ML20033D918
| ML20033D918 | |
| Person / Time | |
|---|---|
| Issue date: | 02/19/1987 |
| From: | Office of Nuclear Reactor Regulation |
| To: | |
| Shared Package | |
| ML20033D829 | List: |
| References | |
| FOIA-92-522 NUDOCS 8703040465 | |
| Download: ML20033D918 (13) | |
Text
[o UNITED STATES g
NUCLEAR REGULATORY COMMISSION y y
.c,f g
5 E
WASHINGT ON. D. C. 70555 ot.,...../
SAFETY EVALUATION BY THE OFFICE OF NUCLEAR REACTOR REGULATION RELATING TO GENERAL ELECTRIC COMPANY LICENSING TOPICAL REPORT NEDE-30996P YOLUME 1
" SAFER MODEL FOR EVALUATION OF LOSS-0F-COOLANT ACCIDENTS FOR JET PUMP AND NON-JET PUMP PLANTS"
1.0 INTRODUCTION
The subject of this safety evaluation report is a modified version of the approved SAFER computer model (Ref. I Volume 2) which calculates the long-term reactor vessel inventory and peak cladding temperatures (PCT) for loss-of-coolant accidents (LOCA), and the COREC00L code which is used for fuel rod heatup calculations during spray cooling.
The topical report (NEDE-30996P, Volume 1) was submitted by the General Electric Company by letter dated June 25, 1986. This revised SAFER includes extensive model refinements to extend its application to the non-jet pump, external loop, BWR/2 plants. An earlier version of SAFER (Ref.1, Volume 2) was approved for jet-pump plants (Ref. 2) in August,1983. An application methodology for SAFER for jet pump plants was approved (Ref. 3) in June, 1984 This report also provides the staff evaluation of the COREC00L code. The primary function of COREC00L is to calculate the PCTs during LOCA transients under spray cooling conditions.
Although the SAFER code is also capable of calculating PCTs under spray cooling conditions, the simplified heat transfer model of the SAFER code is less rigorous than that used in COREC00L.
Therefore, for certain applications which require more realistic PCT
~
predictions, the CORtC00L code is used.
Section 2 of this report is a sumary of the topical report. Modifications made to the already approved version of SAFER (Ref 1) are discussed and a description of ine COREC00L code is provided.
Section 3 is a sumary of the
~.
$7
2 staff evaluation of COREC00L and the revisions to SAFER. Section 4 provides the staff conclusions.
2.0
SUMMARY
OF TOPICAL REPORT The topical report consists of a description of the SAFER and COREC00L codes, and the data bases used for code verification. Additional information, which was provided in response to staff questions, is documented in Reference 4 2.1 REVISIONS TO CURRENTLY APPPOVED SAFER CODE 2.2.1 Hydraulic Model The revised SAFER code changes the hydraulic model nodalization, counter current flow limiting (CCFL) correlations, and the overall momentum equations.
Changes to nodalization and momentum equations are made to specifically account I
for the effect of the external loops for the BWR/2 plants. The CCFL correlations are updated using additional data.
The revised SAFER adds two regions to the reactor system model, i.e., an intact external recirculation loop and a broken external recirculation loop. These two regions are important for non-jet pump reactors for modeling LOCA response.
l The revised SAFER refines the core region from 7 subregions to 12 subrecions y
for a more detailed calculation of the void fraction profile in the core.
The total momentum equations were modified to reflect the addition of the new regions. The methodology used to calculate the effect of recirculation loops i
is similar to that used in the approved LAMB code.
Counter current flow limiting (CCFL) is a phenomenon wherein the downward flow of liquid is limited by an upward flow of vapor in a geometrically restricted CCFL affects fluid inventory within the core significantly. All of area.
the CCFL correlations are unchanged with the exception of the region between f
i
f 3
the upper plenum and central bypass, and the region between the guide tube and r
the bottom of the bypass. These new correlations were obtained from recent l
experiments conducted in Japan (Ref. 5).
Since all BWR-2 plants use an isolation condenser, the revised SAFER includes an r
external flow model to calculate the transient performance of an isolation condenser. An isolation condenser consists of a heat exchanger and inlet line l
fmm the steam dome and return line connected to the recirculation loop. This system removes steam from the steam dome, condenses it on the tube side of t
the heat exchanger, and returns the condensate back to the reactor.
The system functions by natural convection.
The isolation condenser model is simple and uses classical heat transfer models.
t 2.2.2 Heat Transfer Model i
The heat transfer model of the revised SAFER differs substantially from the
{
previously approved model. The SAFER model was revised in order to be cocr.patible with the CORECOOL model. The more significant changes are discussed.
below:
The convective heat transfer model includes changes in the film boiling region and a major change in the spray heat transfer model.
In'both code versions the film boiling regime is assumed when the cladding superheat exceeds the minimum superheat for stable film boiling as defined by the lloeje ccrrelation, or when e
the coolant quality is greater than a specified critical quality. The revised SAFER film boiling heat transfer coefficient is modeled as a function of local void fraction as illustrated in Figure 1.
The previously approved correlation is shown as a dotted line, which is a linear interpolation between the modified l
Bromley coefficient at liquid phase to the modified Dittus-Boelter coefficient j
at vapor phase. The new film boiling heat transfer model was modified on the basis of various data sources, which include GE, ASEA, CISE and others' experiments of rod bundles and simple geometry tests.
j 1
i I
4 i
The spray heat transfer model for the previously approved SAFER code uses a l
l mean convective heat transfer derived from tests on a single heated bundle.
This model is revised based upon additional data obtained from Toshiba's tests j
(Ref. 5). These experiments were conducted on a single heated bundle with core spray over a wide range of key parameters, such as power level, initial temperature, spray flow rates and pressure. Heat transfer coefficients for i
both average power rods and peak temperature rods were developed.
It should be l
noted that spray heat transfer is particularly important for non-jet pump plants since the tow Pressure Core Spray is the sole cooling mechanism during the long term transient.
l r
Another significant change in the revised SAFER is the radiation heat transfer model. The previously approved SAFER considers radiation heat transfer only froi.1 the rod of interest to a single boundary temperature. The new model, however, divides the fuel assembly into two fuel rod regions (interior rods and average power rods) and the channel wall. The effect of wetting the surface of the channel wall and the rods by falling liquid film is considered.
t When the bundle is partially uncovered, water droplets can be entrained into the vapor continuous phase by the steam updraft through the core generated by lower plenum flashing.
Interfacial heat transfer between superheated steam and the entrained droplets is an irrportant phenomenon considered in the new model, t
Because of this heat transfer, the cladding convective heat transfer is l
improved, and consequently, the peak PCT is lowered. This model was developed to be compatible with the COREC00L model.
2.3 COREC00L Model CORECOOL is an one-dimensional model used to calculate spray heat transfer in a
[
fuel channel. The overall model is comprised of a hydraulic model and a heat transfer model. COREC00L provides more a detailed and mechanistic droplet heat transfer calculation than SAFER does. As a result, the PCTs calculated by COREC00L are considered more realistic.
i 5
i Since COREC00L is a model for a single fuel channel only, the' boundary j
conditions for the fuel channel must be provided by another source, such as l
SAFER. CORECOOL uses these boundary conditions to calculate the thennal f
hydraulic conditions in the fuel channel and fuel rod temperatures.
The hydraulic model assumes dispersed annular flows in the fuel channel during spray cooling. The model then determines interfacial heat transfer between steam, droplet and film flows on the fuel rods and channel wall. The effect of i
the CCFL on the downflow of core spray into the top of the channel is also considered.
t j
One of the significant features of COREC00L is its treatment of droplet size distribution dee to sputtering at the quench front. During spray cooling, d
liquid entering the fuel channel will either ferm a falling film on the f
surfaces of the rods and channel wall or fall down through the fuel channel as f
droplets. At the quench front of the film, a portion of the liquid in the film
{
may fonn droplets because of sputtering. The density distribution of the droplet size determines the fraction of the droplets which may be entrained j
into the upward steam flow or drop into the bottom of the fuel channel 0n the basis of data, the model determines the critical size of droplets.
Droplets smaller than this critical size will be entrained upward, and those f
larger will fall downward.
Entrainment of droplets has significant effect on j
I steam cooling, and consequently, on the PCT 'of the fuel rods.
[
2.4 Code Qualification l
i 1
i The approach used to qualify the SAFER and COREC00L codes is to compare SAFER and COREC00L predictions with experimental data and with TRAC benchmark j
calculations. The data bases used to qualify SAFER include the Two Loop Test f
Apparatus (TLTA), the Rig of Safety Assessment (ROSA)-11, FIX-II and the Full
[
Integral Simulation Test (FIST-APWR). The TLTA and ROSA-111 data are the j
same data used to qualify the previously approved SAFER code for jet pump plants (Ref. 1).
FIX-II and FIST /ABWR are new data sources to qualify the I
l t
6 r
I SAFER for non-jet pump applications.
For COREC00L qualification, the data base j
from the Core Spray Heat Transfer (CSHT) facilities at General Electric, AB f
~
Atomenerge in Sweden, and at Hitachi and Toshiba in Japan is used. A summary of l
these tests is presented in the topical report.
i Calculations using the revised SAFER are also compared with TRAC (GE proprietary version) predictions for the same events.
TRAC is used for best-estimate LOCA analyses.
3.0 EVALUATION The SAFER and COREC00L codes and their supporting data bases have been i
satisfactorily documented in the topical report.
Additional infonnation requested by the staff during the course of reviews has also been documented l
in reference 4.
Review of these documents leads the staff to the following regulatory position.
3.1 SAFER CODE
[
3.1.1 Evaluation of Model Changes The SAFER code (Ref.1) was previously approved by the staff for jet-pump plant applications only. The revised version of the SAFER code includes several
[
nodeling changes, which were identified in Section 2.1 of this report. The rodeling changes to nodalization, momentum ecuations and the isolation condenser provide the code the additional capability to evaluate the LOCA events for l
non-jet pump plants. These plants rely solely on external pump loops for coolant circulation and are equipped with an isolation condenser. The formulations and simulation of these modeling changes are straight-forward and acceptable. Refinement of nodalization, particularly in the core region, provides a more detailed calculation of core inventory and heat transfer.
Thus, prediction of the PCTs should be improved, and this modeling change is acceptable.
P 7
The new CCFL correlations, which involve only the region between the upper plenum and central bypass, and the region between the guide tube and the bottom of the bypass, were developed on the basis of data obtained from Toshiba's Eighteen Degree Steam Test Apparatus (ESTA) Guide Tube CCFL (Ref. 5) test These tests were performed by simulating the actual transients, facilities.
i.e., upflowing steam against upper plenum water or inflowing water against a Results of these tests reconfirm the Wallis type correlations steaming core.
However, these new CCFL correlations affect core cooling i
for the CCFL effect.
and, consequently, the PCTs, only for plants with ECCS injection into the There is no significant effect on the bypass region such as BWR 5/6 plants.
F.eview of these tests indicates that the test facilities The new data have therefore been BWR/2 plants.
adequately simulate the actual situations.
The new CCFL used to update the correlations of the previously approved model.
correlations are acceptable.
The new film boiling heat transfer model differs from the previously approved model by adding a transition r'egime between the liquid continuous regime and This transition regime uses a linear interpolation vapor continuous regime.
between the modified Bromley correlation (liquid continuous regime) and the Both Bromley modified Dittus-Boelter correlation (vapor continuous regime).
and Dittus-Beelter correlations have been previcusly approved by the staff.
The Dittus-Boelter correlation, however, has now been modified to include the heat transfer enhancement resulting from tMe presence of droplets. The staff has reviewed the modified film boiling heat transfer model against existing data and finds that it reasonably represents the data and is therefore acceptable.
3.1.2 Comparison with System Test Data Predictions using the new version of the SAFER code were compared with the same Review TLTA and ROSA 111 data used to qualify the previously approved SAFER.
of the comparison ir.dicates that the revised SAFER predicts the rieasured 1
primary system pressure response and the core inventory more closely than t g
8 previous SAFER prediction.
The predicted PCTs are slightly lower than that predicted previously, but are still above the measured PCTs.
For non-jet pump plants, the SAFER code was compared with FIX-II and FIST-ABWR FIX-II contains an electrically heated, single bundle with 36 data.
The test facility simulates a BWR reactor with external full-length rods.
recirculation loops. Comparisons were made against tests which simulated a 200 % guillotine break and an intermediate split break. Vessel pressures, break flow rate, and cladding temperatures were compared. SAFER captures the trend of vessel pressures very well for these tests.
For the large break test, For the measured PCT was 1320 F while SAFER predicted a slightly higher PCT.
the intermediate break, SAFER predicted earlier core uncovery, and consequently a higher PCT by a considerable margin.
FIST-ABWR is also an electrically heated, full-size bundle designed for code qualification and for evaluation of BWR system response to LOCA transients.
The f acilities include an exte nal recirculation loop which is applicable for Data from tests with large breaks and small breaks were used BWR-2 evaluations.
for SAFER code qualification.
SAFER predicts the experimental data of vessel pressures and core inventory closely. However, SAFER notably overpredicted the Assessment of the code indicates that SAFER PCTs for the large break test.
takes no credit for the effect of direct liquid droplets on rod cooling, which Because provides an important cooling effect on the rods and reduces the PCT.
of this limitation, SAFER conservatively overpredicts the PCT for events where core spray cooling is the dominant heat transfer mechanism.
3.1.3 Comparison with TP.AC SAFER predictions were also compared with 1RAC benchmark calculations for 2
non-jet pump plant breaks. A 1001 DBA and a small (0.1 f t ) recirculation line The codes showed good agreement for vessel pressure break were compared.
For cladding temperatures, SAFER predicts a higher PCT than TRAC prediction.
However, SAFER calculates slightly lower PCT than TRAC for does for DBA case.
^
9 the small line break. The slightly less conservative result for small line breaks is not considered significant because the PCT for these breaks is
[
l several hundred degrees less than that for the large break DBA.
i 3.2 COREC00L CODE The staff evaluated the COREC00L code by using data from the Core Spray Heat Transfer (CSHT) test facilities at GE, AB Atomenerge in Sweden, and at Hitachi and Toshiba in Japan. The experiments resulted in PCTs ranging from 1500*F to 2100 *F.
As discussed below, in each case the COREC00L predictions provide a reasonable estimate of rod temperatures. Therefore, the use of COREC00L is acceptable.
~
3.2.1 Comparison with GE/CSHT Tests GE/CSHT tests were conducted with a full-scale (8x8), electrically heated rod bundle. COREC00L predictions were compared with 22 experiments with measured PCTs ranging from 1530 to 1985 *F.
COREC00L predicted well the cladding temperatures distribution within the bundle as a function of time.
In general, COREC00L overpredicts PCTs for most of the tests with bottom venting, which is but more representative for a recirculation line break for a BWR-2 plant, The differences, either underpredicts some of the tests with the top venting.
overprediction or underprediction, however,' are within 100 *F.
3.2.2 Comparison with AB Atomenergi CSHT Tests The AB Atomenergi CSHT is also a 68 full-scale electrically heated rod bundle. The 8x8 assembly was divided into five rod groups and one channel wall. COREC00L was compared with data obtained from 11 CSHT tests.
i Comparison between the calculated and measured temperatures indicates that COREC00L accurately predicts the cladding temperatures history for each rod The PCTs from these 11 tests ranges from 1711 to 2075 *F.
The group.
b
-w
=.. -
.i
~
?
i 10 i
difference between the predicted (either overprediction or underprediction) and I
the calculated values is less than 100 *F in all cases.
l 3.2.3 Comparison with Toshiba and Hitachi CSHT Tests The Toshiba and Hitachi CSHT tests were also performed in a 8x8 full-scale.
l electrically heated rod bundle.
In the COREC00L simulation of these tests, the 8x8 assembly was divided into seven rod groups, and one channel wall.
Comparison of the calculated and measured temperatures indicates that COREC00L l
closely predicts the heatup rate and the temperature distribution within the rod groups.
4.0 CONCLUSION
i Based on the evaluation presented in Section 3.0 of this report, the staff concludes the folicwing:
The SAFER code described in the licensing topical report NEDE-30996P, 1.
Volume 1, is acceptable for use for realistic or best estimate evaluations of loss-of-coolant accidents for the jet pump and non-jet pump BWR plants.
2.
COREC00L, when used in conjunction with SAFER (NEDE-30996P) is acceptable for use for realistic or be'st estimate evaluation of loss-of-coolant accidents for non-jet pump BWR plants and jet pump
{
BWR plants, r
SAFER and COREC00L may be used as a licensing evaluation model only in 3.
conjunction with an application methodology which is specifically approved by the staff for use with these codes. An application methodology for An
[
jet pump plant applications has already been approved by the staff.
application methodology for non-jet pump plants (NEDE-30996P, Volume 2) has been proposed by General Electric Company and is currently being f
reviewed by the staff. Results of the staff's review will be reported in l
t a separate safety evaluation report.
t
11 i
4.
SAFER and COREC00L may be used only in conjunction with the approved short-tem thent.al hydraulic codes LAMB and SCAT, and with the GESTR l
fuel model.
?
?
i P
?
1 6
l b
e t
' I
l 12
]
I References i
i 1.
"The GESTR-LOCA and SAFER Models for the Evaluation of the Loss-of-Coolant Accident" Vol. II, NEDE-23785-1-P, December 1981.
2.
Letter from C. O. Thomas to J. F. Quirk, " Review of NEDE-23785-1(P),
l "GESTR-LOCA and SAFER Models for Evaluation of the Loss-of-Coolant Accident, Volume I and II, August 29, 1983.
3.
Letter from C. O. Thomas to J. F. Quirk, " Acceptance for Referencing of Licensing Topical Report NEDE-23785 Fevision 1. Volume III (P), The GESTR-LOCA and SAFER Models for the Evaluation of the Loss-of-Coolant Accident," June 1, 1984.
4 Letter from H. C. Pfeffer1en to M. W. Hodges, " Responses to Request for j
t Additional Information on NEDE-30996-P SAFER Model," July 30, 1986.
t 5.
H. Nagasaka, "New Japanese Correlations on Core Cooling and CCFL Characteristics during BWR LOCA " Paper presented at Thirteenth Water Reactor Safety Information Meeting, Gaithersburg, Maryland, October 1985.
l t
i k
i r
i
4 I
Il i
t r
l-
?
Mop:FIED BROMLEY g
o N
N UNEAR INTERPOLATION g
w N
N N
e l
t-o g
y CURRENT MODir1ED DITTUS-BOELTER f
CORRELATION N
j
!N z
N E
l l
t T
I E
l l
i I
l I
I l
1 i
i eT
- T + 0.1 1.0 e
VOID FRACTION i
t
\\
Figure 1 Film Boiling Heat Transfer Coefficient Model f
7 d = Transition (Liquid to vapor) void fraction y
1