ML20116P119
| ML20116P119 | |
| Person / Time | |
|---|---|
| Site: | Oconee  |
| Issue date: | 04/30/1985 |
| From: | Jo J, Pu J, Rohatgi U, Saha P BROOKHAVEN NATIONAL LABORATORY |
| To: | NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| References | |
| CON-FIN-A-3266, REF-GTECI-A-49, REF-GTECI-RV, TASK-A-49, TASK-OR BNL-NUREG-51750, NUREG-CR-3703, NUDOCS 8505070497 | |
| Download: ML20116P119 (98) | |
Text
__
I BNL-NUREG-51750 i
ASSESSMENT OF SELECTED TRAC AND RELAP5 CALCULATIONS FOR OCONEE-1 PRESSURIZED THERMAL SH0CK STUDY U.S. Rohatgi, J. Pu, P. Saha, and J. Jo Manuscript Completed - March 1984 Date Published - November 1984 LWR CODE ASSESSMENT AND A'PPLICATION GROUP DEPARTMENT OF NUCLEAR ENERGY, BROOKHAVEN NATIONAL LABORATORY UPTON, LONG ISLAND, NEW YORK 11973 i
f
{), t Prepared for ggi j jj United States Nuclear Regulatory Commission
~
Washington. DC 20555
$80' $8$$ S!8$S069 P
, 7 ],;_
NUREG/CR-3703 BNL-NUREG-51750 AN. R-4 L
l ASSESSMENT OF SELECTED TRAC AND RELAP5 CALCULATIONS FOR OCONEE-1 PRESSURIZED THERMAL SH0CK STUDY U.S. Rohatgi, J. Pu, P. Saha, and J. Jo LWR Code Assessment and Application Group Department of Nuclear Energy Brookhaven National Laboratory Upton. Long Island, New York 11973 Prepared for UNITED STATES NUCLEAR REGULATORY COMMISSION OFFICE OF NUCLEAR REGULATORY RESEARCH WASHINGTON, OC 20555 CONTRACT NO.DE AC02-76CH00016 FIN A-3266
NOTICE This report was prepared as an account of work oponsored by an asency of the United States Government. Neither the United States Government nor any agency thereof,or any of their employees, makes any warrenty, expressed or implied, or soeumes any legel liability or responsibility for any third party's use, or the results of such use, of any information, apparatus, product or procese disclosed in this report, or represente f
that its use by such third party would not infringe privately owned rights.
The views expressed in this report are not necessarily those of the U.S. Nuclear Regulatory Commission.
Available from GPO Sales Program Division of Technical Information and Document Control U.S. Nuclear Regulatory Commission Washington, D.C. 20655 and National Technical Information Service Springfield, Virginia 22161 E
EXECUTIVE
SUMMARY
Several Oconee-1 overcooling transients were computed by LANL using TRAC-PF1 and by INEL using RELAP5/M001.5.
These calculations and the input decks were reviewed by BNL.
Three of these transients were selected f or detailed review as they either had the potential of challenging the integrity of the pressure vessel or highlighting the effect of the code differences. The three transients selected were: (1) main Steam Line Break (MSLB), (2) All Turbine Bypass Valves Stuck Open, and (3) 1-Inch Small Break LOCA.
Comparison of the computations of the MSLB transient indicated that the difference in the minimum downcomer fluid temperature predictions was due to the modeling of the control system that regulated the MFW and EFW pumps, to the multi-dimensional effects resulting in different temperature histories for the hot legs, and to the modeling of the upper head as through-flow or dead-end volume. Both calculations had some weaknesses and a method has been des-cribed to predict the best-estimate downcomer fluid temperature using INEL's early temperature prediction and RCP restart time and azimuthal temperature distribution from LANL's calculation.
The second transient compared was initiated by the f ailure of all four TBVs at the full open position af ter a turbine trip. This transient is like a smf l break in the steamline.
Here the initial conditions (full power for l
TRAE and hot standby for RELAPS) and additional f ailures were different.
The codes predicted the transients reasonably well.
However, the important rea-sons for the differences in the calculations were the controller f ailure to throttle EFW based on the secondary side level and the f ailure of the operator to close the TBVs at 600 seconds in the TRAC calculation.
The third transient compared was the Small Break (2-inch) LOCA in a hot leg.
The ICS was assumed to work as designed. Both TRAC and RELAPS computed a continuous drop in the primary side pressure during the transient. This made the transient less critical for PTS. However, comparison of the two calcula-tions indicated that there were many differences. The codes modeled the reac-tor trip differently. RELAPS correctly based it on the low primary side pres-sure while TRAC assumed it to occur at 0.5 seconds.
Furthermore, af ter the loss of natural circulation due to candy cane voiding, TRAC computed flow oscillations in the cold legs while RELAP5 predicted a stable circular flow between the cold legs connected to the common steam generators.
The loop oscillations in the TRAC calculation warmed up the cold leg and the downcomer fluids.
However, it is not clear if these oscillations are real.
- Moreover, the TRAC calculation was not carried out far enough in time to determine the minimum downcomer fluid temperature with confidence.
The RELAPS calculation, on the other hand, is more complete and looks reasonable.
Both codes were reasonably successful in modeling these transients.
The major differences in their results were due to the differences in modeling the plant, control systems, and event sequences, and to the one-dimensional model-ing of reactor vessel thermal hydraulics by RELAP5.
- lii -
ABSTRACT Several Oconee-1 overcooling transients that were computed by LANL and INEL using the latest versions of TRAC-PF1 and RELAPS/ MODI.5 codes have been reviewed by BNL. Three of these transients were selected for detailed review as they either had the potential of challenging the integrity of the pressure i
vessel or highlighted the effect of code differences.
These are (1) Main Steam Line Break (MSLB), (2) All Turbine Bypass Valves Stuck Open, and (3) 2-Inch Small Break LOCA.
Both codes were reasonably successful in modeling these transients.
How-ver, there were differences in the code results even though the specified scenarios were exactly the same for two transients (MSLB and Small Break LOCA).
This report compares the code results and explains the possible rea-sons for ' these differences.
Recommendations have been made regarding which result seems more reasonable for a specific transient.
-V-
ACKNOWLED(NENTS The authors would like to acknowledge the efforts of Mrs. Ann Fort in typing this report -and preparing the figures.
They are also grateful to Don Fletcher. of INEL and B.
Bassett and J.
Ireland of LANL for providing information and clarifications about their calculations.
9 l
1 l
l i
l-I i
.yt.
E?
- TABLE OF CONTENTS Page Execu ti ve Summa ry...........................
iii Abstract...................-.............
V l
Acknowledgment............................
vi List of Tables................'............
Vii List of Figures............................ viii 1.
Introduction............................
1 2.
Main Steamline Break........................
3 Failure of all Turbine Bypass Valves at Full Open Position.....
28 l
3.
4 Small Break (2-INCH) LOCA in Hot Leg................
37 5.
Summa ry and Concl us i ons......................
55 6.
References.............................
56 i
Appendix..............................
57 LIST OF TABLES Table No.
Title Page l
2.1 Main Steamline Break Scenario..............
4 2.2 Sequence of Events for MSLB Transient..........
5 3.1 Comparison of the LANL and INEL Scenarios for all TBVs Stuck Open Transient............
29 3.2 Sequence of Events for All TBVs Stuck Open Transient...
30 4.1 Small Break LOCA in Hot leg Scenario...........
38 4.2 Sequence of Events in the 2-Inch Hot Leg Break......
39
- vli -
l L
'l
LIST OF FIGURES Figure No.
Title Page 2.1 Prima ry Side Pressure..................
7 2.2 Steam Generator A Secondary Side Pressure........
7 2.3 Steam Generator B Secondary Side Pressure........
8 1
2.4 Steam Generator A Secondary Side Liquid Temperature...
8 2.5 '
" team Generator B Secondary Side Liquid Temperature...
9 2.6 Steam Generator A Secondary Side Vapor Temperature...
9 2.7 Steam Generator B Secondary Side Vapor Temperature...
10 2.8 Hea t Tra ns fe r Ra te i n SGA................
10 2.9 He at Trans fe r Rate in SGB................
11 2.10 HP I Fl ow t o Col d Leg A-1................
11 2.11 HPI Fl ow to Col d Leg A-2................
13 2.12 HP ! Fl ow to Col d Leg B-2................
13 2.13 Downcmer Fluid Temperature...............
14 2.14 Ve n t Val v e Fl ow.....................
14 2.15 Co l d L eg A-1 Fl ow R a t e.................
15 2.16 Col d Leg A-2 Fl ow Rate.................
15 2.17 Cold Leg B-1 Fl ow Ra te.................
16 2.18 Col d Leg B-2 Fl ow Rate.................
16 2.19 Cold Leg A-1 Fluid Temperature.............
17 2.20 Cold Leg A-2 Fluid Temperature.............
17 2.21 Cold Leg B-1 Fl uid Temperature.............
18 2.22 Cold Leg B-2 Fluid Temperature.............
18 2.23 Upper Pl enum Void Fraction...............
19 2.24 Candy Cane Void Fraction in Loop A...........
19 2.25 Candy Cane Void Fraction in Loop B...........
20 2.26 Hot Leg Temperature i n Loop A..............
20 2.27 Hot leg Temperature in Loop B..............
21 2.28 Hot Leg Mass Fl ow Rate in Loop A............
21 2.29 Hot leg Ma ss Flow Rate in Loop B............
22 2.30 Mai n Ste mli ne Break Fl ow................
22 2.31 Ma i n fe edwa t e r Fl ow to SGA...............
23 2.32 Ma i n Fe edwate r Fl ow t o SGB...............
23 2.33 Eme rge ncy fe edwa t e r Fl ow to SGB.............
24
- vili -
1 k
' LIST OF FIGURES (Cont.)
l Eage Figure No.
Title a
l 2.34 Onergency Feedwater Fl ow to SGA.............. 24 2.35 Temperature of Emergency Feedwater to SGA......... 25
[
2.36 Temperature of Dnergency Feedwater to SGB.
25 2.37 Estimate of Lowest Average Dowscomer Fluid Temperature.. 27 3.1 Pres s ur e i n R. V. Down c one r...... -.......... 31 3.2 Fl ow Thro ugh TB V-A.................... 31 3.3 Fl ow Th r oug h TB V-B.................... 32 3.4 Fluid Temperature in R.V. Dow1 comer............ 32 3.5 Ma i n Fe edwa t e r Fl ow i n SGA................ 33 3.6 Mai n Feedveter Fl ow i n SGB................
33 3.7 Steam Generator Secondary Inventory - Loop A.......
34 3.8 Steam Generator Secondary Inventory - Loop B.......
34 3.9 SGA Indicated Liquid Level (Operating and Startup).... 35 3.10 SGB Indicated Liquid Level (Operating and Startup).
35 l
4.1 Down c ane r Pres sure.................... 40 4.2 Ma s s Fl ow Rat e Out o f B re a k................ 40 4.3 Static Quality in the Surge Line Volume With Break.... 41 4.4 Void Fraction in the Surge Line Volume With Break..... 41 4.5 HPI Fl ow to Cold Leg A-1.... -.............
42 4.6 HP! Fl ow to Col d Leg A-2.................
42
(
4.7 HP! Fl ow to Cold Leg B-1................. 43 j
4.8 HPI Fl ow to Col d Leg B-2.................
43 4.9 Candy Cane Void Fraction in Loop A............ 45 4.10 Candy Cane Void Fraction in Loop B............
45 4.11 Void Fraction in Upper Plenum............... 46
(
4.12 Emergency Feed Water Flow to Steam Generator A......
46 4.13 Emergency Feed Water Flow to Steam Generator B......
47 L
4.14 Steam Generator A Secondary Side Pressure.........
47 4.15 Steam Generator B Secondary Side Pressure.........
48 4.16 Steam Generator A Secondary Side Temperature.......
48 4.17 Steam Generator B Secondaar Side Temperature.......
49 4.18 Steam Generator Primary Side Inlet Temperature in Loop A.
49 i
i
- ix -
i i
.~
?
LIST OF FIGURES (Cont)
Figure No.
Title h
j i
4.19 Steam Generator Primary Side inlet Temperature in loop B.
50 4.20 Downcomer Liquid Temperature...............
50 4.21 Cold Leg A-1 Temperature.................
51 I
4.22
' Col d Leg A-2 Temperature.................
51 4.23 Cold Leg B-1 Tempe rature.................
52 4.24 Col d Leg B-2 Temperature.................
52 4.25' Lo op A Col d Leg Fl ow...................
53 4.26 Loop B Col d Leg Fl ow...................
53 i
t l
I i
i 1
I x
n.
b
f I
I 1.
INTRODUCTION Rapid cooling of the reactor pressure vessel during a transient or acci-dent accompanied by high pressure has the potential of producing severe the rmal stresses in the vessel wall and challenging the vessel integrity.
This phenomenon is called overcooling or Pressurized Thermal Shock (PTS). As long as the f racture toughness of the reactor vessel remains high, overcooling transients will not cause vessel failure.
However, the Nuclear Regulatory Commission (NRC) staff analysis (SECY-85-465) showed that certain older plants with copper impurities in the vessel weldments may become sensitive to PTS in a f ew years as the nil ductility transition temperature of the weld material gradually increases.
In late 1981, the NRC designated PTS as an unresolved safety issue and de-veloped a Task Action Plan (TAP-49) to resolve the issue. The NRC has select-ed three plants representing PWRs supplied by three vendors in the United States for detailed PTS study:
Oconee-1 (Babcock and Wilcox), Calvert Cliff s (Combustion Engineering), and H. B. Robinson (Westinghouse Electric).
Oak Ridge National Laboratory (0RNL) is the lead contractor f or the entire PTS study, and they have identified several groups of transients with multiple equipment f ailure and with no corrective operator action which could lead to severe overcooling in these plants.
The thermal hydraulic calculations f or these transients were to be calculated at the Los Alamos National Laboratory (LANL) and the Idaho National Engineering Laboratory (INEL) using the latest versions of TRAC-PWR and RELAPS codes, respectively.
The Oconee-1 transients were divided betscen LANL and INEL, with some transients common to both. The Calvert Cliff s and Robinson transients were assigned to LANL and INEL, re-spectively.
Brookhaven National Laboratory (BNL) was requested by the NRC to review and compare the plant input decks developed at LANL and INEL, and to review the calculation results.
Emphasis would be placed on explaining the dif f e-rences between the common calculations perf ormed at these two laboratories, and on recommending which result seemed more reasonable for a specific trans-1ent.
For transients computed at only one laboratory, a review f ocusing on the reasonableness of the results will be performed. This report presents the results of the BNL review of the selected Oconee-1 calculations perf ormed at LANL and INEL.
In the first part of the task, the BNL staff reviewed and compared the plant input decks f or Oconee-1 as prepared by LANL and INEL.
There were some dif f erences between these decks in the reactor vessel and heat structure de-scriptions. BNL also reviewed the models for control systems as developed by Science Application, Inc. (SAI) f or RELAPS and by LANL f or TRAC-PF1. The com-ments based on these reviews were transmitted to the NRC and the Oconee-1 PTS study participants through several letters and presentations during August to October 1982. For the sake of completeness, a copy of these communications is presented in Appendix A of this report.
Calculations f or twelve Oconee-1 transients specified by ORNL were divided between LANL and INEL (Fletcher, et al 1984 and Basset, et al 1983).
Some of them -- transients such as the Main Steamline Break (MSLB), 2-inch Hot leg....
Small Break Loss-of-Coolant Accident (SBLOCA), and the actual Turbine Trip j
transient at Oconee - were common to both the laboratories. The TRAC and RELAP5 results for these transients were compared at BNL.
It was also ob-served that the MSLB and the Turbine Bypass Valves (TBVs) stuck open tran-sients were relatively severe transients. Theref ore, all f our TBVs stuck open transients were also investigated.
The Oconee-3 turbine trip transient was the first calculation performed by both LANL and INEL.
The review of this transient indicated several differ-ences between the TRAC and RELAP5 code calculations and the data.
- However, af ter this calculation, the codes were modified and the conclusions f rom this transient are no longer relevant to the other transients. Moreover, the plant data are proprietary to Duke Power Company.
Therefore, this transient will not be discussed here.
This report will, therefore, discuss the results of the f ollowing three transients in details.
1.
Main Steamline Break (MSLB).
I 2.
Failure of all (4) Turbine Bypass Valves (TBVs) at full open position.
I 3.
Small Break (2-Inch) LOCA in Hot Leg.
[
The main steamline break transient was initiated at full power by a break of 34 inches in one of the steamlines.
This led to the depressurization of the secondary side. There were also other f ailures which influenced the course of l
'the transient.
The transient was computed by using both TRAC-Pf1 and RELAP5/M001.5 codes.
L The TBVs stuck open transient was initiated by the f ailure of TBVs to close when the pressure decreased in the steamline.
This was similar to a small steamline break and caused depressurization of the steam generator sec-ondary side.
However, the initial conditions (full power for TRAC, but zero hot power for RELAP5) and some additional f attures were intentionally diff e-i rent for the two code calculations.
The purpose of comparing this transient was, therefore, to investigate the eff ect of these differences on the final l
results.
The last transient discussed in this report is a small break (2-inch) LOCA in hot leg initiated at full power.
The transient had no other f ailures or operator actions other than tripping the Reactor Coolant Pumps (RCPs) 30 seconds af ter the initiation of High Pressure Injection (HPI).
Although the primary side did not repressurize, which made it less significant to PTS, the calculation showed the sensitivity of the results to the codes.
These three transients are described in detail in the next three chapters.
In general, both the TRAC and RELAPS codes computed these transients reason ably well.
The crucial differences were not due to the thermal hydraulic models in the codes but to the way the plant, the control systems and the se.
l quences of events were modeled.
The only exception was the multi-dimensional reactor vessel thermal. hydraulics, which was modeled with TRAC but could not be modeled with RELAP5.
l i
l I
l 2-
- 2. MAIN STEAM LINE BREAK This is one of the set of transients in which the secondary side is de-pressurized. The initiating event is a (34-inch) diameter break in the steam-line. The transient scenario is made more severe by an operator delay in iso-lating the feedwater (FW) flow to the affected steam generator coupled with a delay in throttling the High Pressure Injection (HPI) flow and restarting Reactor Coolant Pumps (RCPs) in each loop af ter 75F (42K) subcooling is at-tained in the primary loop.
The scenario of this transient, as specified by ORNL and as shown in Table 2.1, was quite involved with various operator ac-tions for the primary and secondary sides.
This transient was computed using both TRAC and RELAPS codes (Fletcher, et al 1984 and Bassett, et al 1983) and the resu}ts indicated that it could have severe consequence for vessel inte-grity. A comparison of the results of the two codes will also show the sen-sitivity of the code results.
In this transient the primary side loses energy to the steam generators, specifically to the steam generator with a break in the steamline.
The de-pressurization of the steam generator caused a reduction in the saturation temperature which resulted in increased heat transfer f rom the primary side and a larger vapor generation.
The f ailure of the operator to stop or throt-tle feedwater provided additional fluid to the steam generator f or vaporiza.
tion, and cooling of the primary side.
The sequences of events as predicted by two codes have been summarized in Table 2.2.
In the remaining section the results f rom the two calculations will be compared.
Figure 2.1 shows the primary side pressure.
In general, RELAP5 not only computed higher pressures but also repressurized to PORV set point sooner than TRAC.
However, in the very beginning of the transient, RELAPS predicted a f aster pressure drop due to an early Main Feedwater (MFW) pump trip and int-tlation of colder Emergency Feedwater (EFW).
This is confirmed in Figures 2.2 and 2.3, in which secondary side pressures are compared.
RELAPS modeled the control systems on the basis of pressure dif f erential between the top of the tube region and the bottom of the downcomer in the steam generator and was in closer agreement with the plant control system.
On the other hand, TRAC mod-eled the control system on the basis of collapsed water level and therefore, missed the MFW pump trip which depended on the pressure differential. RELAP5, as expected, predicted a lower secondary side pressure than TRAC in the begin-ning of the transient, resulting in a lower Saturation temperature and a lar-ger heat transfer in the steam generator. The secondary side temperatures are shown in Figures 2.4 through 2.7.
The early rapid pressure drop calculated by RELAP5 also caused the initia-tion of HP! flows earlier than in TRAC.
As the RCPs were tripped 30 seconds af ter the HP! flow, the loop flows were in the natural circulation mode. The heat transf er decreased and the system started to repressurize. However, when the subcooling in the hot legs reached 42K, the RCPs were restarted in loops Al and 81 in both the calculations, which caused volds in the primary system to collapse, and the pressure dropped, as shown in Figure 2.1, at 300 seconds f or RELAPS and 526 seconds f or TRAC.
Af ter 600 seconds, when both the steam generators were isolated, the heat transf er f rom the primary side decreased.
This is shown in Figures 2.8 and 2.9 for RELAP5.
It seems that af ter RCPs were restarted, there was some reverse heat transfer in Steam Generator B for RELAPS.
This resulted in repressurization of the primary side, Also RELAP5 Table 2.1 Main Steam Line Break Scenario Initial Conditions:
1.
Full reactor power.
2.
Nominal temperatures and pressures in primary / secondary.
3.
Decay heat: 1.0 times the ANS standard.
4 Pressurizer spray / heaters operate as designed.
Sequence of Events:
1.
Reactor trips, coincident with break of (34") steam line.
2.
Turbine trips; TSVs close.
3.
ICS functions as designed.
4.
Protection systems on hotwell, condensate booster, and MFW/EFW pumps function as designed.
5.
HPI actuates at set point (1500 psig).
6.
Operator trips RC pumps 30 s after HPI actuation.
- 7.
EFW pumps start when low MFW pump discharge pressure is sensed.
- 8.
MFV/EFW system attempts to maintain 240" SG level.
9.
Core flood tanks actuate at set point.
- 10.
LPI actuates at set point.
11.
(a) Operator isolates feedwater to both steam generators 10 min into the transient (close MFW, start-up-FW, EFW, and TBV systems).
(b) Operator begins refilling unaffected steam generator to 240" level with EFW at maximum rate 15 min into transient.
(c) Turbine bypass system on unaffected steam line opens at 15 min and maintains 1000 psi pressure.
12.
Operator restarts one RC pump in each loop af ter attaining 75'F subcooling and throttles HPI to maintain 75 + 25'F
~
subcooling.
13.
PORV opens at set point (2450 psig).
- 14.
SRVs oper at set point (2500 psig).
15.
PORV/SRVs reseat at their set points (2400 psig).
- 17.
EFW surge tank capacity (72,000 gal) is exhausted; and the two motor-driven EFW pumps trip.
- 18.
Turbine-driven EFW continues to draw from hotwell.
- Event is phenomenology dependent.
1,
. - ~.
i l
Table 2.2 Sequence of Events for MSLB Transient UPDATED TRAC-PFI RELAP5/M001.5 ITEM (LANL)
(INEL);
- 1) Reactor & Turbine Trip 0.5 sec 0.0 sec
- 2) HPI Initiated 21.2 sec 5.3 sec
- 4) EFW to SG-A (Mfected) 29.4 sec, based on low level.
4.4 sec, based on low MFW pump discharge pressure.
e
- 5) EFW to SG-B (Unaff ected) 48.7 sec, based on low level. 4.4 sec, based on u.
a Tow MFW pump discharge pressure.
- 6) MFW Pump Trip 47.8 sec. Iow suction 0.3 sec, high level pressure.
in SG secondary.
- 7) Condensate Booster Pump Trip 51.2 sec
- 8) Vent Valve Flow 112 sec to 526 sec.
Valve did not open.
- 10) EFW to SG-8 terminated (or 346.7 sec, valve closed.
320 sec throttled) as a level of 240 inches is reached.
l l
1 I
i.
r-i
)
l l
1 Table 2.2 Sequence of Events f or MSLB Transient (Cont'd)
UPDATED TRAC-PF1 RELAP5/M001.5 ITEM (LANL)
(INEL)
- 11) Hot well surge tank empty. Motor 513 sec driven FW pump stopped.
- 12) Accumulator On 530.9 sec
- 13) Accumulator Off 537.9 sec
l 2354 sec
- 16) Pressurizer level reaches top.
- 17) PORY set point for opening reached.
4678 sec on 2432.0 sec l
i
- 18) Calculation terminated.
7200 sec 2697 sec
- 19) Minimum downconer fluid 405K at around 526 sec.
493K at around temperature.
600 sec.
MAIN STEAM LINE BREAK 1.8 i
i i
i i
)i $f
_ l.6 8
E f
~ 1.4 4
14 8
W
/
[ 1.2
/
/
d
/
E107
/,,/
~
5,,o PRIMARY SIDE PRESSURE,
B
- RELAPS y
i
~
y 6.0 I /,
4.0 - r alO' 2.00 2000 4000 6000 8000 REACTOR TIME (s) i Figure 2.1 Primary Side Pressure i
M18
,A1:4 STEAM Ll:#E BPEAK I
I i
IRAN 5!ENT f!T (t. IO)s)
_ _ _ taAc f
5.0 1
2.9 4.0 4
lete 2000 3000 4000 IIT (s)
Figure 2.2 Steam Generator A Secondary Side Pressure l
t MAIN STEAM LINE BREAK 8
i i
i STEAM GENERATOR 8 SECONDARY SIDE PRESSURE y,,,
s
- 6 s
o i
/
~
(
/
t M
I
- 4 i
f
/
- RELAPS
~!
f
--- TRAC -PFI
~
/
/
2 -i i /
w/
,j O
2000 4000 6000 8000 REACTOR TIME ts)
Figure 2.3 Steam Generator B Secondary Side Pressure i
Mall STIAM tlNE BREAK i
i i
TRMSIEli 114 (t WAh)
~
RELAP5
_5
--- ruc 1
5-a i
h i
-!e N
W W
i i
1 g
e sees sees asse
.ees fl4 (t)
Figure 2.4 Steam Generator A Secondary Side Liquid Temperature !
L
F-u MAIN STEAM LIE BEAK saa 5
i
,---v-,.--
/'
I
/
\\
/
/
s i
/
.I
[
$, s20
/
3 B
/
acues
's *
/
_ _ __ mc n
I i
3
/
l j/
48::
s
}
0 1000 2000 3000 8000 5000 6000 1000 8000 REACTOR fift (s)
Figure 2.5 Steam Generator B Secondary Side Liquid Temperature MAIN SitM LIE BREAK see I
l l
IRAN 51[NI IlE (t - 400s) f',
~,s-L 9m g
/
$s
\\
~
I-i i
I m
e 300s 3g33 3gge 4ggg I!K (s)
Figure 2.6 Steam Generator A Secondary Side Vapor Temperature n
i 4 -
MIN $iEAM LINE BREAK 600 RELAPS sa0
--- IRAC se0
- v,..-
/
/
5840
,/
G
/
/
o
\\
~
/
$s20 '1 e'
I
/
{sm I
/
l
/
I
/
\\/
\\/
(/
N60 0
1000 2000 3000 4000 5000 6000 1000 8000 REACTOR IlE (s)
' Figure 2.7 Steam Generator B Secondary Side Vapor Temperature l
MIN STEAM LINE BREAX i
i i
IRANS!ENT I!E (t - f 00s)
RELAPS/M001.5 e
y -2000
-eees s
I I
I
-sees e
lese sees zees esoe
!! T (s)
Figure 2.8 Heat Transfer Rate in SGA i
i i
10 -
l MAIN STEAM LINE BREAK l
i I
i TRANSIENT !!ME (? - 4 % ):
RELAPS/n0DI.5.
t<_
O E
. tees a
Ms I
I I
-2ees e
1000 2000 3000 4000 ilME (s)
Figure 2.9 Heat Transfer Rate in SGB MAIN STEAM LINE BREAK l
HPl Flow To COLD LEG A 1
_4 I
[
l RELAP5 I
TRAC-PFI W
i N
I
$10 l
9 i
I 4
1 2
i j
i 4
i o
-400 o 1000 2000 3000 TIME (s)
Figure 2.10 HPI Flow to Cold Leg A-1 computed an earlier repressurization as the RCPs were started when the primary side pressure was higher than in the TRAC calculations.
Furthermore. the HPI flows were initially higher in TRAC, but RELAPS continued HPI much longer as shown in Figures. 2.10 through 2.12.
This additional mass also increased the pressure in the primary side. The pressurizer became full much earlier in the RELAPS calculation than in TRAC, and this is reflected as a rapid increase in pressure in Figure 2.1.
One of the most important parameters for PTS is the downcomer fluid tem-perature. A comparison of these temperatures as predicted by the two codes is shown in Figure 2.13.
The first observation is that TRAC predicted a multi dimensional behavior, and the azimuthal spread in the downcomer temperature was about 30K.
RELAP5 predicted only the average fluid temperature, because of the one-dimensional modeling.
The downcomer fluid temperature is a com-bination of the cold leg and vent valve fluid temperatures. Figure 2.14 shows a comparison of the vent valve flows as predicted by the two codes.
The vent valve did not open in the RELAPS calculation because there was sufficient flow in' both the cold legs due to natural circulation, as shown in Figures ' 2.15 through 2.18.
Thus there was no warming effect due to the vent valve flow in the RELAPS calculation, but the cold leg temperatures, as shown in Figures 2.19 through 2.22 were higher than in the TRAC calculation and were the cause of a higher downcomer fluid temperature in the RELAPS calculation.
Further more, the downcomer fluid temperature as well as fluid temperatures in cold legs A-1 and B-1 had jumps in both the calculations at the time of RCP restart due to the mixing of hotter fluid f rom the hot legs. There were also similar jumps in the fluid temperatures in cold legs A-2 and B-2 in both the calcula-tions due to a circular flow between the cold legs with the common steam gen-erator.
The temperatures for all the cold legs and the downcomer started to increase af ter the steam generators were isolated in both calculations at 600 seconds. Thus the minimum downcomer fluid temperature was reached just before the RCPs were restarted in the TRAC calculation.
In the RELAPS calculation the downcomer fluid temperature started to decrease again af ter the jump at RCP restart time.
The lowest downcomer fluid temperature was reached at 600 seconds, which was very close to the temperature at 300 seconds.
The RCP restart time, which depends upon achieving 42K subcooling in each hot leg of the system, is a critical event'in this transient.
It occurred at 300 seconds in the RELAP5 calculation, whereas it occurred at 526 seconds in-TRAC.
This difference is due to the way the voids were distributed in the primary system and the time when natural circulation st'arted.
Figure 2.23 shows a comparison of the predicted void f ractiores in the upper.
It seems that there was an early void accumulation in the RELAP5 calculation but none in the TRAC calculation. Figures 2.24 and 2.25 show a comparison of the candy cane void f ractions as predicted by the two codes.
In loop A, the natural circulation was strong because of depressurization of Steam Generator A secon-dary side, and no voids accumulated in the candy cane. However, the natural l
circulation was slower in the unaffected loop (f.e., loop B), and most of the voids in the TRAC calculation accumulated in this candy cane, which' resulted in the termination of natural circulation there. On the other hand, the voids accumulated in the upper head and upper plenum instead of candy can es in the I
RELAPS calculation, and the natural circulation was maintained. This resulted in a good and uniform cooling of both loops in RELAP5.
As RELAP5 also pre-dicted a higher primary side pressure, both hot legs achieved the required subcooling of 42K at 300 seconds and the RCPs were restarted.
However, TRAC had no natural circulation in Ipop B and the hot leg there remained warmer..
MAIN STEAM LINE BREAK
'~_ i 2e j
i I
8' l
TRANSIENT TIE (t - 400s)-
RELAPS 1
TRAC I
_j te W&
.\\
=
I i
I l'l I
e O
1900 2000 3000 4000 TIME (s)
Figure 2.11 HPI Flow to Cold Leg A-2 MAIN STEAM LINE BREAK 15 8
8 jl i
I i
TRANS!ENT ilME (r - 4COs) e l
RELAPS ie l
TRAC l
I i
g s
l %
s a
E 1
o l
t e
seee e
n 4.
TIME (s)
Figure 2.12 HPI Flow to Cold Leg B-2 13 -
F
-r,,
s l
DOWNCOMER LIQUID TEMPERATURES l
600 R TH Z l
MAIN STEAM LINE BREAK 216
- 226 g
E550 6236 W
5
+246
$525 s
--gRELAP5
=256 g
g 0266 E 500 4
~~
475 H
==
. y450 O
3425
- TRAC
~
400 VESSEL i
i i
i i
i ID =M 375 100 300 500 700 900 TIME (s)
Figure 2.13 Downcomer Fluid Temperature MAIN STEAM LINE BREAK 500 VENT VALVE FLOW o.
RELAPS
~
i
- -- TRAC-PFI W
.dvn E
l O
9w 9
-500 i
-400 0-1000 2000 3000 TIME (s)
Figure 2.14 Vent Valve Flow M14 STEM LINE BREAK 7000 tu
~
I RELAPS
- jm l
__ _ _ rRAC i
~
I l
5000 Em I
1000 W
s 0
0 1000 2000 3000 4000 5000 6000 7000 8000 REACTOR flT (s)
Figure 2.15 Cold Leg A-1 Flow Rate MAlli STEM LlhE BREAX S000 RELAPS
TRAC e 2000 W
E 1000 E
C' km l l != !jj" C::::jliliii;;::l l:lll !l!:
.il isn'.i 33.i ::3 i. "!!!
E
~
i I:
-=
saamma dalL==
t
-2000 g
I 3000 0
1000 2000 3000 4000 5000 6000 7000 8000 REACTOR ilE (s)
Figure 2.16 Cold Leg A-2 Flow Rate i
MAIN STEAM LlY BGEAK 7000 N~~~_'~~~.________
6000 h
m
. _ _ _ _ eEws i
--- TRAC 3
1 h2000 1000 l
jp.l k
0
~
0 1000 2W 3000 m 5000 6000 7000 8%0 REACTOR TIME (s)
Figure 2.17 Cold Leg B-1 Flow Rate MAIN STEAM LINE BREAK 5000 WO
-- RELAPS
. - -- IRAC 3e i
E 2n h 2
1T)3
- a l
1.
' # ~~
200:
0 1m 2000 3000 GO 5000 6m 7e 8000 PEACTOR Tl"E (s)
Figure 2.18 Cold Leg B-2 Flow Rate MAIN STEM Ll%E BREAK 580
,/
,s' 540 520
,/
500
/
b
/
RELAPS C
i e
,/
- - - - TRAC t ggo
.I
'\\ l/
t 460
\\
5
.\\ l 3 44 g j 420 ' if\\t 0
12 2000 3000 4000 5000 6000 7000 8000 REACTOP TIT (s)
Figure 2.19 Cold Leg A-1 Fluid Temperature MAIN STEAM LikE BREAX 580
/
540
,e
/
520
,/
- RELAPS
/
-- - - TRAC p
e
/
4
/
}
/
480
.\\ l 460
\\ t 1
420 LI\\l I
O 1000 2000 3000 4000 5000 6000 7000 8000 REACTOR II"I (s)
Figure 2.20 Cold Leg A-2 Fluid Temperature
~
MIN SIEM LINE BREAK l
580 560
)
s'
/
L
,/
l o
520
,l
,/
/
/
g 5500 l
/
RELAP5 l
/
g
--- TRAC
.;,g
,/
480 g
e
\\Qj
" 460
\\
l>\\
I 420 K X) 0 1000 2000 5000 4000.5000 6000 7000 REACTOR TIE (s)
Figure 2.21 Cold Leg B-1 Fluid Temperature-MIN STEM LINE. BREAK 580 560
,--v<<<<-
s' I,
540
.l 3
520
/
s' o$500
/
- RELAP5 I, i
/
TIE
- 480.t !
i n'/
af
. u,u s
460 11
- 440. I I
420 0
1000 2000 1000 4000 5000 6000 7000 8000 REACTOR TIE (s)
Figure 2.22 Cold Leg B-2 Fluid Temperature
..i
. ~.
~
MAIN STEAM LINE BREAK 0.75 UPPER PLENUM VOID FRACTION g--
.g-g a50 RELAPS
~
~
a h
e 0.25 2
s-TRAC-PFI I
O
-400 O 1000 2000 3000 TIME (s)
Figure 2.23 Upper Plenum Void Fraction MAIN STEAM LINE BREAK O.075 i
CANDY CANE VOID FRACTION IN LOOP A z9
~U
@ GOSO RELAPS 9
3 g 0.025 a>
l i
i i
o
-400 0 1000 2000 3000 TIME (s)
Figure 2.24 Candy Cane Void Fraction in Loop A d t rs-w-y E,
c.,
,.--mc
,-v,
-.y
MAIN STEAM LINE BREAK l.2 CANDY CANE VOID FRACTION IN LOOP B 1.0 7 1 i-g I RELAP5 E*
I I TRAC-PFI
{
l I h 0.6 -l I 8 o5 I I
> O.4 1 '
I s
l O.2 4 1 I I I O'
O 2000 4000 6000 8000 REACTOR TIME (s)
Figure 2.25 Candy Cane Void Fraction in Loop B MAIII STEM LillE BREAK 580 560
,/
,/
/
520 1,4
/
9
\\\\
,/
- 500 i
/
I
/
RELAPS p
e a s
,i
we
\\/
1/
Y l
440 0
1000 2000 3000 4000 5000 6000 1000 8000 REACTOR TIE (s)
Figure 2.26 Hot Leg Temperature in Loop A L_
1 MAIN STEAM LINE BREAK 600 i
i i
i i
i i
HOT LEG TEMPERATURE IN LOOP B 580 560
/~~~~~~"
/
/
- 540 f'
E
'g
/
2 8
/
f' 520 - I I
s O
/
y500 i
l RELAPS 480 - I
-_- TRAC-PFI
~
f I
/
/
460 - l' 00 2000 4000 6000 8000 REACTOR TIME (s)
Figure 2.27 Hot Leg Temperature in Loop B MAIN STEAM LINE BREAK i
i i
e i
i i
3o000 HOT. LEG MASS FLOW RATE IN LOOP A
--- TRAC LOOP HOT LEG TRAC LEVEL 7 EXIT 8000
~
--- TRAC LEVEL 8 EXIT w
$ 6000 RELAP5 HOT LEG m.
3B 3
m 4000l i
B l,I' 2000}
l bf -----
hJ A
1 I
i i
t t
00-2000
-4000 6000 8000 REACTOR TIME (s)
Figure 2.28' Hot Leg Mass Flow Rate in Loop A -
PAIN STEAM LIME BREAX TRAC TOTAL HOT LEG 8000
---TRAC LivEL 7 EX1T
~*-TRAC LEVEL 8 EXII 6000 7.4 a 4000 i
w i
f
' ' ' - - ~ ~ - - ~ ~ _ _ _ _ _ _ _ _,
h2000 E
[
~ ~. _ _ _ _.
0 O
1000 2000 3000 4000 5000 6000 1000 8J00 REACTOR 11T (s)
Figure 2.29 Hot Leg Mass Flow Rate in loop B MAIN STEAM LINE BREAK 5000 i
i i
-10000 7 4000 b
MAIN STEAM LINE BREAK FLOW
-8000 h3000_-
RELAPS
~-6000 h2000_-
-- TRAC-PFI 4000 v,
g 1000-
--2000 s
0-
--O
--2000
-1000 0 1000 2000 3000 4000 TIME (s)
Figure 2.30 Main Steam Line Break Flow k-
i-PAIN STEAM LINE BREAK 1500 g
TRANSIENT TIME (r - 400s) -- RELAPS
~ -- TRAC tese k
2 W
~
& See B
l i
M 4
2 e
'\\_,
I I
-See e
1000 2000 300e 4000 TIME (s)
Figure 2.31 Main Feedwater Flow to SGA MAIN STEAM LIE BREAK teos TRANSIENT TIME (t - 400s)
RELAPS
---TRAC j
500 w
3 E
.- k e
E I
-50e e
1000 2000 3000 4000 TIME (s)
Figure 2.32 Main Feedwater Flow to SGB 23 -
mlh STEArt Lite BREAK
-1Q f
140 REM
.___1.
120 100 O
l.
4
'l J
lI em on
=
ei ll ll il :ll s aa.
V1 iau n
e ll1 q:l
'I l
20 ll lL hl it i
lIL a
J 0
i
- 20
-40 0
1000 2000 3000 g 5000 6000 7000 8000 g
Figure 2.33 Emergency Feedwater Flow to SGB MAIN STEAM LINE BREAK 600 EMERGENCY FEED WATER FLOW TO SGA -
N
[400 A
RELAP5 w
--- TRAC-PFI 1
D "2@
E O
5 i
1
-200-400 O 1000 2000 3000 TIME (s) i Figure 2.34 Emergency Feedwater Flow to SGA L
MIN STEAM LIE BREM 600 550 ELAPS
___ ypAc 500 C450 m o
B e 400 83 I
350 i
300 0
1000 2000 3000 4000 5000 6000 7000 8000 REACTOR TIE (s)
Figure 2.35 - Temperature of Emergency Feedwater to SGA '
Mit STEAM LIE BREM 600 550 ELAPS
-- TRAC 500 Q
- 450 b
a 5400
. 3 l
350 [
l i
L --'- -
300 l'
0 1000 2000 3000 4000 5000 6000 7000 8000 REACTOR TIE (s)
Figure 2.36 Temperature of Emergency Feedwater to SGB t
F 25 -
,-,n
.,,-r-_
.,_.--,v--
..-n,
.y
,,,,, - - - -..-s
than in loop A.
The hot leg in loop A achieved the 42K subcooling even bef ore it did so in the RELAP5 calculation, but, the loop B hot leg was slower to cool and took longer to achieve the required subcooling. This delayed the re-start of RCPs in the TRAC calculation and resulted in a lower downcomer f.luid j
temperature.
There were some multidimensional effects as reflected in the differences in the two hot leg behaviors.
Figures 2.26 and 2.27 show a ccm-3 parison of the hot leg temperatures as predicted by the two codes.
The l
RELAP5 calculation did not show much difference between the two hot leg tem-
)
peratures as they started f rom the same branch but the TRAC calculation showed significant differences, which seems to be more reasonable.
Beside the multi-dimensionality of the transient, there were also impor-tant differences in the way the upper head was modeled in these two calcula-tions.
In the TRAC calculation the upper head had an artificial connection to the hot leg and there was no volume at the top of the reactor pressure vessel with a dead end. There was significant flow through the upper head to the hot leg, as shown in Figures 2.28 and 2.29.
(The level 8 exit in these figures represents the upper head connection.) This flow probably prevented the accu-mulation of voids in the upper head in the TRAC calculation. However, if the upper head in the TRAC had been modeled as a dead end space, the voids pro bably would have accumulated there instead of migrating to the candy cane, and the natural circulation would have continued in loop B.
This would have re-sulted in a lower hot leg fluid temperature in this loop, and the diff erence between the times of achieving the subcooling in two hot legs would also have been less. The RCPs would have started earlier in TRAC, resulting in a higher downcomer fluid temperature. Changing the RELAP5 model to have a flow through the upper head would probably terminate the natural circulation in loop B, but the cooling of hot leg B would not be delayed as much as in the TRAC calcula-tion as both the hot legs were connected to the same branch component.
The RELAPS approach of modeling the upper head as dead-end volume is more realis-tic.
The performance of the steam generator as a heat sink can be assessed by comparing the EFW. MFW, and break flow rates calculated by the two codes.
Figure 2.30 shows a comparison of the break flow rates as predicted by the two codes. TRAC computed a smaller break flow rate than RELAP5.
Figures 2.31 and 2.32 show the main feedwater flows as computed by the two codes.
The time scale is too large to show the effect of the early MFW pump trip and realign-ment of MFW.
Figures 2.33 and 2.34 show the EFW flow rates.
RELAPS had larger EFW flow than TRAC.
However, the EFW temperature was much higher in RELAPS than in TRAC, as shown in Figures 2.35 and 2.36.
Furthermore, the break flow quality or enthalpy was not provided by the TRAC calculation, and it was difficult to compare the heat transf er in the steam generator. On the basis of limited information, it can be concluded that the codes were consis-1 tent in their prediction of flow parameters for the primary and secondary I
sides.
The conclusion f rom this transient is that the most crucial event in de-termining the minimimun downcomer fluid temperature was the RCP restart time which should have been somewhere between the RELAP5 (300 sec.) and TRAC (526 sec) calculated values.
The initial MFW trip and EFW modeling in RELAPS were more appropriate than in TRAC.
Therefore, for a conservative yet rea-listic estimate one should delay the RCPs restart time in the RELAPS calcula-tion until the TRAC RCP restart time.
This would yield the lowest average m
M AIN STE AM LINE BREAK 600 1,..
i.,i i
RELAPS l-Loop A-1 Cold Leg (of vessel)
C 550 2-Loop B-l cold Leg (at vessel) 3-RV Downcomer (at cold leg elevation)
$a 500 E
2 2
U 450 ea 400 g,
'4465 K l379F Average downcomer fluid
*P
350 I
1 TRAC RCP RESTART TIME o
1000 2000 3000 TIME (sec)
Figure 2.37 Estimate of Lowest Average Downcomer Fluid Temperature downcomer fluid temperature of 465K or 379F, as shown in Figure 2.37. It was found from the TRAC calculation that there was an azimuthal temperature
' distribution in the downcomer and the spread was around 30K. This multidimen-sional effect should also be taken into account in estimating the. lowest down-comer fluid temperature, which would then be ISK lower than the average down-comer fluid temperature of 465K.
Thus the minimum fluid temperature in the down comer would probably be 450K.
Even this value is slightly conservative as the actual time of RCP restart could be earlier than that predicted by TRAC if the voids can accumulate in the upper head as discussed earlier..
o I
3.
FAILURE OF ALL TURBINE BYPASS VALVES AT FULL OPEN POSITION The steam generator secondary side can be depressurized either by a steam-line break or by failure of turbine bypass valves at open position. Steamline break yields the largest break; TBV stuck open produces a smaller break. Two different transients initiated by all four TBVs stuck open were specified.
INEL was assigned to calculate a transient starting from the hot standby con-dition (9 MW + RCPs power) whereas the transient assigned to LANL started from the full reactor power.
Both scenarios had further operator failures of not throttling the HPI and not restarting the RCPs when needed.
Additionally, in the LANL case, the Integrated Control System (ICS) failed to runback FW, and the EFW level control failed.
In the INEL scenario, the feedwater did not align to the EFW header.
The hot standby condition assumed in the INEL sce-nario also implied that initially there was no steam supply to the feedwater heaters and the main feedwater was going through the start-up line. Further-more, the INEL scenario also required closing of all TBVs at 600 seconds.
These differences between the LANL and INEL Scenarios are summarized in Table 3.1.
The differences in the initial conditions and scenarios resulted in a quite different response during the transient.
The purpose of comparing the two calculations is to indicate the effect of the differences in the initial conditions and scenarios on the transient. The sequences of events as compu-ted by the two codes are sumnarized in Table 3.2.
Figure 3.1 shows a compari-son of the primary side pressures.
The pressure in the TRAC calculation ini-tially decreased faster than in the RELAPS case probably, because of a larger energy transfer to the stean generator.
This is reasonable, as in the TRAC calculation the steam generators had a larger liquid inventory (
200 inch) and also a larger mass flow rate at the TBVs, as shown in Figures 3.2 and 3.3, than in the RELAP5 calculation.
Figure 3.4 shows a comparison of the down-comer fluid tempertures. Here again the fluid cools down at almost the same rate in both cases except that it started at a higher temperature in the TRAC calculation.
HPI began earlier in the TRAC calculation than in the RELAPS because of a faster decrease in the pressure in the TRAC calculation.
Main feedwater was lost earlier -in the TRAC scenario than in the RELAPS case as shown in Figures 3.5 and 3.6.
This caused a slower rate of energy transfer in the Steam Generator (SG) in the TRAC calculation than in the RELAPS calculation. This is reflected in the change in the slope of the pres-sure curve in Figure 3.1.
The secondary sides of steam generator,s were almost full by 500 seconds, and the primary side had also cooled sufficiently to re-duce the heat transfer in the steam generator. The secondary side inventories as predicted by TRAC are given in Figures 3.7 and 3.8.
Similar information for RELAP5 is given in Figures 3.9 and 3.10, where secondary sides were filled by 1000 seconds.
This led to a rapid increase in the primary side pressure, as shown in Figure 3.1, due to a decrease in heat transfer rate in steam gen-e rato rs.
The TBVs were closed in RELAP5 calculations at 600 seconds as per operator action, and the secondary side inventory started to increase. This also caused reduced heat transfer in the stean generator for the RELAP5 cal-
- culation, l
l - ---
. Table 3.1. Comparison of the LANL and INEL Scenarios
.for All TBVs Stuck Open Transient UPDATED TRAC-PFl RELAPS/ MODI.5 ITEM (LANL)
(INEL)
- 1) Initial-Canditions a) Core Power.
Full Power (2568MW)
Hot Standby (9MW+ power of 4 RCPs) b) Steam to FW Heaters.
Yes No c) FW temperature at SG.
510K 305K n
- 2) Failures a) All Four TBVs f ailed open.
Yes Yes b) Operator f ailed to throttle Yes Yes HPI and restart RCPs when needed.
c) SG liquid level controls Yes No for EFW f ailed.
d) FW f ailed to realign to EFW No Yes header af ter RCP trip.
3)OperatorAction a) TBVs closed af ter 600 sec.
No Yes l
i f
Table 3.2 Sequence of Events f or All TBVs Stuck Open Transient UPDATED TRAC-PF1 RELAP5/M001.5 ITEM.
(LANL)
(INEL) 1
- 3) MFW Pump Trip 91.2 sec. due to high 168.5 sec. due to SG-B liquid level low suction (6.2m).
pressure.
to low level in SG.
- 5) Accumulator Injection Did not come on.
383.5 sec. to 391.6 sec.
- 6) TBV Isolated Did not.
600 sec.
- 8) EFW Off SG level control f or Between 1010.2 and EFW did not work; 1030.4 sec. f or SG-B EFW always on.
Between 1070.5 and 1074.6 sec for SG-A.
- 9) Lowest Downcomer Fluid 350K (170.6*F) 402.6K (260*F)
Temperature m
4 TURBINE BYPASS VALVES Fall 20000 1
5
-2500 15000
,' PRESSURE IN R.V.DOWNCOMER -
m
-2000 S
RELAPS E
m 0000~
--1500 o
y i,/
--- TRAC-PFI
- 1000 w ct 5)
E w 5000 s@
-500 0
O O
2000 4000 6000 8000 TIME (s)
Figure 3.1 Pressure in R.V. Downcomer t
4 TBVS STUCK OPEN i
i i
i i
i 300 r l
FLOW THROUGH TBV-A I
'r
]
l RELAP5 l
--- TR AC-PFl
< 200 T
8 3
I g
i.
\\
4, 'p
- ffT4gfr Its 4 10 0
\\[g l)(
2
\\ s s q l
1 i
t i
e 0
0 400 800 1200 REACTOR TIME (s)
Figure 3.2 Flow Through TBV-A 1
- i. !
I l
I I
I 4 TBy$ $ TUCK OPDt 350 300 RELAPS l
f
--- inAt I
- 250 I
'd td 200 6
I
" 150 i *'*\\/
E t
,/ !
I'I T 100 f
\\l 50 0 0 200 400 600 800 1000 1200 1400 REACTOR T!E (s)
Figure 3.3 Flow Through TBV-B 4 TURBINE BYPASS VALVES Fall i
i i
_ soo DOWNCOMER LIQUID TEMPERATURE 550 k E
t
- 500 $
$ oo (
i w
gS 4
\\
- 400 $
450
\\' TRAC y
2 E
RELAP$'
.s
- 300 g 400
~~~.~s'-200 350 s
o 200o 4000 sooo sooo TIM E (s)
Figure 3.4 Fluid Temperature in R.V. Downcomer
)
{
l I i
4 TBVS STUCK OPEN l
3" i
i i TRAC FROM 350 as TRANSIENT TIME (t - 500s) 1 Ob g00 w
RELAPS g,00 0
I O
2000 4000 6000 0000 TIME (s)
Figure 3.5 Main Feedwater Flow to SGA 4 TBYS STUCK OPEN s
s a
TRAC TRANSIENT TIT (t - 500s)
T W TO SGB W3 200 Ed M2 RELAPS I
I I
0 e
2000 4000 6000 0000 TIME (s)
Figure 3.6 Main Feedwater Flow to SGB 33 -
80000 70000 -
TRAC-PF1
-140000 3 60000 -
~,
Y
.1200003 g,
y 0
-100000UE 5 40000 -
tE
- 800005
" 30000 -
y
- 60000 20000 -.
0 Ida 200 360 4b0 SD0 6'00 /00 8'009'00 l'000 TIME (s)
Figure 3.7 Steam Generator Secondary Inventory - Loop A 90000 80000 TRAC-PF1 180000
/. 160000 _
70000
[
I 360000 U U/ / / ("t
/
120000y
/
~
a50000 N
b40000
/
100000 5 80000s W
s30000' 60000 20000 40000 10000 0 100 200 300 400 500 600 700 800 900 1000 TIME (s)
Figure 3.8 Steam Generator Secondary Inventory - Loop B I (
4 TURBINE BYPASS.WLVES FAIL AT REACTOR HOT STANDBY 10 i
i i
SG A INDICATED LIQUID LEVEL
-30 (OPERATING ANO STARTUP)
--25 5 E8
.,_,,y._,.
a p
W w
3 6--
--20g J
4 l
-15 O ag4 (l:
OPERATING y
il
--- STARTUP
-10 8 I
@2 I
E O
O-2000 0
2000 4000 6000 8000 TIME (s)
Figure 3.9 SGA Indicated Liquid Level (Operating and Startup) 4 TURBINE BYPASS VALVES Fall AT EFACTOR HOT STANDBY 10 i
RELAPS/M001.5
[fg 8.'
.25 l
'm_ :.
E
.- 20 36i i$ !
[
" 15 d i
,'i, Q
g4' y1 l
l
/
10$
?2-
,Iy
?.
i
- 7. 31 5
w.'
-22 0
2000 4000 6000 8000 TI4 (s)
Figure 3.10 SGB Indicated Liquid Level (Operating and Startup) 2
In stanma ry, the prima ry side was repressurized, while the tenperature kept decreasing for. some tine in botn the calculations.
The primary reasons for the two different temperature predictions were the differences in the emergency feedwater control and the time of isolation of steam generators.
In the case of the RELAPS calculation, the stean generator was lost as a heat sink, and any possible cooling was due to the HPI and PORV flows, which main-tained a stable temperature of 402.6K.
In the TRAC calculation, the EFW flow controller, which was based on the secondary side water level was assumed to fail, and that caused the emergency feedwater (EFW) to continue until the con-
{
densate tanks were empty. As a result TRAC calculated a continuous decrease in the primary side temperature to 350K, which was lower than the RELAPS pre-diction. Therefore, although the differences in the initial conditions caused some differences in the early part of the transient, the failure of the EFW control based on the SG secondary level and the failure to close the TBVs after 600 seconds were the major contributors to the lower downcomer fluid temperature in the TRAC calculation.
In general, both the TRAC and RELAPS calculations look reasonable for the specified transients.
Note that there were no multidimensional effects for all TBVs stuck open transients discussed above.
1 4.
SMALL BREAK (2-INCH) LOCA IN HOT LEG A primary side small break can initiate an overcooling transient if the only allowed operator action is to trip the RCPs at 30 seconds after the HPI initiation.
Such a transient was specified by ORNL, and both LANL and INEL computed the same transient using TRAC and RELAP5, respectively.
The speci-fled scenario is presented in Table 4.1.
Note that the ICS is assumed to work as designed.
There were several differences between the TRAC and RELAPS results.
The first difference was the criterion for reactor trip. TRAC tripped the reactor at 0.5 seconds while RELAPS tripped it on the basis of the low primary side pressure and was more realistic.
Both codes ran back MFW pumps after the re-t actor trip as designed in ICS.
The purpose of modeling the same transient with two codes was to determine the sensitivity of the results to the codes.
The transient was initiated by assuning a break in the pressurizer surge line, and an asymmetric loop behavior was expected.
Table 4.2 summarizes the timings of various events such as HPI initiation, RCP trip, etc. for two cal-culations.
It also indicates that the method of modeling the plant and the modeling differences of codes do affect the results.
In this transient, the primary side lost energy through the break and the steam generators.
The primary side fluid temperature further decreased when the cold HPI water mixed with the primary coolant.
Most of the differences between the two calculations could be explained in terms of these heat sinks.
Figure 4.1 shows the primary side pressures as computed by TRAC and RELAP5.
The TRAC calculation showed a faster drop in pressure than RELAP5 during the first 300 seconds, and a slower drop thereafter.
The initial rapid pressure drop in TRAC was consistent with the early reactor trip and with the larger break flow rate prediction than in RELAPS, as shown in Figure 4.2.
- However, during the time period between 300 and 1100 seconds, TRAC predicted a higher break flow rate and a higher primary side pressure than RELAPS. This could be consistent only if there was either more HPI flow or lower energy loss through the steam generators.
The energy loss through the break was not provideo.
However, the void fraction at the break for the TRAC calculation and the sta-tic quality at the break for the RELAPS calculation were provided, and are shown in Figures 4.3 and 4.4, respectively.
During the first 1000 seconds, TRAC computed a very low void fraction while RELAPS predicted a high static quality and consequently a high void fraction.
On the basis of these results and rough estimates of energy loss, it can be concluded that the energy loss through the break in the TRAC calculation was higher than that in the RELAP5 cal cul ation.
The TRAC break ficw rate was also approximately twice that in RELAP5. The specific energy at the break in the TRAC calculation was only, at the most, 25% less than in the RtLAPS calculation.
The HPI flows, as expect-ed, were initiated slightly early in TRAC calculations, as shom in Figures 4.5 to 4.8.
However, the differeqces between the HPI flows in the two calcu-lations wre insignificant.
Therefore, the cause of the apparent inconsis-tency between 300 and 1100 seconds lies wi th the steam generator heat transfer.
7 _-.
Table 4.1 Small Break LOCA in Hot Leg. Scenario Initial Conditions:
1.
Full reactor power.
2.
Nominal temperatures and pressures in primary / secondary.
3.
Decay heat: 1.0 times ANS standard.
4.
Pressurizer spray / heaters operate as designed.
Seq'uence of Events:
1.
SBLOCA.'
2.
Reactor trips, turbine trips, TSVs close.
3.
HPI actuates at setpoint (1500 psi).
4.
TBVs/SRVs in secondary f unction as designed.
5.
Operater trips RCPs 30 seconds af ter HPI.
6.
i i I
fT t:
L Table 4.2 Sequence of Events _in the 2-Inch Hot Leg Break UPDATED TRAC-PF1 RELAPS/MODl.5 ITEM (LANL)
(INEL)
- 1) Break 0.0 sec.
0.0 sec.
45.2 sec.
Pump Runback
- 3) TBV Opens 4.2 sec.
47.0 sec.
/
- 5) SRV Opens No 50.0 sec.
SRV Closes No 69.0 sec.
- 6) HPI Initiation 43.1 sec.
78.5 sec.
- 10) Vent Valve Opens 100 sec.
55' sec.
- 11) EFW Trips 1000 A 350 sec.
loop A 503 sec.
loop B 400 sec.
loop B 500 sec.
- 12) Loss of Natural loop A 750 sec.
loop A 815 sec.
Circulation loop B 600 sec.
loop B 1020 sec.
- 13) Circular Flow and Flow loop A 1200 sec.
loop A 872 sec.
Oscillation Between Cold loop B 1200 sec.-
loop B 1100 sec.
Legs.
(No oscillation)
- 14) Accumulator Injection None 2215 sec.
H 5124 sec.
- 15) LPI'
- 16) Minimum Downcomer Fluid 470K at 750 sec.
355-361K at 7200 Temperature (based on calculation sec. (estimated) up to 1800 sec.) J
I 1
2 INCH HOT LEG BREAK
_ 20000 i
DOWNCOMER PRESSURE
-2500j M
,T
--2000f 15000 8
5 C
$ ooco-
-- 1500 g i
E W
[
TRAC-PFI q
-1000 3 5000 RELAP5 8
-500 3 o
i i
o o
o 2000 4000 6000 TIME (s)
Figure 4.1 Downcomer Pressure
- 2 INCH HOT LEG BREAK 300 RELAP5 -
-600 TRAC-PF1---
k
-500f
- 2m~
' TRAC-PFI 3
{
{
,RELAP5 5
h if et
-300h
$ 10 0_-
(h kl
\\
li d
-200 f
_ioo RELAP5' o
o
-1000 o
1000 2000 3000 4000 TIME (s)
BREAK FLOWRATE Figure 4.2 Mass Flow Rate Out of Break I
l l
r 2 IEH HOT LE6 I
i i
RELAPS TRANSIENT TIE (T - 400s)
C I
8.s e
s t;
I
' E 1
I s.e e
ages 4eos sees sees TIE (s)
Figure 4.3 Static Quality in the Surge Line Volume With Break 2 INCH HOT LEG BREAK 1.2 TRAC-PFI 1.0 l
I
.8 i
IA I '
i II di
.E.6 f
f 1
1f,.
g3 g
y l
J l
l l
y i
i l
I
>.4 i.
'2 I,
Jl
..,..d, t
' ' ; '! T i
j.
0 l
0 200 400 600 800 1000 1200 1400 1600 1800 REACTOR TIE (s) l Figure 4.4 Void Fraction in the Surge Line Volume With' Break 9
.)
4 4
i l
I I
2 INCH HOT LER BREAK 30 l
i I
- RELAPS
(
TRANSIENT TIME (r - 400s)
--- TRAC 1
ae
}
g 3
e W
i M
I 3:e l
m N
I I
\\
l I
I I
g e
2000 4000 4000 0000 TIME (s)
Figure 4.5 HPI Flow to Cold Leg A-1 2 INCH HOT LEG BREAK 30 t
I L
TRA!!SIENT TIME (T - 400s) 1 ELAPS
--- TRAC Qae 5
/
U l
E l
e
[te
-I E
l 1
I i
I I
I I
e e
2000 4000 6000 000e TIME (s) l Figure 4.6 HPI Flow to Cold Leg A-2 I
l.
I I
l' i
U i
m-
't l.
t 4
2 INCH NOT LEG BREAK 2e TRANS!ENT TIE (t - 400s)
,/ /
O
)
- RELAPS f
y
--- TRAC E
se b
l 1
1 M
i 1
4 r
)
1 1
I e
seee
<eee seen eeen TIM (s)
Figure 4.7 HPI Flc,w to Cold Leg B-1 2 INCH HOT LEG EREAK 20 t
i I
TRMSIENT TIPE (T - 400s) ff j'
RELAP5
];
--- TRAC i
git a
i d
\\
M 15 t
's
- I I
I e
e 2006 '
4000 stee '
sees TIME (s)
Figure 4.8 RCL Flow to Cold Leg B-2 I
s T
s 4
't
$g d
The heat transfer in the ste.am generator was governed by the primary side flow and temperature and by the secondary side fluid conditions. The RCPs were tripped in both the calculations, and the primary side was in natural circula-tion mode.
The natural circulation lasted until the candy cane voided as shown in Figures 4.9 and 4.10.
Figure 4.11 compares the upper head voiding as predicted by the two codes. The RELAP5 computed complete voiding of the upper head by 300 seconds while TRAC calculated only 50% voiding in the upper head.
More vapor went to the candy cane than to the upper head and caused earlier termination of natural circulation in the TRAC calculation than in the RELAPS
{
calculation.
(This is consistent with the main steam line break transient discussed in Chapter 2.)
The steam generator secondary side conditions were controlled by the feed-water conditions.
After the reactor trip, the MFW pumps were run back to maintain proper flow, and the main feedwater was aligned to the EFW header through the SUFCV (Start Up Flow Control Valve). The main feedwater was lost in the RELAPS calculation because of MFW pep trip on the high discharge pres-sure at 70 seconds while in the TRAC calculation the MFW lasted until 350 seconds when the SG secondary level control was exceeded and the SUFCV was closed. The EFW was started at the time of RCP trip in both calculations, but TRAC terminated it earlier, as shown in Figures 4.12 and 4.13.
The EFW was more than 100*C colder than MFW.
Consequently, the steam generator secondary side had wamer fluid in the TRAC calculation. The colder fluid in the RELAPS calculation caused the secondary side pressure to be lower, as shown in Fig-ures 4.14 and 4.15.
This was also confimed in Figures 4.16 and 4.17, where SG secondary exit temperatures are compared.
Thus the warmer fluid in the SG secondary side caused less heat loss in the TRAC calculation and therefore resulted in a higher pressure in the primary side.
In fact, when the steam generator primary side inlet temperatures, as shown in Figures 4.18 and 4.19, were compared with the SG secondary exit temperatures, TRAC had reverse heat transfer in the steam generator after 300 seconds while RELAPS had heat trans-fer in the normal direction.
Both codes predicted a continuous decrease in the primary side pressure as the break flow rate exceeded the HPI flow rate.
This made this transient less severe for the PTS consideration.
Both codes computed comparable downcomer fluid temperatures, as shown in Figure 4.20.
This fluid temperature was a function of cold leg and vent valve flows.
Figures 4.21 through 4.24 show a comparison of the predicted cold leg t empe ratures.
TRAC computed lower cold leg temperatures than RELAP5 even though there was some reverse heat transfer in the steam generator. This is due to the mixing of cold HPI with the cold leg flows. As the cold leg flows in the the TRAC calculation were small, the effect of HPI flow was more pro-noun ced. This cold leg fluid mixed with the wamer vent valve flow. The net effect was the initially colder fluid in the downcomer in the TRAC calcula-tion. Furthemore, fluid tenperature in the downcomer in the TRAC calculation recov! red when the code predicted flow oscillations between the cold legs, the steam generator, and the downcomer after natural circulation was lost, as shown in Figures 4.25 and 4.26.
The RELAPS calculation did not show any os-cillation, but a stable circular flow between the cold legs with the common steam generator and the downcomer.
The downcomer fluid temperature kept de-creasing in the RELAPS calculation as the primary side energy continued to be lost through the break and the steam generator.
2 INCH HOT LEG BREAK
-l l
1.0 i
l
(
b c
E i!
90.5 9 - TRAC-PFI O
8 jjl c
RELAP5 f
CANDY CANE VOID FRACTION IN LOOP A ff O
=d a
-400 0 2000 4000 6000 TIME (s)
Figure 4.9 Candy Cane Void Fraction in Loop A 2 INCH HOT LEG BREAK I.2 i
i i
i i
i i
i CANDY CANE VOID FRACTION IN LOOP B 1.0 e,- - - - - -
z G8 p
9 i
Y I
S RELAP5 e
y E O.6 e
O I
N a4 l
e 0,2 fr\\
s 0 g
O O
400 800 1200 1600 REACTOR TIME (s)
Figure 4.10 Candy Cane Void Fraction in Loop B 2 INCH HOT LEG BREAK 1.0
.i_
i i
m, VOID FRACTION IN UPPER PLENUM z9
(
h RELAP5 E'
TRAC-PFI
,, q 90.5 e.
l Q 9
8 ti 8
a 9
/
l b^'d <likleilu.
O i
i i
-400 0 2000 4000 6000 TIME (s)
Figure 4.11 Void Fraction in Upper Plenum 2 INCH HOT LEG BREAK 200 i
i EMERGENCY FEED WATER FLOW 2
TO STEAM GENERATOR A i
RELAP5
[ 10 0 - _ I' ti
--- TRAC-PFI g
f i 3:
l '!
I 3
u.
o Jt h _t IS i
-10 0
-400 0 2000 4000 6000 TIME (s)
Figure 4.12 Emergency Feedwater Flow to Steam Generator A 2 INCH HOT LEG BREAK I
l l
TRANSIENT TIE (t + 400s)
RELAPS
--- TRAC jlee
~
W
- (/
f 3
i E"; e w<
g l
i I
I
. gee e
2000
-4000 6000 0000 ilT (s)
Figure 4.13 Emergency Feedwater Flow to Steam Generator B B
2 INCH HOT LEG BREAK
,'(,,1 TRA1SIENT TIME (r
- 400s)
-RELAPS
\\
- - - T RAC
] 5.0 at d
?
$ 2.5 d
i I
I e.e e
2ees does 6000 esse TIME (s)
Figure 4.14 Steam Generator A Secondary Side Pressure 47 -
v Mte' i
2 INCH HOT LEG BREAK 7.5
\\ _
TRANSIENT TI E (? - 400s)
RELAPS
}
%%j
- - -TRAC 2,.8 N
S Wi g2.5 W
l I
I I
e 2000 4000 Sete 8000 TIE (s)
Figure 4.15 Steam Generator B Secondary Side Pressure 2 INCH HOT LEG BREAK
-600 i
g STEAM GENERATOR A SECONDARY SIDE TEMP.
C 2
- 550
'5, RELAP5-9
--- TRAC-PFI 8a y500 S
9 450
-400 0 2000 4000 6000 TIME (s)
Figure 4.16 Steam Generator A Secondary Side Temperature i
e 2 INCH HOT LEG BREAK TRANSIENT TIM (r - 430s)
RELAPS
---TRAC 2 SSe
' ~ ~ ~
\\
s esz see W
G I
I I
45e e
em 4eee me see.
TIME (s)
Figure 4.17 Steam Generator B Secondary Side Temperature 2 INCH HOT LE6 BREAK I
I i
TRANSIENT TIME (t - 400s)
RELAPS
- - - TRAC
=
-c W
i (b
E=
e8s I
I I
49.
m.
m.
TIE (s)
Figure 4.18 Steam Generator Primary Side Inlet Temperature in Loop A 49 -
I 2 I KH HOT LEG BREAK i
i i
i TRANSIENT TIME (T - 400s)
RELAP5
---TRAC g
g u
I b
N s
im a
i i
i 1
4gg e
2000 4000 6000 0000 TIE (s)
Figure 4.19 Steam Generator Primary Side Inlet Temperature in Loop B 21NCH hot LEG BREAK g6%
U
- soot DOWNCOMER LIQUID y
TRAC-( AVG)
"550 7 TEMPERATURE h
g 5
,'~~~
g 3500 y~
q
- 4ooW 5450 3
RELAP5
_ g W400 y
3 2
0 200 d 350o 1000 2000 3000 4000 5000 sooo 7000 TIME (s)
Figure 4.20 Downcomer Liquid Temperature L.
2 INCH HOT LEG BREAK 600 2
COLD LEG A-1 TEMPERATURE e
- s. t l
'N b
', nlj 'l pf,
" 500
--- TRAC-PFI 9
S if' g
o 3
61 t
i to 400 n
E v
i i
300 i
i i
-400 0 2000 4000 6000 TIME (s)
Figure 4.21 Cold Leg A-1 Temperature 2 INCN HOT LEG BREAK I
I I
IRANSIENT TIME (t - 400s)
- RELAPS i
I
-- TRAC 17 ef Ws
~
3 I
l l,
i I
l ll<
\\
f l
W I
t i
i l
m
,, g e
m.
i Figure 4.22 Cold Leg A-2 Temperature i
i t
l t
- $1 -
b
2 INCH HOT LEG BREAK 600 i
i i
l
~
5 COLD LEG B-l TEMPERATURE 1
(
RELAP5 n/
W500 5
8 g,
--- TRAC-PFI o
i 4
[
5o 8 I
]
l I w400
', l k
~
Ea 9
300
-400 0 2000 4000 6000 TIME (s)
Figure 4.23 Cold Leg B-1 Temperature 2 INCH HOT LEG BREAK I
i i
TRANSIENT TIME (T - 400s)
REL/P5 C
t 'l
--- T RAC i=
I,h, $ /d h
- 5 lI f)
Y
,,i, I
e t
g...
k,g I
]
g
\\
d I
i i
n.
m.
,,y y Figure 4.24 Cold Leg B-2 Temperature 2 (MCH HOT LEG BMAK ALLAP5 ~ COLD LE6 A-1 13000 e
4000 J
--- COLD LIG A-2
' l TRAC-Pfl --COLD trr, a.1 8000 3000 J
COLD LIG A 2 m
-"l mi.
-i=
$ 10m -
mg i
h
'~ (N55*'h N
0 O
-1000
-2000 0
1000 2000 5000 m 5000 6000 7000 flME (s)
Figure 4.25 Loop A Cold Leg Flow 2 IllCH HOT LE6 LEAK 5000 ;
h10000
~j
-COLD LI6 B-1 U#
m,4
-.- COLD LI6 B-2 -'
L 8000 2 o ';l faAC-ni----COLD tr6 B-i i
3 COLD LIG B-2 }6000 5mjl
'"^C I
2
'p m '!
'[ m b i
4,5 utAe5
--l' - % C '.rJ:pc E
-5 0
0 k-m
.}aan 0 1000 2000 3000 m 5000 6000 7000 IIT (s)
Figure 4.26 Loop B Cold Leg Flow.
In stama ry, both codes camputed reasonable results for this transient.
I There were dif ferences in the break flow rates, reactor trip criterion, upper
[
head voiding and flow oscillations. The flow oscillations in the TRAC calcu-lation were very important as they caused the downcomer fluid temperature to increase.
As the TRAC calculation was tenninated at 1800 seconds it is dif-ficult to guess the downcomer fluid temperature at 7200 seconds into the j
t ransient.
Also, it is not clear whether the loop oscillation predicted by TRAC is real.
The RELAPS calculation, on the other hand, was carried out until 6100 seconds and looks more reasonable.
4
)
i l l l
k____.______.
5.
SUMMARY
AND CONCLUSIONS Three of the several transients computed by LANL and INEL using the latest versions of TRAC-PFl and RELAP5/M001.5 have been reviewed in detail at BNL.
Both the codes were reasonably successful in modeling these transients.
The major differences in their results were due to the difference in modeling the plant, control systems, and event sequences, and to the one-dimensional mo-deling of reactor vessel thermal hydraulics by RELAPS.
Comparison of the computations of the Main Steamline Break (MSLB) tran-sient indicated that the difference in the minimum downcomer fluid tempera-ture predictions was due to the modeling of the control system that regulated the MFW and EFW pumps, and to the multi-dimensional effects, which resulted in different temperature histories for the hot legs and the RCP restart times.
The RELAPS model of the control system, which was based on the secondary side pressure drop, was closer to the plant control system than the TRAC model based on the collapsed water level in the SG downcomer. The other major dif-ference was the way the upper head was modeled in the two calculations.
TRAC had no dead end volume for the upper head; therefore, the void accumulation did not occur in the upper head, but instead migrated to the candy cane which resulted in the termination of natural circulation in the unaffected loop B.
However, if this natural circulation was maintained a little longer in loop B, the diff erences in the two hot leg fluid temperatures would be less and the multi-dimensional effect would be less significant. The RELAPS model of the upper head was more appropriate. Based on the comparison of the two calcula-tions, a reasonable yet conservative procedure would be to delay the restart-of RCPs in the RELAPS calculation until IRAC's RCPs restart time of 526 sec-onds, which would result in a minimum downcomer fluid temperature of 450K.
The second transient compared was initiated by the f ailure of all four TBVs at the full open position af ter a turbine trip. This transient is like a small break in the steamline.
Here the initial conditions (full power for TRAC and hot standby for RELAPS) and additional failures were different. The codes predicted the transients reasonably well.
However, the important rea-sons for the differences in the calculations were the controller f ailure to throttle EFW based on the secondary side level and the f ailure of the operator to close the TBVs at 600 seconds in the TRAC calculation.
The third transient compared was the Small Break (2-inch) LOCA in a hot leg. The ICS was assumed to work as designed.
Both TRAC and RELAPS computed a continuous drop in the primary side pressure during the transient.
This made the transient less critical for PTS. However, comparison of the two cal-culations indicated that there were many differences.
The codes modeled the reactor trip differently.
RELAPS correctly based it on the low primary side pressure while TRAC assumed it to occur at 0.5 second. Furthermore, af ter the loss of natural circulation due to candy cane voiding, TRAC computed flow os-cillations in the cold legs while RELAPS predicted a stable circular flow be tween the cold legs connected to the common steam generators.
The loop os-cillations in the TRAC calculation warmed up the cold leg and the downcomer fluids.
However, it is not clear if these oscillations are real.
- Moreover, the TRAC calculation was not carried out far enough in time to determine the minimum downcomer fluid temperature with confidence.
The RELAPS calculation, on the other hand, is more complete and looks reasonable. _
M
- 6. REFERENCES 1.
Bassett, B., Boyack, B., Burkelt. N., Ireland, J., Lime, J., and Neiton, R.,
(1983), " TRAC Analysis of Severe Overcooling Transients for the Oconea-1 PWR," LA-UR-83-3182 November, 1983.
t 2.
Fletcher, C.
D.,
Bolander, M.
A.,
Stitt, B.
D.,
and Waterman, M.
E.,
)
"RELAPS Thermal-Hydraulic Analysis of Pressurized Thermal Shock Sequences for the Oconee-1 Pressurized Water Reactor," NUREG/CR-3761, June 1984.
l t
s L
APPENDIX This section includes copies of all Brookhaven National Laboratory formal i
communications with the Nuclear Regulatory Commission regarding the Oconee-1 PTS study during the period August - October 1982.
1
\\
))
BROOKHAVEN NATIONAL LABORATORY
{(l ASSOCIATED UNIVERSITIES, INC.
j Upton, Long Island, New York 11973 (516)282s2438 Department of Nuclear Energy FiS 666' August 31, 1982 Dr. Louis M. Shotkin Analytical Models Branch Mail Stop 1130 SS Office of Nuclear Regulatory Research U. S. Nuclear Regulatory Commission Washington, DC 20555
Subject:
Quality Assurance of TRAC-PF1 Calculations for Oconee Pressurized Thermal Shock (PTS) Study Dear Lout This letter presents the BNL comments on the TRAC input listing and steady-state calculation for the Oconec plant as received from Dr. Nelson S. DeMuth of LANL with his letter dated August 20, 1982. The information was received at BNL on August 23, 1982.
In view of the tight schedule and as agreed upon between the NRC and BNL staff on July 30, 1982, the main emphasis at BNL was to check the consistency of the input parameters such as correct component connections, symmetry of the two primary loops, order-of magnitude of the component volumes and flow para-meters, and the correct input options. The steady state output was che: ked for the symmetry, correct thermal-hydraulic conditions and any unexpected or unusual nunbers. The steady-state thermal-hydraulic parameters, i.e.,
pressure, temperature, flow rate, etc., and the overall steady state heat balance were compared to the 100% operating condition for the Oconec plant an reported in the Oconce FSAR.
The code used for the LANL calculation was an interim version of TRAC-PF1/ MODI. Although BNL was provided with an interim input manual, some sec-tions of the manual were not consistent with the input listing. A more com-plete input description would have been more helpful for the Q/A activity.
Although no gross error was discovered either in the input listing or in the steady state output, there are several areas which need further attention before the transient calculations should begin. They are grouped in two parts: (1) input listing, and (2) steady state output.
L. M. Shotkin 8-31-82 1.
Comments on the Input Deck a)
The Oconee plant has two loops with two cold legs in each loop. In the LANL input deck the two cold legs of loop B were combined into one single loop while the two cold legs of loop A were left separate.
This in turn caused a non symmetric nodalization of the 3-D vessel module. However, one of the ORNL specified event sequences (SBl to SBA2 in the letter f rom R. C. Kryter dated June 9,1982) would re-quire restart of one reactor coolant pump in each loop. Therefore, we suggest that both loops be modeled with two separate cold legs with separate RC pumps.
b)
For some components, the volume was not equal to the product of the length and the flow area. This may be due to the presence of inter-nal structures in the components like plena, steam generator, pres-surizer, etc. The same type of discrepancy also exists in the feed water train components such as condensate booster pumps (component 50), hot well pumps (component 51), branch between C and D heaters (component 52) and condenser (component 55). Also, the volumes of component 14 and 19 were interchanged by mistake.
c)
For some components the input data of the two loops were not sym-metric.
Specific examples aret (1) TWTOLD = 0.8 for component 9 in Loop B
= 0.9 for component 19 in Loop A
= 0.8 for component 39 in Loop A This may cause unexpected non symmetric boundary conditions when the HplS is triggered.
(11) GRAVITY = 0.419 in the first ec11 of primary tube of component 12 (steam generator) in Loop A.
= 0.5655 in the first ec11 of primary tube of component 2 (steam generator) in Loop B.
This appears to be due to the actual difference of the two loops.
(iii) FRIC = 0.02 in the last cell of secondary tube of component 6 (TEE).
= 0.0 in the last cell of secondary tube of component 16 (TEE).
(iv) The maximum rates of valve flow area adjustment (VARMX) for the accumulator check valves (component 90 and 80) are dif ferent (1.0 and 4.0, respectively).
L. M. Shotkin 8-31-82 d)
Exit flow area of the PORV (component 25) on the pressurizer was set to zero. Since the pressurizer was isolated f rom the rest of the system during the steady-state calculation, it did not affect the steady-state results. However, this error must be corrected for the transient calculation, otherwise, no fluid will leave the system even if the PORV opens.
e)
Initial void fraction in the feedwater lines, i.e., the secondary pipes of components 47 and 57 should be 0.0 instead of 1.0 as used in the input deck. Although a void fraction of zero was obtained in the steady-state, correction of this initial value of void fraction might help in the steady-state calculation.
f)
The input deck specified the steam generator exit conditions for the steam and the hot well exit temperature for the liquid in the FILL component at the condenser inlet. Although this is inconsistent and incorrect, it did not significantly affect the steady-state result as
.the code employed the liquid temperatere for the steam for the con-stant mass flow option used in the FILL. Specification of correct inlet conditions is, however, recommended for the reason given later (see Items 2(f) and 2(g)).
2.
Comments on the Steady-State Output a)
A reasonably good steady-state was obtained. The mass, energy and heat balances for the primary and the secondary sides matched within 1%.
However, the calculated steady-state gave about 5% larger flow rate and 5% less temperature rise across the core than the 100% oper-ating condition given in Oconee FSAR. However, the actual plant operating condition could be slightly different than that in the FSAR.
b)
The vapor entering the steam generator downcomer through the aspirator did not completely condense even at the bottom of the downcomer. This resulted in a reduction of water inventory in the downcomer and may cause higher primary fluid temperature in the case of a loss of feedwater transient. In that case, the code results will be non-conservative. The same phenomenon was observed in the BNL TRAC-PFi calculation of the B&W OTSG tests and it was suggested that the rate of condensation be increased.
c)
The steady-state results of the two loops were not very symmetric.
In some components (for example, in two steam generators) dif ferences of a few degrees in temperature were observed. However, this may not have any significant effect on the transient calculation.
I d)
The fluid velocity in the surge line was still substantial ( vt " be 0.4777 m/sec). Considerably longer steady-state calculation may required to reduce this further. Again, this may not have any significant effect on the transient calculation.
1 l
l !
I
L. M. Shotkin 8-31-82 e)
A significantly high liquid velocity (4.2 m/sec) was observed at the liquid-vapor interface in both accumulators or core flood tanks.
This did not affect the steady state calculation because the accumu-lators were isolated from the rest of the system during the steady-state calculation. However, the cause of thir, irregular behavior should be studied before the transient calculation begins.
It may be related to the interface sharpener.
a f)
The calculated steady-state pressure at the top of the condenser was approximately 1.1 bar, whereas that at the hot well exit was ap-proximately 0.48 bar. Both are considerably higher than the desired values. From the Oconee FSAR and the SAI supplied information, it can be determined that the pressure in the condsenser should be in between 0.228 and 0.1 bar.
Further attention should be paid to the condenser calculation since one of the ORNL-specified event sequences (FW13 to F012) would require an accurate prediction of the condenser pressure for the Turbine Bypass Valve (TBV) closure.
g)
The calculated feedwater conditions were close to the SAI specified values. Various parameters such as tha condenser heat transfer coef-ficient, the ambient temperature, the feedwater heater heat sources and the form loss coef ficients were adjusted to achieve this. How-ever, if the steam and liquid conditions at the condenser inlet were corrected and the correct condenser pressure were obtained, the feedwater conditions would have been different from the desired values with the current parameters. Therefore, further adjustment of the various parameters, mentioned above, is needed for the desired steady state conditions at both the condenser and the feedwater inlet to the steam generator.
Most of the above comments have been communicated to Mr. Britt Bassett of LANL over the telephone on August 27, 1982.
It must be reiterated that the above comments are not exhaustive and a more thorough quality assurance effort would require more time.
In addition, we would require information on all model changes between Version 7.0 and the version being used for the LANL cal-culation to perform an in-depth review of the end results.,
,n
_r-y-,
L. M. Shotkin 8-31-82
)
Please feel free to contact me at FTS-666-2438 if you need any further clarification. We are now reviewing the RELAPS input listing for the Oconee plant and a separate letter will follow shortly. We received the RELAPS input listing from INEL on August 30, 1982.
With best regards, Sincerely, fjsb',.
JJ/USR/PS taf Pradip Saha, Group Leader LWR Code Assessment & Application cct F. Odar, NRC N. Zuber, NRC C. E. Johnson, NRC N. S. Delluth, LANL B. Ba s se t t, LANL R. C. Kryter, ORNL A. C. Peterson, INEL W. Y. Kato, BNL R. J. Cerbone, BNL J. II. Jo, BNL U. S. Rohatgi, BNL e
)
BROOKHAVEN NATIONAL LABORATORY
((
ASSOCIATED UNIVERSITIES, INC.
Upton. Long biond. New York 11973 (516)282s Department of Nuclear Energy FTS 666/ 2438 September 15, 1982 Dr. Louis M. Shotkin Analytical Models Branch Office of Nuclear Regulatory Research Mail Stop 1130 SS U. S. Nuc1 car Regulatory Commission Washington, DC 20555
Subject:
Quality Assurance of RELAP5 Calculations for Oconee Pressurized Thermal Shock (PTS) Study
Dear Lou,
This letter presents the BNL comments on the RELAPS input listing for the Oconee plant as received at BNL on September 1, 1982.
This input listing supersedes the previous RELAP5 input listing received on August 30, 1982. Our comments are, of course, based on the latest input.
In view of the tight schedule and late arrival of the input listing, we concentrated on checking the consistency of the input parameters such as correct component connections, symmetry of two loops, order of-magnitude of the component volumes, flow parameters and the correct input options. Efforts were also made to compare the TRAC-PF1 and the RELAPS input listings and to look for any significant differences in modeling various components.
Trips relevant to the steady state calculation were checked; however, the remaining trips related to the transient calculation will be reviewed after the final input listing in received.
We were told by the INEL staf f (M. Waterman and D. Fletcher) that the input listing was of preliminary nature and it was still being modified at INEL.
Furthermore, the code version to be used for the PTS calculation at INEL is different f rom the RELAP5/ MODI code, and no updated manual was sent with the input.
The task of reviewing and comparing the RELAP5 and TRAC listings became even more time consuming because different units were used in these listings.
The LANL staff used the SI units for TRAC, whereas the INEL staf f used the British units for RELAP5.
It is recommended that both labora-tories (LANL and INEL) provide results in the same units. Otherwise, a direct comparison between the TRAC and RELAPS results would be very time consuming.
It is our understanding that RELAPS results could be obtained in either the L. M. Shotkin 9/15/82 British or the SI units, whereas the TRAC results are always in the SI units.
Because the plant data are usually in the British units, the RELAP5 results should probably be obtained in both SI and British units.
No major error was discovered in the RELAPS input listing. However, some differences between the TRAC and RELAPS inputs were found and there are several areas which need further attention before the transient ca'1culations should begin.
Our comments, given below, are in two parts.
The first part deals with the RELAPS input and the second part describes the dif ferences be-tween the TRAC and RELAPS inputs.
1.
Comments on the RELAP5 Input Deck a)
Primary Loop:
(i) There is a connection between the lower outer upper plenum (com-ponent 545) and the downcomer inlet annulus (component 565). It is not clear what feature of the reactor is being modeled with this connection since the vent valve is already being modeled by another connection (component 536).
(ii) There is a big jump in the hydraulic diameters for the last two volumes of the downcomer (component 570).
The reason for this is not clear.
(iii) The accumulator model is using a junction flag of 0020 while the recommended value in the RELAP5/ MODI manual is 0000.
It is not clear whether this is due to the change in the code version.
(iv) The homogeneous equilibrium option is being used in the secon-dary side of the steam generators. This is not appropriate for the expected flow situation in the steam generator secondary side.
b)
Feedwater Trains (1) The input deck does not have a model for the condenser and the hot well.
The secondary loop will, therefore, not be closc1.
Either the flow or pressure boundary condition along with the fluid temperature and quality at the exit of the hot well has to be specified.
It is not clear how the code can calculate some of the ORNL-specified transients without a condenser and hot well model.
L. M. Shotkin 9/15/82 (ii) The hot well exit pressure used in the input is 5.7 psi (0.393 bar) which is larger than the SAI recommended value of 0.103 bar.
(iii) D and E heater drains have been modeled with time dependent j unctions with velocity boundary conditions.
It will be more accurate to use the flow boundary condition which is allowed in the code.
(iv) Input description contains some apparent errors in the pump models.
The rated torques for the hdt well, booster and main feedwater pumps are specified as 10-6 lb-ft which is unrealistically low.
Also, the rated fluid density and motor torque have been specified as 0.0 for all three pumps. These should be changed prior to the steady-state calculations.
c)
Control System:
The control system for the feedwater, emergency feedwater, and the turbine bypass valve in the RELAPS input deck was reviewed and com-pared with the SAI version of the control system which was reviewed earlier at BNL. Several discrepancies between the two were detected.
Ilowever, we were informed on September 7,1982 by.he INEL staff (M.
Wate rman) that the RELAPS control system was substantially modified from the input listing received by us.
Therefore, our review of the RELAPS control system does not cecm to be applicable anymore, and it will not be discussed here.
2.
Comparison 3etween the RELAPS and Tiuc Inputs Both inputs are quite similar in their description of the plant.
liow-ever, there are a fe.i exceptions where either some components are missing or there is more than 5% difference in the dimensions of the same component in the two inputs. These are discussed belows a)
TRAC-PFL input has a condenser and hot well model which RELAP5 input does not have, flowever, the RELAPS deck accounts for the safety re-lief valve on the pressurizer which is missing in the TRAC input.
b)
Both codes are using different characteristics for the same primary loop recirculation pumps.
The TRAC input contains the pump model based on the LOFT pump, while the RELAP5 model is based on the Westinghouse pump. The recirculation pump in the Oconee plant has a specific speed of 4000 which is higher than that of the LOFT pump (specific speed 3300) but lower than that of the Westinghouse pump in RELAPS (specific speed 5200). It will be more appropriate to use the.
L. M. Shotkin 9/15/82 Bingham pump characteristics of RELAP5, as it has a specific speed of 4200.
Furthermore, the TRAC model will compute coastdown while RE-LAP 5 will use a table.
c)
RELAPS accounts for the heat stored in the shell wall of the steam generator which is not possible in TRAC.
This is an additional en-ergy available to the feedwater and will make the primary loop cool-ing rate less severe. Ilowever, the quantitative impact of neglecting this stored energy in TRAC will depend on the scenario and it is not kr.;wn at this time.
d)
The feedwater heaters have been modeled differently by the LANL and the INEL staff.
The TRAC input combines the A & B heaters and does not account for the energy input into the demineralizer.
The fol-lowing table presents a comparison of the heater powers used in the two inputs.
llea te r TRAC Input RELAPS Input Number MW MW A
79.9 l
[ 227.5 B
77.4 C
213.0 251.1 D
113.6 146.4 E
79.7 130.8 F
78.4 120.7 Demineralizer 9.3 TOTAL 712.2 815.6 Because of the tight schedule, we could not determine which input better represents the plant conditions.
e)
There are some differences in the component sizes which are num-marized in the following table along with the values suggested by the Duke-Power in their letter dated January 12, 1982 to R. C. Kryter of ORNL:
L. M. Shotkin 9/15/82 i
Component TRAC RELAPS Duke-Power Data Vessel Downconer 21.89 m3 21.36 m3 Volume Lower Plenum 24.45 m3 19,39 3 Volume Core Volume 22.91 m3 23.21 m3 Upper Plenum 47.21 m3 43.62 m3 Volume Total Vessel 116.5 m3 107.6 m3 114.9 m3 Volume Downconer 0.55 m 0.29 m Hydraulic Dia.
i Downcome r Inner Wall 0.0762 m 0.0507 m Thickness 2
Pressurizer Pressurizer 42.51 m3 45.15 m3 42.47 m3 Volume Level 5.6 m 6.47 m 5.59 m i
Surge Line 0.56 m3 0.59 m3 0.57 m3 Volume Surge Line 0.254 m 0.222 m Diame ter Steam Generator Feedwater inlet 0.04 m2 2.164 m2 Area Secondary side 4.345 m2 1.688 m2 Flow Area Hydraulic Dia.
0.006 m 0.02 m Aspirator area 2.167 m2 0.972 m2 i
1
^l L. M. Shotkin 9/15/82
-l Both codes uso approximately equal volumes ' for similar components.
The above table shows the major differences found in the geometric parameters in j
the two inputs. The differences in the volumes will affect the liquid inven-tory and. therefore, the primary _ side cooling rate. - The dif ferences. in the steam generator feedwater inlet area, hydraulic diameter and aspirator area are probably due.to adjustments of the additive friction factors to achieve the correct steady-state pressure drops.
The ' differences in the secondary side flow area and other volumes may be due to the differences in the internal structures used.
The quantitative impact of the above differences is scenario-dependent and not. easy to determine a priori.
It is recommended that the INEL and LANL staff resolve these differences before they proceed with the transient calculations.
Most of the above comments have been transmitted to the INEL staff (Mike Waterman and Don Fletcher) on September 9,1982.
It must be emphasized that-these comments are not exhaustive and a more in-depth quality assurance will require more time.
For the future review and quality assurance of the TRAC and RELAP5 calculations it is strongly recommended that: (1) all inputs and-calculations are in the same units, (2) all pertinent information on the code /model changes since the last released version be supplied, and (3) no work is sent for Q/A unless it is reasonably complete.
Please feel free to contact me at FTS-666-2438 if you need any further clarification. With best regards, Sincerely yours, USR/JJ/PS taf Pradip Saha, Group Leader LWR Code Assessment and Application cc F. Odar. NRC N. Zuber, NRC C. E. Johnson, NRC T. Charlton, INEL A. C. Peterson, INEL M. Waterman, INEL D. Fletcher, INEL N. S. DeMuth, LANL R. C. Kryter, ORNL W. Y. Kato, BNL R. J. Cerbone, BNL J. H. Jo, BNL U. S. Rohatgi, BNL QUALITY ASSURANCE PROGRAM FOR PTS CALCULATION PREPARED BY J. J0, U. S. ROHATGI AND P. SAHA DEPARTMENT OF NUCLEAR ENERGY BROOKHAVEN NATIONAL LABORATORY UPTON, NY 11973 PRESENTED AT OCONEE PTS ASSESSMENT MEETING OAK RIDGE NATIONAL LABORATORY OAK RIDGE, TENNESSEE SEPTEMBER 22, 1982 BROOKHAVEN Nail 0NAL LABORATORY l)l)l A5500ATED UNIVERSITIES, INC(l(Il j
I i
OBJECTIVES OF BNL Q/A PROGRAM e
REVIEW NEW THERMAL-HYDRAULIC MODELS e
REVIEW PLANT DECKS e
REVIEW THE THERMAL-HYDRAULIC CALCULATIONS e
COMPARE TRAC AND RELAPS CALCllLATIONS FOR OCONEE I
I BROOKHAVEN NATIONAL LABORATORY l} ggl ASSOCIATED UNIVERSITIES, INC.(IIll s
i
4 REVIEW 0F PLANT DECK e
ASSURE THE CORRECT AND COMPLETE MODELING OF THE PLANT
- ACCOUNTING FOR ALL COMPONENTS NEEDED FOR A t
SPECIFIC SCENARIO
- DIMENSIONS, LENGTH, AREA AND VOLUME, ELEVATIONS
- INTERNAL STRUCTURES
- HEAT STRUCTURES
- COMPONENT CONNECTIONS
- CORRECT OPTIONS
- VALVE SETTINGS - PRESSURIZER, ACCUMULATOR, HPI, LPI, VENT
- PUMP CHARACTERISTICS
- LOSS COEFFICIENTS
- CONTROLS t
i i
- BOUNDARY CONDITIONS, HPI AND LPI FLOW AND TEMPERATURE, VALVE EXIT PRESSURES, ETC.
/
i BROOKHAVEN NATIONAL LABORATORY l} g)l A5500ATED UNIVERSITIES, INC.(Illl
REVIEW 0F THERMAL-HYDRAULIC CALCULATIONS 1
1 STEADY-STATE e
SYMMETRY OF TWO LOOPS e
MASS AND ENERGY BALANCES FOR PRIMARY AND SECONDARY SIDES 4
e COMPARIS00 WITH PLANT CONDITIONS
- PRESSURE (PRIMARY AND SECONDARY)
- TEMPERATURE
- FLOW RATE
- PRESSURIZER LEVEL
- STEAM GENERATOR LEVEL, STEAM GENERATOR EXIT TEMPERATURES
-i
(
BROOKHAVEN NATIONAL LABORATORY l}gyl l
A5500ATED UNIVERSITIES, INC.(Itil l
i i-l :
L._
PRESSURE e
FLUID TEMPERATURE AND H.T. COEFFICIENT IN DOWNCOMER e
COLD AND HOT LEG TEMPERATURE AND FLOW RATE e
PRESSURIZER LEVEL e, SAFETY / RELIEF VALVE FLOW RATES e
HPI, ACCUMULATOR, LPI FLOW RATES SECONDARY LOOP e
PRESSURE e
S.G. INLET FLOW RATE, TEMPERATURE AND QUALITY o
EMERGENCY FEEDWATER FLOW RATE e
S.G. LEVEL e
S.G. EXIT FLOW RATE, TEMPERATURE AND QUALITY e
TURBINE BYPASS AND RELIEF VALVE FLOW RATES e
HOT WELL PRESSURE, TEMPERATURES AND LEVEL e
FEEDWATER HEATER EXIT TEMPERATURES 1
BR00KHAVLN NATIONAL LABORATORY l} g)l ASSOCIATED UNIVERSITIES, INC.(1Ill i.
r
---.e n.,- -. - - -,..--,.. - - ~-
,-n.--
w--.....--
- -s 1
COMMENTS ON TRAC INPUT e
NO MAJOR ERRORS e
MINOR ERRORS TRANSMITTED TO LANL-(BNL LETTER DATED 8/31/82) e SAFETY RELIEF VALVE ON PRESSURIZER MISSING e
MODEL MAY BE IMPROVED WITH 4 COLD LEGS COMMENTS ON TRAC STEADY STATE e
REASONABLY GOOD STEADY STATE e
LOW WATER INVENTORY IN S.G. DOWNCOMER DUE TO lNSUFFICIENT CONDENSATION e
SIGNIFICANT LIQUID VELOCITY AT THE INTERFACE IN THE ACCUMULATOR e
CONDENSER MODEL NEEDS FURTHER ASSESSMENT.
COMPUTED PRESSURE T00 HIGH.
BROOKHAVEN NATIONAL LABORATORY l} gy l A5500ATED UNIVERSITIES, INC.(EIll -
t_
COMMENTS ON RELAP5 INPUT e
NO MAJOR ERRORS e
EXTRA CONNECTIGN OTHER THAN THE VENT VALVE BETWEEN UPPER PLENUM AND DOWNCOMER e
HOMOGENEOUS EQUILIBRIUM OPTION FOR SECONDARY SIDE OF THE STEAM GENERATOR e
NO CONDENSER /H0T WELL MODEL e
PUMP MODEL HAS ERROR IN RATED TORQUE AND VALUES.
i-I l_
BROOKHAVEN NATIONAL LABORATORY l)l)l l
A5500ATED UNIVERSITIES, INC.(IIll l t
1 COMPARIS0N OF TRAC AND RELAP5 INPUTS e
MOST OF THE COMPONENT MODELS ARE SIMILAR.
e FEW DIFFERENCES ITEM TRAC RELAPS PLANT 1)
R.C. PUMP LOFT WESTINGHOUSE SP SPEED 4000 SP SPEED 3300 SP SPEED 5200 2)
FEEDWATER 712 2 MW 815 6 MW HEATER TOTAL POWER 3)
PRESSURIZER 56M 6 47 M 5 59 M LEVEL 4)
VESSEL 116 5 M3 107 6 M3 114 9 M3 VOLUME 5)
S.G. WALL NOT MODELED MODELED STORED ENERGY BROOKHAVEN NATIONAL LABORATORY l}l}l ASSOCIATED UNIVERSITIES, INC.(llll SAI CONTROL SYSTEMS i
1 SEVERAL ERRORS DETECTED AND COMMUNICATED TO SAI WHICH HAVE BEEN CORRECTED BY sal.
2 MODEL ONLY APPLICABLE DURING THE PTS TRANSIENT AETfLg REACTOR AND TURBINE TRIP.
3 DISCONTINUITY IN SOME CONTROL VARIABLES IS EXPECTED WHEN CONTROL SYSTEM IS ACTIVATED.
- 4. 'PRESENT CONTROL SYSTEM CANNOT MODEL OPERATOR ACTION AND EQUIPMENT FAILURES.
l BROOKHAVEN NATIONAL LABORATORY l}l)l A5500ATED UNIVERSITIES, INC.(IIll., _..
RECOMMENDATIONS 1
TRAC CONDENSER MODEL SHOULD BE ASSESSED.
A CONDENSER i
MODEL SHOULD BE ADDED IN RELAP5 2
COMPLETE CONDENSATION SHOULD BE ACHIEVED IN THE STEAM GENERATOR DOWNCOMER AT STEADY STATE.
3 INPUTS AND CALCULATIONS SHOULD BE IN THE SAME UNITS FOR BOTH TRAC AND RELAP5 4
ALL THE DESCRIPTIONS OF MODEL CHANGES AND UPDATES SINCE THE LAST RELEASED VERSION SHOULD BE PROVIDED.
5 NO WORK SHOULD BE SENT FOR 0/A UNLESS IT IS REASONABLY COMPLETE.
4 BROOKHAVEN NATIONAL LABORATORY l}ggl A5500ATED UNIVERSITIES, INC.(IIll
b -} I BROOKHAVEN NATIONAL LABORATORY
{' (
ASSOCIATED UNIVERSITIES, INC.
Upton, Long Island. New York 11973 (516) 282s Departmentof NuclearEnergy FTS 666-2438 October 14, 1982 Mr. J. D. White ORNL PTS Integration Oak Ridge National Laboratory P.O. Box X Oak Ridge, Tennessee 37380
Subject:
Control System for the PTS Study
Dear Mr. White:
The purpose of this letter is to clarify the BNL concents made at the PTS study meeting at ORNL on September 22, 1982, concerning the SAI control sys-tem. -We (BNL) stated that:
1.
The SAI control system is applicable only after the reactor and/or turbine trips.
It is not applicable during the initiating events, i.e., between the steady-state and the reactor and/or turbine trips.
2.
Discontinuity in some control variables may occur when the control system is activiated.
Clarification on Comment 1 According to the SAI [1], some components and subsystems were not included in the control system model. Among them are:
Unit Load Demand Development subsystem, including loss of feed-water pumps or reactor coolant flow, since "these conditions were judged to be the initiating events." This subsystem was replaced by the Unit Load Demand Trip. Unit Load Demand ramps down at a rate-limited 20% per minute on the triggering of this trip from full power.
Turbine control since all cases of interest involved early turbine trip. The only aspect of turbine control modeled was the turbine by-pass valve control.
[1] R. A. Hedrick, Letter to R. C. Kryter of ORNL on July 23, 1983.
r:
I l
J. D. White 10/14/82 The above statements clearly indicate that the control system developed was intended to be activated at the point of reactor and turbine trips and not to be used during the initiating events, i.e., between the steady-state and the reactor / turbine trips. We were merely pointing out that the users (INEL and LANL) should be aware of this, limitation and be prepared to provide proper boundary conditions during the initiating event calculation, i.e., between the steady-state and the reactor / turbine trips.
Among the eleven scenarios or event sequences proposed by ORNL [2], for the Oconee PTS study, seven involved immediate reactor and turbine trips (3 steamline breaks, 2 turbine bypass valve failures and 2 feedwater overfeed transients). For these transients, the real time between the steady-state and the reactor / turbine trip is either zero or very short so that the SAI control system is applicable. However, in four other scenarios (2 small break LOCAs and 2 loss of main feedwater), the real time between the steady-state and the reactor / turbine trip may not be very short and the " proper" boundary condition may not be obvious. The overcooling transient which occured at Rancho Seco on March 20,1978 [3] is similar to two ORNL-specified loss-of-feedwater trans-ients with subsequent restoration. This incident was initiated by the loss of main feedwater which was triggered by the loss of power to the control system.
The loss of feedwater caused the reactor coolant temperature and primary side pressure to increase. After approximately 15 seconds fr'om'the loss-of-feed-water, the reactor was tripped by the high pressure trip,9hich in turn trip-ped the turbine. During this p'eriod (~15 seconds) one needs the boundary condition at the steam generator exit side. We do not know how the users (INEL and LANL) will provide these boundary conditions and whether they will significantly impact the final results.
Clarification on Comment 2 This is a general caveat for numerical instabilities which may be trig-gered when the control system is suddenly activated at some mid-point of cal-culation. Again, we do not know if this will occur in this particular con-trol system or what effect it will have on the calculation if it does. How-ever, we felt that this might be a useful point to keep in mind if the calcu-lation exhibits difficulty in converging or produces unreasonable results at the activation of this control system.
l
[2] T. J. Burns, " Status of Event Sequence Specification," Presented in the meeting at ORNL on September 22, 1982.
[3] R. Lobel, " Summary of Meeting held at Rancho Seco Nuclear Power Plant on June 10, 1978 to Discuss a Recent Cooldown Event," Memorandum for Paul S.
Check, Chief, Reactor Safety Branch, DOR, (1978)
J. D. White 10/14/82 Should you have any further questions on this issue, please call Dr. Jae do at FTS-666-2337 or me at FTS-666-2438. With best regards.
Sincerely yours,
[7hdf-i JJ/PS :a f._
Pradip Saha, Group Leader LWR Code Assessment and Application cc: C. Johnson, NRC T. Lee, NRC L. M. Shotkin, NRC F. Odar,.NRC R. C. Kryter, ORNL R. A. Hedrick, SAI-0R N. DeMuth, LANL B. Basset, LANL T. Charlton, INEL A. C. Peterson, INEL F
f-1 l
DISTRIBUTION LIST NRC_
NRC Roy Woods, GIB Fuat Odar, RSRB Edward Throm..RSB William Beckner, RSRB Newton Anderson, RIB Ralph Landry, RSRB Normal Lauben, RSB Harry Tovmassian, RSRB Raymond Klecker, MTEB Jose Reyes, RSRB (3)
' Pryor Randall, MEBR Paul Boehnert, ACRS Donald Kirkpatrick, EGRB Guy Vissing, ORB Harold Ornstein, AE0D William Johnston, RIV Wayne Lanning, AE00 Ashok Thadani, RRAB Demetrios Basdekas. ICBR Sanford Israel, RRAB Jack Strosnider, MGBR
- 0. E. Bassett, DAE Joseph Murphy, DRA Harold Denton/Edson Case, NRR Frank Schroeder, AD/GP G. Lainas, AD/0R Karl Kniel, GIB Thomas Murley, Region I Milton Vagins, MEBR Lawrence Shao, DET James Clif ford, PSRB George Knighton, LB3 Lambros Lois, CPR John Austin OCM Charles Serpan, MEBR James Buzy, LQB Felix Litton, GIB James Milhoan, OCM Merrill Taylor Dennis Ziemann, DHFS Andrew Szukiewicz, GIB Daniel Garner, OCM Charles Kelber, DAE Richard Johnson, GIB Novak Zuber, RSRB Edward Goodwin, PE Denwood Ross, RES Thomas Novak, AD/L Robert Bernero, ASTP0 Stephen Chestnut, OCM Malcolm Ernst, DRA Alan Rubin, GIB
' Ga ry Burdick, RRBR Charles Morris, RRAB Leonard Soffer, ASTP0 Alfred Spano, RRAB Frederick Manning, RRBR Martin Virgilio, ICSB Pradyot Niyogi, ASTP0 Thomas Dunning, ICSB Charles Overby, HFSB Charles Rossi, ICSB Thomas Ryan, HFSB Surin J. Bhatt, MTEB Rubert Curtis, CSTB
' Robert Senseney, IP Elpidio Igne, ACRS Public Document Room Darrell Eisenhut, DL James Van Vliet ORB 4 Roger Mattson, DSI Sydney Miner, ORB 4 Richard Vollmer, DE Steven Varga, ORB 1 Hugh Thompson, DHFS Glode Reque, ORB 1 Themis Speis, DST Edmond Tourigny, ORB 3 Louis Shotkin, RSRB James Miller, ORB 3 Carl Johnson, RRBR (2)
John Stolz, ORB 4 l '
Other'0rganizations L. T. Peterson M. Modarres Pacific Northwest Laboratories University of Maryland Post Of fice Box 99 Department of Nuclear Engineering Richland, Washington 99352 College Park, Maryland 20742 Robert Gill Theofanis G. Theofanous Duke Power Company Purdue University 422 South Church -Street 132 Pathway Lane Charlotte, North Carolina 28242 West Lafayette, Indiana 47906 Sol Levy Steve Mirsky Sol Levy, Incorporated Baltimore Gas and Electric Charles Center 1999 South Bascom-Avenue Post Office Box 1475 Suite 2200
- Baltimore, Maryland Campbell, Cali fornia 95008 Rudy Oliver Paul Rothe Carolina Power and Light Creare Incorporated Post Office Box 1551 Post Office Box 71 411 Fayetteville Street Hanover, New Hampshire 03755 Raleigh, North Carolina 27602 Bart Daly Avtar Singh Los Alamos National Laboratory Electric Power Research Institute Post Of fice Box 1663 Post Office Box 10412 Los Alamos, New Mexico 87545 Palo Alto, California 94304 Pradip Saha Dan A. Prel ewicz Brookhaven National Laboratory NUS Corporation Building 130 4 Research Place Upton, New York 11973 Rockville, Maryland 20850 Jan Koenig D. A. Peck Los Alamos National Laboratory Combustion Engineering P. O. Box 1663 P. O. Box 500 Los Alamos, New Mexico 87545 Windsor, Connecticut 06095 James D. White Debu Majumdar Oak Ridge National Laboratory U.S. Department of Energy P. O. Box Y Idaho Operations Office Oak Ridge, Tennessee 37830 550 Second Street Idaho Falls, Idaho 83401 Yih-Yun Hsu Daniel Speyer University of Maryland 160 West End Avenue Department of Nuclear Engineering New York, New York 10023 College _ Pa rk, Ma ryla nd 20742 George Irwin R. L. Turner-University of Maryland.
Westinghouse Electric Corporation Department.of Nuclear Engineering Monroeville Nuclear Center - Bay 319 College Park, Maryland 20742 P. O. Box 355 Pittsburgh, Pennsylvania 15230 Jerry Phillips Frank Walters Carolina Power.and Light Babcock and Wilcox Post Of fice Box 1551 Post Of fice Box 1260 411 Fayetteville Street Lynchburg, Virginia 24505-1260 Raleigh, North Carolina 27602 Donald Ogden EG&G Idaho, Incorporated 550 Second Steet Idaho Falls, Idaho 83401 4
l
- U""
U.S. NUCLE AR REGULATORY COMMeSSION BIBLIOGRAPHIC DAT,A SHEET BNL-NUREG-51750 4 TITLE AND SUtsTITLE (Ader n'otume Na, et esanpaa
- 2. (Leave arme/
Assessment of Selected TRAC and APS Calculations e
for Oconee-1 Pressurized Thermal ck Study 1 RECIPIENT'S ACCESSION N
- t. AU THOHIS)
- 5. D ATE HEPORT COMPLE[D
" " November [ *" 1984 U. S. Rohatgi, J. Pu, P. Saha, and. Jo 9 Pk HFOHMING OHGAN3/ATION N AME AND MAILING A RESS tractuae I,p Code /
OATE REPORT IS[ED Department of Nuclear Energy
/
Ive^a
=o~'"
Brookhaven National Laboratory J
[
Upton, Long Island, New York 11973 5"'*'*'*"*
.. ae y.,
it. SPONSOHING OHGANelATION NAMk AND M AILING ADORE (tactuae Inp Coael Reactor Systems Research Branch 10 PR[O "CTITASK/ WORK UNaf NO Division of Accident Evaluation i i.
NO.
Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Commission A-3266 Wathinatnn_ nc 205R4 IJ. 8 YPE OF Hk PUH T PE RIOD COV E fE D finclussrv deses/
Formal Report
(
li SUPPLEMENTAHY NOIkS 34 It'd** cd*
- i 96 ArsTH ACT (200 waras er truJ Several OConee-1 overcooling [ansients that Were Computed by LANL and INEL using the latest versions of TRAC-PF1 and AP5/M001.5 codes have been reviewed by BNL. Three of these transients were selected for iled review as they either had the potential of challenging the integrity of the pre r vessel or highlighted the effect of code differences. These are (1) Main Steam Line reak MSLB), (2) All Turbine Bypass Valves Stuck Open, and (3) 2-Inch Small Break LOCA.
f li nEY wuMUS ANu OUCUMkNT ANALYSIS 1 74 DE SCHiP TOHS Pressurized Thermal Shock DOwncomer Fluid Temperature DOwncOmer Pressure IIPI Vent Valve Flow i
+
t he IOk N TI F IE RS. OPE N E N DE D TE RMS IS AV AILABILITY ST ATEMENT 19 SECURITY CLASS iThes eeporrt 2 9 NO Of P AGE S UNLIMITED toSEcuH Ye a ra..,2w,
// PHICk N C. o.. m......
)
_... __ _-._. - -_. = _,.... _ _ _.. _. _..__., _ _.... _ _.... -
4' r
1 r
e.
2 1,
d
+
a 4.
s 1-i -
4
'l i
l l
L.
i i
4
.i.
t k
t r
4I e
t-1 i
I d
7 li i la 1-.
j r
., s
- Tl L
( If
'-P f
)<-
4 1
1' y
i 4
1~
L L -
L (
i I
e e
P E _
i f
e c-.
t t.
i i.
t e-
- l t
l l
l.'
i-I
. 1 f
-l t
l
[ --
t i'
T 4
t I
, j
=
l
. 1 1,.
1.
i t.
I
+
i '
s I
f l '
5 L-