7079 1.27 -6.730 > Pump discharge 3 1 0.4465 1.359 -6.603 Component 3 contains the pump dis-pumping 2 0.8220 2.501 -6.429 charge isolation valve. Low 4 1 0.7427 2.260 -4.406 pressure coolant injection
?-
2 1.177 3.581 - 1.4 86 piping discharges into Cell 2. u) 3 2.500 3.4 50 0.560 Cell 3 represents the recirculation 4 1.0 98 3.327 2.478 nanifold. Cells 4, 5, and 6
- 5 0.7535 2.550 4.599 represent the jet pump risers.
6 0.4504 1.771 5.738 7 0.4504 1.771 7 .50 9 Jet pumps 5 1 0.7516 2.261 8.522 Primary Cells 1, 2, and 3 represent (primary) 2 2.339 2.1 64 8.123 the throat, nozzle and diffusion 3 1.008 0.4850 6.801 res pectively. Secondary Cell 1 (secon- 1 0.05415 0.3833 4.589 represents the nozzle. Secondary dary) 2 0.1249 0.4987 3. 2 64 Cell 2 represents the driveline between the riser and the nozzle.
- a. The elevation presented is that of the center of the cell. The elevation of the (inside) bottom of the reactor vessel corresponds to 0.0 m elevation.
the' piping connected.to the valves. The isolation valve in the pump discharge piping Q s modeled explicitly as a TRAC-BD1 VALVE component
-because this valve'may-close during a LOCA.
Each recirculation pump was modeled.as a TRAC-BD1 PUMP component. The
' ~
PUMP component represented on1y the fluid volume between the inlet and
-outlet nozzles of:the pump; none of the connecting loop piping was represented ir. the PUMP component. The actual geometry of the flow path
' ~
- within the pump was'not known and therefore~was'not modeled accurately.
1The. pump was modeled so that negligible wall friction would be calculated within the PUMP component because internal friction is accounted for by the pump? homologous curves. The single phase pump homologous curves were based son Exxon's RELAX model of Dresden 3. The. RELAX homologous curves were based on data proprietary to the. Commonwealth Edison Company in those
. quadrants where:such data was available. In those quadrants where proprietary data wa's not available, the single phase homologaus curves were bas'ed on data for a Bingham pump.A-2 'The two phase homologous head curves were based on-Semiscale data.A-2 The' fully degraded two phase
~
pump head curve'was obtained by. subtracting the Semiscale two phase homologous head difference curve from the Dresden 3' single phase homologous head curve. The two phase' head multiplier vs void fraction curve was also
~
.taken from the Semiscale' data.A-2 Pump torque was assumed to not be degraded by two phase flow.
Dresden 3 has'a manifol'd in the pump; discharge piping of the.
~
recirculation loops. The manifold is an. incomplete (<360 ) torus wh h feeds:the det pump risers (see Figure A-2). There are two normally closed isolatibn valves in the manifold which divide it into two hydraulically . separate pieces; each piece is connected to only one of the recirculation
. loops. Each piece of the manifold was represented by a single cell in the
- TRAC-BDI model. -The. length of the manifold cell was the average distance-along.the axis;of the torus.that fluid travels while within the manifold.
Each recirculation loop has five risers which connect the manifold to the jet pumps. These' five parallel flow paths were combined into a single flow path in the TRAC-BD1 model. A TEE component was used to represent the
~
4 A-10
-1 ten jet pumps in each recirculation loop. The TEE component was extracted from the JETP component. The three-dimensional TRAC-BD1 model currently
. represents the act'ual jet pump geometry although the jet pump nozzle is relatively.small.and can. limit time step size.
The head losses due to bends in the recirculation loop piping and due to expansion or contraction effects at the vessel outlet nozzle and the
- manifold were modeled. Default form lo'ss coeffecients were used in the jet m.
pump components.
-2 .' 1. 2 Reactor Vessel
.The reactor' vessel can be represented as a three-dimensional thermal-hydraulic' component in TRAC-801. A cylindrical coordinate system is employed in which the user defines the radial, angular, and axial coordinates of mesh cell boundaries. The user-input radial coordinates
- define. rings in the~ vessel which TRAC-BD1 uses to calculate radial gradients in thermal-hydraulic parameters. The angular coordinates define azimuthal segments'which TRAC-BD1 uses to calculate thermal-hydraulic gradients in:the theta direction. The axial coordinates define levels which TRAC-BD1 uses to calculate axial gradients in thermal-hydraulic
- parameters.-
A three-dimensional thermal-hydraulic model of the Dresden 3 vessel was developed as illustrated by the nodalization diagram shown in Figure A-3. The vessel was represented-with two azimuthal segments, four radial rings, and 13 axial levels. The azimuthal and radial nodalization of the vessel is illustrated by e Figure A-4 which shows a cross section of the vessel. The TRAC-BD1 vessel model_.is axisymmetric. For example,.the input volume for Cell 1 is identical to that for Cell 2 at each level in the vessel as illustrated by Table A-2 which shows the volumes and face flow areas of the vessel cells. Two azimuthal sections were modeled so that the two recirculation
~ '
. loops could be connected to different vessel cells. The TRAC-BD1 model can i- 1
-A-ll
CE:ITERLDE SECTfCN A-A PRESSURE VESSEL 'dALL RI?G 4 e
' RI!G 3 00WNCCMER s
. RING 2 ,
~.-
"%, s t
N \
' ' \ t RIris 1 4 e 1
i
-~"
i I l I T e y1 LOW-POWERED CHANNEL
AVE 4 AGE-PCWERED CHANNEL i,,,
/.
i .:IGH-POWERED CHANNEL BYPASS i D I LEGEND:
' ADS- Automatic depressurization system
' CORE SHROU,0 J HPCI- High pressure coolant injection 9 LPCS- Low pressure core spray i
e 20.920
- i .
l i STEAM l ; :. tait STEAM LEVEL 13 i DOME ___.___a.,,_..,'_.,__. e , 17.564 -' '
, i i e i STEAM i
' CRYERS-LEVEL 12 I 4--- ADS 13.500 l - '- , - - - -l - - - , - - - -
LEVEL 11 3 SEPARATOR l 4 HPCI 11.036 ,___s___, ,,_g,_,l. LEVEL 10a- 10.281 ' l- - n P~eR 'UD 4- MAIN FEED mLfu*k'w - - - l
' 4--- LPCS LEVEL 9 .9.509 _, , _ _ _ _ . .
I LEVEL _7 d 7.933 l. k i t 1l -l1 g3 8 h .[/ '9 c; z . , 6.500 ,- , % g$ g= -d- / -- LEVEL 6~ A 5.267 ' a:;b ED' ' lg-'.]-9 '
-/"1
,, JET PUMP LEVEL ' 5 , i ,
4,43l LEVEL 4 i --->TO RECIRCULATICil PUMP i
'~ '
-LEVEL 3 3.023 8
- r_ -
LEVEL' 2 2.591 '- 2- - ' - - - r - 1.295 l-
- i - - - d' - -
' I LEVEL 1. 0.0 0.0 0.7891 2.0128 2.6051 3.1877 PADIUS (m)
Figure A-3. TRAC-BD1 three-dimensional vessel model. A-12
Y n e e 8 a
~
.d %e z e m E 5 5 * .9
~
se 53
" 6 m
5 m 1
)
g O @ .5 g ~ 3 y = w wA { e a e e - 5 8 -. 5
.E ARBS @W W
' i+0 8BM:Hgg& W2 c .! ,
%=l$2M8EEBij8Edsai$a 8 eE88888SEEss's28 a8 :
8BBBB8889R9aFEE8Sm a s :E
.k ' 8!)E2888BE"7#3BE8 s -f2 s O
s
, .. .~ ,
E 88 38,38,88,0c- bE,-.--, EE,,-5, 5585E 88 - - - - , - ,, 1 s s - E c N 1% OQ W CC i A A LJLLA J CC CO w w Q Q OO D% w , g. 88 %8888EBMER 888EEM BB i
*z
< @REEES8288E8BSSEEE gg a gggrB888888EC"j s SB 31 c-888-88f .
4o ' u M nw 45 .
- 88888 E8 B s
s
' s g -- -
\
5
= '
Nxxxx 5 "s - I e = d== ss aeas e =e m 8 e ~ O " e 8 O C E EdGd a s 555d d 53 ) Sc+ev 1 l A-13
c M
' TABLE- A-2. : CELL VOLUMES. AND FLOW AREAS IN' THE VESSEL COMPONENT. ..
~
Flow Area Volume Radial . Azimuthal Axial Cel l .. Level Nunber (m ) (m )- (m ) (m ) Comments 1 1, 2 0.9613' 2.078 . 0.515.' - 0.3387 Levels 1, 2, and 3 and Cells 1 through 6 of 3, 4 3.800. 1.770 0.5786 1.883 Levels 4 and 5 represent the. lowers plenum. - 5, 6 - -1.070 0.0 0.1184; '2.263 7, 8 '0.0. 0.0 0.0 0.0
- 2 1, 2 0.4387 0.3802 0.09508- 0.3387-3,~4 2.439 0.9892- 0.1521 1.883 5, 6 4.004 10.14 0.4104 3.095-7, 8 3.440 0.0 0.3911 .4.742-
> 3 1, 2 0.1462 0.1267 0.03169 ' 0.3387- .L 3, 4 0.8131 0.3297 ~0.05071 1.883.
- 5, 6 1.335 0.9355 0.1368 2.268 7, 8 2.178 0.0 0.2404 .0.0 4 1, 2 0.4771 0.4131 0.1033 0.3388 3, 4 2.652 1.075 0.1653 1.884 5,'6 3.194 0.0 ~0.3454 2.269 7, 8 5.245 0.0 0.3195 .4.504 5 1, 2 0.2661 0.2304 0.05761 0.0 The top of Level 5 represents the core plate.-
3, 4 1.479 0.5989 0.09218 0.0. 5, 6 1.724 0.0 0.1853 0.0 7, 8 3.517 0.0 0.1352 4.194 6 1, 2 0.2156 .0.2836 0.07091 0.1346 The outer surface of Ring 3 represents the 3, 4 1.210 0.7372 0.1134 0.7565 core shroud. The outer surface of Ring 4 5, 6 3.823 0.0 0.4197 2.552 represents the pressure vessel wall. 7, 8 5.849 0.0 0.3787 4.422 7 1, 2 0.1795 0.2749 0.06872 0.1346 3, 4 1.008 0.7146 0.1099 0.7565 5, 6 3.402 0.0 0.4188 2.552 < 7, 8 5.895 0.0 0.3787 4.422 s 1
~
a :...
. .- u
, :i
,4
. TABLE A-2. (continued)'
Flow Area Cell V lume. Radials Azimuthal: . Axial 3 Level Number ~ (m k _ ( ,2). (,2) .(,2) . Commen ts' 8 -. 1, 2 0.~1933- 0.2515 0.06288 0.09873- , 3, 4 1.083. 0.6539 0.1006 0.5453-5, 6 3.910 .0.0 0.3833 2.470-7, 8 7.182 0.0 0.6231 3.721 9 1, 2 0.7527 2.351 0.5879- 0.9755- Levels 9 ;and 10 . represent the uppcr plenum. 3, 4- 4.158 6.117- 0.9406- 5.388 5, 6 '4.017 0.0 0.5857 5.207 7, 8 2.871 0.0 .0.3057 '3.721; , 10 1, 2 1.001 1.989 0.5239 0.1677 ? 3, 4 4.010 2.709 0.6361 0.1677 G 5, 6 1.321 0.0 0.1337~ 1.016 7, 8 7.949 0.0- 0.5006 0.8576-11 1, 2 0.9998 6.101 0.0 0.4003 Level 11 represents the steam separators- . 3, 4 6.0 54 15.58 0.0 2.424 5, 6 5.110 3.430 0.0' 2.046 7, 8 27.52 0.0 3.381- 0.5047-12 1, 2 3.590 4.174- 1.537 0.2566 3, 4 19.94 10.45 2.389 1.191 5, 6 30.26 0.0 1.155 1.82. 7, S 6.241 0.0 0.2065 4.684-13 1, 2 3.062 7.647 2.486 0.0 Level ~13 represents the steam dome. 3, 4 15.04 15.63 3.474 0.0 5, 6 9.323 15.03 1.293 0.0 7, 8 6.4 94 0.0. 0.7277 0.0 i
V represent azimuthal effects.in~.the vessel due to asymmetrical behavior of
^ the recirculation loopsLsuch as might be caused by a pipe rupture.
TheLvessel was divide'd into four radial rings as shown by Figure A-4.
~
~The outer-r'ing was<used to represent the downcomer. The region inside the core shroud was modeled with three. radial rings to allow rearesentation of thigh powered, average powered, and low powered fuel bundles. The six inner
' L VESSEL. cells represented the bypass which is the region inside the shroud and outside the' fuel bundles. TRAC-BD1 CHAN components, which will be described'later, were used to represent the region inside the channel walls ofLthe' fuel bundles. The radii of.the inner two rings were determined in
~
- the following manner. Two polygons were used to separate the core into thre'e~ radial zones. Polygons were used rather than circles so that each ring would contain an integral number.of fuel bundles. The ring radii were input so that tne, cross sectional area of the ring would be equal to the
~ area of-the. polygon.
.The: TRAC-BDI.model represented the vessel with 13 axial levels as
- shown in . Figure' A-5. Rationale for the placement of levels within the vessel is provided below. TheLportion of the lower plenum below the baffle plate was-noded with three levels to model those transients in which~the
' jet pump. discharge uncovers. The top of Level I was located at the top of the control rod drive housing. The top of Level.3 was placed at the baffle plate. Level 3 was relatively short to allow the model to represent transients'in which a mixture level forms near the jet pump discharge.
Level 4 was placed so that the model could represent uncovery of the
' reactor vessel outlet nozzle. The top of Level 5 was placed at the core plate to represent the separation between the~ fluid in the lower plenum and
~
.the bypass. The top of Level 6 was placed halfway between the core plate and the jet. pump' suction to provide resolution of the calculated bypass and *
-downcomer liquid levels. The top of' Level 7 was placed at the jet pump suction so that the' model could represent jet pump uncovery. The top of n Level 8 was placed at the upper core grid to represent the boundary between Tthe bypass and the upper plenum. The upper plenum was modeled with two
_ levels to provide ' resolution in the calculated void fraction and fluid
. temperature _ distribution _s after the initiation of core spray. The region A-16
A STEAM DONE > LEVEL 13
- ---- V STEAM DRYER PANEL d' A j -
FLOW SKIRT j , LEVEL 12 STEAM LINE
~
[
" NORMAL WATER
' -!" ~ _ " " . "
- 7. - . . - . . '. '-
. . . - . . . - - LEVEL STEAM SEPARAT0" >
w LEVEL 11 FEEDWATER LIN E - m r STAND PIPE - TOP OF UPPER f LOW PRESSURE CORE SPRAY
,,/ ' .%
~~ Y D' r?E.M _
'* N , LEVEL 10 UPPER PLENUM - C r' -- - - - - - - - J-l LEVEL 9 e _ __-AVVLPPI,R, CORE GRID DOWNCOMER "
LEVEL 8 CORE SHROUD > m BYPASS ' m JET PUMP SUCTION
' * ~- * #" --~
FUEL CHANNEL- -F- -- -- - "- -" n VESSEL WALL- - . --- - - - - -- - - y LEVEL 7
--A JET PUMP _ LEVEL 6 l
- I CORE P.L. ATE
_,J I,l l . - -- [,,, LEVEL 5 REACTOR OUTLET --D' N0ZZLE LEVEL 4 _, y.,,jAFfLE PLA.TE_ _ _ r.- - _ __JL. LEVEL 3 - - - s k LEVEL 2 LOWER PLENUM m.. '- . .- .:i-i-- -
-- -} _ - - - - - -
LEVEL 1 e GUIDE TUBE
-- - --- N.. ,- ,,
/ .,_ _ .- _ Y_ _ _ _.
Figure A-5. Axial nodalization of the three-dimensional detailed vessel model. A-17 i
between the top of the upper plenum and the normal downcomer liquid level was represented by Level 11. The region between the normal downcomer liquid level and the top of the steam dryer assembly was represented with Level 12. The steam dome was represented by Level 13. The steam separators and dryers are physically present in the region represented by Levels 11 and 12. Level 11 serves as a perfect separator in - the TRAC-BD1 model. Several compromises were made in modeling the steam
~
separators and dryers. First, the distribution of downc ::ner fluid which surrounds the separators was not modeled properly because of a current limitation of the TRAC-BD1 computer code for representing steam separators. This limitation required either lumping all the downcomer fluid into the outer ring, even though some of the liquid should be in the inner three rings, or distorting the velocity profile in the upper plenum. The' velocity profile in the upper plenum was thought to be more important than the distribution of downcomer fluid. Consequently, all the downcomer fluid was lumped into the outer ring while the fluid inside the separators was distributed between the inner three rings. The upper plenum velocity profile and the distribution of downcomer fluid cannot both be properly represented with TRAC-BD1 until a separator component is developed. Levels 11 and 12 should be modified to better represent the separators and downcomer when a separator component becomes available. Second, geometrical details of the separators and dryers were not available during the development of the Dresden 3 model. The input flow areas and hydraulic diameters of the separators were based on data applicable to a ROSA separator.A-3 The gross dimensions of the ROSA ar.d Dresden 3 separators were similar which justified using the ROSA data in the absence of any better information. The details of the flow path through the steam dryers - were not represented in the model. Third, the model does not have the capability to accurately represent a transient in which the mixture level
- rises above the top of the separators. In the plant, a sufficient level increase would allow water on the inside of the flow skirt of the steam dryer assembly to spill into the top of the steam separators. In the model, the fluid inside the flow skirt and outside the separators was represented in the outer ring where it is prevented by the flow skirt from spilling into the separators until the liquid level reached the top of the A-18
steam dryer assembly. The Dresden 3 model would have to be modified by placing an extra level at the top of the steam separators and adding separator components to properly represent a transient which would cause < liquid to spill into the top of the separators. Finally, the model does not allow carryover of liquid into the steam dome or the carryunder of steam into the downcomer and thus may not accurately represent separator
. performance during some transients.
2.1.3 Fuel Channels Six TRAC-BD1 CHAN components-were used to represent the 724 fuel bundles in the core. The entire core was assumed to be loaded with Exxon fuel bundles although some General Electric assemblies will remain in the plant after the first Exxon reload. Each azimuthal half of the core model contained one high powered, one average powered, and one low powered CHAN component. -Each average powered CHAN component represented 232 fuel bundles. Components 41 And 42 each represented 42 high powered bundles which were located near the center of the core. Components 43 and 44 each represented 232 average powered fuel bundles. Components 45 and 46 each represented 88 low powered fuel bundles which were located near the periphery'of the core. Each high powered and average powered bundle in the plant contain a large side entry orifice. The low powered bundles in the plant contain either a large or a small orifice. An average orifice area was used in the low powered CHAN components. Each CHAN' component was modeled with 10 cells as illustrated by F.igure A-6. Two cells were used to model the unheated length between the
. side entry orifice and the bottom of the active fuel. The active length of the fuel was modeled with seven cells. The lower and upper thirds of the
. active fuel were each modeled with~two cells. The center third of the active fuel was represented with three cells. The more detailed nodalization which was used in the center third of the active fuel provided better resolution of the maximum linear heat generation rate than could be
'obtained with two cells. A single cell was used to represent the unheated length between the top of the active fuel and the top of the channel.
A-19
t
.d V -
g -- UPPER CORE GRID
-10 TOP OF THE ACTIVE FUEL LEVEL 8 9 f.
8 v -..... n 7.- LEVEL 7' 6 5-aL 4 LEVEL 6-l 3 3: I LEAKAGE M 2 /-- BOTT0f10F THE ACTIVE FUEL
- . . . . _ _V. PATH. _ _ _ d - - - SIDE ENTRY ORIFICE
, Figur'e A-6. Axial':nodalization of the CHAN component.
A-20
' Leakage'f' rom the fuel assembly to the-bypass was modeled in the cell
'below.the bottom of the heated length. Two cells were used to represent the: unheated length below.the bottom of the active fuel because unrealistic leakages-were calculated when a single cell was used.
~ Frictional head losses due to the . side entry orifice, grid spacers, and a' lower and upper tie plate were represented based on information 4'
-obtained from Exxon's RELAX model of Dresden 3. Geometrical details of the grid spacers and upper tie plate were not ainilable during the development of-'the TRAC-BD1 model.
Assumptionsincorporatedintothehhdrodynamicmodelofthefuel
-assemblies are described below. The side entry orifice was assumed to be at the elevation of the core plate although physically it is about 10 cm below the core plate. An additional vessel level would have to be placed in the model to represent exactly.the elevation of the side entry orifice.
.The lengths of the unheated cells were adjusted-so that the elevation of the' bottom of the active fuel was modeled correctly. The upper tie plate was modeled at'the top of the active fuel. The channel was assumed to end at the top of the upper core grid.
2.1.4 Guide Tubes Six TRAC-801 TEE components were used to represent the distribution of
-control . rod- guide tubes.within the lower plenum. Dresden 3 contains 177 control rod guide tubes as shown in Figure A-4. Components 51 and 52 each represented half of the 21 guide tubes located in the firs; (inner) vessel
.: ring. Components _53 and 54 each represented half of the 116 guide tubes located in the second vessel ring. Components 55 and 56 each represented
~*- half of the 40 guide tubes lccated_in the third ves.sel ring. The guide tubes were modeled between the top of the control rod drive housing and the
. core _ plate. The volume of the guide tubes and vessel cells were each computed assuming that the control rods were fully inserted into the core.
Three cells were used to represent the primary side of each guide tube TEE. A single cell was used to represent the secondary side of each guide tube TEE. A-21
t i f The major flow path between the lower plenum and upper plenum is through the. fuel channels. However, many.rclatively small leakage flow paths directly.or indirectly: connect the lower plenum and bypass. The { geometrical details-of the. leakage' flow paths in Dresden.3 were not
- available during theidevelopment'of: the TRAC-BD1 model. The primary source of information for modeling theileakage flow paths was obtained frem
~
. Reference - A-4'. The two largest leakage flow paths described in the -
Ireference were modeled. 'These flow-paths represented leakage from the
" lower tie plate to.the-bypass and from the fuel support piece into the z
l control; rod guide ' tube.- Le'akage from the lower tie plate to the-bypass was modeled with a flow path from the CHAN component to the' VESSEL: component. The; remainder of.the-leakage was modeled with.a flow path from the lower plenum to'the secondary side of the guide tube TEE. The flow area of each path was' adjusted to obtain the- desired-leakage. 2.1.5L Balance' of-Plant The detailed TRAC-BD1 model of Dresden 3 represented portions of the feedwater steam,1 control rod drive (CRD) hydraulic, and safety systems.
;The models of these systiems are de' scribed below.
Twoffeedwater lines penetrate the primary containment of Dresden 3. Each.of these.feedwater lines divide into two smaller lines which penetrate
~
'the: pressure vessel and are connected to the'feedwater sparger. In the
' TRAC-BD1_model, feedwater was introduced symmetrically into the.two azimuthal halves of the.downcomer through Components 71 and 73. Each component represented an arbitrary. length of the feedwater lines which penetrate the vessel. Geometric details'of the feedwater system were not- .
.available.during the development of-the model preventing an accurate 4 representation.of the feedwater system. The feedwater lines physically -
Lpenetrate the' vessel 12.3 m above the bottom of the lower plenum in the
- region represented by . Level 11. . However, feedwater was modeled entering the vessellin Level 10.in order to prevent unrealistic condensation which would>beccalculated if.the downcomer mixture level:was between the
- feedwaterusparger'and the top of Level 11. FILL components 72 and 74 were usedIto. inject'feedwaterintothe.feedwaterlines.
A-22
, _. _. _ _ _._ .- _ _ _ _ _ - . _ ~_
G
.Two VALVE components, one for each azimuthal segment, were 'used to
,. represent the four steam lines which penetrate the Dresden 3 vessel.
. Components 76 and 78'each represented two. steam lines. The first VALVE cell represented the portion of the steam line between the pressure vessel and the inboard main steam isolation valve (MSIV). .The second VALVE cell
' represented the steam lines between the inboard MSIV and the turbine stop
- valve. The inboard.MSIV was represented by the junction between.the first ir and second VALVE cells. The MSIV was used in the model to isolate the steam lines during a LOCA and to control reactor pressure during initialization. The' steam lines physically penetrate the vessel 15.7 m
-above the bottom of the vessel in the region represented by Level 12.
However,.the steam lines were connected to Level 13 in the model in order to prevent the calculation of unrealistic entrainment of liquid out the steam lines when the mixture level was between the bottom of Level 12 and the steam line nozzles. The development of a mixture level tracking model and/or the use of-more vessel levels would allow the steam and feedwater lines to be connected to the correct level. TRAC-BD1 BREAK components were connected to the steam lines to impose a pressure boundary condition on the model. The CRD hydraulic system was represented by-FILL components 61 through 66. -Each FILL component was connected to one of the six guide tube components. The automatic deoressurization system (ADS), high pressure coolant injection (HPCI), and low pressure core spray (LPCS) safety systems were
' represented in the TRAC-BD1 model. Each safety system was represented with
.- two FILL ~ components so that the flow from each safety system would be symmetrically introduced to the two azimuthal halves of the vessel. The geometric details of the safety systems were not incorporated into the model because these details were not available during the development of the model. An additional safety system, low pressure coolant injection (LPCI), was not represented because LPCI was assumed to fail due to a filed admission valve in the transients for which the model was developed.
A-23 l
-The ADS valve's were represented with two FILL componets as shown in Figure ~A-1. The FILL componets were connected.to PIPE componets which in l turn were connected to the vessel. The PIPE components were required because-TRAC-BDI- does not; allow a FILL component- to-be directly connected
;to a VESSE'l component- . The ADS valves are physically located in the main steam lines. -:The ADS was connected'to the correct vessel level even though ithe steam . lines were artifically.. raised one level asl described previously.
- y TheHPCIlandLPCSsystemsweremodeledasFILLcomponentsconnectedto -
VALVE; components'which were connected to:the vessel. The VALVE components
~
'wereiused to' represent the opening times of-the HPCI and LPCS admission valves. HPCI. physically enters the vessel-.through'the feedwater spargers.
-The.HPCI system was connected.to the correct vessel level even though the
- -feedwateriline:was artificially lowered one level in the model as described
'previously. ._LPCS physically enters the vessel through spargers-in the L . upper' plenum. 'The LPCS system was attached to the third vessel ring which
' '; representsithe' region containing the core spray spargers. 2.2- Heat Transfer Nodalization-
.Three different' types of. heat stuctures are available in TRAC-BD1. A rod. heat structure is available to represent heat transfer from fuel rods.
~
'A, wall-heat slab, called a. double-sided slab ~in the-VESSEL component, is available to. represent heat transfer through a cylindrical wall. The heat conduction. equation is solved for both rod and wall heat slabs. A lumped
~
parameter heat slab is ~availabe to represent internal vessel structures that cannot be readily. characterized by a. cylindrical heat conduction model.
~
2i2l1 ' Recirculation Loops- .
. I
- A pipe wall was attached to each cell in the recirculation loops to
= account for. heat transfer from the pipe to the fluid. The outer surface of the pipe. wall wastassumed to be adiabatic except in the jet pumps where the
- outer; wall was' attached to the'downcomer fluid. The walls of the isolation ivalves and reactor inlet and outlet nozzles were assumed to be of the same geometry as the adjacent. loop-piping. Adjustments for loop penetrations L
A-24
rf and/or irregular geometries were not made. 'The recirculation pump wall was assumed to be an' average of the walls of the adjacent suction and discharge piping. The diameter and thickness of the recirculation loop piping wall downstream of,the pump is not constant in Dresden 3. For example, the 4 . thickness of the manifold wall.is different from the thickness of the jet pump.. risers. 'However, TRAC-BD1 assumes the same wall thickness and radius for~each cell within a component. An average wall thickness and radius was
~ calculated for Components 4 and 14. .The total heat transfer surface area and metal volume for Components 4 and 14 were correct but the distribution of a'rea and volume within the components was different from the actual
. distribution in the plant. A. portion of the jet pump risers are physically inside the vessel and can communicate thermally through the riser wall with the downcomer fluid.' The TRAC-BD1 heat transfer option which allows pipe cells to thermally communicate with vessel cells was not employed for the risers because the code would'have assumed significantly more riser metal volume would be present in the downcomer than actually exists. An improved
. representation of the distribution of the wall mass and heat transfer area in the pump discharge piping coulo be obtained by modeling the pump discharge piping with more components but would increase the computational expense. The present representation of the piping wall is believed adequate for most transients.
'2.2.2 ~ Reactor Vessel The major metal. structures located within the reactor vessel were represented as' heat slabs in the TRAC-BD1 model. A double-sided heat slab
~
j was used to represent the pressure vessel wall at each vessel level in the
... - mo' del . : TRAC-BD1 solves the heat conduction equation for each double-sided heat slab. The Dresden 3 pressure vessel-is primarily carbon steel although the vessel is lined with a thin stainless steel cladding. The pressure vessel was assumed to be entirely carbon steel in the model because TRAC-BD1 allows only a single material for each double-sided slab.
The outerisurface of the-pressure vessel wall was assumed to be adiabatic. The' pressure vessel was assumed to be of constant thickness; no adjustments A-25
.to the. heat' slab area-or thickness were made because of penetrations or flanges. The' core shroud, which separates the bypass.from the downcomer,
~
was also modeled as a double-sided' heat slab in Levels 4 through 9. F f The distribution of heat' transfer area and mass for a hemispherical wall,'such as exists'in the lower plenum, cannot be readily described with the TRAC-BDI. double-slab model because the code assumes that a slab attached' to an inner. ring forms the boundary between rings. The code then calculates heat conduction from one ring to.the next through the
-double-slab. The code's treatment of a double-slab does not correctly represent the geometry.of a hemispherical vessel wall. An~ approximate solution was obtained by doubling the thickness'of the double-slabs in the inner ~ rings of. Level _1:and inputting the area such that the total wetted
. area equalled the' geometrical area. This technique conserves the heat
_ transfer. area, metal mass, stored energy,'and provides a close
. approximation to the correct conduction solution.
Lumped: parameter heat slabs were used to represent the vessel internals listed in Table A-3. TRAC-BD1 does not solve the heat conduction
^
equation for lumped. parameter heat' slabs but instead assumes that the temperature-is uniform within the heat slab. The assumption of uniform temperature within the= heat slab is valid only if the slab is thermally
" thin". Representing structures which are' thermally " thick"'with lumped parameteriheat slabs requires ~the user to model an " effective" thickness which is'less than the actual-thickness. The_ structures represented by lumped-parameter heat slabs in the Dresden 3 model were thermally " thin" because all"the metal ins less than 2.54 cm from fluid.' Thus, the total thickness of each heat structure was.' represented in the model. The lumped -
, parameter heat slabs were assumed to be 304 stainless steel.
2.2.3' Fuel Channels CHAN components 41 and 42 were used to_ represent the high powered fuel assemblies located near the center of the core. Components 45 and.46 represented the low powered fuel' assemblies located near the periphery of the core while components 43 and_44 represented the remaining A-26
~ ~
(1 TABLE A-3. LLNPED PARAMETER liEAT SLABS s Heat Slab Surface Area Vessel Heat Slab Mass 2 Structures Represented By Lumped Level '(kg) . (m ) Parameter Heat Slabs l- 8224 82.11 Control rod drive housings In-core instrument housings
, 2 -1019 17.95 In-core instrument housings
- 3. 5360 25.61 In-core instrument housings Lower shroud 4 -1108 15.39 In-core instrument housings 5 3730 23.04 In-core instrument housings Half of the core plate 6: 10800 239.6 Half of the core plate In-core instrument assemblies Control rods
- 7. 8279 241.0 In-core instrument assemblies Control rods
~
8 14590 278.3 In-core instrument assemblies
. Control rods Upper core grid 9' 907 11.68 Core spray spargers 10 10330 81.11 Shroud dome Stand pipes 11 35050 861.2 Stand ' pipes Steam separators Holddown bolts Lift rods
* ~
Bolt rings Flow skirt
, 12 44950 330.1 Steam separators Holddown bolts Lift rods Bolt ring Flow skirt Steam dryer assembly
-13
- 0. O. --
l A-27 L - - !
faverage powered fuel. assemblies. Five rod groups were.used to represent
- ithe fuel
- rods"in the high powered channels. Specific rod groups were used (to represent the. rods adjacent to:the channel. wall, the high powered rods,
.an'dlthe water rod. The average powered and low powered channels ~were
{modeledwithtwo'rodgroups;onegroup~ represented'the.waterrodandthe
, -other_ group' represented an average fuel rod. The channel. walls were ~
modeled:as heat . slabs to represent: the heat transfer from the channel to' - _ the bypass. TRAC-BD1 does-not have a: mechanical _ fuel rod model. Consequently, nominal radial dimensions of the fuel rods were input an'd the gap conductance adjusted ~so.that the initial fuel centerline temperatures in
.the TRAC-BD1 model were consistent with detailed! fuel rod calculations performed by Exxon.
~2.'2.4 . Guide Tubes.
.A heat' slab was attached to.each guide tube cell to model the heat transfer:between the fluid in the guide tubes and lower plenum. The heat
._ slabs. represented.the guide tube walls ~ and the fuel support pieces.
2.2.5- Balance of Plant ~ The walls of the feedwater,. steam, CRD hydraulic, and ECC systems were not represented with heat slabs because'the.model generally did not represent the actual geometry of these systems. 2.3 Code Options- . The~ user-selected options which are available~in TRAC-BD1 can be -
-separated linto three general classes;.those pertaining to main program control,. components, or reactor. power. Descriptions of the three classees of options are provided b'elow. The selected options are consistent with current recommendations for use of TRAC-BD1.
i A-28 i i l
p i
~
'2.3.1 Main Program Control' The'covergence criteria used in'the Dresden 3 model were the values suggeste'd in the TRAC-BD1 manual.A-1 Suggested values for the maximum
-number of the various types of iteration were also used. The fully implicit-option for the conduction boundary condition was 'used. The water
-m ' packing option was not used.
- r
'2.3.2 . Component
.The TRAC-BD1 critical flow model was applied at the jet pump driveline nozzle;in:the broken' recirculation loop. The critical flow model will also be' applied at the BREAK junctions which will connect the broken loop-and the p containment. The critical flow model was not applied at any other junction in the model.
The TRAC 4BD1 countercurrent flow limiting (CCFL) model was applied only at-those junctions representing a side entry orifice, the upper tie
-plate,-the upper core grid,.or a jet pump suction. The CCFL model with side entry orifice coeffecients was-used at the junctions representing .a side ' entry orifice. The CCFL model with upper tie. plate coefficients was used at all the other junctions mentioned above.
Critical heat flux (CHF) was calculated with either the Biasi or the Bia'si critical quality correlations A-1 in the CHAN components. CHF was
~
calculated.with'the Zuber/Biasi correlation' A-1 in the guide tube components. CHF-calculations were not performed for the other components. a The minimum film boiling temperature was taken as the maximum of the. temperatures obtained from.the homogeneous nucleation and Iloeje
. .: correlations for-CHAN components.
TRAC-BD1 calculates heat transfer due to radiation in CHAN compon'ents. The radiation of heat to steam and water droplets was modeled. View factors were; corrected for anisotropic radiation. The
' suggested valueA-1 of the~ threshold void fraction for performing radiation: calculations, 0.9, was used.
A-29
~
A
- x. ['
~
'The;: TRAC-BD1 model for. metal-water reaction was~ used.
-2.3.3 Powir a
Reactor power- was_' calculated with a point. kinetics model which i45 -included theDeffects:of. reactivity insertion due to scram and reactivity
% feedback. . Reactivity feedbacks due to changes in fuel temperature, *
-d imoderator-temperature,:and moderator void fraction were represented. The power: squared weighting option' for reactivity feedback was used. Default values for;the' fuel temperature and moderator ~ void fraction reactivity c'oefficients were :used. The moderator temperature reactivity coefficient
- was based on Figure 3.3.'11 of the Safety Analysis ReportA~4 (SAR). The
~
scram-reactivity insertion was based on information obtained from the Exxon
-RELAX model. . Scram was assumed to be caused by 10% closure of the MSIV.
Decay heat was calculated based on fission products of uranium-235, uranium-238, and plutonium-239_ assuming an infinite reactor operating time. 1The decay of:the heavy: elements uranium-239 and neptunium-239 was
;also represented. 'The-decay heat power wa's multiplied by 1.2 to be' consistent:with Exxon's licensing calculations.
All: core power appeared as heat flux from the fuel rods. Direct
~
. moderator. heating was n'ot mcdeled.
=2.4 -
Boundary Conditions o Bounda'ry conditions were used to represent the interaction between the portion of_Dresden 3' represented by the TRAC-BD1 model and the balance of plant. The. boundary conditions were developed to represent the behavior of the plant following a large-break LOCA with a simultaneous loss-of-offsite - l power.and a_ failure of the LPCI system. The' assumed boundary conditions are described below. 2.4.1.~ ADS Five pressure relief valves are located on the main steam lines in Dresden 3. The_ actuation of ADS opens these relief valves. Table A-4 A-30 L ".
L-
~'
TABLE A-4. ADS' FLOW Pressure. Velocity
- Mass Flow Rate
- (Pa)~ 'm/s) (kg/s) 0.0 .0.0 0.0 4.461E5 0.0 D 0.0 1 1.136E6 -14.35~ - 54.38 1.379E6 -14.41 -65.80
; 3. 2.758E6 -14.45 -130.1
--14.35 -194.7
-. 4.137E6 7.929E6 -13.77 -378.0 9.653E6 -13.42 -464.8 a'. : . Negative velocities indicate flow .out of the vessel .
~ b. The ADSLvalves close by internal-spring tension at low reactor pressure.
S 9 e a A-31.
i shows the flow vs pressure curve used in the model to represent 'Ai?S. The table shows both velocity, which is the input required for TRAC-BD1, and the corresponding mass flow rate. The flow vs pressure curve, which was based on the assumption that all five relief valves open, w'as obtained as described below. The total relief valve capacity was obtained frcn the SAR and was assumed to represent choked flow of saturated steam. The total relief valve capacity was converted to a choked flow area using the - homogeneous equilibrium critical flow model.A-2 The choked flow area and
~
the critical flow model were used to obtain a critical steam mass flow rate vs pressure curve. The steam mass flow rate was converted to velocity by dividing by saturated steam density and the total flow area of the steam lines. The velocity vs pressure curve used in the model was based on the assumption that the steam lines were filled with saturated steam. The representation of ADS would not be valid if significant amounts of liquid were entrained into the steam line. The Dresden 3 ADS is actuated on coincident signals of low-low reactor water level and high drywell pressure if at least one LPCI or LPCS pump is running. A maximum time delay of 120 s is used to allow either of the initiating signals to clear before ADS is initiated. The ADS valvas will open when the initiating signals are satisfied and the reactor pressure is sufficient (greater than 1.136 MPa) to overcome the spring tensi9n of the valves. The TRAC-BD1 model does not represent 'the drywell as a cor rol volume and hence cannot represent the high drywell pressure condition. The user should verify that high drywell pressure would have occurred before allowing ADS to be initiated in the model. 2.4.2 HPCI . Dresden 3 contains one turbine-driven HPCI pump. The HPCI pump takes - suction from either the Condensate Storage Tank or the suppression pool and delivers liquid through the feedwater spargers to the downcomer. Table A-5 shows the flow vs pressure curve used in the model to represent HPCI. The curve was based on the minimum HPCI flow rate allowed by the Technical Specifications.A-6 The HPCI pump should actually deliver a higher flow rate than that assumed in the model. The HPCI flow presented in the SAR is A-32 1
4 r* / TABLELA-5. HPCI' FLOW '
- ' Pressure c Velocity"' -
Mass Flow Rate
- -(Pa)- (m/s) (kg/s) 0.0 0,0. 00 3.447 x 105 0.0 o0 11.136 x<106 1.296 '313' 8.030 x 106 :1.296- 313.
i
. a. ;The've' locity'was based on the combined ' flow ' area of the four feedwater
~
lines which penetrate the pressure vessel. x 1
- v
- f.
~
, i c .Y jL. @
v E. . [- ,
-A-33 1<
m
aboutil2% higher than the minimum flow allowed by the Techr.ical Specifications.- Information on the' performance of the HPCI system at high reactor pres'sure, above 8.030 MPa, and low pressure, below 1.136 MPa, was not available during the development of'the: TRAC-BDI model. Consequently,
-the. assumed performance of the HPCI system may not be correct at either high or low reactor' pressure. The HPCI water' temperature was assumed to be 310.9 K. zThe line which supplies steam to the HPCI turbine was not modeled. -
~HPCI is' initiated by either low-low reactor water level or high drywellopressure-in'Dresden 3. The TRAC-BD1 model does not represent the
.drywell as a control volume and hence can't represent the initiation of HPCI due to high.drywell pressure. A 20 s. delay was modeled between the occurrence of the' low-low level initiating signal and the opening of the HPCI admission valve. The 20 s delay consisted of a 10 s delay for the
~
HPCI. turbine to reach operating speed and a 110 s delay for the admission
, valve to.open.
2.4.3 LPCS-Dresden~3 contains two LPCS pumps. The LPCS pumps take suction from the suppression pool and discharge the water throu5h separate core spray
.spargers in the, upper plenum. Table A-6 shows the flow vs pressure curve used in-the model to represent.LPCS. The flow vs pressure curve, which represents the output of'both LPCS pumps, was based on the minimum LPCS flow rate allowed by the Technical Specifications.A-6 The LPCS pumps should actually deliver significantly more flow than that represented by the model. The LPCS flow' presented in the.SAR is at least 30% higher than the flow rate required by the technical specifications. The LPCS water .
' temperature was assumed to be 308.2 K which corresponds to the maximum suppression pool _ temperature allowed by.the Technical Specifications during -
normal power operation.
- Either low-low reactor water level or high drywell pressure actuates
[the"'.PCS pumps in Dresden 3. Low reactor-pressure causes the LPCS admission valve to open provided that an LPCS pump is running. The TRAC-BD1 model does not represent the drywell as a control volume and hence l A-34 n
- m. -
4 2 TABLE-A-6. LPCS FLOW
' Pressure Velocity" Mass Flow Rate
.(Pa)' (m/s) (kg/s)
.0.0 7.383 707.
.1.014'x.105_ .7.383 707.
;7.219 x-105 5;880 563.
2.170 x 106 0.0 0.0
- n a. The velocity was based on the combined flow area of the two core spray
- nozzles which penetrate *.he reactor -vessel.
i
)
.h '
,.9 i
A-35
cannot represent'the initiation of L'PCS due.to-high drywell pressure. A
~
3D sldela'y for the diesel' generator to start the LPCS pumps after the receipt-of the low-lbw level initiation signal.was modeled. The LPCS
. admission valve was assumed to open 8.5 s after the reactor. pressure
-dropped below the low pressure.setpoint of 2.17 MPa.
!2 . 4.'4 LPCI.
.Dresden 3.contains four LPCI pumps. :The LPCI pumps.take suction from the~ suppression pool and discharge into the recirculation loops between the pump and the. manifold. .The LPCI system has logic to detect the occurrence of a-recirculation loop break and select the intact loop for LPCI injection. The TRAC-BD1 model of Dresden 3 did not represent the LPCI
-sytem becauseLin the transients for which the model was developed the LPCI admission valve in the intact loop was assumed to fail closed.
2.4.5 -Steam, Feed, CRD, and' Recirculation Pump Systems-
-The TRAC-BD1 model of.Dresden 3 represented portions of the main steam, main feedwater, and CRD systems. These systems contain valves which isolate the. reactor from th'e balance of the plant following loss of offsite
~
power. :The model-repre'sented the isolation of the vessel from the balance of. plant following, loss of offsite power. The MSIV's were assumed to close cithin 5 s of the loss of offsite power. The check valves on the feedwater and CRD lines were assumed to close within 0.5 s and 3.0 s respectively, of
;the: loss'of offsite power.
The' recirculation pumps were tripped and allowed to coast down , simulataneously with the-loss of offsite power. All the recirculation pump
. isolation-valves were assumed to remain open except for the discharge valve -
.onLthe' intact loop recirculation pump. The discharge valve began closing 5 s after the loss of offsite power and was fully closed 27.5 s later.
2.5 Initial Conditions The initial conditions used in the TRAC-BD1 model were as close as possible to the initial conditions used in the Exxon licensing A-36
~_
i. calculations. . The initial conditions generally represent " conservative" rather than "best-estimate"' values. Table A-7 shows the initial conditions used in the TRAC-BD1 model. The TRAC-BD1 control system was used to made a steady state run which provided the initial conditions incorporated in the model. The pressure, temperature, velocity, and void distributions obtained from the steady state run were checked against available data and were generally found to be reasonable. However, insufficient data were available to completely check.the accuracy of the calculated initial conditions. The fuel assemblies were assumed to be filled with beginning-of-life Exxon fuel. The beginning of life assumption was thought to be the most
~
appropriate for the model because TRAC-BD1 does not have a detailed fuel model to represent the mechanisms, such as ballooning, which may be important at higher burnups. A " conservative" rather than "best-estimate" core power distribution is incorporated in the TRAC-BD1 model since the model was developed to audit Exxon's licensing calculations. The maximum average planar linear heat generation rate for the model equals the proposed Technical Specification limit for Dresden 3. The radial power peaking factors were input to the model so that the results of the TRAC-B01 LOCA calculations for the high powered and average powered assemblies would be directly comparable to Exxon's corresponding calculations. The specification of the radial peaking factors for the high powered and average powered channels caused the radial peaking factor for the low powered channels to be about 50% too high compared to a "best-estimate" power profile. The assumed
, axial power profile was symmetric about the center of the core. A slightly different axial power profile was used in the high powered channels than . was used .in the other channels so that the TRAC-801 results for the high powered and average powered channels would be directly comparable with the Exxon results. Gap conductances were adjusted so that the initial fuel centerline temperatures in the TRAC-BD1 model were consistent with the results of detailed fuel rod calculations performed by Exxon.
A-37
a .--
-.- TABLE A-7. . INITIAL '. CONDITIONS _
' Parameter Initial Value Reactor powera 2577.54 MWt Steam dome. pressure 7.033 MPa .
lFeedwater flow ratea 1251. kg/s
.Feedwatar temperature 441.0 K -
. Total recirculation loop flow . 4560. kg/s
- Recirculation pump speed- 170.7 rad /s r _
~
Total . jet pump discharge flow 12348. kg/s Bypassiflowb 10.66% Channel leakage b 5.01% Guide tube leakage b- 5.66% CRD flow. 2.268 kg/s CRD_ temperature Lower plenum temperature 548.9 K Downcomer liquid levele 10.479 i
- a. Correspondsf to 102% of rated thernal power.
- b. Expressed as a percentage.of the total jet pump discharge flow.
- c. Corresponds to the. level above the bottom of the baffle plate which
~ separates the downcomer and lower plenum. The baffle plate is 3.023 m above the bottom of the lower plenum.
e c A-38
b,_
- 3. OTHER MODELS
- A two-dimensional detaile'd model and a one-dimensional scoping model of Dresden 3 were. developed in addition to the three-dimensional detailed model described in Section 2. The two-dimensional. detailed model and the one-dimensional scoping model were based heavily on the three-dimensional Edetailed model. .The two-dimensional detailed model-and-the one-dimensional
-scoping'model are described in Sections 3.1'and 3.2, respectively.
3.-1 Two-Dimensional Detailed Model
.The nodalization diagrams for the two-dimensional-detailed model of
~
Dresden 3:are shown in Figures A-7 and A-8. The two-dimensional detailed model was identical to the three-dimensional model-except for changes
.resulting .from representing the vessel with one rather than two aximuthal segments. .The radial and azimuthal nodalization of the two-dimensional
' detailed vessel model is illustrated in Figure A-9. 'Each vessel level in the two-dimensional detailed model contained half.as many cells as in the three-dimensional model. Consequently, the two-dimensional model used half as many CHAN and TEE _ components-to represent the fuel assemblies and guide tubes as the'three-dimensional detailed model. The two-dimensional model also used half as many components to represent the steam lines, feedwater lines, HPCI, LpCS',' ADS, and CRD systems as_the three-dimensional model. It
'was 'a relatively easy task to collapse the three-dimensional detailed model
~into a two-dimensional model because the three-dimensional model was
.azimuthally' symmetric.
_,. 3.2 One-Dimensional S:oping Model
= ~Nodalization diagrams for the one-dimensional scoping model of Dresden 3 are shown in Figures A-10 and A-11. The one-dimensional scoping i model_was designed to achieve relatively fast running times. - Consequently, the scoping model represented significantly less detail in the recirculation -loops and the vessel 'than either of the detailed models.
A-39
2 lireak (Main steam) I15 F80-Ll6_)(_ JE 49 Fill fill (ADS) i 50 iM 13_ _._JM F 111. (llPCI)' l 3a jEC2 I 41'- fill (ftnin feed) F- 7z _ 82r- p ,- , 8J, ,
~ ' '
4 43 41'
~
6, 7 r-- 18 37 i
.I I 5 4! O 4: 15
'1 15 '3 11 I i
Intact Recircula tion > m -.tc-- Broken Recirculation Loop i ~ '
'" {m-f fj [ j Loop 8 as u >> i, 4
g , s g ,3 g, o , a i i I I 65 I 6'[I 6'l I , l
,5 ,3 rij Notes:
I
}4 i l 1 Junc tion ' t
}6 1 Component 1 iIL 1 2 MM i, 3 gin X Valves LEGEND:
ADS - Automatic depressurization system O aecircuiatioa no ns CRD - Control rod drive llPCI - liigh pressure coolant injection LPCS - Low pressure core spray Figure A-7. TRAC-BD1 two-dimensional detailed model of Dresden 3. t t t Q 9
^
- _ _ _ _ _ - _ _ _ _ . _ _ - - - - - - - -- - - - - - - - - ^ - - - -
CENTERLINE SECTICN A-A RI!G 4 PRESSURE VESSEL WALL RI?G 3 , s
., OCWNCCMER N-RI?G 2 g
\
- \
RIIG 1- 's k
.i ' j= g . ,
. g i
** o I
< ! k
LCW-PCWERED CHANNEL f
' AVERAGE-PCWERED CHANNEL SYPASS ---> -
,' HIGH-PCWERED C'4ANNEL LEGEND:
CORE ACS - Autcmatic decressuri:stion system IHRCth HPCI - Hfgn pesssure coolant injection LPCS - Low pressure core scray M 20.920 i
; I 5- . - 7.,gg3 377;g LEVEL 13 , ,
i.- - I. 17.564 ---e l STEAM LEVEL 12 , CRYERS I
- I : 205 13.500 .---i---J----- - - ---
', SEPARATCR$ s HPC:
LEVEL 11 7 ~~~
~~j LEVEL 10$ uiPER ?EET:03~~~ -
4 I.'s FEE 3
- 10.231 -- , ----e---- - - - .
L?CS LEVEL 93 9.509 ; . w LEVEL 8 ::: a lo It T*i' 1
. v e "-
5 6 v4
- A 7*733 -
2 -" w - - '- LEVEL' 7 r: *Sd "
- ' ig :!
6 6.600 _ eE ,, d 3 ~ es 53- . 'ig s > EC C
. LEVEL b 5.257 f *: C - ...
JET P'#P LEVEL 5' 4.431 -- - - - - - ' - - - - - - -- I
- 'ECIRC'J:.ATICM 'et:'P LEVEL 4 - - 6. - -
LEVEL 3 7.b1 - - 1 - ---.r--' LEVEL 2 1. 295 - - -r- - - r - ' - - - - LEVEL'1 0.0 O.3~ 0.7381 Z.3123 2.5051 3.1877 5 RADIUS (t't) Figure A-8. TRAC-BD1 two-dimensional vessel model. A-41
l 8 S d 5 5 y e nu d
- 5. a . E c; $ l s s e m c
s 5
=.
i s s x G
- s ,
E 8 -
? 3 -
e e f C e N a a"EB8EESEE88tse
* * =*uF' a 8Ey[SBEEEE8888888Ek s R i 3
a5 EB888888BR85WBBB(2Sco!, ; C 85 3588BEEsP++isF+2EE88* 8E 8888888 8EB88 E~ uss88e e e : s :
& 8 88888 585E888$ 888888E @ '
s 5 SE EB8888si WEB 882 ESBESW + i
> .9 86' 88E8%E8488883E E88&s888EF m i
8 o EEEEEE88& cs 3 l 888SEEBAE li oc 28SRBR- ic& o li i ; k Sc3EE && 33
;z g
m , .
=
3 s g s s x s s en 8 - E - d%= E 8 E EeW N S g
$ $ $5 Qm E i
=S '
$c$555 o +e , A-42
as Break (Main steam) l 7s FBIi.176 X P ] 49.- Fill Fill (ADS) i so I E *L__57 11' Fill (llPCI) ( 38 l 'lWY !
- 8 Fill (Main feed) i ic d ii i _1, 43
- ,-- 18 6} 7 37
.l' l!
5 O IS 1- '3 11
.A, o -- c Intact Recirculation 8 - 33 19 Broken Recirculation Loop i3 Loop
'{
5 1 1 i I I 6I *
~
l I . 33 Notes: l l i A Junction 1 . f4 - {6 1 Component 1 hell 3-]; X Valves
-h -
LEGEND: O aecircuiation Pu ns ADS - Automatic depressurization system CRD - Control rod drive llPCI - liigh pressure coolant injection LPCS - Low pressure core spray Figure A-10. TRAC-BD1 one-dimensional scoping model of Dresden 3.
l l'~ CE:ITERL!NE
'SECTION A-A L
[ h' RII.G 2 % PRES $ uke VESSEL 'JALL
\
\
\
\ CO'JNCCMER
\~
v . RI:4G 1 ; I . 43-3YPASS I AVERAGE-PC'JERED CHANNEL LEGEND: CCRE ADS - Automatic depressuri:ation system SiiROUC i:PCI - Hign pressure coolant injection LPCS - Low pressure core scray 20.920 , . LEVEL 8 ---> - MAIN STEAM 17.564 --------_-l---- STEAM CRYER LEVEL 7 p ADS 13.036 - - - - - - - . - - - - - - - LEVEL 6 SEPARTORS 4---- H P C I 2 11.036 _ _ _ - - _ _ _ - _ . . _ - . . LE'!EL 5 g UPPER PLENUM , 4 HAIN FEED
@ 9.509 _ _ _ _ -_ _ - _ . 4 LPCS A, A LEVEL 4 $ y
- 7.913 V_, . _ _ , , , ,, _
, ~ LEVEL 3 3ggk ,
5.267 1 < c- v s
-, 4- JET PUMP LEVEL 2 - - > TO RECIRCUALTION PtNP 3.023 . . . . . . . - - . . . ,
LEVEL 1 l 0.0 . 0.0 2.6051 3.1877
> RADIUS (m)
Figure A-11. TRAC-BD1 one-dimensional vessel model. A-44 I
f 7.. F'igure A-12 is a schematic of the intact recirculation loop which illustrates the-regions represented by the various TRAC-BD1 components and cells; The broken recirculation loop was modeled similarly. The vessel was' represented with one azimuthal-segment,.eight axial levels, and two
~
h radial rings in th'e scoping model. The use of only one azimuthal ~ segment precludes the scoping model from-representing azimuthal gradients within
+- - the . ve s sel'. The~ axial nodalization of the vessel ~is illustrated by
- Figure A-13. zThe scoping model represented the vessel with eight axial levels whereas the detailed models used 13 levels. The scoping model provided-lessfaxial. detail in the lower plenum, bypass, and upper plenum regions than the detailed models. Two radial rings were used to represent the-vessel in the scoping mo' del as illustrated by Figure A-14. One ring represented the region inside the core shroud; the other ring represented the region between the core shroud and the vessel wall. The scoping model cannot. represent radial gradients within the core shroud. Consequently, one averaged powered CHAN component was used to represent all the fuel channels. A single rod group was used to represent all of the rods in the fuel channels. One TEE component was used to represent all the guide tubes. Two cells, corresponding to two vessel levels in -the lower plenum, 4 were.used to represent the primary side of each guide tube TEE. The main steam, main feedwater, ECC, ADS, and CRD systems were represented with the same nodalization as used in the two-dimensional detailed model.
- 4. MODEL-LIMITATIONS Many assumptions were made during the development of the TRAC-BD1 models of Dresden 3. These assumptions may limit the applicability of the
- y. models for certain transients. Some of the major limitations of the model are described below.
1 The initial and boundary conditions incorporated in the models were based on the_ initial.and boundary conditions used in~ Exxon's licensing calculations. The initial and boundary conditions incorporated in the models generally represent " conservative" rather than "best-estimate"
.. values. Thus, the initial conditions incorporated in the models may not be appropriate for certain applications.
A-45
- DRIVELZNE Y
'f 10ZZLE
-REACTOR OUTLET. JET PUMP N0ZZLE d TYPICAL OF 10) p REACTOR INLET- p i THROAT
/
g 'N0ZZLE g ,r
# #- J + DIFFUSER
... + VESSEL WALL 5 .
1 -DISCHARGC'
.E7 72 __ . M -
NORMAL FLOW N AFFLE B PLATE DIRECTION JET PUMP RISER (TYPICAL DE 5) 1 _. y L _...
)
MANIFOLD PUM ION p RECIRCULATION PUMP 4 PUMP DISCHARGE 4---- PIPING
~
y l J l 2' \ 3$ A j J' LEGEND 1 COMPONENT 1
-- CELL BOUNDARY ISOLATION VALVES. - COMPONENT BOUNDARY Figure A-12. Scoping nodalization of the intact recirculation loop. .
^ A-46
STEAM DOME LEVEL 8 ____a--_ V _____,__m STEAM DRYER PANEL
%s g
.i --
FLOW SKIR*- K _. LEVEL 7
.- STEAM LINE 7 .
NORMAL WATER lm
-'-_.-r.'-
- _ .. _ . . _Y A
LEVEL. _
, STEAM SEPARATOR FEEDWATER LINE y - LD EL 6 TOP OF UPPER STAND PIPE y ~--
-3,,,- - - -
x -; - -
< ---y PLENUM LOW PRESSURE CORE SPRAY P LEVEL 5 UPPER PLENUM 4. -
,. _Y ._ _ _ _ _ y_UPP__ER CORE _GR_ID D0'.!NCCMER -'**
CORE SHROUD I LEVEL 4 BYPASS % 'j M _ FUEL CHANNEL- I tr
.x- _ _. _ ._ __.. _ _ a . _ _ _ r J_ET_ PUMP 5,UCT,, ION VESSEL WALL
/ ^
LEVEL 3 UU N~
- g. . .
J L m _ -
- -[ CORE LEvEt z PLATE
~
REACTOR OUTLET ---->~ N0ZILE I -- --
- _ - - {_ BAFFLE PLATE LOWER PLENUM , , ,, ,
LEVEL 1 ., GUIDE TUBE N - _ j.__ _ _ . _ _ - _ _ - - y Fiaure A-13. Axial nodalization of the scoping vessel model. A-47
4 4 m 5 5m SS
~
W ia e. ~ss s E e n g
~~~ _
N obE$8Ea mEEEEEEEEEm ; s cEEEEEEEEEEEm s , EEEEEEEEEEEEE ; l eB8 E E E E E E E E E E E Ea i ! MEEEEEEEEEEEEEE s l 88EEEEE8BEEEEEEEE s ; EEE8888E8E88EEEEEEE , s 1
/ 88E88EEEEBBEEEEEEEE : !
MEEEEEEEEEEEEEE 5 l k EEEEEEEEEEEEE $ i EEEEEEEEEEEEE i e EEEEEEEEEEE 5 ' EEEEEEEEE 5 e EEEEE l
~
~
~ i 4 i E . l N N ' .2m
~ -
d$a w s a
!ss o e bh5 ac+e A-48
The TRAC-BD1 models do not represent the entire Dresden 3 plant. The turbine, condensor, feedwater heaters an'd pumps, isolation condensor, and containment were not represented. Some safety-related systems, such as standby liquid control and LPCI, were not represented. The relief and safety valves were not represented although the ADS function of the relief valves was represented. -The control systems which determine plant behavior s during normal operation were not modeled. The Dresden 3 models may need to be modified to represent some of the above-mentioned systems for certain transients. Geometric details of some of the components represented by the TRAC-BD1 models were not available during the development of the models.
-The best available information regarding these components was incorporated in the models. Components for which geometric details were not available include the steam separators and dryers, main steam and feedwater lines, safety systems, upper and lower tie plates, upper core grid, and grid spacers. The geometric details of the DresJen 3 core bypass flow paths were also not available during the development of the models.
The TRAC-BD1 models do not have the capability to accurately represent a transient in which the steam separators are flooded due to a high downcomer liquid level. The models should be modified prior to running a transient in which the steam separators are flooded. The computer run time required to calculate a transient with the models may be relatively large because the models represent the actual geometry of the jet pump nozzles. TRAC-BD1 limits the time step size based
. .on a Courant criterion, cell length divided by velocity, in order to resolve spacial gradients of temperature, density, etc. The jet pump nozzles are relatively short and generally experience high velocity flow.
Consequently, the jet pump nozzles will Courant limit the time step during most transients. Some transients may require a renodalization of the jet pumps in order to obtain an acceptable run time. The TRAC-BD1 models of Dresden 3 are thought to be adequate for performing the LOCA calculations for which they were designed even though A-49
f- 1 +: , y , l $hemodels..have'.someinherent' limitations.-Themodelsmaybeused-to-
~
> represent other.' thermal-hydraulic transients although some modifications-
[ Q ay be' req'u' ired. c h-s
) v g 6 P
L t
)
- l. t
(.. ?', e i , i l* s I i
?
p. I '
* ?
?
l i-
- i i
f :lx . 1. [ ! - [ e A ; i
v.
'r, REFERENCES
. : v-A-1. _J. W.lSpore et al'.,' TRAC-BD1: An Advanced Best Estimate Computer
- / = Program for Boiling Water Reactor Loss-of-Coolant Analysis, F ;NUREG/CR-2178 EGG-2109, (October 1981).
A-2. S; ' R. Behling et al . , RELAP4/M007 - A Best Estimate Computer Program 1to -Calculate- Thermal and Hydraulic Phenomena in a Nuclear Reactor or
^
- . Related System,.NUREG/CR-1998 EGG-2089, August 1981.
,, ) ~
? A-3.' ;V.'Anoda, ROSA-III-System Description for Fuel Assembly No. 4, JAERI-M_9363, February 1981.
A-4. A. :F? Ansari et a1. ,' FIBWR: A Steady-State Core Flow Distribution
' Code for' Boiling Water Reactors Code Verification and Qualification
--Report, EPRI NP-1923 Project 1754-1, -July 1981.
'A-5. Dresden Nuclear Power Station Units 2 and 3 Safety Analysis Report,
' Commonwealth _ Edison Co., Chicago, Illinois, November 1967.
- A-6. Dresden Nuclear Power-Station No. 3 Technical Specifications, Facility-Operating License DPR-25.
c- ! e-1
, A-51 i
L-.
g,_.... . . ._ 4 {- . V l t-i: 1 I APPENDIX B i 200% LARGE-BREAK LOCA I y I l i, h-l i-l 4 l- t i I i [' i ) i l B-1 i )
APPENDIX B 200% LARGE-BREAK LOCA
.This-appendix describes the results of the TRAC-BD1 calculation for a
' hypothetical 200% large-break loss-of-coolant accident (LOCA). The accident analyzed was assumed to be initiated by a 200% double-ended offset shear break ~in the rec'irculation pump suction piping. The transient is -
. defined more completely in:.ection'2 of the main body of this report. The
~
-calculation was performed with.the two-dimensional detailed model of Dre'sden 3 which is' described in Section 4 of the main body of this report.
The:results or.the calculation are documented in Figures B-1 through B-49.
- Some of the figures are shown at the-request of the Nuclear Regulatory Commission in_ order to audit specific results of the LOCA calculations performed by the Exxon-Nuclear Company.
Table B-1 lists the sequence t.f significant events occurring in the calculation. The' table illustrates the trips assumed in the calculation.
-The-times.of significant-thermal-hydraulic events, such as. lower plenum
' flashing, are also presented. Times at which the model and/or code options were. changed are also'noted. ,
l The calculated core power lvs time is shown in Figure B-1. The power began ~ decreasing at the start of the transient due to the pump trip which decreased the core flow and increased the void fraction causing a negative ;
. reactivity-insertion into the core. The scram signal was generated at I 0.5 s, due to a 10% closure of the main steam isolation valve (MSIV).
Control rod insertion.was begun at 0.77 s and the-rods were fully inserted by 4.5,s. The control rod insertion effectively shut down the nuclear
~
. reaction reducing the core power to decay heat levels. A conservative l multiplier of 1.2 was applied to the decay heat calculation. The core -
power decayed slowly for the remainder of the calculation. Calculated upper plenum pressure vs time is shown in Figure B-2. The .
. loss of . fluid out the breaks caused the calculated pressure to decrease
.during the first 3 s of blowdown. A slight pressure increase was B-2
s TABLE B-1. SEQUENCE OF EVENTS Time-(s) Event 0.0 Break opened; loss of offsite power 0.5 Scram signal generated; feedwater terminated
'?
1.3 Low water level signala 3.0 CRD flow terminated 2.2 Low-low water level signala 4.5 Control roos . fully inserted 5.0- MSIVs closed 6.7 Jet pump suction uncoveredb 11.0 Recirculation lines uncoveredD 12 Lower plenum flashing 23.2 HPCI initiated 32.5 Intact loop isolation valve closed 37 Dryout at the peak power zone c 44.8 LPCS initiated 54.4 Rated LPCS deliveredd 72 HPCI-isolated 72.5 HPCI system deleted; jet pumps renoded; intact loop pump speed set to zero 75 Lower plenum refill started
~
110 Backflow from containment to vessel 160 Reflood initiated 163 Fine mesh and water packing models turned on B-3 i
a 4 4
. TABLE;B-1. l(continued) i
-Time (s): Event 1 1175 Peak? cladding temperature obtained
; . ?l90 Calculation terminated
' Y --a. cThe low and low-low water level l signals correspond to collapsed down-l comer.; liquid levels of -12.80 and 11.28 m, respectively, above the bottom of . . theLlower plenum. Ib.1The times.at which the jet pump' suction and recirculation lines uncovered correspond to the times the collapsed ~downcomer liquid level dropped to' the elevation .of thefjet pump suction and the top of the reactor
. vessel. outlet nozzle, respectively.
- c. Dryoutl corresponds to the time the void fraction in the center of the-heated length of _the high-powered channel approached unity.
Ed. Rated LPCS corresponds to a flow ofL4500 gpm per pump which is the flow , required by;the Dresden 3 Technical Specifications. 1 1 1 i Lf d I B ^ L _
g .- l l L TOTPOW000001
.2500 -
m
.R -
2 2000' - v c '; *.
~t 3 g 3300 _
4 5 1000' o.
- 0.
Control rods fully _ inserted 500 - k 0 O 50 10 0 15 0 200 Time (s) Figure B-1. Core power. 8000 . , i P300901 [ MSIVS clo::ed -1000 Q
- m o .
S 6000 - g
* - -800 v 0 e
' L 3 3-M g
-600
$ 4000~ m
, m 1 c.
- e . 400 ,
E- E
."O 2000' -
3
- o
~> - -200 >
O 0 O SO 10 0 15 0 200 Time (s) Figure B-2. Upper plenum pressure. B-5
-calculated between.3 and 7 s. This pressure increase resulted from closing
-the MSIV'while steam was being' generated in the core due to the removal of
~ decay. heat and stored energy from the fuel rods. The vessel began
;depressurizing-again at 7 s when the volumetric flow out the vessel-side break increased due~ to' increasing void fraction in the downcomer which
~
supplied the break. The increased void fraction in the downcomer was c'aused by numerical diffusion of-steam,'due to the code's assumption of - Lh omogeneous fluid within a' cell, as the code attempted to represent the uncovery of the recirculation line. A more realistic calculation of the uncovery of the recirculation line would have been calculated if the code had a mixture level mocel. With a mixture level model, the calculated uncovery of the recirculation line would have occurred at about 11 s; _hence, the depressurization of the. vessel should not have begun 'until about 11's. In addition, the calculated pressure curve was more rounded between 7 'and 11's than it should have been due to the lack of a mixture level model.
-The calculated mass flow rate.out the vessel-side break is shown in
' Figure 8-3. The mass flow rate rapidly increased after the break opened to values representative of subcooled critical flow. The break flow transitioned to two phase saturated flow at 4.2 s when.the void fraction in the cell just upstream of the~ break reached 10%. The break mass flow rate decreased sharply, although the volumetric flow increased sharply, when the recirculation line uncovered at= 11 s and allowed nearly dry steam to exit the break. Flashing.in the lower plenum and bypass, which started near 12 s, caused liquid to flow into the downcomer and out the break and was
' responsible for the increase in break mass flow near 15 s. The mass flow l rate then decreased until 65 > when the flow unchoked. Reactor pressure .
dropped below containment pressure at 110 s causing back flow of steam from the containment through the break and into the vessel. - The calculated mass flow out the pump-side break is shown in Figure.B-4. The flow rate out the pump-side break was limited by the small
. area of the jet pump driveline nozzles and hence was significantly smaller than the flow through the vessel-side break shown in Figure B-3. Only B-6 ,
1 l
15000 i i i MFLOW190001 - 30000 m.
- g. 10000 M
~ m .
- 20000 N Recirculation' lines unc0vered v
;p v g 5000'.-- h ueD to flashing 10000 g_
m' g n- e o ^ a 2 o._
~ 0 2
* ' ' ' ~
-5000 0 50 10 0 15 0 200 Time (s)-
' Figure B-3. Moss flow rate out the vessel-side break.
-4000 i i i MFLOW200001 - 8000
. 6000 m S ue to pump-2000 ,
. 4. -
s - 3' O
. Y - 2000 3 i O
. . .*- y 0-- --
--0 y
.o a 2 2
. - -2000
-2000' O 50 10 0 15 0 200 Time (s)
Figure B-4. Moss flow rate out the pump-side break. B-7
E
. , e s-o .- , ' s.y* ,;.y. g: .E'- -
. q.
( .
. 1 2 fg , . 7'f
-w " - .. L . . ' & .7 ' , . . % "-
.- v.. . . . . ..
?.
~, , . . . ' . , .u . , . . , . .%..
, z- ..
. . .7 ..
_ . _ - ~
......,,,s.. , - . . . _ . .
about 20% of the total mass lost to the containment flowed through the pump-side break. The pump-side break flow transitioned to saturated two phase critical flow within 0.01 s of the opening of the breaks. The transition to saturated flow on the pump-side break occurred earlier than on the vessel-side break because of a large pressure drop associated with reverse flow through the recirculation pump. The large pressure drop across the pump caused the fluid just upstream of the pump-side break to - saturate earlier than on the vessel-side break. Apparent " discontinuities" in the break flow were calculated at 10, 19, and 27 s. These
" discontinuities" were caused by switching from one homologous head curve to another. Even though the homologous curves were continuous, the flow / head characteristics of the different curves varied sufficiently to alter the hydraulic resistance of the pump which changed the fluid conditions upstream of the break and affected the break flow. The break unchoked near 41 s. Backflow of steam from the containment occurred after 110 s.
The calculated high pressure coolant injection (HPCI) and low pressure core spray (LPCS) flow rates are shown in Figure B-5. The HPCI system was activated by a low-low downcomer level signal which occurred at 3.2 s. A 20 s delay was assumed prior to adding HPCI to the reactor to represent the time for the HPCI turbine to reach operating speed and the time for the admission valve to open. The HPCI flow remained constant until 48 s when the flow began decreasing to represent the isolation of the HPCI system which occurs at low reactor pressure. The HPCI system was completely isolated, with zero flow delivered to the reactor, after 72 s. The LPCS l pumps started 30 s after the low-low downcomer level signal was calculated l but flow was not delivered to the vessel until 44.8 s when the reactor . pressure dropped below the shutoff head of the pumps and the admission valves opened. The LPCS flow then increased with the decreasing reactor - pressure and by 54.4 s the rated flow required by the Dresden 3 Technical B-1 Specifications was delivered to the reactor. The total emergency core coolant flow exceeded the total break flow af ter 55 s. Two additional safety systems, the automatic depressurization system (ADS) and low pressure coolant injection (LPCI) were not activated during the calculation. ADS was not activated because the reactor pressure was too l B-8
e V 800 i i i Hi'C I O LPCE g c c O O -O -1300 m m 600 - J2I' 7 N o E
.x .o V
-1000 C f 400 3 o
~
m
- m I e I -500 a 2 200 -
4 2 0 , j
'T HPCI isolated
\ ' '
00 O kO 0 0 50 10 0 15 0 200
) Time (s)
Figure B-5. ' HPCI and ' LPCS mass flow rates. ' ( 2000 i , ,
- 4000 MFLOW150005 1000 _- -
2000 m n m
" N
).x 0--- - - -
--O
-Q g
v - v
-- -2000 g
{ -1000 ~- o i := m -
- -4000 w -2000 - - m O m 3 i O
1 2
- -6000
-3000 - -
-8000
-4000 O 50 10 0 15 0 200 Time (s)
Figure B-6. Moss fl ow rate through the broken loop Jet pump driveline nozzle. B-9
% * -Q .. . , :; - , .: .<
._f .g ' . _. ..: , : '
' ) : . , . . ;$h~ :,,. ,., [; ;_,. n, ~p;_
- n. g. .
-( . -
=
. gh . .
, y ,.: .i w
+ . :
~ . _ ey . e l , . . . :. v .y
;..v.
.- e;
y W a 3 ph, .y
- low to overcome the relief valve spring tension when the ADS time delay was i exceeded. The LPCI admission valve to the intact recirculation loop was.
7 u-
- assumed to fail closed which prevented any LPCI from reaching'the reactor
% vessel. j CalculatW iuss flow rates in the broken and intact loop jet pumps are
~
fshown in Figures' B-6 through B-11. Mass flow rates-are presented for the -
;jdriveline nozzle,' discharge,'and suction of'each jet pump. Positive flow j Jin each figure corresponds to flow in the normal, steady-state direction.
'L: The, flow through;the broken loop-jet pump driveline nozzle'is shown in Figure B-6 and, as expected, is similar except for sign to the flow out the
~
I pump-side t reak which 'was shown in Figure B-4. .The' mass flow through the discharoe of.the broken loop-jet pump, shown in Figure B-7, is also similar to the_ flow through the driveline nozzle during blowdown which indicates h ,thIt.the primary flow path to the pump-side break was from the lower plenum No'.k,hthebrokenfloop'jetpump.'Theflowatthejetpumpofscharge
-b#c'a' m b oscillatory shortly after lower plenum refill began at 75 s. The
; flow at.the broken loop -jet pump suction is shown in Figure B-8. The mass fiow rate was generally qear zero between 50 and 160 s. However, the flow cagnitude increased %ortly after 160 s. At the end of the calculation, thefpwlfrombothbrokenandintactloopjetpumpstothedowncomer ry-nded to 83% of the LPCS flow'. The-large mass flow from the jet corresM.x pumps?at the end of the calculation was thought to be unrealistically caused by condensation in the upper plenum which will be discussed more
' fully later.
The mass flow rate through th'e driveline nozzle of the intact loop jet 'v. (pump:is shown in Figure B-9. The coastdown of the intact recirculation
~
I' . pump caused the ficw to coastdown until 11 s when flashing in the rel:ircul'ation:loopicaused a flow surge into the jet pump. The flow rate - tthen decreased ~ and remained near zero after. the recirculation pump 1-discharge valve was closed at 32.5 s. The mass flow. rate at the discharge m.y
\
' C of-the intact loop jet pump'is shown in Figure B-10. The flow coasted down Ywiththerecirculationpumpearlyinthetransient. Flow oscillations simiiar to those in the broken loop jet pump were calculated during the refill portion of the transient. The mass ~ flow rate at the suction of the
, W!
(. V B- k
2000 , , '
- 4000 MFLOW150004 1000 .- ~- 2000 7 A
\
\ 0- -
I E
.- cn
[e
, i 4 ! }I o 6
- - _.- -2000
$ -1000 - - o
~ '~
- m g -2000 Start of refill 2 a
~
- -6000
-3000 - _
- -8000
-4000 O $0 10 0 15 0 200 Time (s)
Figure B-7. Moss flow rate through the broken loop let pump discharge. 2000 , , '
- 4000 MFLOW150001 1000,- ~- 2000 7 -
7 N N o__ "r' --O v _. v D Flow incrense _. 2000 0 -1000 '- 3
*- - -4000
$ -2000 ' - _ m 3 a
- -6000
-3000 - _
* - -8000
-4000 -
0 50 10 0 150 200 Time (s) Figure B-8. Mass flow rate through the broken loop jet pump suc tion. B-11
3000 ' ' Low 050005 - 6000 m 2000 - 7 o .
- 4000 N N E y
v
.a .
v
.g ' 1000 _- -- 2000 3 - = e
- a
$ 3 - = --0 $
Due to flashing
. , , , - -2000 0- 50 100 15 0 200 Time (s)
Figure B-9. Moss flow rate through the intact loop let pump driveline nozzle. 7500 , i i MFLOWO50004 - 15000 l n 5000 7 m - 10000 N N F E i v G g 2500
~
5000 g Start of refill C
'N : -
0-- vi= .7,= h ffh k A j,A
--0 $
, , , - -5000 0 50 100 15 0 200 Time (s)
Figure B-10. Moss flow rate through the intact loop let pump discharge. B-12 i
.. s Lintact loop jet pump is shown in Figure B-11. Flashing in the recirculation loop at' 11 s caused a surge of flow into the downcomer. The flow then decayed and remained near zero until near the end of the calculation when flow-into the downcomer. increased similarly to that i . . previously shown for. the broken loop jet pumps. The flow increase represer.ted.entrainment of LPCS liquid from the lower plenum, through the
< jet pump and.into the downcomer. As was the case for the broken loop, the il increased flow into the downcomer was not thought to be realistic.
The calculated downcomer liquid level is shown in Figure B-12. The level corresponds to a collapsed liquid level computed from the void Tfractions of the downcomer cells. The liquid level decreased rapidly due to loss of liquid out the breaks until 14 s when a small increase in level was calculated. The small increase in-level was caused by flashing in the
. lower plenum, bypass, guide tubes, and intact recirculation loop which forced liquid into the downcomer. The primary flow path into the downcomer for this liquid was through the separators while a secondary flow path was through the intact l loop jet pump. The downcomer level peaked at 18 s and then decreased until the downcomer was essentially voided at 40 s. The calculated downcomer voiding may not be realistic because a small mixture level, less then 0.7 m, could physically be maintainad between the bottom of the reactor outlet nozzle and -the bottom of the downcomer. The code had no choice but to calculate complete voiding of the downcomer because TRAC-BDI Version'12 does not have an explicit mixture level model and because the ' region between the reactor outlet nozzle and the bottom of the downcomer was coarsely noded with only one axial level. This potential limitation of the calculated results was not thought to be of major 7.., importance. The downcomer remained voided until 175 s when the liquid
' level: began increasing due to liquid entrainment from the lower plenum. A i- subsequent discussion will indicate that the calculated entrainment from the-lower. plenum was not thought to be reasonable. The total fluid mass in the downcomer, including the mass of liquid and steam, is shown in
' Figure B-13. The trends of the downcomer fluid mass were similar to the trends in liquid level shown in Figure B-12.
B-13
4 7500 . . . MFLOWO50001 - 15000
~
m 9 _
- 10000 k N E o .o -
6 I C g 2500_- ~
- 5000 g
.. Due to flashing a N E 2 o__ --0 2 Flow increase
, , ,- - -5000 0 50 10 0 15 0 200 Time (s)
' F i gu r e B-11. Mass flow rate through the infect loop jet pump suction.
15 a i LLEV300001
-40
}v 10 _
~
-30 v 0 8 c
.@ 0
-20 %
3 5 - - C
-10 *'
Due to flashing 0 O O 50 10 0 15 0 200 Time (s) Figure B-12. . Downcomer liqu.d level. i B-14
r 100000 . i i DCMASS
- 2.0 0'10' 75000 - -
- 1.5 0*10' e
. ^-
O E
-p., .x' 4_
50000 N - 1.0 0*10' U E 2 O 2 25000 -' - 5.0 0*10' 0 O.00 O ~50 10 0 15 0 200 Time (s) Figure B-13. Downcomer fluid moss.
'60000 . . . ~
LPMASS.
- 1.2 0*10' ower plenum flashing
- 1.0 0*10'
'40000 - -
m
^
CD E
- 8.0 0*10 o 6 Refill arrested -
a 'N -
- 6.00*10' $
U
'3 O y
20000 - -
- 4.0 0*10 F
- 2.0 0*10' Start of refill 0 O.00 0 50 10 0 15 0 200
., Time (s)
Figure B-14. Lower plenum fluid moss. B-15
The total fluid mass in the lower plenum'is shown in Figure B-14. The. lower plenum mass decreased rapidly at 12 s when the lower plenum began flashing. .The lower plenum mass ~ continued to decrease until the minimum
' inventory was calculated at 75 s. The lower plenum then refilled at a nearly constant rate, corresponding to about 60% of the LPCS flow, until J160 s. The calculated lower plenum refill was arrested at 160 s when the fluid' mass first leveled off,. then decreased, and then increased sharply. -
The calculated refill behavior was thought to be unrealistic after 160 s.
-The perceived unrealistic behavior after 160 s appeared to be related to condensation of steam in the upper plenum due to LPCS. Prior to 160 s, the differential pressure between the lower and upper plena was primarily due to gravitational head. After 160 s, the differential pressure began exceeding the value due to gravitational head. The increased differential pressure after 160 s. forced liquid out of the lower plenum, arresting lower
. plenum refill, and into the core and downcomer. The increased differential pressure between the lower _and upper plena was apparently caused by LPCS-induced condensation in the_ upper plenum. A sensitivity calculation was performed to verify that condensation-related effects were responsible for the perceived unrealistic lower plenum. refill results after 160 s. The sensitivity calculation,-which was restarted from the ba'se case calculation at 163 s, represented warm (375 K) LPCS rather than the cold (308 K) LPCS represented in the base case. In the sensitivity calculation, the lower plenum refill rate after 160 's was similar to the base case rate prior to 1601s. The sensivity calculation indicated that the lower plenum would be completely-full at 195 s which is consistent with the refill time obtained
~
by extrapolating the prior-to-160 s refill rate shown in Figure B-14. In addition, the'_ increase in mass flow from the jet pumps to the downcomer near the end of the calculation, shown in Figures B-8 and B-11, and-the . subsequent increase in downcomer liquid level, shown in Figure B-12, did
.not occur in the sensitivity calculation. -
Comparisons between experimental data and calculations made in the
- independent assessment of TRAC-BD1 showed that the code consistently
- underpredicted the amount of fluid mass in the lower plenum at the time of minimum lower plenum inventory. An evaluation of the results presented in I
l 4 B-16 l
^
w Ref'erences B-2,18-3, and B-4 showed that the minimum lower plenum fluid mass for each calculation was about half of the corresponding data for both large-break and small-break experiments. The cause of the underprediction
~
of minimum lower plenum. inventory is not known. An extrapolation of the L results of the' assessment. calculations yields that the minimum calculated lower plenum fluid mass shown in Figure B-14, 11,000 kg, would also be a
. factor of two low. Based on the assessment calculations performed to date, I
an improved estimate of the minimum lower plenum fluid mass in a 200*4 break
~
in Dresden 3 would be 2 x 11,000 = 22,000 kg. Based on a minimum mass of 22,000 kg, lower plenum refill would have been completed at about 170 s instead of the 195 s estimated for the base case. The calculated guide. tube fl'uid mass is shown in Figure B-15. The I fluid mass represents the combined fluid inventory of the three guide tube components in the TRAC-BD1 model. The mass remained nearly constant until 13 s'when the guide tube fluid began. flashing. The mass decreased until 60 s when liquid from the bypass first. started dropping through the core
, plate into the guide. tubes. The fluid mass then increased irregularly until.150 s when the guide tubes were liquid full. The guide tubes remaine'd-full until 185 s when some liquid was calculated to slosh back across the core plate into the bypass. This sloshing was representative of oscillations throughout the system and may have been an unrealistic consequence of the-previously discussed problems with LPCS-induced
-condensation.
'The ' total fluid mass in the bypass is shown in Figure B-16. The mass decreased rapidly when.the fluid in the bypass began flashing. The mass
. began increasing a few seconds after LPCS was initiated as ECC began flowing from the upper plenum into the bypass. The bypass mass tended to increase, although irregularly, for the remainder of the calculation. The fact that the bypass mass did not increase smoothly indicated that the calculated. refill process was relatively violent with fluid sloshing-back and forth between components.
The calculated fluid mass in the upper plenum is shown in Figure B-17. .The mass initially decreased and then increased in response B-17 i
ti 140000- i '. '
-MGT
~ Guide tubes full -80000 N q 30000
~
Flashing g-
. 60000 m g -
9 e 6 v 20000 n ) 83 -40000 $
.j O
2 10000
~
, -20000 O
O 50 10 0 15 0 200
. Time (s)
Figure B-15. Guide - tube fluid mass. 30000 i i ' BYMASS
. -60000 f
-50000 20000 g- .
1
-40000 j.o U ,
O ch - -30000 $
$ n 1
2 - 10000 -
/ -20000 sN .
-10000 k -LPCS initiated i i i 0 0 0 50 10 0 15 0 200 Time (s)
F i gu r e B-16. Bypass fluid mass. B-18
w- ' Lto the pump coastdown and the reactor scram.' The mass increased sharply near-10 s when the-fluid'at the top of the bypass began flashing which
'l pushed' liquid into the_ upper plenum. The mass increased until the steam
~
~ velocity could no longer-' support the liquid which then drained into the F.
^
. bypass and the' low powered channels. The upper plenum was almost g
completely. voided at 44.8 s when LPCS was initiated. The mass in the upper
. plenum then increased because the LPCS flow rate into the upper plenum 3"fg
~
- O exceeded the-flow rate _from the upper plenum to the bypass allowed by~the
~~
_ countercurrent _ flow-limiting'(CCFL) model. .The upper plenum mass increased until about 75 s when CCFL breakdown began to' occur at the upper tie plate
~1n the. low powered channels and at the upper core grid in the bypass. The LCCFL' breakdown caused upper plenum liquid to sporadically dump into the
-bypass-_and low powered channels. The upper plenum was again nearly voided
, at 110's. The upper plenum remained nearly voided, with LPCS flowing
--through the upper plenum into the low powered channels and the bypass,
.until near~the end of the calculation.
The c: ore fluid mass,.which includes the' mass of the high powered, average powered, and low powered fuel channels, is shown in Figure B-18. The core fluid mass, which decreased, then increased, and then decreased again,in the first 12 s.of.the~ blowdown, was affected by the combination of
~
core flow'and power which in turn were influenced by the pump trip and the reactor scram. Lower plenum flashing at 12's. caused a surge of flow into I' the core and.a momentary-increase in core fluid mass. The core fluid mass I
~
_ then decreased,Ldue to draining and boiling away of the liquid inventory, until 75 s when the mass began increasing due to CCFL breakdown at the upper; tie plate in the low powered channels. The core fluid mass tended to
.. . increase' for the remainder of the calculation. The " spikes" in fluid mass . .after 75 s were caused by sporadic dumping of LPCS into the low powered or 3 average powered fuel channels.
The calculated mass flow rates at the inlet and outlet of the
-low powered channels are shown in Figures B-19 and B-20, respectively. The
. channel inlet corresponds to the side-entry orifice while the outlet
. corresponds to the up'per tie-plate. The mass flow rate at the inlet and outlet decreased following the pump trip at 0 s. Lower plenum flashing, B-19
I: } 15000 , i i h- UPMASS
-30000
-25000 tart of CCFL breakdown _
^ -
E l 5
-20000 o a f
-15000 $
{ 0
.2 2 5000 -
g -10000 k
\ -5000 0
HPCS initiat d dkm,= 0
-0 50 10 0 15 0 200 Time (s)
Fi gu r e B-17. Upper plenum fluid mass. I 20000 , , , I 40000 15000 - -
-30000
$ j D 10000 -
Start of CCFL breakdown _ O
} -20000 $
2 S . 9 5000 - -
-10000 .
)
0 O O 50 10 0 15 0 200 Time (s) Fi gu r e B-18. Core fluid mass. B-20
s 4000 , , ,
~
MFLOW450001 - 8000 3000 - -
- 6000 m m
7 \
\ 2000 - -
j -
- 4000 f_
4 v
$ 1000. -
-- 2000 3:
C O j -- y3y spWf -- j
-1000 - j -- -2000 Lower plenum flashing
' ' , - -4000
-2000 0 50 10 0 15 0 200 Time (s)
Fi gu r e B-19. Moss flow rote at the inlet of the low-powered channeI. 5000 -
-- MFLOW450010 f 10000 I - 5000 7 l ,
7
) 0-- -
k . . LW Q; __[ LJ b --o 6 - ,,9,1 , t o a
~ 2
- -5000
$ Start of CCFL ,
y c breakdown o
- -10000 m 0
_5000 ' - _
- n 2 0
- -15000
- -20000
-10000 O 50 10 0 15 0 200 Time (s)
Figure 8-20. Moss flow rate of the outlet of the l ow-power ed channel. B-21
which started'at.12 s, caused a surge of flow into the core. The mass flow then' increased until all the fluid in the lower plenum was flashing near
'14 s. The mass flow at the inlet and outlet then decreased and remained near.zero u'ntil 75 s when CCFL breakdown began to occur at the upper tie
-(plate'in the low powered channels. CCFL breakdown, which was evidenced by large. downward flow spikes shown in Figure B-20, allowed LpCS to sporadically drop.down through the low powered channels into the lower
' plenum and was thus responsible for the start of lower plenum refill shown
.in Figure B-14. The primary' LPCS flow path to the -lower plenum during
' refill was through the low powered channels; 60*4 of the LPCS reaching the lower plenum between 75 and 150 s. flowed from the. upper' plenum down through
.the low powered channels. Countercurrent flow, with steam flowing up and liquid flowing down,- was generally calculated at the upper tie plate in the
~
Llow powered channels during. refill. The liquid downflow prior to 75 s was limited by the CCFL model; the liquid downflow after 75 s was generally not limited by the model. The liquid flow at the side-entry orifice was infrequently limited by the CCFL model. The calculated mass flow rates at the inlet and outlet of the av'erage powered fuel channels are.shown in Figures B-21 and B-22. The mass flow rates prio" to 75 s have similar characteristics to the flows in the low powered channels discussed previously. Countercurrent flow, with the CCFL model limiting the liquid downflow, was generally predicted at the upper tie' plate in the average powered channels after 75 s. However, the CCFL model let only:a small amount of liquid down through the upper tie plate. The steam mass flow rate up generally exceeded the liquid mass flow rate down. Countercurrent flow was also generally-calculated at the side-entry orifice of the average powered channels during refill. The . liqbid downflow at the side-entry orifice was generally not limited by the
.CCFL model~. Leakage from ~ the bypass was the_ primary source of liquid for -
the average powered channels during refill. Some of the liquid leaking into-the-channels was entrained upwards past the heated length where it was boiled away. However, most of the liquid leaking into the average powered channels ; fell downwards through the side-entry orifice into the lower plenum. The calculated flow rates in the average powered channels were quite " noisy," with large flow spikes, after 170 s as shown in B-2 2
c_ _ 10000 , , , MFLOW430001 - 20000
- - 15000 m n- m 5000 -
- .' j - - 10000 f v
l'
) - 5000 3 o -
O C i .' L IN 0--- - - - - - -- f k --0 j
*[f 2
- - -5000 lf
'~
~
-5000 0 50 10 0 15 0 200 Time (s)
F i gu r e B-21. Moss flow rate of the Inlet of the overage-powered channel. 10000 , , ,
~ - 0000 MFLOW430010 g _5000_
-- 10000
\
.N E m i d5 v e
,Ap( --0
{ -=
= gi i
y 0---
' c ?
2 -5000 ~ - 000 g I
'- - -20000
-10000 O 50 10 0 15 0 200 Time (s)
Figure B-22. Moss flow rate of the outlet of the overage-powered channel. B-23 l
Figure B-22. These large flow spikes were indicative of condensation-related= problems occurring in the calculation after 170 s and were probably not realistic.
-The calculated mass flow rates at the inlet and outlet of the high powered channels are shown in Figures B-23 and B-24, respectively.
The calculated flow at the upper tie plate, .shown in Figure B-24,
- alternated between cocurrent and countercurrent during refill.
Insignificant amounts of liquid downflow were calculated when the flow was countercurrent because the steam mass flow generally exceeded the liquid downflow. Countercurrent flow was generally calculated at the side-entry orifice between 75 and 160 s. The liquid downflow was generally not limited by the CCFL model. Leakage from the bypass was the primary source of-liquid for the high powered channels during refill. About half of the leakage from the bypass drained down through the side-entry orifice into the lower plenum; the remainder of the leakage was entrained up through the high powered channels. The mass flow rate in the high powered channels increased significantly near 160 s. This increase in flow was caused by an increased differential pressure between the lower and upper plenum which forced liquid into the high powered channel and thus effectively initiated reflood. The flow into the high powered channel after 160 s corresponded to' a very high reflood rate of about 0.2 m/s in the heated length. The
-reflood did not occur in the sensitivity calculation with warm LPCS water and thus appeared to be caused by condensation effects. In reality, reflood would probably not start until after the lower plenum refill was completed. The "best-estimate"' time for the completion of lower plenum refill was 170 s. This estimate was obtained by extrapolating the prior-to-160 s refill rate and accounting for the code's possible -
underprediction of minimum lower plenum inventory. Thus, by a fortuitous combination, the time reflood actually started in the calculation (160 s) - agreed well with the "best-estimate" reflood time (170 s). The calculated thermal-hydraulic response of the core is affected by the CCFL model used at the upper tie plate and side-entry orifice. The upper tie plate CCFL model affects the amount of LPCS which flows down into the fuel channels. The side-entry orifice CCFL model affects the B-24
r. 2000 , , , T MFLOW410001 - 4000
- - 3000 m
= m n 5 00 ' ~
E
- 2000 i s e v
~
- - 1000 ?
c .s
~'0
- 0 ~ "
wig J v 0}Yff 2 l Start of reflood
- -1000
~ ~~
-1000 0 50 10 0 15 0 200 Time (s)
Figure B-23. Moss flow rate et the inlet of the high-powered channel. 2000 , , , MFLOW410010 - 4000
- - 3000 m n a M %'
\08 1000 -
~ ~
- 2000 E 3 g C
R - 1000 *
, o -
C + l
$ o.- '
Q -& --o m M O
-3 0 2
- - -1000
'~
-1000 '
0 50 10 0 15 0 200 Time (s)
. Figure 3-24 Moss . flow rate of the outlet of the high-powered channel.
B-25
- i. -
. calculated liquid holdup within the channels. Relatively few refill calculations have been performed during the assessment of TRAC-801.
Consequently b the accuracy of the code's calculation of refill and CCFL behavior is not well- known which ' introduces uncertainty in the calculated
-refillJ foDresden 3. In particular, evidence exists that the TRAC-BD1 CCFL model, with default coefficients, may underpredict liquid downflow through
-the upper. tie plate.B-3 .
The collapsed liquid levels in the low powered, average powered, and high powered fuel: channels are presented.in Figures B-25, B-26, and B-27, respectively, to provide an indication of the relative liquid inventory
, etween b channels'. Each collapsed liquid level was computed based on the cell. void-frac'tions in the corresponding channel. The general trends of the collapsed li_ quid levels were similar to those of the total core mass shown in Figure B-18. As expected, the liquid level in the high powered channel;was' generally less than or equal to the level in the average powered channel. For example, the collapsed liquid level in the high powered channel dropped to the bottom of the heated length nearly 40 s earlier than~in the average powered channel. Similarly, the level in the low powered channel was generally less:than the level in the average powered _ channel. The sharp increase in level in the low powered channel at 75 s was. caused by the start of CCFL breakdown at the upper tie plate-which allowed: liquid from the upper plenum to dump into the
- low powered channel. The dumping from the upper plenum almost filled the low powered channel with liquid two times during the period when the upper
'p.lenum was draining (75-to 110 s). The low powered channel filled with
. liquid near-the end of the calculation.
Th~e fluid velocities at the vessel -faces representing the upper core grid, which separates the bypass and upper plenum, are shown in - EFigures B-28, B-29, and B-30. The three figures show fluid velocities for g :the three vessel rings inside the core shroud. Fluid velocities are shown in lieu.of mass flow rates which are not available as plot parameters for
'the: TRAC-BD1 VESSEL component. The third vessel ring, the outermost ring cinside.the' core shroud,-contains the low powered channels. The second vessel iing containsLthe average powered fuel bundles. The first (inner) +
B-26 r 4 + - 4 , -. . - . , , , _ , ,,, , -m-
P' G i i i CHANLL45
-15 Channel ful L , 4
'*~'
^ 4 - _
O l C
.~.
.g -
-10 g
'O O n , m O'2 -
a 5 h I
.0
, ad of M Weadown 0 0 50 10 0 15 0 200 Time -(s) _
Figure B-25. Collapsed liquid level in the low-powered channel. 6 i , , CHANLL43
-15
^
4 _. - O O g' -
-10 g O' O Y : Y 82 -
3
-5 W r
-0 O O 50 10 0 15 0 200 Time (s)
Figure 9-26. Collapsed liquid level in the overage-powered channel. B-27
6 , , , l CHANLL41
-15 y4 -
Top of channel A - p
, Top of heate ,
g - length -10 g , D 0 m m a 2 - - c
-5 Bottom of heated length 0 O O 50 10 0 15 0 200 Time (s)
Figure B-27. Collapsed liquid level in the high-powered channel. 50 , , ;
- - 15 0 LIQUID 40 - ~
O VAPOR g 30-- -- 10 0 g N N E 20 - - v - h, - 50 v
>. 10 - -
-0 h f,I O
g
$ -10 - I c
- - -50 y y
-20 -
Start of CCFL breakdown O 3
- u. -3 0 -- -- -10 0 L. -
-40 -
- - -150
-50 O 50 10 0 15 0 200 Time (s)
Figure 8-28. Fluid veloci11es at the upper core grid (third vessel ring). B -28
d ' F E - $2 x vENg F Egy $2oOx vENg i - - - - - i - - - . g 5 4 3 2 t 2 3 4 5 g 5 4 3 2 1 1 2 3 4 5 u r 0 0 0 0 m 0 o 0 0 0 0 u 0 0 0 0 0 o 0 _0 0 0 0 O - r O - -
'n - '
e
- 0 -
~
e ' 8 B 0
- O -
3 2 0 9 VL VL AI AI PQ PQ rF i OU rF i OU n u l RI n ul
- RI g i D g D
)i -
)d
. 5 0 '
~
. d 5 0 ' i v v e
l e l o o c i c i t ti i e ) j e L, s l s B a T o T i i t f m0 1 2 .t t m10 1 9 he0 ' fC
,. l h e.0 '
e e pu( )s pu( )s. p ) p g d e e Ih r 1 r (lr\ J c c o - o ' r r { e 1 l
@,i s e 1 5 5 g 0 ' , g 0 '
r i r i s d h d .. f jI!jf0 ( V, ( l, s s f i r l e c h l - I N , t " o , n v lI{ d e v s 2 2 s 0 e 0 0 - - - - - - - - ~ s 0 - - - - - - ~ el - - - - - - ~ s e
- - - . _ - ~
- - - 0 5 1 0
l
- - - o s 1 0
1 5 1 0 5 0 0 1 5 1 0 5 o 0 0 0 0 0 0 0 E y$2 xv:Ng E g y $ 2 o.t x v :Ng
y g - - - - '
/^
vessel ring contains~the high powered fuel channels. Figure B-28 shows
'that larg'e negative-liquid velocity spikes, indicative of CCFL breakdown, began at 80 s in the third vessel _ ring. The beginning of CCFL breakdown
' allowed liquid to dump from the upper-plenum into the bypass. The third vessel ring was-the preferred LPCS flow path.between the upper plenum and bypass. Only a relativelyismall fraction of the LPCS flowed down th-ough~
'the inner two. rings. The upward steam velocity tended to decrease and the -
1 downward. liquid velocity tended to increase as.the radial distance from the vessel' centerline increased. Multidimensional flows were thus calculated
'during refill. The multidimensional flows were primarily caused by the
~
, distribution of flow area in the bypass. -The inner tv'.,.ings had a
~
'relatively small ana while the third vessel ring had a relatively large flow area due to a '0.3 m space between the outermost fuel channel and the shroud.
, Three TEE components were used to represent the distribution of guide
. tubes.within the lower plenum. The calculated mass flow rates between the
.three.modeled guide tubes and the corresponding cells in the bypass are-shown in Figures B-31 through B-33. A flow path from the lower plenum
.through each guide tube TEE to the bypass was used to represent the leakage across the core plate and the leakage into the guide tubes occurring in the plant. .The mass flow rates were relatively low until 13 s when the fluid near the top of the guide tubes began flashing which caused a surge of flow into the bypass. A second flow surge was calculated near 18 s when the f.luid in:the lowest guide tube cell-began flashing. The different flashing times. wi_ thin the guide tubes ware a consequence of modeling the injection of cold control rod drive (CR , fluid into the bottom of the guide tubes.
-The mass _ flow generally remained from_the guide tubes to the bypass until 50 s. _The guide tubes began refilling rapidly but irregularly at 75 s.
Oscillations in the mass flow rates after 75 s were indicative of liquid - sloshing back and~forth between the guide tubes and bypass in the calculation. The guide tubes filled with liauid near 150 s. The flow rates'after 150 s were generally due to liquid leaking from the bypass through the guide _ tubes and into the lower plenum. B-30 e
'2500 ,
- 5000
MFLOW550004
^
N- " --0
'O-
- c _-
=*'
$ Due'to flashing *
- 2
- - 000 ,
E --2500 ~ - S 1
- - -10000
-5000 O 50 10 0 15 0 200 Time (s)
Fi gu r e B-31. Mass flow rate at the inlet to the bypass (third vessel ring). 4000 , , , MFLOW530004 1 I - 5000 m 2000 7
\
$m l ( .o E
.x
" O g 0-L gg !
I 4
- I ! l --0 s
y e o ! l ,\ '
- g. Due to flashing -
o m
\
1 -2000 -
) -
. -5000
-4000 O 50 10 0 15 0 200 Time (s)
Figure B-32. Mass flow rate at the inlet to the bypass (second vessel ring). B-31
1000 ' i ' _ - 2000 MF1.OW510004 m 500,- q 1000 k E -
! n e
v
, c g 0-L h - - N I J _ _ _
--0 y -
',9 c N Due to flashing 6
W,
] l j
2 E n
- -1000 2 -500 ~ - -
l
- -2000
-1000 O 50 10 0 15 0 200 Time (s)
Fi gu r e 9-33. Mass flow rate at the inlet to the bypass (first vessel ri ng). 400 , , . MFLEAK450001 - 800 300 - -
- 600 m n m
" N
) 200 _ 400 E 6 '
e v j 100.- i
-- 200 3
= l 2 l\ 1 I
-100 ~
p f yn$ , k
-- -200
~
' ' ' ' ~400
-200 0 50 10 0 15 0 200 Time (s)
Figure B-34. Leckcge from the low-powered f uel channel. 1
)
B-32
p g . t i L F , c_ 1The. calculated leakage mass flow rates from the low powered, s A average powered,'and high powered' fuel channels are.shown_in Figures B-34,_ i; .. 4B-35, and B-36, respectively; The figures represent the flow from the {' - channels to the bypass including the. flow'due to the holes drilled in the lower tie plates. The leakage into the channels was modeled between the
- side-entry orifice'and the bottom of the heated length. The operation of M the~ recirculation pumps-caused steady-state. leakage from the channels to
!, :the bypass. .The; flow rapidly reversed'after the pump trip at 0. s. The
*~
Lleakage was. generally from the bypass.to.the channels for the remainder of
'the~ calculation. : Leakage was the primary source of liquid for the i
average-rowered and.high powered channels during refill. The leakage flow p also contritiuted to the refill of the lower plenum as most of the leakage dropped down through .the- side-entry orifice into the lower plenum. The channel _ leakage flow paths were' responsible for'about 30% of the LPCS which reached.the lower, plenum during refill. The thermal response of the low powered, average powered, and Thigh powered fuel; channels are illustrated in Figures B-37, B-38, and B-39,
~
respectively. Each figure shows temperatures of fusl rod cladding, steam, liquid, and the Inner. wall of the' channel box at the' center of the heated _ length of.one~of 'he t channels. 'The center of the heated-length corresponds to the locat. ion of the maximum axial power peaking factor in each channel. The_ cladding temperature presented for.the low powered and avera~ge powered channels _ corresponds to the surface temperature of an average powered fuel trod within-the channel. The averaging process excluded the water rods
-which.were modeled as a separate rod group. The cladding temperature ' L presented for the high powered channel re' presented the surface temperature j _
of the maximum powered rod within the core. j C' ~ The cladding surface temperature in the low powered channel, shown in Figure B-37, followed the fluid temperature until 36 s when critical heat
' flux (CHF)-was calculated:and the fuel rod. began heating up. The maximum
~
post-CHF cladding temperature was 511 K and occurred at 50 s. The flow of LPCS down through the upper tie plate was sufficient to turn the cladding
. temperature around even though the liquid downflow was limited by the CCFL
~ mods 1. The fuel rods were effectively quonched at 76 s when the cladding B-33
500 ' s i _ - 1000 MFLEAK430001 400 - 300 -
^ - - 500 "
N (m200 E ' _ 5 , 10 0 '- S
- 2. v 0-- --0 y .
[ O
- -10 0 Is !
I n j.-200, i l -- -500 20 J
-300 -
-400 -
- - -1000 )
' ' ' l
-500 0- 50 10 0 15 0 200 Time (s)
Figure B-35. Leckage from the overage-powered f uel channel. 100 , . .
~ - 00 MFLEAK410001
.n 50 _- -- 10 0 5
--0
{- 0- - g 3 ( l h -50 ~ -- -10 0 1
- - -200
-10 0 O 50 10 0 . 15 0 200 Time (s)
Figure B-36. Leckage from the high-powered fuel channel. 7 B-34
~
800 ' i i _ -600 X CLADDING O VAPOR IL - v A Lioulo 550 O CHANNEL WALL ggp n -500 g b' v G
- 500. - -
a. 3
. . -400 2 O 0 u u
- 450 - _ o Q-
[ E d '-
-300 ,_
400 - _ - f _ - f f +~-M
. -200
-, i 350
.o 50 10 0 15 0 200 Time (s)
Figure B-37. Thermal response of the lo*-powered fuel channel. 600 i i ' _ -600 v X CLADDING 0 VAPOR t CHF
;; A LlOUID 550 - ,
O CHANNEL WALL f n . [ -500 g v u 500 ~ u h 3 400 5 o , 0
' u e -- _ o O 450 E
; / E t3 300 0 H
400 - A d - kI 1 3- 3 , k7 -Id
-200 i '
350 . 10 0 15 0 200 Time (s) Figure B-38. Thermal response of the overage-powered IueI channeI. B-35
t'emperature dropped to within 5.K of'the saturation temperature. The cladding remained near. saturation temperature for the remainder of the calculation. The liquid temperature closely followed the saturation temper'ature for the entire transient. The steam temperature also generally followed the saturation temperature although as much as 20 K superheat was calculated when the fuel rods were in' transition boiling. Even though steam superheat was calculated,othe thermodynamic quality never reached -
'+ unity and.some liquid, although at times less than 1% by volume, always remained at-the center of the heated length. The inner wall of the channel
' box followed the saturation temperature closely for the entire calculation. During refill, the wall temperature was generally within 1 K of saturation temperature. .Due to the low cladding temperatures, radiative
.' heat flux from the fuel rods to the channel wall was only a few percent of the convective heat flux-from the rods.
The thermal response of the average powered channel is illustrated by Figure B-38. The response of the average powered channel was similar to the'. low powered channel described previously except that the cladding-
' temperature was generally higher due to the radial power profile. The fuel rods began heating up at 43 s when CHF was calculated and the transition boiling regime was entered. The' fuel rods heated up until 94 s when a maximum cladding' temperature of 589 K was calculated. The turnaround in
~ cladding. temperature:at 94 s was apparently caused by a small amount of LPCS liquid which leaked into the bottom of the channel and was then entrained up through the channel. All the axial elevations in the
-average powered channel were quenched by 165 s. The liquid temperature at the center of the average powered channel was within 3 K of the saturation temperature.for the entire calculation. The steam temperature also .
followed the saturation temperature except between 43 and 165 s when the fuel. rods were in transition boiling and steam superheat was calculated. - '
'Although the steam superheat reached 70 K, some liquid was always present
.and.the equilibrium thermodynamic quality never reached unity. The inner j wall of the channel box was generally within 3 K of the saturation temperature. The radiative heat flux from the fuel rods to the channel wall'was generally less than 10% of'the convective heat flux from the rods.
1 B-36
.,- 4 4
The calculated temperature response nf fuel rods during transition ' boiling is sensitive to small changes in void fraction. For example, the turnaround in cladding temperature at 94 s in thcJr ; rage powered channel was caused by a drop in local void fraction from cbove 0.999 to 0.98. Since it is difficult for any computer code to accura ely resolve differences in void fractions between 0.98 and 1.0, considerable - uncertainty exists in the calculated turnaround time of S4 s. The
; uncertainty in the calculated turnaround time for the average powered channel is estimated to be 50 s which could cause an uncertainty in maximum cladding temperature of about 100 K. i The thermal response of the high powered channel is illustrated by Figure B-39. The fuel rod began heating up at 25 s when CHF occurred. A rewet was calculated at 36 s and a final dryout was calculated at 37 s.
The cle; ding temperature then increased until 175 s when a peak cladding temperature of 1021 K was calculated. The turnaround in cladding temperature at 175 s was caused by the high mass ficw rate, shown in Figure B-23, associated with the start of reflood. The cladding temperature then decreased at about 1 K/s until the calculation was terminated at 190 s. The calculation was terminated at 190 s due to a code execution failure which occurred when the low powered channel filled with liquid. Since the peak cladding temperature had been obtained, the calculation was terminated. Experience indicated that it would be difficult and expensive to extend the calculation. The cladding temperature turnaround at 175 s was.a consequence of using the fine mesh (reflood) model which was turned on at 163 s. One fine
, mesh node was used per cell in the high powered channel. The fine mesh model contains a quench front velocity model which represent.s heat-transfer . effects due to axial conduction in the cladding near the quench front and precursory cooling above the quench front. The improved heat transfer due to modeling axial conduction and precursory cooling was responsible for the calculated temperature turnaround at s. A code update, which improved the criterion for quench front e '^ :c t r, was required in order to obtain reasonable results with the fit + c3 i Code development personnel B-37
,,e
)
'i 1200 , , ,
X CLA0 DING Turnaround
- -1500 0 VAPOR 1000 - #~
.O C ANNEL WALL -
m n M -p. v v
- 800--
--1000 m -
e b . 3 -,- O g . 6 (, -500 400- - O ; ; ; ; 3 g-
- .o 200 O 50 10 0 15 0 200 Time (s)
F i gu r e B-39. Thermal response of the high-powered fuel chcnnel. 600 , , ,
-600 X CELL 1 (BOTTOM)
O CELL 2 A CELL 3 _ 550 _ O CELL 4 (CENTER) g -
+ CELL 5 -500 *C v 0 CELL 6 e _
8 CELL 7 (TOP) , , _' Soo s t 3 \ 3 g - -400 g
' u 1450 g i E ',. E
- -300 0 400 -
O O O O_ ;m
- -200 350 O 50 10 0 15 0 200 Time (s)
Figure B-40. Axial clodding temperature distribution for the low-powered chonne!. B-38
r_ 1
-plan to. incorporate this update in future versions of the code. The fine mesh model will eventually be replaced by a more sophisticated moving mesh reflood model.
A sensitivity calculation was performed in which the fine mesh model was not used; the reflood heat transfer was calculated with the normal e TRAC-BD1 heat-transfer package which had been used prior to 163 s in the i base case. In this calculation, the cladding temperature at the center of
~
the heated length in the hot channel did not turnaround at 175 s but continued to increase until the end of the calculation. The cladding _ temperature increase in the sensitivity calculation was not thought to be reasonable because the calculated heat-transfer coefficients were thought to be too low based an informal comparison with heat-transfer coefficients measured in FLECHT SEASET Run 31701.B-5 Run 31701 was a reflood experiment. performed with boundary conditions similar to those calculated for Dresden 3 and hence provides an experimental indication of the heat-transfer coefficients expected for a reflood rate near 0.2 m/s. Heat-transfer coefficients were measured at the core mid plane in the dispersed flow region well in advance of the quench front. The measured heat-transfer coefficients were two to three times higher than the TRAC-BD1 calculated heat-transfer coefficients. Thus, the TRAC-BD1 heat-transfer package may underpredict heat-transfer coefficients in the dispersed flow regime. In the sensitivity calculation, an increase in the convective heat-transfer coefficient of only 25% would have turned the cladding temperature around.
- The liquid and vapor temperatures of the fluid at the center of the
, beated length in the high powered channel are also shown in Figure B-39.
The liquid temperature was within 1 K of the saturation temperature during
, the entire. calculation. Steam superheat was calculated after the final dryout of the fuel rods at 37 s. The calculated superheat was generally between 80 and 170 K. The fluid quality was generally unity between 40 and 75 s. After 75 s, the steam superheat was caused by nonequilibrium effects associated with the interfacial heat transfer model; the equilibrium fluid quality was less than unity after 75 s. Comparisons with data have shown that the interfacial heat-transfer package in TRAC-BD1 B-39
f I Version 12.significantly overpredicts steam'superheat in the droplet flow regime.8-6 Consequently, it is expected that the steam superheat illustrated by Figure B-39 is unrealistically large. Since-steam was the
; primary' heat sink.in the high powered channel, the calculated peak cladding temperature was probably too'large because the steam temperature was too large due to excessive superhet.t. The expected error in peak cladding
-temperature due to the interfacial heat transfer model is not known but is -
~
estimated to be less than 50 K. The code developers plan to improve the
~
interfacial heat-transfer model in. future code versions by applying the Saha correlation,8-7 which predicts significantly smaller superheat in the' droplet flo'w regime. The channel, box inner wall temperature for the high powered channel, shown in Figure B-39, was'always within 7 K of the saturation temperature. Radiation to the channel wall:was a significant heat-transfer mechanism in the high powered channel because of-the high cladding temperatures. The radiative heat flux-from the rods adjacent to the wall was 40% of the
' convective heat-flux:at 140 s. The radiative heat flux from the high powered rods was about 20% of the convective flux at the same time.
The axial distributions of fuel' rod cladding temperatures for the low powered, average powered, and high powered channels are shown in Figures B-40,.B-41, and'B-42,' respectively, to document the calculated thermal ~ response of the core. A cladding temperature is shown for each of the seven. hydrodynamic cells used to represent the heated length of a
. channel. Cell 1 is at the bottom of the heated length. Cell 4 contains the center of the heated length and represents.the peak axial power
. region. Cell.7 is at the top of.-the heated length. Figure B-40 shows that .
a small heatup was calculated.through most of the bottom half of the low powered channel while no~ appreciable heatup was calculated above - Cell:4. However,.the maximum cladding temperature in the low powered channel occurred at'the start of the transient. The low powered channel
- was completely quenched by 76 s. Rod heatup occurred through most of the average powered channel as shown by Figure B-41. A channel maximum cladding 1 temperature of 589 K occurred at 94 s in Cell 4. The average powered channel was completely quenched by 165 s. Rod-heatup also
}
800 , , , X CELL 1 (BOTTOM) O CELL 2 A CELL 3 O CELL 4 (CENTER) 800
^ + CELL 5 ^
U 650 - O O CELL 6 _ b
, , CELL 7 (TOP) ,
' u i 3 -
-600 3
. O I o u
e' - e E & E 500 - k - E 1 f -400 $
. . . ~
' ' ' -200 350 O 50 10 0 15 0 200 Time (s)
Figure B-41. Axial cladding temperature distribution for the overage-powered channel. 1600 , , , X CELL 1 (BOTTOM)
' O CELL 2 -2000 A CELL 3 0 CELL 4 (CENTER) g + CELL 5 C v -
O CELL 6 -1500
. O CELL 7 (TOP) S ,
' t t 3 3 3' 800-- '
-1000 g L
3
- e
' e a
. E l -500 E
' r O H a ~- 4
- ud ._
.o 0 ' ' '
O 50 10 0 15 0 200 Time (s) Figure B-42. Axial clodding temperature distribution f or the high-powered channel. B-41
cccurred through'most of the high powe'ed channel as shown by Figure B-42.
, A peak cladding temperature'of_1021 K occurred at 175 s in Cell 4. The lower three~ cells;were completely quenched by 140 s. The uppermost cell was' quenched by 130 s due to a small, CCFL limited downflow through the upper tie plate.
The calculated void: fraction at the center of the heated length in the -
, high. powered channel is shown in Figure B-43 to illustrate the fluid response. The void-fraction ~ increased to unity at 37 s which caused the
. final fuel. rod heatup shown in Figure B-39. The void fraction generally
-remained at unity.until 75 s. The void fraction decreased from unity after 75 s but stayed.near 0.999 until the reflood of the hot channel was initiated at 160 s.
. Cal'culated convective heat-transfer coefficients from the fuel rod to the liquid and vapor phases'at the center of the heated length in the low powered, average powered, and high powered fuel bundles are shown in Figures B-44 through B-49. The heat-transfer coefficients correspond to the rods for which the cladding temperatures were shown in Figures B-37 through;B-39. The calculated heat-transfer coefficient to liquid in the low powered channel was relatively large because the rod was generally calculated to be in_ nucleate boiling. The void fraction exceeded 0.96 at 33 s which started the calculated dryout and caused the liquid Lheat-transfer coefficient to begin decreasing and the vapor coefficient to begin increasing. CHF occurred shortly thereafter and the rod was in transition boiling until 76 s when-return to nucleate boiling and quench were calculated. The liquid heat-transfer coefficient then returned to a
'high value while the vapor coefficient returned to zero. The heat-transfer .
. coefficients for the average powered channels, shown in Figures B-46
. and B-47, are similar to those for: the low powered ' channels except that the -
departure from nucl_eate boiling lasted longer. The liquid heat-transfer coefficient actually went to zero near 80 s when a complete rod dryout was calculated because the void fraction exceeded 0.999. In the high powered channel, the liquid heat-transfer coefficient decreased to zero and the
. vapor heat-transfer _ coefficient increased from zero at 25 s when CHF was calculated as shown in Figures B-48 and B-49. The liquid heat-transfer B-42
4 ( 1 < , _.
' ' 'y p, Dryout Start _of reflood c
. I. .
o
.. o 0
.' M t
- ALPHA 410006 o .I j..
6-
- 0.4 200
-~ 15 0 0 50 10 0 Time (s)
Figure B-43. Vold fraction at the center of the heated length in the high-powered channel. r-e i l i I-I l-i- i B-43 g i
id = _
~
, i
=..
RODHL450104 i
.. N 'I c
e N
,l
%.a f '-
j g : : . , o . un uE 16 = = .
*)
m x c v. 5 5 0 -
=
10-' = 0 : :-
.e -
2 - gg2 0.0 50.0 10 0.0 15 0.0 200.0 Time (s) Figure B-44. Heat-transf er coefficient to liquid in the low-powered channel. 1 i i 6 RODHV450104 e 0.8 - o C
"m
$y 0.6 -
un
;u E Omu N 0 .4 -
cv . O
.h
~
]e 0.2 -
-I O
O 50 10 0 15 0 200 Time (s) F i gu r e B-45. Heol-tronsf er coeificlent to vapor in ihe low-power ed channel. B-44
10' -
. . i =
~
RODHL430104 - N-g 1d r p%yddi3 l
$$y'::
- eg-_ :
Og . - l Un
. uE 16 y 7 i
- k. ? . i
?
g6 : ! . , o - 1 - 1 , Q 10" :- a 5 ! 3 A ! l ! 10-' O.o 50.0 10 0.0 15 0.0 200.0 Time (s) F i gu r e . B-46. Hect-trcnsf er coefficient to liquid in the overage powered channel. 1 i i i RODHV430104 E 0.8 - - Y 0.6 - - on uE -
*muh 0 .4 - -
. g-A
% 0.1 - -
0 O 50 10 0 15 0 200 Time (s) Figuro B-47. Hect-transf er coefficient to vapor in the overage-powered chonnel. B-45
1d ; _ i i i = RODHL410304 3 1d =- = 0 - O : :
%m - -
eg. - - . Og _ - s: um t E' 16 y 7 .
*k, i
?
Cv 0 - - 10" r I. - 0 5 5
= :
- A 10-*
0.0 50.0 10 0.0 15 0.0 200 .0 Time (s) Figure B-48. Heat-transf er coefficient to liquid in the high-powered channel. 1 i i i RODHV410304 i E 0.8 - - o , C l
*m
$y 0.6 - -
un uE Omu 0 .4 - l 8" A ( - 0.2 - i, 0 b 4 A.p yp.z;w 0 50 10 0 15 0 200 Time (s) F i gu r e 9-49. Hect-transf er coeficient to vapor in the high-powered channel. B-46
coefficient remained zero until 36 s when a rewet was calculated followed by: a second CHF at 37 s. The liquid-beat-transfer coefficient generally remained zero until 167 s~ .The heat-transfer to liquid increased at 167 s
, _ due to the axial heat-conduction'and prdcursory cooling effects computed by
.the fine mesh model during reflood. The primary heat-transfer mechanism
.between 37 and 167's was convection to vapor.
1 The following paragraphs' describe some of the code usage experience gained during the' 200% large-break calculation. ~Several time periods in
^
-the transient were identified in which the code was likely to fail during execution. The calculation generally failed when the low powered channel
- filled with liquid. The water packing option and a code update, which sets k ' an upper limit on condensation rates, were employed after 163 s in an attempt to eliminate the failure which~ occurred as the low powered channels filled. The combination of water packing option and code update did not p completely eliminate the code failure for Dresden 3 although the same combination has been successfully used-to eliminate similar fiilures in
, other calculations.B-8,B-9 Since the peak cladding temperature for Dresden 3 had already been obtained, the code failure was not thought to be critical. . Code failures were also likely to occur when HPCI was either initiated or terminated. The code failures encountered at HPCI initiation were generally-solved by resetting the time step to its minimum value. An alternative solution would be to open.the HPCI admission valves at a slower rate. The code failures associated with tripping off HPCI were solved by deleting the HPCI system after the HPCI flow terminated. Code failures due to high vapor temperature in the intact loop recirculation pump were also i- encountered after the pump voided. This failure was eliminated by setting o- the pump velocity to 0.0 rad /s at 72.5 s. The pump impeller was rotating
, in nearly stagnant steam at 72.5 s because the pump discharge isolation
-o' valve had been closed earlier. Consequently, setting the pump speed to 0.0 rad /s was not thought to have a significant effect on the results other than to allow the calculation to proceed at a reasonable time step. It is recommended that the code developers continue to improve code reliability by decreasing the probability of code failures.
B-47 L
The 200% large-break LOCA calculation was performed with a CDC-176 computer located at the Idaho National Engineering Laboratory. The computer run time for the model is illustrate-d by Figure B-50 which shows the ratio of central processor seconds used to calculate each real second during the transient. The average run time was 170 central processor seconds for each real second. The average time step size during the calculation was 0.007 s. The model of Dresden 3 consisted of - 155 hydrodynamic cells and 206 heat structures. Convergence studies were not performed in order to determine if more favorable running times could be obtained by reducing the size of the model. Faster running times could also probably be achieved by-judiciously increasing the maximum time step size although the probability of code failure during execution would also be increased. Some model optimization was performed in order to decrease the running time, as described in Section 4 of the main body of this report, but further work in this area might be fruitful. 4 l l l l B-48 l
- s,
'h i 300 , , ,
cm- - RUN TIME 1
^'
' 4:
C
;9'
*-~
p ' g '" . . c z200 -- o.
.g.
-; . \ ~. -
.y .
.g -
- 4. ' .o V.
i . ig . '10 0 - - - E
..C 2 3 Qll -
0' 0- c'50 10 0 . 15 0 200-Time (s)
' Figure B-50. Cornputer. run time.
t, f S
+
1 !, 'p[ 1 L Q 4, 4 i 2 l B-49
-REFERENCES
- B-1. Dresden Nuclear Power Station No.- 3 Technical Specifications,
' Facility Operating License DPR-25, January 1971.
B-2. R. J. -Dallman, TRAC-BD1 Calculation and Data Comparison of
' International Standard Problem 12, EGG-CAAD-5860, May 1982.
B-3~ ' M.'A.'Bolander, TRAC-BD1 Computer Code Assessment Calculations of - General Electric TLTA DBA and Small Break Tests, EGG-NTAP-5984, August 1982. , B-4. E..H31 comb, TRP
'91 Version 11 Code Assessment Calculations of General Electri 'ITA ECC/NO ECC Tests 6425 and 6426, EGG-CAAD-5857, May 1982.
B-5. PWR FLECHT SEASET Unblocked Bundle, Forced and Gravity Reflood Task
~ Data Report, Volumes 1 and 2, NUREG/CR-1532, EPRI NP-1459, WCAP 9699, June 1980.
B-6. R. B. Phillips and R. W. Shumway, TRAC-BWR Heat' Transfer: Model Description and Steady State Experimental Assessment, WR-CD-82-064,
-May 1982.
B-7. P.' Saha, B. S. Shiralkar, and G. E. ' Dix,'"A Post-Dryou't Heat Transfer Model' Based on Actual Vapor Generation Rate in Dispersed Droplet Regime," ASME-77-HT-80. B-8. P.' O. Wheatley, Scaling ' Analysis of the FIST Facility Using
; TRAC-BD1, EGG-NTAP-6073, October 1982.
' B-9. 'R. W. Shumway, TRAC-BD1-Version 12 BWR/6 Large Break LOCA
' Calculations, EGG-CDD-5995, to be published.
e h' i. i. B-50 l t l
[ s s. b s a I - ? E t 2 . { W' s APPENDIX C 2 e 160*." LARGE-BREAK LOCA B 4 i i e ? ?- i C-1
s
'I w
., APPENDIX.C:
160% LARGE-BREAllLOCA
>This. appendix describes the'results.of the TRAC-BD1 calculation for a
;hypothetica'1,1160%L large-break, loss-of-coolant accident. The accident was
- assumed to be initiated by.a 160% double-ended, offset shear break in the recirc61ation bloop piping. The calculation was performed with an input -
~
b Ed eck.' identical'to the one use'd to made the calculation described in
~
~
jAppendixBexceptthat'he' t break'areawas. reduced.
- Table C-l'. lists the sequence.of significant events occurring in the calculation. The results of the calculation are documented in Figures C-1
~
through1C-49. The results are similar to those discussed in Appendix B excep_t_ that.the events generally occurred slightly later because the g emperatur s 0 Ka ccurr 174 as sh Figure C-39. A detailed discussion of the results will not be presented
'becaush_ofthesimilarityof_theresultstothosepresentedinAppendixB.
tThe'160%-large-break LOCA' calculation was performed with a CDC-176 computer. TheLcomputer run' time for.the model is illustrated by
. Figure. C-50 which shows' the ratio' of central _ processor seconds used to
- calculate each real second during the transient. The average.run time was 175 cent'ral~ processor seconds for each real second. The average time step '
~
-size during the.calculatton was 0.007's. ~The user experience relative to code'veliability-for'the calculation was similar to that described in L Appendix B. The calculation was terminated at 196 s when the low powered
- channel was 11 quid full which caused a code execution failure.
~
L C-2 e
+ '
a n-1 TABLE-C-1. SEQUENCE OF EVENTS Time (s) Event 0.0 Break opened; loss of offsite power
~
. ;* 0.5- Scram signal generated;'feedwater terminated-1.5 Low water level signala 3.0 ' CRD flow terminated 3.2 Low-low water level'signala
~4.5 Control rods fully inserted 5.0 MSIVs closed 7.1 Jet pump suction uncovered b 11.9 Recirculation lines uncoveredb
.13 Lower plenum _ flashing
'23.2 HPCI initiated 32.5 Intact loop isolation valve closed 40 Dryout at the peak power zonec 48.1 LPCS initiated 59.6 Rated LPCS deliveredd 76 Lower plenum refill started
., 77.5 HPCI isolated and system deleted; jet pumps renoded, intact loop pump speed set to zero
. 130 Backflow from containment to vessel 160 Reflood initiated 165 Fine mesh and water packing models turned on C-3
TABLE C-1. - (continued)
. Time (s) Event 174 Peak cladding temperature obtained 196- Calculation terminated
~
- a. The low and low-low water level signals correspond to collapsed down-
. comer liquid levels of 12.80 and 11.28 m, respectively, above the bottom of the lower plenum. -
- b. The times at which the jet pump suction and recirculation lines uncovered correspond to the times the collapsed downcomer liquid level dropped to the elevation of the jet pump suction and the top of the reactor vessel outlet nozzle, respectively.
- c. Dryout corresponds to the time the void fraction in the center of the heated length of the high-powered channel approached unity,
- d. Rated LPCS corresponds to a flow of 4500 gpm per pump which :is the flow required by the Dresden 3 Technical Specifications.
4 e C-4
. .s -
3000 i . . TOTPOW000001 2500' - - n +
= 3:
f ' 2 ..-- 2000
. . . 3
- @'1500 -: -
u e-
- 3. 1000 -
o
-Q.
~
SM - - C 0 0 50 10 0 15 0 200 Time (s) Figure C-1. Core _ power. 8000 , i i P300901
-1000 m
^ 0 0 ._
v S-6000 - - { 800 v
- 6. .e e
' L 3 3 00
~h.4000"-
' u
- O. Q e 400 e E
- [O 2000 . - -
3
>. o
;'s .
-200 >
0 O O 50 10 0 15 0 200 Time (s) Figure C-2. Upper plenum pressure. C-5 t-
.e-a 15000 i i .
MFLOW190001 - 30000 n a m 10000 - M o .
- 20000 N
-N E E
. v e
- v
'5000 - -
10000 g -
=
m U $ 2 o__ - 0 2
' ' ' '~
-5000' 0 50 10 0 15 0 200 Time (s)
F i gu r e C-3. Mass flow rate out the vessel-side brook. 4000 . i 6 MFLOW200001 - 8000
.) - 6000 m T !
.2000 ,-L ]- -
- 4000 f
v c
# 3
?,
- 2000 O
g
~Q- o__ 0 g O
2 U 2 ' l
- -2000
-2000' O 50 10 0 .. 15 0 200 Time (s)
F i gu r e C-4. Mass fl>>w rate out the pump-side break. l C-6
A E 800 , , , HPCI O LPCS - O C O. O -1500
, ;soo .
q
,. g" N o E 1 .x .O t-V - -1000 C
- .. {,400 -
3 o
.1 w
E M
-500
$ 200-2 a
~00 0 'O O O' O
O 50 10 0 15 0 200 Time (s) Fi gu r e C-5. HPCI and LPCS mass flow rates. 2000 , , ,
- 4000 WFLOW150005 1000 _ -
2000 m A M
^
M g g.x 0-- -- _z_=- 0
.a E
V G 5 _ioco -. _- -2000 3 e - o
, M -
- -4000 m -2000 - - M U M 2 0 1
2
. - - -6000
-3000 - -
- -8000
-4000 O 50 10 0 15 0 200 Time (s)
Figu r e C-6. Mass flow rcte through Ihe beoken Icop j e t pump
'driveline nozzle.
C-7
t 2000 i i '
- 4000 WFLOW150004 1000 -
-- 2000 m- Q
) k Ilh~
-1000 - -
-- -2000
$ _2000 - -
- -4000 2- ! o
- -6000
--3000 -
- -8000
--4000
.0 50 10 0 15 0 200 Time (s)
Figure C-7. Moss flow rote through the broken loop jet pump discharge. t 2000 , ' '
- 4000 MFLOW150001 1000 -
-- 2000 7
0- -.
- , =, y - - - . ___4;. . ,, , --0 6 e v
-- -2000 ,
- [ -1000 --
= 2
- -4000 N -2000'-
-O I m , 2 0
- -6000
-3000 -
- -8000
-4000 -
O SO 10 0 15 0 200 Time (s) Figure C-8. Moss flow rate through the broken loop lel pump suc tion. C-8 i
,p..
.. a 2 3000 i i MFLOWO50005 - 6000 m
.M..
2000 - 7
- 4000 N E
e.. .O 3 g: ,
'V O
..'- { - .1000 h -- 2000 g
$ M
.O 2 - @
o-_' ----- - 0 2
-1000 ' -
O 50 10 0 15 0 200 Tim. (s) Figure C-9. Moss flow rate through the intact loop let pump driveline nozzle.
.7500 i i 6 MFLOWO50004 - 15000
- n 5000 -
7
- m. . - 10000 N N E 2 .o v
c l 2500
-- 5000 g c
$ m 0-- ci.iiii >%
i y , Q fkl!j- -0
' ' ' ~~
-2500 ~
0 50 10 0 15 0 200 Time (s) Fi gu r e C-10. Moss flow rate through the intact loop [et
. pump discharge.
C-9
a.
- 7500 i i i WLOWO50001 - 15000 m
nto 5000 - - m
- 10000 N N E j .0 *
~
c
'( 2500_- 5000
{
.C -
$ M
'O.
2 $ o-_ ._ _
.-__-.- -- --0 2
' ' ' ~ ~
-2500' 0 50 10 0 15 0 200 Time (s) ,
F i gu r e C-11.~ Wass - flow rate through the intact loop 'let , pump suction. 15 i i i LLEV300001
-40
^ 10 -
O
+
-E v -
-30 v e
- o 0 C
.C O O q - -20 g 8 5 -
5 '
- -10 -
0 O O 50 10 0 15 0 200 Time (s) F i gu r e C-12. Downcomer liquid level. C-10 q x - , y - , ,--p m , - - - , - , , -
( a-1 v. a *: 100000 i i .
- - 2.0 0*10'
'75000 - -
- 1.5 0*10' n g E o
.g -- .x -. _
v v- 50000 - -
- - 1.0 0*10' $
'2 $
15000- -
-- 5.0 0*10' 0 .0.00 0- 50 10 0 15 0 200 Time (s)
F i gu r e C-13. Downcomer fluid mass. 60000 i , i LPMASS
- - 1.2 0 *10'
- - 1.0 0*10' 40000 -' -
p v-
- - 8.0 0*10' .o e
'I g - - 6.00 **0' E 2
y 20000 -
- - 4.0 0*10
- 2.0 0 *10' O.00 0
0 50 10 0 15 0 200 Time (s) Figure C-14 Lower plenum fluid mcss. C-Il
h-h [ ~. 40000 g , I t- WT 80000
? _
3o000 -
. -60000 n'@. .
-6
- 20000 - ~
._ -40000 $
y.
~
10000 - _ -20000 0 O
.0 50 10 0 15 0 200 Time (s)
F i gu r e C-15. Guide tube fluid ' mass. 30000 i i i BYMASS
-. -60000
. -50000 20000 -
^
[1 4' A -40000 m {] n- \ e i -30000 o
$' I _
1 ' 10000- -
', s / -20000
.- -10000 0 O O 50 10 0 15 0 200 Time (s)
Fi gu r e C-16. Bypass fluid mass. l
.C-12
~
~
f: 1 i e
~ 15000,
' .' i UPMASS-
-30000
-25000 n
~*' -~
-10000
-20000 E' .
\
1- . 3
-15000 $
N i o 0 - 2 2 ' j 5000 k -5000 k' k V k h'I ' O O. 15 0 200
- 50. 10 0 O
' ' Time (s)
Figure C-17. Upper plenum flu.d mass. i i
-20000
-CMASS -40000 I. _
15000 -
-30000 m
)
E e-
^
E V ( i 10000 -
-20000
{
, . y -
, 2 s I
-.. --10000 5000, w Il Y
{
' ' 0 0 15 0 200
[ 50 10 0 O Time (s) rigure c-ts. core fluid mass. C-13
4000 , , ,.
~
MFLOW450001 - 8000 3000 - -
- 6000 m
^ m
~
2000,
- 4000 s e v =
[ 1000 . 7 -- 2000 e 3 .
~~ ~
TTV M G F'% [
-1000*- ( f -' ~2000
' ~
-2000 0 50 10 0 15 0 200 Time' (s)
Figure C-19. Moss flow rate of the inlet of the low-powered channel. 5000 - -
' ' ' - 10000 MFLOW450010 m
2 0-
'-l-1<I1 Wl E 3
y s a 0 7 l hr f N .N 1 E m-5 i v e 3 -10000
.-5000'- -
g O c O m - 0 - -20000 m 3 -10000 - - - S
- -30000
-15000
- 0 50 10 0 15 0 200 Time (s)
Figure C-20. Mcss flow rate at the outlet of the low-powered < channel.
)
C-14
.. - . . . . .. - . . . . . . - . . .. ~- ;. .. .
i 10000 , , , MFLOW430001 - 20000
- 15000 m m m 5000,- + -
, - 10000
, v _.
- )o -
- 5000 3 t
C O h O-- -
;p 7 r- I i hi} y)'f- -0
, 2 d - -5000
' ' ' - -10000
-5000 , \
0 " 50 10 0 150 200 Time (s) Fi gu r e C-21. Moss fl ow rate of the inlet of the overage-powered ~ chonne l. 10000 , , ,
- 20000 MFLOW430010 5000 ^
n - e .10000 N \ cn i E 6 .\ .; .
% ) e v
( -{! F8l%" 1 { 0-- -- c 't --0 y o c
. m O $
3 -5000'- -- -10000 2o
- -20000
-10000 ' ' '
O 50 100 15 0 200 Time (s) Fi gu r e C-22. Mass flow rate of the outlet of the overege-powered channel. C-15 1
2000 , , , MFLOW410001 - A000
't
- 3000 7 \
\ 1000 - -
- 2000 j ~
f_ s v v N 3
-}. - 1000 '
0-- - 3,, y g l} _.o 2
- -1000
~
' , , - -2000
-1000 0 50 10 0 15 0 200 Time (s)
Figure C-23. Moss flow rate at the inlet of the high-powered channel. 2000 - , , , MFLOW410010 - 4000
- 3000
^
m 7
\
\ 1000 - -
j -
- 2000 f v __
v D y o -
- 1000 O
G: n . 8
) -
- m. 0-- --
- --- l 2 --o n O m 2 0 1 .
- -1000
]
, , , - -2000
-1000 0 50 100 15 0 200 Time (s)
Figure C-24. Moss flow rate at the outlet of the high-powered channel. C-16
z - c:. 6 . , , CHANLL45
-15
^
E 4 -
\ k- O -A '
i 7
, 'y -
-10 g O f 0 M M 32 ; n .- 8
~ '
b bl i hjd l 0 O O 50 10 0 15 0 200 Time (s) rigure C-25. Collapsed liquid level in the low-powered channel. 6 , , , CHANLL43
-15
^4 '
O O g- -
-10 g i 0 0 M M
'82 - -
E ._
-5
/) pV 0 . O O 50 '
10 0 15 0 200 Time (s) Figure C-26. Collapsed liquid level in the overage-powered channel. C-17
6 , i i CHANLL41
- -15 ^ 4 - -
O
+
E v . v e e o - -10 0 C C - 0 0 M M o'2 -- - c
-5 0 O O '50 10 0 15 0 200 Time (s)
Figure C-27. Collopsed liquid level in the high-powered chonnel. 50 , 4 i
- 15 0 LIQUID 40 - ~
O VAPOR g 30-- -- 10 0 g N N E 20 - -
=
v v -
- 50 l
x M - - x f 06 !I j ' A.tn }-0 2> - 10 - 1 h I i ; - 2>
- -50 y
-20 - -
3 3 L. - -- -10 0 L
-40 - -
- - 15 0
-50 O 50 10 0 150 200 Time (s) l Figure C-28. Fluid velocities of the upper coro grid (third vessel
[ rirg).- t l r Y C-18 L
i ' .
~
- 15 0 40 -
t.100l0 - o VAPOR g 30-- -- 10 0 g N ' N v E 20 - v l j - 50
>. 10 -
V' i k . OC j --0
~'
e i . o
> -10 - '4 i -
l IU - -50 y
. -20 - -
E ? 6 -
-- -10 0 L
-40 - -
- -15 0
-50 O 50 10 0 15 0 200 Time (s)
Fi gu r e C-29. Fluid velocities of the upper core grid (second vessel ri ng). 50 , , i 40 - Ll0UID ~ O VAPOR g 30-- -- 10 0 g N I N v E 20 - f - v
- 50 3 10 lM.
e , 3 8 OC - 1 ( ' --0
$ -10 - -
+ -
- -50 y y
-20 - -
E - -
-10 0 E
-40 - -
I
-50 ' ' ' f -150 O 50 10 0 15 0 200 Time (s)
Figure C-30. F!uld velocities at the upper core grid (first vessel ring). C-19
m 2500 ' i ,
- 5000 MFLOW550004
^
0-4'(jgI - J --O
/
a A n l G
}
.h i 3 -
2 g
~ ~~ , -2500 - -
2 $
- - -10000
-5000 O 50 10 0 15 0 200 Time (s)
Figure C-31. Moss flow rate at the inlet to the 'oypass (third vessel ring). 4000 i i i MFLOW530004
- 5000 m n 2000 - -
m
- N h E
- 5
{ 0- N 1 yk g l l \_
-J k
h- - --0 t \ ; =
=
0
}
M
-2000 -
l -
- -5000 .
-4000 O 50 10 0 15 0 200 Time (s)
Figure C-32. Moss flow rate of the inlet to the bypass (second vessel ring). C-20
F 1000 e i i MFLOW510004 _ g 500 -
-,- 1000 N
E
.E
- ! < e p1- g i v
, 0-C h- ~- - --0 C
E 0 $
- -1000 0 2 -500 ~ -
1 2
- -2000
-1000 O 50 10 0 15 0 200 Time (s)
Figure C-33. Mass flow rote at the inlet to the bypass (first vessel ring). 400 . i i MFLEAK450001 - 800 300 - -
- 600 m n
7 N
) 200 , - -
- 400 s f I
100.-
$ ] l0 -- 200 g c ;
c o. E . 1 { d
--0 y l
^fI -- -200
-10 0 -h f
, , , - -400
-200' O 50 10 0 15 0 200 Time (s)
Figure C-34. Leckage from the low-powered fuel channel. C-21
500 i , '
- 1000 MFLEAK430001 400 - .300 - -
T 200' _)- E 5 100 - - N 0- -
--0 y .
A O
-2 -
l; - -500
-300 - -
ff
-400 -
4-
- -1000 -500 O 50- 10 0 15 0 200.
Time (s) Figure C-35. Leakage from the overage-powered fuel chonnel. 100 i . i
~
- 200 MFLEAK410001 g 50 ,- -. goo
)_v E 4 G 3 o. .
--0 3
\ \ \
3 -50' -
~~
2 { ,
- -200
- 10 0 O 50 10 0 15 0 200 Time (s)
Figure C-36. Leokage from the 'high-powered f uel channel. C-22
7 600 -
> i '
-600 X CLADDING
) O VAPOR A LIQUID 550 . _
O CHANNR M n .
-500 ^
b- P-v
. e
[500 - w 3
.e. -
-400 3
- 0 0 w u
- 450 - .'\ - o a
f E s
-300 g 400 - -
a' r --%bu u- ru
-200 350 0 50 10 0 15 0 200 Time (s)
Figure C-37. Thermal response of the low-powered f uel channel. 600 , i
-600 y X CLADDING 0 VAPOR A Ll0010 550 -
O CHANNR M m _
-500 C c
[500 - - E 3
-400 U 0 L L L Mf S 1450 -
V g a E E I '300 4
. 400 -
bls 29 %
-200
, t I 350 o 50 10 0 15 0 200 Figure C-38. Thermal respons o the overcge-powered f uel channel.
C-23
1200 , , , X CLADDING 0 VAPOR -1500 1000 - A LIQUID O CHANNEL WALL - m m M U l'- v
- 800 -- --1000 e .
L 2 3 o - E .
' u
- 600 e - e
.g Q 400 -
- ; 3 ; y
-0 200 O 50 10 0 15 0 200 Time (s)
Figure C-39. Thermal response of the high-powered fuel channel 600 , , , [ X CELL 1 (BOTTOM) j . O CRL 2 50 - A CELL 3 ~ O CELL 4 (CENTER) Q + CELL 5 -500 C v 0 CELL 6 L
- 500 - O CEL 7 (TOP) _ ,
3 $ g -
-400 g ' L 1450 - -
g E E -
-300 g 400 -
^
w _- - _ m , j
-200 350 ' ' '
O 50 10 0 15 0 200 l Time (s) i Figure C-40. Axial clodding temperature distribution f or the I low powered chonnel. l C-24
u
- E c. g, v2 [E&eLo3L, 6^ 8 1 1 F 5 6 1 1 4 6 3 5 0 F 4 6 8 0 2 i 5 0 0 i 2 0 0 0 0 0 g 0 0 g 0 0 0 0 0 0 u 0 -
0 0 0 0 0 O . - l-u - , r - r O , "a
- _ _ e e - '
C 4 _r_ O0+0AOX C 4 1 2 CCCCCCC EEEEEEE
~
LLLLLLL t A LLLLLLL hx A ei hx i 7654321 c . g a l
- ( oi 5 ,
5 ( ( B v 0 '
- c. er c hl 0 ' T C
- C E O p c P N T T aa l .
ol ) T g d wa E O M ed ed R ) - i 4 rd ) pn ei " og dng w C c C et r eT
- ht T em i 2
5 ae i dp m10 nmm1 c e r e0 n p 0 , e e e0 ' - ' l
. r a( 0 ,
ha( o n t n u r )s t s u) r e e
.l e OO+OAOX e
d o i CCCCCCC d s EEEEEEE i s t . LLLLiLL t r i 1 LLLLtLL r b 5 i b 1 5 u 0 7654321 , u 0 i
' . (
=
t i ( ( B t o T C O i o n O E T n P N T f ) T O f o o r , E R ) M r = 3 ) t h 2 e 2 0 e _ g y O - , 0 - . _ 0 - - - - 0 - - - 4 6 8
- 1 2 2 0 0 0 5 1 0 5 0 0 C 0 0 0 0 0 0 0 0 0 0 0 0
$ECcLo3L, b^
LC yEc . g,
.i.
I ' ' '
'1 C
'O
=
u ' O
.h .
T. 0
> j ALPHA 410006 0.4 O 50 10 0 - 15 0 200 Time '(s)
F i gu r e C-43. Void fraction at the center-of the heated length in the high-powered channel. I O I C-26 I
i 1d -- i i i =
- RODHL450104 h c
N w s
~
id I wW E
*n
- !?lybyN=!
. #M o on uE 1d = 3 Ehn h ? 5"a ~. - 16' -
?
o - e . - 1 - _ ~ 10-' O.0 50.0 10 0.0 15 0.0 200.0 Time (s) Figure C-44. Heat-transfer coefficient to liquid in the low-powered channel. 1 i i i ROOHV450104 e 0.8 - - o C
*m
$y 0.6 - -
ou uE Ous N 0 .4 - ~ . ga A
]. 0.2 - -
t t t 0 0 50 10 0 15 0 200 Time (s) F i gu r e C-45. Heat-transf er coefficient to vapor in the low-powered chonnel. C-27
1d -
; i i i ;
RODHL430104 : N : I. wl:l -
" 3d -
. 5
- }l$l h ..Mbl. 5
.% m - - ex O L oln uE 11 =. I l =
*h b : l b s : ,
a , o - i , - f:
~
10 -
% l '
7 E J lE i T -
, i
; I 10'*
0.0 50.0 10 0.0 15 0.0 200.0 Time (s) Figure C-46. Heat-transf er coefficient to liquid in the overage-powered channel. 1 e i i RODHV430104 e 0.8 - - o. C
*m
$M 0.6 - -
ou uE
*m2) 0.4 -
Cv O ,
.h ]. 0.2 - -
l DAT l ll l 0 0 50 10 0 15 0 200 Time (s) Figure C-47. Heat-tronsf er coefficient to vapor in the overage-powered channel. C-28
1d -
. . i
~ -
RODHL410304
~ -
~ ,
c - - e u 1d =
=
C,
. og - -
O uln uE 11 = =
*)
5 5 m x Cv 0 - - 10~' r 3 O : ' p I -
, i
- i -
' ' ' l ' '
10-* o.o 50.0 10 0.0 15 0.0 200.0 Time (s) Figure C-48. Heat-transf er coefficient to liquid in the high-powered channel. 1 i i i RODHV410304 0.8 - -
.h_
o i C I
*n
$y 0.6 - -
un 6 ON* o.4 ua Cv i . 0 h h . o.2 - 1 -
= i
$,+, %Y=r; O 50 10 0 15 0 200 Time (s)
Figure C-49. Hect-transf er coeficient to vapor in the high-powered channel. ,, C-29
-400 , , ,
% :RUN< TIME n
i'. ;
.c
.,'l 300 - _
'n C
O
- 6. .
.N
' n . 200 - _
,._. a' t- ;:: .u
.v-
. (
E;
- - 20 -
' j.
oc 0
'0' 50 -10 0 15 0 200 Time (s)
. Figure C-50. ' Computer run time.
m' A 9 4 l l l
?
z C-30 =. = - - - -.~,.-.- .. .. \
7-b h.. 3-3 7
.a s -:s f
l APPENDIX 0 120% LARGE-BREAK LOCA i I .- E e D-1
%E APPENDIX D 120% LARGE-BREAK LOCA This appendix describes the results of a TRAC-BD1 calculation for a hypothetical,120% large-break, loss-of-coolant accident. The accident was assumed to be initiated by a 120% dotole-ended, offset shear break in the recirculation loop piping. The calculation was performed with an input -
i deck identical to the one used to make the calculation described in Appendix B except that the break area was reduced. Table D-1 lists the sequence of significant events .ccurring in the calculation. The results of the calculation are documented in Figures 0-1 through D-49. The results are usually similar to those discussed in Appendix B except that events generally occurred later because the smaller break area resulted in a slower blowdown. Only the results which do vary significantly from those described in Appendix B will be discussed below. The upper plenum pressure, shown in Figure D-2, did not drop below containment pressure during the calculation. Consequently, the break flows, shown in Figures D-3 and D-4, always remained towards the containment. The calculation was not run long enough to obtain the backflow from the containment which had occurred in the calculations discussed previously. The mass flow rate at the inlet of the high powared fuel channel is shown in Figure D-23. The calculation was'not run long enough to obtain the increase in flow rate which initiated reflood in the previous calculations. The calculation was not run through reflood because the peak - cladding temperature had already been obtained. The calculated thermal response at the center of the heated length in the high powered fuel - channel is shown in Figure D-39. The main heatup of the fuel rod occurred near 58 s when the rod dried out because the local void fraction, shown in Figure D-43, reached 1.0. The rod heatup was arrested at 68 s when the void fraction dropped to 0.997. The corresponding small amount of liquid D-2
TABLE D-1. SEQUENCE OF EVENTS Time (s) Event O.0- Break opened; loss of offsite power
>. j. 0.5 . Scram signal generated; feedwater terminated 1.5 Low water level signala 3.0 CRD flow terminated 3.4 Low-low water level signala 4.5 Control rods fully inserted 5.0 MSIVs closed 7.6- Jet pump suction uncoveredb 13.6 Recirculation lines uncoveredb 15 Lower plenum flashing 23.4- HPCI initiated 32.5 Intact loop isolation valve closed 54.6 LPCS initiated 59 Dryout at the peak power zonec 70.1 Rated LPCS deliveredd 3 90.8 HPCI isolated and system deleted; jet pumps .- renoded; intact loop pump speed set to zero t
+ D-3
TABLE D-1. (continued) Time (s) Event 92 Lower plenum refill started 120 -Peak cladding temperature obtained
~
132 Calculation terminated - i-
~ a. SThe ion arid 16.-low-w&ier level signals correspono to collapsea down-comer liquid levels of 12.80-and 11.28 m, respectively, above the bottom of the lower plenum.
- b. The times at which the jet pump suction and recirculation lines uncovered correspond to the times the collapsed downcomer liquid level dropped to the-elevation of the jet pump suction and the top of the reactor vessel outlet nozzle, respectively.
- c. Dryout corresponds to the _ time the void fraction in the center of the heated length' of' the high-powered channel approached unity.
. .d. Rated LPCS corresponds to a flow of 4500 gpm per pump which is the flow l_ required by the Dresden 3 Taciiaical Specifications. 4 I I i D-4 _ _ -~
- was-able to partially rewet the rod surface and stop the rod heatup. The rod heatup began again near_78 s when the void' fraction increased to 1.0.
The cladding. temperature' increased until 94 s when the rod was partially rewetted.again. A' quench was calculated near 110 s when the void fraction decreased to 0.96. The.rewet and quench were caused by liquid which leaked from the bypass--into the high powered channel and was then entrained up ,
. through the channel. The liquid was able to partially wet the rods because i- . the surface' temperature was less than'the Leidenfrost temperature. In the E'fl calculations doenmented in Appendicae 8 and C, dryout occurred earlier resulting.in cladding temperatures that were greater than the Leidenfrost temperature.which thus precluded the rewet for those calculations.
The calculated thermal response was very' sensitive to void fraction
.when the surface temperature was less than the Leidenfrost temperature.
Many' code parameters probably could alter the void fraction enough to either cause or eliminate a rewet. For example, the time at which the jet
-pumps were renoded was found to influence the rewet. (The jet pumps were renoded during the calculations in order to obtain a more realistic lower plenum refill rate as described in Section 5.3 of the main body of this report.) In the. base calculation for each break size, the jet pumps were renoded near the time'the HPCI system was isolated by low reactor pressure which was generally also near the start of lower plenum refill. In the base calculation for the 120% break transient, the jet pumps were renoded at 90.8 s. A sensitivity-calculation was made in which the jet pumps were renoded at 115 s instead of 90.8 s. Figure D-50 shows the sensitivity of the cladding' temperature at the center'of the heated length in the j high powered channel to the time at which the jet pumps were renoded. In f'I the sensitivity calculation, the temperatura leveled off momentarily at 95 s but did not rewet like the' base case. The cladding temperature then ., increased above the Leidenfrost temperature and did not turn around until
- 170 s. The different rewet behavior in the two calculations was caused by small differences in'the calculated void fractions. The base and sensitivity calculations show that cladding temperature was sensitive to small changes in the model. -Variations in models of CCFL, wall heat transfer, or interfacial momentum transfer within their uncertainty limits could also:probably either promote or eliminate the rewet. Consequently, D-5
,c - ,1 e tthh ba'se and sensitivity calculations:can only be interpreted to mean that
'the potential'for a rewet. exists near-100 s in this transient but a.: considerable uncertainty must be accepted in the reported cladding
,itemperatures. However, the peak cladding temperatures for both the base L
' ar.d~sen'sitivity' calculations'were well below the' temperatures reported in
-Appendix B for.the 200% large-break calculation.
I' The peak cladding temperature for the base' case calculation was 637 K andoccurrcd'atl120Lsinthecelljustabovethepeakpowerzoneasshown
~
-in Figure D-42. A rewet at this cell was calculated at 125 s.
The1120% large-break LOCA calculation was performed with a CDC-176
# computer. The_ computer run time for the model is illustrated by Figure D-51.which shows the ratio of central processor seconds used to calculate each real second during the transient. _The average run time was 150 central processor seconds.for each real second.= The average time step size'during the calculation was 0.008 s. The code. executed more reliably for this transient than the ones described in Appendices B and C. No execution failures ~ occurred'during the calculation.
e e y, D-6
~ _ - _ _ _ . . . _
I 3000 , , ,
< TOTPOW000001 2500 - -
n-3;. v 2 2000 - - t s 1500 -
.e .
3 1000 - o n. 500 - - C 0 O 50 10 0 15 0 200 Time (::) Figure 0-1. Core power. 8000 , , , P300901
-1000 ^
m 0 0
$ 6000 - -
*E V -
300 O
' 8 u
s ' rn 3 e 4000~ -' - -600 $ u 8
. Q.
1 g -
-400 e s E 2000 - - 3 o ~
200 $ c 0 O O SO 10 0 15 0 200 Time (s) Figure D-2. Upper plenum pressure. D-7
n __ r 15000 i * ' MFLOW190001 3 coco
-25000 7 7 \
Ym 00001 E x . -20000 -Q
- v v
'g o
) *
-15000 o
.C 3
.. m a 5000 -
W O . -10000 O
.2 2 L
. -5000 o
0 50 10 0 15 0 700 Time (s) Figure D-3. Moss flow rote out the vessel-side break. 4000 t i 6 MFLOW200001 - 8000
- 6000 m u
7 \
-:\a'2000 ,
- ~
- 4000 E
.x v
.D v
3 - 2000 W o - o
, C ~
a M
--0 e 0-- M C 0
~2 2
- -2000
, , i i -4000
-2MO O 50 -10 0 15 0 200 Time (s)
Figure D-4. Moss flow role out the pump-side breck. , D-8
800 i i i HPCI O LPCS 1500 m n -.600 n N N E 06
* .o 6 . -1000 C
- D 400
.O D
O .v; y f y o .
-500 o 2 200 - O -
2 i ' 00 0 0 0 O _i O O 50 10 0 15 0 200 Time (s) Figure D-5. HPCI and LPCS mass flow rates. 2000, i e i 4000 MFLOW150005 1000.- -- 2000 7 7 N N o._ _ _ , - = -
--0 E 03
.a b O
-3 _1000 -- -- -2000 O 3 O
.i j .
- -4000 g w -2000 _
n O l-3 o 1 2
. - -6000
-3000 - -
- -8000
-4000 0 '50 10 0 15 0 200 Time (s)
Figure D-6. Moss flow rate through the broken loop jet pump driveline nozzle. D-9
2000, ' i *
- 4000 MFLOW150004 1000 . - -- 2000 e .
7 YM
~
~ ~
I v
-- -2000 3
~ l -1000 -
=
-[ 2
-- 00
$ -2000 ~
o
-$ - -6000
-3000 -
- - -8000
-4000-O 50 10 0 15 0 200 Time (s)
FI gu r e D-7.. Moss flow rate through Ihe broken Ioop jet pump discharge. 2000, , , ,
- 4000 MFlow150001 1000 -
- 2000
^
7 0- % --0 5 v e
-- -2000 ,
$ -1000 ~-
= 2
~~
- m .
$-2000 O
M
- -6000 ,
-3000 -
- -8000
-4000 O 50 10 0 15 0 200 Time (s)
Figure D-8. Moss flow rate through the broken loop jet pump suc tion. D-10
.,W
- 3000 . i i MFLOWO50005 - 6000 m
m.
-2000 - -
7
- 4000 N N E o o
$ 1000,- -
2000 5
! = g
- 0 a 8
2 o.. -
--0 2
. , , , - ~2000 0 50 10 0 15 0 - 7,00
-Time (s)
Figure D-9. Moss flow rate through the intact loop jet pump driveline nozzle. 7500 . i i
' MFLOWO50004 - 15000 l
n m 5000 " - 7 N
- 10000 N E
j- .o v c l 2500 -
-S000 g
, E C l
0-- W" j --0
' ' ' '~
-2500' 0 50 10 0 15 0 200 Time (s)
Figure D-10. Moss flow rote through the intact loop [of pump discharge. D-ll r--y
+
L 7500 , , , WFLOWO50001 - 15000
, n 5000 - -
T
- 10000 N
._ N. ,' . E x .Q .
O j 2500,- -- 5000 3 -
= .
o E. m O M
'2 o_. _- _.o j
' ' ' - -5000
-2500' 0 50 10 0 15 0 200 Time (s)
F i gu re D-11. Moss flow rate through the intact loop lat purnp suc11on. 15 i , , LLEV300001 40
'A * -
^
E v -
-30 v e e o o C C 0 0
+-
- a
-20 m 35 - -
3 ~ *
-10 .
l 0 0
^0 50 10 0 15 0 200 ;
Time (s) l Figure 0-12. .'Downcomer liquid level i D-12 ~ _ -
I
't 100000 i i i
- 2.00'10' 75000 - -
- 1.50*10' ' m
- ^ E cn
! v X S 50000 - -
.I ' a a
~
- 1.0 0*10 s o 0
w 0 1 , 2 25000 - ' - 5.00*10'
? '
' ~
0 O.00 O 50 10 0 < 1bD 200 ( Time (s) F i gu r e 0 ,13. Downcomer fluid mass. '
.)
-i
/
60000 , i i i LPMASS ~
- - 1.2 0*10'
.00'10' 40000 -
j n
^
E p -
- 8.00*10' .o
, O o
4 - 6.00*10'
. I 20000 - -
- 4.0 0*10'
- 2.0 0*10' 0 . I 0.00 0 , 50 10 0 15 0 200 Time (s)
Fi gu r e 0-14. Lower plenum fluid mass.
?
/- /
D-13
' m mas -
,, i
?)
40000 8 i i MGT
-80000
?(lb s 30000 - -
-60000 m g .E a
u.x n -
,(200C0 m -40000 $
j
~
U 2
~
10000 -- -
-20000 r-'
' i i 0 0 0 50 10 0 ' 15 0 200 T\q Time (s)
Figure D-15. Guide tube fluid mass.
)]E 30000 i i i BYMASS
-60000 20000 - ~
n Qu
-40000 o E
v - C e
' l 30000 $
y" '*IY O
.? y ,
10000 - , h -20000
"'0000
;f
., t i ,
..y> 5,0 0 0 50 10 0 15 0 200 Time (s)
. F l gu r e ' D-16. Bypass fluid mass.
Y. ?
.< >J D-14
15000 i i ,
, UPMASS
- -30000
- -25000
-10000
']!
n --. ^ E
- j. f. - f -20000 o
- O 3
e y - -15000 $ E 5000-j -10000 0
- -5000 0 O O 50 10 0 15 0 200 Time .(s)
- F i gu r e D-17. Upper plenum fluid mass.
20000 , , , CMASS
- -40000 15000 -
- -30000 m o,
E 5
- 2
. , 10000 -
# y - -20000 $
.! 3 $
5000 - Lj -10000 4 4
' ' ' O 0
O 50 10 0 15 0 200 Time (s) Fi gu r e D-18. Core fluid mass. D-15
9 4000 -' i MFLOW450001 - 8000
' ~
3000 -
. - 6000 m 7 -
-2000, --
- 4000 f
~
V-I 3 1000 -
-- 2000 3 .
?
O
.2-0--
(- ,. -' g'
'l y --0 3
-1000*- _- - 000
-2000 O 50 10 0 15 0 200 Time (s)
Figure D-19. Moss flow rate of the Inlet of the low-powe red channel. 5000 .
- 10000 MFLOW450010
. - 5000 7
7 N N m o._ n i g h % --0 n E
.x w v
- y. . - -5000
( , 0 0
- C' e .
~
- -10000 '
m -5000 -
$ 1 O. O 3 2
. - -15000 ,
. - -20000
-10000.
0 50 10 0 15 0 200 Time (s) ! Figure 0-20. Moss flow rate of the outlet of the low-powered channel. D-16
~8000 , , ,
MFLOW430001'
-15000 6000 - -
7
-(
n-N' 10000 E
- m-'
5 4000 - - e
,v
)
o 3 2000'- -
-5000 $
M M M O M 2 0 2 0-- . i * --O
-2000 O ~50 10 0 15 0 200 Time (s)
F i gu r e D-21. Moss flow rate at the inlet of the overage-powered chonnel. 8000 , , , MFLOW430010
-15000
-6000 - -
^ 7 C N <
.m- _
-10000 E 6 _4000 - -
3 v 5 3 o j U ' 00 [ 2000 ~ -
*- E m a
0- - W_ - -
-r -- -
--O i
-2000 O $0 10 0 15 0 200 Time (s)
Figure D-22. Moss flow rete of the outlet of the overage-powered chonnel. D-17
[ -2000 ., , , MFLOW410001 - 4000
^
- 3000 m m
r 1000 -
- en .-
E
- 2000 5' O '
' v
)
o - 1000 3 - C .O 0-- '" t
-g ,Q T U F --O g 2 0 2
- -1000
' ' ' '~
-1000 0 50 10 0 15 0 200 Time (s)
Figure D-23. Moss flow . rote. ot the inlet of the high-powered chonnel. 2000 , , , MFLOW410010 - 4000 n
- 3000 m
.n 1000 -
j
- 2000 f V
3-o -
- 1000 3 C O C
m-a 0--
% k '- u 0 --O m -
2 W 0 2
-1000
~
-1000 ' ' ' '~
0 50 10 0 ' 15 0 200 ' Time (s) t-Figure D-24. Mcss . flow rate at the outlet of the high-powered channel. D-18 i-
t- . t
, 6 i , ,
CHANLL45 15
^
E - l - O-
- c. v i v
e e 0 -
-10' o C
1
. \
0-C O e m
.3'2 I -
8
-5 k kl O O O 50 10 0 15 0 200 Time (s)
Figure D-25. Collapsed liquid level in the low powered channel. 6 , , , CHANLL43
-15
^
E 4 - O
+
v v e o g -
-10 g O O e-I.
e m 8 '2 - 3 5
.=
0 O O 50 10 0 15 0 200 Time (s)
. Figure 0-26. Collapsed liquid level in the overcge-powered channel.
D-19 m.
-6 , , ,
CHANLL41 f
- -15
- ^4 -
O v E. - v .
- e e u .
.jo o C C .
O O m m o 2 - a
- -5 0 O O 50- 10U 15 0 200 Time (s)
Figure D-27. Collapsed liquid level in the high-powered channel. 50 , , ,
- - 15 0 40 -
LIQUID _ O - VAPOR g 30- - -- 10 0 g N- N
'g :20 v - - 50 v x 10 -
x Of j --0
. -10 - O -
- -50 y
.n
-20 -
3 _3
- u. - -- -10 0 L
-40 -
- -15 0
-50 O 50 10 0 15 0 200 Time (s)
Figure D-28. Fluid velocities at the upper core grid (third vessei ri ng). i l D-20
I III_ F k _3 y I @ > - vENg F L.3.y .
. >e oUO> vENq . _
i - - - - - i - - - - - g 5 4 3 2 1 1 2 3 4 - 5 g 5 4 3 2 1 1 2 3 4 5 u 0 0 u r e O 0 0 0 0 0 0 9 0 0 0 1 0
.0 r e
O 0 0 0 0 0 0 0 0-0
~
- 0_ g{
0 D
- O O -
. O 3 2 0 9 VL . VL AI AI PQ PQ rF i OU rF i OU n u l RI n l RI g i D g iu D
)d
. 5 0 ' i
)d
. 5 0 ' '
v v e l el o o c i ci t t i i D e s e s 2 1 o T i a T i f ~ t - m0 m0 1 1
~
t t = he0 ' , he0 ' 3 i e e p u (s IP p u( )s p) .
. p e '
e r r c c o o r r e 1 e 1 5 5 g 0 ' , g 0 ' i r i ri d d ( ( f s i r e s c t o n v d ~ e 2 v s 0 e 2 s 0 e 0 - - - - - - - - ~ s 0 - - - - ~ l
- - - - - - . s - - - -
e 1 1
- - - 0 5 1 0
1 5 l
- - - 0 5 0 5 1
5 1 0 5 0 0 0 1 5 1 0 5 0 0 0 0 0 0 0 0 6 L. 3y$ 8 > vONg L.3_..
. y >e o0 O > v ONq ,
w - - i
'2500 , , ,
- 5000 WFLOW550004 m Y Q
$ o-W [' , q
--o 1 5 l e v
-) 5 o -
C *
~~
E -2500 ' - - m O M 2 0 2
- -10000
-5000 O 50 10 0 15 0 200 Time (s)
F i gu re 0-31. Moss flow rate at the inlet to the bypass (third vessel ri ng). 4000 , , , MFLOW530004
.3000 - -
n -2000 - - 7
.( N j 1000 v
h { 0-C >
--0 y e
4y ? m -1000 - - c $ 3 -2000 -
- -5000 .
-3000 - -
-4000 ' ' '
0 50 - 10 0 15 0 200 Time (s) Figure D-32. Mass flow rate of the inlet to the bypass (second vessel ring). D-22
i
~
+{ Va)E 2UM* = Ng F _ F 1 1 i 2 1 2 3 4 i 0 5 5 0 g 0 0 0 0 0 g 0 0 -
0 0 u r 0 0 0 0 0 u 0 0 o 0 0 0 ~ . , . - r O- . . - , ~ e - - - e - L - D D 3 3 4 3 L r iM e a no _ k gs e 5 )s
. 5 ; 0 ' ; 0 ' i f
e l o f r w o D m r t a b;
- t e 2
3 h eT i oT- w_ p f i p om1 l we0 0 t m10 g i he0 e
' i
- \
p( i( os q ns w) l) e e , r , t e l d (I t o l f u t e 1 5 M h 1 l 0 F e 5 M L i 0 ' F i c E y b L h A O a K p n n 4 a s W 5 5 e s 1
.l 0 0 0 ( 0 0 f i
0 1 r 4 s t 2 l 2 0 0 0 -
~
v e s 0 - 2 4 6 8 s - - 0 1 2 0 0 0 0 e 2 1 0 0 0 0 0 0 l 0 0 0 0 0 0 0 0 0 0 0
?3 vee m 20nm = E Mm
l ~
! !; I 1.
2O$ $* $ n g$ -{ 6a(m F - - - -- - 5 F - 5 4 3 2 1 1 2 3 4
- 1 i 0 0 0 0 0 0 0 i 1 0 5 5 0 g '0 0 0 g 0 0 0 0 0 0 0 0 0 u 0 0 o 0 0 u r
0 0 r O ~ _ , ~ O - , e
. - e - - .- - - -
D D - 3 3 - 6 5 -- M F L L E L e A e K a o k k 5 A a 5 a 0 3 , g 0 ' i g ' 0 e e 0 0
~
f f 1 r r o . o m : m ~
~
t t h h D eT e T
- i i
2 4 hm1 i g e0 0 i i om1 v e e0 0 i h r p (s ga (s ' o) e)- w p e l o r e w d e r f e u 1 M d 1 5 e 5 F 0 i l 0 ' L i f ' E u c A e h K l c 4 n 1 c n 0 h e 0 a
.l 0 n 1 n e
.l 2 2 0 0 - - - - - -
- - - 0 - - -
0 - - - - - - - - - 1 0 1
- - 0 5 0
- - 0 0 2 1 0 1 5 0 0 0 0 0 0 0 0 0 0 0 0 0 0
g Mo %. *, -OE "m 20 M *23 v"EN* O 1 ; l,
600 e i i
-600 X CLADDING y O VAPOR g
A LIQUID 550 - ~ O CHANNEL WALL m .
-500 f
v
-*. *L 500 -- >
_ e
<- u i 3
-1 0
-400 2
-* i O
' L
[450 - _ o. E E
.y - .
-300 e 400 -
-200
' i i 350
[. 0 50 10 0 15 0 200 Time (s) Figure D-37. Thermal response of the low-powered f uel channel. 600 i i i
-600
' X CLADDING
' C VAPOR g
A LIQUID 550 - - O CHANNEL WALL n
-500 Q d
- 500 - - e 3
O
-400 2 t O
' 6 u
h [450 - - 1
-E E 0
3 - l -300 S
- 400 -
-200 350 8 e 00 10 0 15 0 200 Time (s)
Figure D-38. Thermal response of the average-powered fuel channel. L t D-25 l
800 i i i X CLADDING O VAPOR A LIQUID 700 -~ --800 O CHANNEL WALL m m M v h v e - Partial rewet o , _~' i
} 600l& % ' *
} -
E ig' e X o [500 -
/ oryout
-400 g E
nch N 400 - -
-200 300 O 50 10 0 15 0 200 Time (s)
Figure D-39. Thermal response of the high powered fuel channel. 600 , , ,
~
- 0 X CELL 1 (BOTTOM) l .
O CELL 2 A CELL 3 1 0 - O
~
CELL 4 (CENTER) g -
+ CELL 5 -500 C v O CELL 6 L e 500 -
( - o u 3 3
-400 +
0 0
' 6 1450
{ E E -1 0
-300 j 400 - -
l
-200 350 0 50 10 0 15 0 200 Time (s)
Figure 0-40. Axial cladding temperature distribution f or the low-powered chcnnel. D-26 I J
r ' 800 i , , X CELL 1 (BOTTOM) O CELL 2 A CELL 3 O CELL 4 (CENTER) -800
- g. + CELL 5 O CELL 6 p
v v 650 - -
.. , O CELL 7 (TOP) ,
' s.
3 .- -600 2 o I o 5- t e o O- a. E 500 - - E
-400 y A - $
350 O. 50 10 0 150 200 Time (s) F i gu r e . D-41. Axial cladding temperature distribution f or the overage-powered chcnnel. 900 , X CELL- 1 (BOTTOM) O CELL 2 800'- A _-1000 CELL 3 . 0 CELL 4 (CENTER) g + CELL 5 C v 70o .- O CELL 6 --800 L
, G CELL 7 (TCP) ,
b t 5 600 ( W --600 o We
~ ' ! % t ! O. o.
E 500 - 37 o - E 4 -400 y
-2^
400 - -
-200 300 O oO 10 0 15 0 200 Time (s)
Figure D-42. Axiol cicdding temperature distribution for the high-powered channel. D-27
)
J ke a f j l { Dryout: Partial rewet
.-- f-o e
.O -
y
'O .
ALPHA 410006
*5
> ,)
.i 0.4' O. 50 10 0 15 0 200 Time (s)-
Figure D-43. Vcid fraction at the center of the heated length in - the high-powered channel. D 6 I' j. a 1 0-28 I
1d =
. i i =
~
RODHL450104 c e N. s
' ' ~
1d g g
%m _
.- ey -
, O g- _
un
.- uE e
16 3 3 mu ... .c v - o o .
~
=
~
10 ' = E E
.g - : :
1 _ l ' ' 10-* O.0 '50.0 10 0.0 15 0.0 200.0 Time (s) Figure D-44. Heat-transf er coefficient to liquid in the low powered channel. 1 , , , RODHV450104 e 0.8 - u-C "m
$y 0.6 -
un u E-
'Oh 0 .4 m u
. .1 Cv g
.h 3e
- 0.2 -
r 0 O 50 10 0 15 0 200 Time (s) Figure D-45. Heot-tronsf er coefficlent io vapor In the low-powered chonnel. D-29
10' ;_ i i i
=_
- . RODHL430104 2
.,_ N w -
C- - o (g
..I d . g .
I
.C' : :
;g :- :
oi Un 1 I i uE 1(f '
*n2)_ E 1
E
~
Cv - Q - l 10-' r 3 i f : S : 11 :
~
~
. I .
= 10~'
o.o 50.0 10 0.0 15 0.0 200.0 Time (s) . Figure D-46. Heat-transf.er coefficient to liquid in the overage-powered channel. 1 i i i RODHV430104 E 0.8 - - o C
~n.
$y 0.6 - -
um uE Owa N 0 .4 - - Cv 0 -
.h 3 0.2 -
e \. 0 0 50
, M 10 0 15 0 200 Ilme (s)
Figure D-47. Heat-transf er coefficient to vapor in the overage-powered channel. D-30
*i 1([ ' ;
i i i :
~ ~
N, RODHL410304 - c 1d ' Q) '
\
-{
- {
(%\ :
~
- m2 O
u n g. t l
, .] . LE. 1(f } -
j' .
'r - eh 3:
[,[ l g3 : , . 1 : : o - [ - l i. ,
~
10~' = - f :i
=
- i
, j 2 - I
)l j
" l L '
lj, l 10~* O.0 50.0 10 0.0 15 0.0 200.0 Time (s) Figure D-48. Hect-trensf e'r coefficient to liquid in the hlgh powered channe1. 1 i i i RODHV410304 E 0.8 - - 0.6 - - ou uE - {d
.3 m2
) 0.4 -
Y 1
@v- [
s 3
. g 0.2 -
0 =I , ) , 0 50 10 0 15 0 200 Time (s) Figure D-49. Heat-transf er coeficient to vapor in the high-powered channel. D-31
) ( Shi i i i SENSITIVITY CALCULATION O BASE CALCULATION _-1000
-m. m
.M 12-uv 700-- --800 v e- o .
- L 6 O i 3 j 600 0 -600 o
,ai u- u e e a Q.
E 500 - 0 - E
-400 e 4-400 - -
-200
. 300 O 50 10 0 150 200 Time (s)
. Figure D-50. The eff ect of the jet pump nodolization on cladding temperature of, the center o f t he hea t ed leng t h in the high-powered channel.
.300 , , , 300 m RUN TIME n-
-250 C
e
.'cN 200-- --200 0
.h
-N m- -
-15 0 c.
7 O e- 10 0 --' --10 0 -
-E c -
-50 3
(2: 0 O O 50 10 0 15 0 200 1 Time (s) ! F i gu r e D-51. Compu t er run time. l D-32 J
e.l.. APPENDIX E 80% LARGE-BREAK LOCA i G l' 4 E-1
{ .
- ~
APPENDIX-E-80% LARGE-BREAK-LOCA
. This. appendix describss the results of the TRAC-BD1 calculation for a U ihypothe'tical, 80% 1arge-break, floss-of-coolant. accident. The accident was 1
assumed to be initiated by a 80% double-ended, offset shear break in the recirculation loop piping. The calculation performed with an input deck .
, 4 , identical to the'one used to make.the calculation. described in Appendix B
' ~
except that'the break area was reduced. '
, 4
. Table E-1.'11sts the ' sequence of significant events occurring in the J ' calculation. The results of the~ calculation'are documented in Figures E-1 Dthrough E-49. The:results are' generally. similar to those discussed in
~ Appendix D except that events generallyJoccured later because the smaller
- break area-resulted in a slower blowdown. .Only those results which do vary
-significantly'from those described in Appendix D will tse discussed below.
' Lower plenum refill began at 98 s when the fluid mass, shown in i
Figure E-14, reached its minimum value. The start of refill was caused by leakag'e from:the bypass,.through the channels, down into the lower plenum. CCFL breakdown lat the upper tie plate in the low powered channel did not occur until near 130 s. (In the calculations described previously, lower' plenum refill-was started by CCFL breakdown in the low powered channel.)
< ! Also, CCFL~ generally was not calculated at :the upper core grid, which separates the bypass and upper plenum, during the LPCS injection period.
' Consequently, the: upper plenum. mass, shown in Figure E-17, did not rapidly
-increase after~the start of LPCS as most of the. liquid flowed down into the bypass. .
The thermal respons~e of the high powered channel, shown in - Figure'E-39,-was similar to that described in Appendix 0; rewets were
- calculated prior to the start of reflood. The peak cladding temperature was.699 K-and occurred at 140 s in the cell just above the peak power zone as shown in Figure- E-42.
e E-2
~
'h% A g 4m ,.- - .
- s:
sj.. o TkBLE(E-1. . -SEQUENCE,0F EVENTS Time , (s) ,
- Event
[.; i;
}. '0. 0 - Break: opened; loss of offsite power
_0.5: Scram signal generated; feedwater terminated 2.0 Low water level signala 3.0 .CRD flow : terminated 4.3 Low-low water level signala- _ 4.5 Control . rods fully inserted
' 5.0 - .MSIVs closed-9.1' . Jet pump suction uncovered b-17.l' Recirculation lines uncoveredb-20- Lower.-plenum' flashing 124.3 'HPCI initiated-t, 32.5 . Intact loop. isolation-valve closed
~68.8 -LPCS initiated 78 Dryout at the peak power zone c 94.3- ' Rated LPCS deliveredd 98~ Lower plenum refill started
. .. 128.5 HPCI isolated
. 135 Fine mesh turned on; HPCI system deleted; jet pumps renoded; intact loop pump speed set;to zero p
F n E-3
p n b [ I TABLE.E-1.. (continued) iTime (s) ' Event s 140 -Peak cladding temperature obtained 160 Calculation terminated
; -a. :The low and low-low water level signals correspond to collapsed down-1-
comer liquid levels of 12.80 and 11.28 m, respectively, above the bottom of the, lower plenum.
- b. The , times at which the jet pump suction and recirculation lines uncovered correspond to ~ the times the collapsed downcomer liquid level dropped to the elevation of the jet pump suction and the top of.the reactor vessel . outlet nozzle,-respectively.
-c. Rated LPCS' corresponds to a flow of 4500 gpm per pump which is tha flow
~
required by the Dresden 3 Technical Specifications.
-d.- - Dryout corresponds to the time the void fraction in the center of the heated length of the high-powered channel approached unity.
l i l l 9 ! E-4 [ L ..
The 80% large-break LOCA calculation was performed with a CDC-176 , computer. The computer run time for the model is illustrated by Figure _E-50 which shows the ratio of central processor seconds used to calculate each real second during the transient. The average run time was 125 central processor seconds for each real second. The average time step size during the calculation was 0.009 s. The code executed more reliably a for this transient'than the one described in Appendix B. Only one
- execution failure, which was eliminated by momentary reduction in time step size, occurred during the calculation. A faster run time could have been obtained by increasing the maximum time step from the input 0.010 s.
e E-5
t'
'3000 . .
! TOTPOW000001 2500 - - m
'3
- 2 .2000 - -
.v 3
@ 1500 ;
L 3: . 1000 - o n. 500 . .. 0 O 50 10 0 15 0 200 l Time (s) Figure E-1. Core power. 8000 i i i P300901
' -1000 m
^ o o .
$;6000 _
E v .. .
-800 v 9 o L L 3 3 m n
- n. _-600 n e 4000 -
e L L Q. Q. . e . -400 o E E 3' - 2 ' 2000 - - ! o. o
> . -200 >
o O l 0 50 10 0 150 200 l Time (s) F i gu r e E-2. Upper plenum pressure. r i I. i E-C
J 8000 , , , MFLOW190001 n 6000 - 7 N
- o(. E
.o x '.
,l"
* - -10000 C g 4000 .
y e .s
-M M M
2 2000 - 0 0 0' 50 10 0 15 0 200
-Time (s)
Fi gu re E-3. Moss fl_ow rate out the vessel-side break. 4000 , , ,
~
MFLOW200001 - 8000
- - - 6000 m
-n m 2000 -
- 4000
. ,3 f
" a 7' 2000 )
o - C 3
, ^
--- M m 0--
--0 w c 0 2 2
^ - -2000
~~
-2000'
'0 50 10 0 15 0 200 Time (s)
F i gu r e E-4. Moss flow rate out the pump-side break. E-7
l l 1
-800' ., , ,
HPCI ) O LPCS -
-1500 n 600 - -
Q
\
E a -. . x .o
-1000 C D ~ ~
O - 5 c.. O C m m m o ~
-500 m
'2 200 -
00 0 0 C 'O O ---O O 50 10 0 15 0 200 Time (s) F I gu r e E-5. HPCI and LPCS mass flow rates. 2000 ' i
- 4000 MFLOW150005 1000 _- -- 2000 m- 7 0-- --0 E oi
! x .o v _.
v
.- _- -2000
{ _1000 - O C m -
- -4000
^m -2000 - - m +
c m ! s 5- o i
- -6000
--3000 - -
- -8000
-4000 O SO 10 0 15 0 200 Time (sh Figure E-6.
Moss flow rate through the broken loop jet pump driveline nozzle. E-8
(.i + p -
'i 1 !
( ; i J 2000 ... I- 4000 V/f.0W150004 1000 - _ 2000 n a N N 0 -- [< l i, lh ,p!W! I !
--0 Cp .
.o 3 g p v l- _ -2000 ,
A
,. g : .iogo - .
= 2_
m -
'. - - -4000
- M m -2000 ,
e c (~ ) - .g 2 - -6000 ,
-3000 -
> -8000
-4000 O 50 10 0 15 0 200~ '
Time (s) . Figure E-7. Moss . flot rate through the broken loop let ,,'
?
pump dischorgo. /
j j 2000, -r ,
i i.
- 4000 *
. MFLOW150001
- 2000
.1000 -
t- E
^
a 0- (- --0 >s'
] myv y E,
'b i ,
-2000 g l -1000 -- ' 2 '
; c. %
e - ' '
- -4600 p
, g -2000 ,, c 3
- - -6000
. - 2
-3000 -
i- -8000
-4000 '
O , 50 10 0 15 0 200
, Time (s)
Figure E-8. Mass flow r o t'e# through the broken loop jet pump secJlon; ,
,r i* , . .
e 4 /' 3* *
/ ;
- E-9
3000 , , , MFLOWO50005 - 6000 n 2000 - 7
- 4000 N
_m(- E . 5 e v
- i ej -// { '1000 ,- -- 2000 A o
5 o t/s
-O 2 0-- - ^ --0 $
' ' ' - -2000
-1000 '
0- 50 10 0 15 0 200 Time (s) Moss flow rate through the intact loop jet pump
-! Figure E-9
./. driveline nozzle. e 7500 , , , MFLOWO50004 "iOOO i \
^
5000 n
- 10000 N
, ,Q E
o
'5 v e
,I- { 2500,- -- 5000 2
3 b
,/ *
/: m
'h'. J
. .J$'*
1 0-~ fffy b$ ~'o E -
,1 I
' ' ' ~ -sc00
_2500 - 0 50 10 0 15 0 200 Time (s)
.,. . Fi gu r e E-10. Moss flow role through the intact loop let
'; pump dischorge.
t. E-10 a . -
1; 1 I 7500 .i i .. MFLOWO50001 - 15000 m 5000 -' - 7
- 10000 N
. :a E
.x
.Q
! O
{ 2500, -
-- 5000 3 e .o M
e m 0 M 2 o._ __
=,_. ._ --0 $
, , , - - 000
-2500 0 50 10 0 15 0 200 Time (s)
Figu re E-11. Moss flow rote through the intact loop let pump . suc flon. 15 , , , LLEV300001
-40
^ - -
O v E .10 -
-30 v e e o o C C
.0 0
+- *
. n
-20 m o 5 - -
o
.to O 0 0- 50 10 0 15 0 200 Time (s)
Fi gu r e E-12. Downcomer liquid levet. E-ll ._ . . . .. ---.
U 100000 , , i DCMASS
- 2.00*10' 75000 -
- - 1.5 0*10' ^
- E 9 e
-s -
,.; . - 50000 -
- - 1.0 0*10' $
h j 2-25000 ,
-- 5.0 0*10' 0 0.00 0 50 10 0 150 200 Time (s)
Fi gu r e E-13. Downcomer fluid mass. 60000 , , i LPMASS
- - 1.2 0*10'
- - 1.0 0*10' 40000 - n
-^ E
- 8.00*1r* a f -
v n
- - 6.00*10' E
.$ j .
s -
-20000 -
4.0 0*10' tart of CCFL - 2.0 0*10 Start of refill breakdown
' ' ' O.00 0
0 50 10 0 - 150 200 Time (s) Fi gu r e E-14. Lower plenum fluid mass. i l E-12 l 1
. . , . - =,
D. t L 40000 , , i lI' MGT-
-80000 30000 - -
-60000 m n -
E i t' E
.V e y
. 20000 - -
S . [ -40000 $ C000 - -
-20000 0 O O 50 10 0 15 0 200 Time (s)
Fi gu re E-15. Guide tube fluid mass. 30000 . . . BYMASS
-60000
-50000 20000 - -
n
^
E
.]
. v
-40000 .o O
# f -30000 $
r -
/1 f
$ f \ \ &
10000 - - ii s j -20000
-10000 0 ' ' '
O O- '50 10 0 15 0 200 Time (s) Figure E-1G. Bypass fluid mass. E-13
15000 , , , UPMASS
~
-30000
~
-25000 9 '
-20000 6_
E . j y 1 -15000 $ 5000 1 r -10000 LPCS initiated I l
~
jM -5000 V h 0 ' i i 0 50 10 0 15 0 200 Figure E-17. Time (s) Upper plenum fluid mass. 20000 + i , CMASS
~
-40000 15000 -
~
g -30000 m 6 5
, 10000 -
j f{\ -20000 $ e U
- 2 h
5000 - I M _
,i
. 4 i
-10000 0 ' r .
0' 50 10 0 15 0 200 Fi gu r e E-18. Time (s) Core fluid moss. E-14
I: i.~ i l. 4000 . . i t MFLOW450001 8000 3000 - -
- 6000 m n m
" N NOn2000 1- - 4000 ,a E
! v
^* 1000 .
.{
-- 2000 3
- o 0 ~~0 l 2 WWf \"
f i
)
I 2 O
-1000 --
' -- -2000
' '~
-2000 -
0 50 10 0 15 0 200 Time (s) F i gu r e E-19. Most flow rate at the inlet of the low-powered channel. 5000 ' ' ' 10000 MFLOW450010
- 5000 m m
" I N
)
6 0-- y. ..;WI t l j. --0 2 E v y - I - -5000 o I 3
- 0
. n -
- -10000 g -5000 - -
g 2 O
- -15000 I -
- -20000
-10000
- O 50 10 0 15 0 200 Time (s) i Figure E-20. Moss flow rate of the outlet of the low-powered l chonnel.
L 1 E-15 L
8000 , , , MFLOW430001
-15000 6000 - -
9(- N m -
-10000 E '
0 6 4000 - - v 3 3 O 2000'- --5000 $ M N $ 2 0 2 A --O II '9f " nl 5 0-- r -- - - (
-2000 O SO 100 15 0 200 Time (s)
Figure E-21. Moss flow role at the inlet of the overage-powered chonnel. 8000 , , , MFLOW430010
-15000 6000 - -
^ $ N 4m -
-10000 E D
6 4000 - v 3 5 0
- 00 E 2000'-
a . 2 0 0-- ^ = = ^- \
--O
-2000 O 50 100 15 0 200 Time (s) l Figure E-22. . Moss flow rate at the outlet of the overage-powered channel. l l
)
E-lE
7-7 2000 i i i MFLOW410001 - 4000
- 3000 m m n M- g
\D 1000 -
E
- 2000 1 ~ -
-f -
1000 3 O 0- - -
--0 4
7,q j, y iy 'tg
, o 2
- -1000
~
-1000 0 50 10 0 15 0 200 Time (s)
F i gu r e ' E-23. Moss flow rate at the inlet of the high-powered channel. 2000 , , , MFLOW410010 - 4000
- 3000 ,
^ m 1000 - -
f -
- 2000 j v -
v 3
> o -
- 1000 3 C 2 o._ J- -
-fd 4 7 --0 y 1 0
-* . 2
- -1000
, , , - -2000
-1000 0 50 10 0 f.50 200 Time (s)
Figure E-24. Moss flow rate of the outlet of the high-powered channe1. E-17
6 i . i CHANLL45
- 15
^ 4 -
O
~
v E \ v e o . g_ - -10 g
. . n g e
a m
- 3: 2 l 3
- -5 k
0 O
.0 50 10 0 15 0 200
. Time (s)
..s. Figure E-25.. Collapsed IIquid level in the low-powered channel.
6 . .. i CHANLL43
- -15
^4 _
O
+
E
- m.
e e
'y - -to g O O n m o 2 c
- -5 l l
s .,
' ' ' O 0
O 10 10 0 15 0 200 Time (s) F i gu r e E- ,.d.
' psed liquid level in the overage-powered channel.
.E.18 . _ _ _ . _ _ __ _
\
6 i , , ,- CHANLL41 -
- -15
^
4 - - O
.- .E -
v
.i e o e o . .jo o C C 0 0 n m 32 - -
3
-5 0 O O , 50 10 0 15 0 200 Time (s) ,
Figure E-27.- Collapsed liquid level in the high-powered channel. 50 , , .
- 15 0 LIQUID 40 ' ~
O VAPOR g - -- 10 0 g N N v E 20 - - e v
- 50 x 10 - -
x-
, 0 O- O -O- --0 L e I e
> -n - -
y
- -50 y
. -20 - -
. 2 2 La. -
-- -10 0 L
-40 - -
- -15 0
-50 O 50 10 0 15 0 200 Time (s)
Fi gu r e E-28. Fluid velocities at the upper core grid (third vessel ring). E-19 u
i 50 , , .
- 15 0 LIQUID 40 - ~
O VAPOR g 30-- -- 10 0 g N N E 20 v-v
- 50 >. 10 - -
x o 0 0= C
~
-- 0 0 o o o e
> -10 - -
- -50 y
-20 - -
E ? L ' - -- -100 L
-40 - -
- -15 0
-50 O 50 10 0 15 0 200 Time (s)
Figure E-29. Fluid "-locities of the upper core grid (second vessel ring). 50 , , , , 40 - LIQUID ~ O VAPOR g 30-- -- 10 0 g N N v E 20 - a v g - 50
>. 10 - -
O \ = 8 09 C -- O g 2 -10 - - a - -50 y
.- -20 - -
E E
- 6, - - _10 0 L
-40 - -
- -15 0
-50 O 50 10 0 15 0 200 Time (s)
. Figure E-30. _ Fluid velocities at the upper core grid (first vessel ring).
E-20
1 2500 , , ,
- 5000 MFLOW550004 m !
i Q 0 ^ M _7 --b; -- ,, Q
--0 ? b i S
- i I Y
=
- -5000 E -2500 ~ - - y, 2
1
- -10000
.-5000 0' 50 10 0 150 200 Time (s)
F i gu r e E-31. Moss flow role at the inlet to the bypass (third vessel ring). 4000 , , , MFLOW530004 00 n 2000 - - q
\
cn E 5 e Ih h ) f m
= =
. -2000 ,
- -5000
-4000 O 50 10 0 15 0 200 Time (S)
Figure E-32. Moss flow rate of the inlet to the bypass (second vessel ring). E-21
1000 ' , i _ - 2000 MFLOW510004 n 500,- 1000
" \
N- , E cn a ) , n - 0- k W l - 0 ' i C
- ~~
-500 ~ - -
l
- - -2000
-1000 O 50 10 0 15 0 200 Time (s)
Figure E-33. Moss flow rate of the inlet to the bypass (first vessel ring). 400 . , , MFLEAK450001 - 800 300 - -
- - 600 m a m " N 200 - -
E
- 400 5 \ E v $ 100_- l -- 200 3:
- l S
, ~
E 0-- --O m l .
-10 0 - I -- -200 j
' ' ' - -400
-200 0 50 10 0 15 0 200 )
l Time (s) Figure E-34. Leakage from - the low-powered fuel channel. E-22
500 i i i
- 1000 400 - -
300 - - m n
- n 200
- -- 500 \
N
-. m E
.o j 5 100 - -
. ) o_ .
--0 y o o
- i 0 -200 2 - '
i _- -500 0 2
-300 - -
-400 - -
- -1000
-500 O 50 10 0 15 0 200 Time (s)
Figure E-35. Leokage from the overage-powered f uel channel. 10 0 i e a
' ~
MFLEAK410001 n g 50 , -
-_ 10 0 e N
m E
.o U C D o_,. --0 y O
+ o
- E y
3 . I
- -10 0 0
, 2
- -200
- 10 0 O 50 10 0 15 0 200 Time (s)
Figure E-36. Leakage from the high-powered f uel channel. E-23 L
600 , , , X -00 CLADDING Y O VAPOR I A LIQUID 550 -
~
O CHANNEL WALL n -
-500 ^
M V-v v
- 500
~ '
- e 3 3
+
-400 O '
o
. t- u
[450 - { E E y -
, -300 y 400 -
-200 350 O 50 10 0 15 0 200 Time (s)
Figure E-37. Thermal response of the low-powered f uel channel. 600 , , ,
-600 y X CLADDING O VAPOR E
0 - A LIQUID ~ O CHANNEL WALL g v
-500 C L
- 500 -
A - e 3 3
+
-400 U
L- U g _ [450 - { , E - E [ -
-300 0 400 -
-200 350 ' ' '
O 50 10 0 150 200 Time (s) Figure E-38. Thermal response of the overage-powered f uel channel. 1 E-24
, 3
+
*E 1 b0+3ue vg $E $* v2 3 4 4 5 5 6 2 3 4 5 6 7 5 0 5 0 5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
- - - - O - - E)
F 0 - l F
- - - 5 -
- - - - i i
g g u r u r AOX e e E LVC E I AL 4 3
- ^QPA UOD 0 9 I RD D I 5 5 N A 0 ' , T 0 '
G , l x h o i e ^ r wal
- m p a oc l
'A E wi 'S
- e c d r
e 2 r 5 e d s T i T p d ni oi c h t g m1 0 nm1 s e e0 0 a e e0 n ep m o(s n (s f le) t) r a 0+OAOX h e s t u h r CCCCCC EEEEEE g i X o LLLLLL h d 1 5 LLLLLL - 1 5 i 0 p 0 , s ' 654321 . o t r ( ( w i B e b C O r u E N T e t i T T d o E O n R M f
) ) u f e l
o r 2 c 2 0 h 0 - ~ - - t 0 - - - ~ 0 h - - - - b a . - - - e 2 3 4 5 6 n o 2 4 6 8 0 0 0 0 0 n 0 0 0 0 0 0 0 0 0 e l 0 0 0 0 SE {6c+3 S LC $E $* vh
800 , , i X CELL 1 (BOTTOM) O CELL 2 A CELL 3 O CELL 4 (CENTER) -800 ^ + CELL 5 ^ b 650 . O CELL 6 _ b *
, 8 CELL 7 (TOP) ,
3 -
-600 2 0 $ c 5 s o.
o. E 500 - - E t -. -400 g 350 O 50 10 0 15 0 200 Time (s) F i gu r e E-41. Axici clodding temperature distribution for the overage-powered channel. 900 , i i X CELL 1 (BOTTOM) O CELL 2 A CELL 3 --1000 800 - - O CELL 4 (CENTER) g + CELL 5 C v O CELL 6 L e 700 6 _ @ %7M _.gog o L
? ?
o o L ' f --600 $ " E600l-E e
. E o
H I--
; r g .
500 - o a
'N -400 400 O 50 10 0
' oI'O
~
15 0 200 Time (s) Figure E-42. Axiol clodding temperature distribution for the high-powered channel. E-26
7 _-- 1 i. -r- - i P
~
c yh
.2 a
O o ^E 2o ALPHA 410006 0.4 L' O 50 10 0 15 0 200 Time (s) Figure E-43. Vold fraction at the center of the heated length In the high powered channel en
/
E-27 m
1d = i i i = RODHL450104 !
~
_~ -
'~
1d \
.3 5- A- - - M 'h 5 Q b b
- oi on _
uE 11 7 2 3
%h m 5
E Cv : o _
~'
10" r i o . : e . y . 10-' o.o 50.0 10 0.0 15 0.0 200.0 Time (s) Figure E-44. Heat-transf er coefficient to liquid in the low-powered channel. 1 i i i RODHV450104 e 0.8 - o C
'm
$y 0.6 -
os uE
* ) o.4 mu - -
gv ' A E 0.2 -
, , N ,
o 50 10 0 15 0 200 Time (s) Figure E-45. Heat-trans f er coefficient to vapor in the low powered channel. . E-28
r 1d = . . . =
- RODHL430104 g
-N ~ -
u 1d -
? -5 e :. 1_ . {l :
$g t o -
oi un uE 11 g- 1 7 Oh ( : 8 5 10" 5 r if 5 3 S i 3 x - 10 - - 0.0 50.0 10 0.0 15 0.0 200.0 Time (s) Figure E-46. Heat-tronsf er coefficient to liquid in the overage-powered channel. 1 i i i RODHV430104 E 0.8 -
$y 0.6 - -
ou uE Oa2 h 0 .4 - - gv 2
] 0.2 - -
0
, i l ,
0 50 10 0 15 0 200 Time (s) Figure E-47. Heat-transf er coefficient to vapor in the overage powered channel. E-29 u-.
1d ; s , > = N \ RODHL410304 - E* - % ( ( - c 1d r 5
)\'
I ,%r t 2 E .mog O ) olu uE 16 r " 3 EN i t ? m - 8 _v : - 10" r i g : : 1 - - 10-* O.0 50.0 10 0.0 15 0.0 200.0 Time (s) Figure E-48. Heat-trcnsf er coefficient to liquid in the high-powered channel. 1 e i i RODHV410304 E 0.8 - -
~m
$y 0.G - -
on uE , 2m _y) 0.4 - l 5*
.h 3 0.2 -
h(4 N Y f' & 0 0 50 10 0 15 0 200 Time (s) Figure E-49. Heat-transf er coeficient to vapor in the high-powered channel. E-30
- J 4
'300
' i
- m
, n RUN TlWE C
..e. .
M C 200 -
, :O
.A:
.O.
N e e a, v o e .- 100 E -
- c C
. 3.
. ct:
0 , 200 TINE Figure E-50. ' Computer run time. (s)
'O i
E-31. - - . - - - . - . . __._}}