ML20059D191
| ML20059D191 | |
| Person / Time | |
|---|---|
| Site: | Comanche Peak  |
| Issue date: | 12/30/1993 |
| From: | Office of Nuclear Reactor Regulation |
| To: | |
| Shared Package | |
| ML20059D048 | List: |
| References | |
| NUDOCS 9401070058 | |
| Download: ML20059D191 (8) | |
Text
- .
gn Rtc e 'o UNITED STATES 8 ,. g o NUCLEAR REGULATORY COMMISSION ,
j WASHINGTON, D. C. 20S55
%...../
SAFETY EVALUATION BY THE OFFICE OF NUCLEAR REACTOR REGULATION RELATED TO TOPICAL REPORT RXE-91-005
" METHODOLOGY FOR REACTOR CORE RESPONSE TO STEAMLINE BREAK EVENTS" TEXAS UTILITIES ELECTRIC COMPANY COMANCHE PEAK STEAM ELECTRIC STATION. UNITS 1 AND 2 DOCKET NOS. 50-445 AND 50-441
1.0 INTRODUCTION
The topical report entitled " Methodology for Reactor Core Response to Steamline Break Events," RXE-91-005, dated May 1991 (Ref. 1), describes the steamline break (SLB) analysis methodology developed by Texas Utilities Electric Company (TV Electric) for the Comanche Peak Steam Electric Station Units 1 and 2 (CPSES) using RETRAN-02 H0D005.0 and VIPRE-01, each of which are NRC reviewed and approved codes (Refs. 2 and 3). Additional information was '
provided in Ref. 4.
The stated objective of the topical report is for TU Electric to demonstrate ,
that its SLB analysis methodology is acceptable for licensing applications by
- providing descriptions of methodology, justification of analysis assumptions ,
and techniques, and methods of determining limiting cases for Comanche Peak Steam Electric Station.
TU Electric demonstrated its use of RETRAN for analyses other than loss-of-coolant accidents and Chapter 15 type transients, except certain transients including the SLB event, in a separate topical report (Ref. 5), which was reviewed and approved by the NRC. Similarly, TV Electric's generic departure from nucleate boiling ratio (DNBR) methodology using VIPRE was reviewed and approved during the review of RXE-90-001 report.
This evaluation focused on the adequacy of TV Electric's SLB methodology for reload analysis, using the RETRAN code, by determining if the models were adequately qualified through extensive parametric sensitivity studies.
Transient analyses were reviewed st'ictly from the perspective of determining '
if TV Electric's RETRAN model was adequate for such applications.
Topical Report RXE-91-005 also contains TU Electric's DNBR methodology specifically for SLB applications using VIPRE. Because the NRC staff has already approved TU Electric's generic VIPRE methodology (Ref. 6), only those changes necessary for the SLB DNBR analysis were reviewed. The evaluation of the adequacy of the actual computation, i.e., computational conversion coefficients, mesh inputs, input data, etc., for the DNBR in the limiting transients also was beyond the scope of this review.
9401070058 931230 PDR ADOCK 05000445 P PDR
a.
i 2.0
SUMMARY
An overview of TU Electric's techniques to predict core response during SLB events at power and hot-shutdown is provided in the topical report. Extensive sensitivity studies were performed to determine and qualify models and assumptions for each SLB event. ,
In addition to the RETRAN models, TV Electric provided descriptions of VIPRE-01 models for prediction of core responses. Also provided are differences between the already approved generic VIPRE model and the SLB VIPRE model.
Analysis results were presented and discussed.
3.0 EVALUATION TV Electric's methodology as documented in topical report RXE-91-005 for prediction of the CPSES core responses during SLB events was reviewed.
3.1 Method of Analysis Using the guidelines contained in Regulatory Guide 1.70 and NUREG-0800, TV Electric developed an analysis methodolsgy for SLB events to identify the conservatively limiting cases. TU Electric identified key parameters and assumptions, such as reactivity feedback and core power distribution, which are important and have been found to affect the transient consequences of the SLB event. These key conditions and assumptions as well as break sizes and potential single failures were varied in sensitivity studies to determine the most limiting values.
The transient is analyzed using three computer codes. RETRAN-02 MOD 005.0 is used to generate the system transient response. On the basis of system ;
transient response predicted by RETRAN-02, VIPRE-01 is used to perform the DNBR analyses. The SIMULATE-3 three-dimensional full-core neutronics code is used for reactivity checks and calculation of core power distribution.
3.2 RETRAN CPSES Plant Models CPSES consists of two 3411-MWt Westinghouse four-loop pressurized-water reactor plants. The steam generator (SG) design inclupes an integral steam flow restrictor with an equivalent flow area of 1.4 ft . This constriction in flow area will regulate flow in main SLB events. The four main steam lines are interconnected through a header located downstream of the main steam line isolation valves (MSIVs). Each steam line is provided with one power-operated atmospheric relief valve and five spring-loaded safety valves, which are located upstream of the MSIVs.
I TU Electric developed two plant nodalizations: (1) a single-loop model to be used for at-power cases where symmetric blow down of four loops is expected before MSIV closure because of the constant steam flow to the turbine maintained by the turbine load controller, and (2) a multi-loop model for ,
simulation of asymmetric blow down of SGs for transients at hot shutdown conditions.
The plant nodalizations include the primary and secondary representations that are described in detail in CXE-91-001 (Ref. 5). The secondary side nodalization includes the steam piping to the turbine, main and auxiliary '
feedwater, and various isolation and relief valves.
For the SLB analysis, the SG secondary side model is represented by a single saturated volume to allow for a constant steam exit quality of 100 percent throughout the transient.
3.2.1 General Transient Initial Conditions and Assumptions To maximize the aggregate SG blowdown, and thereby aggregate cooldown, for the ,
hot zero-power (HZP) cases the initial SG mass is conservatively estimated to !
be 173,500 lbm, which corresponds to 110 percent of the design HZP SG mass.
For the full-power cases, the nominal SG mass of 104,450 lbm was used. TU Electric stated that the trarsient scenarios are insensitive to the initial SG inventory because during the time period of interest, the SG mass is not j depleted enough for tube uncovery. This is reasonable.
The primary systems initial conditions include minimum pressure, maximum coolant average temperature, and minimum reactor coolant system (RCS) flow rate. ~t TV Electric stated that the input assumptions used in the analyses are typical of those used in final safety analysis report (FSAR) analyses and reflect the current CPSES design bases. !
Although a low value of gap conductance was used for the demonstration analyses, TU Electric stated that in future cooldown event analysis, a ,
conservatively high value of the conductance will be used to lower the Doppler ,
coefficient reactivity defect, resulting in a higher peak power.
3.2.2 Core Power Distribution !
Core average, hot assembly, and hot pin axial power distribution for use in the DNBR analysis are provided through the three-dimensional neutronics code .
prediction of the reactor core characteristics for the MSLB on a t cycle-specific basis using a full-core model. Assumptions for such calculations include end-of-life, HZP conditions with all control rods inserted except the most reactive control rod being withdrawn. It is further assumed that the MSLB occurs in the reactor loop adjacent to the core quadrant with the stuck rod.
4 i
^y -
3.2.3 Break Spectrum .
A spectrum of. break sizes ranging from 0.0252ft to 1.4 ft 3was considered for each power level of 102-percent, 70-percent and 30-percent -for the at-power -
cases. Similar break sizes were analyzed for the hot-shutdown cases. ,
3.2.4 Single Failure- :i TU Electric stated that for the at-power cases, no single active failure in any system or component required for mitigation will adversely affect the- i consequences. However,.for the hot-shutdown cases, the failure of one safety injection train is assumed for the single failure. l 3.2.5 Reactivity Feedback Modeling 1 A point kinetics model is used by the RETRAN computer code to simulate the l core power response to reactivity changes resulting from variations in control rod position, fuel temperature, coolant density, and boron concentration.. TU :
Electric generated the reactivity coefficients and the core power distribution ;
used in such physics calculations within RETRAN using the NRC-approved three-dimensional physics methodology-as described in Section 3.2.2 of this ;
report.
To maximize the reactivity insertion and, therefore, the power increase, the ,
most negative reactivity coefficients associated with end-of-life (EOL) i conditions are used (i.e., maximum moderator and ~ minimum Doppler reactivity >
feedback). .j The Doppler coefficient reactivity defect is computed as a function of the .
core weighted average fuel temperature, and the moderator density reactivity y defect is calculated as a function of the core weighted average moderator !
density. Reactivity weighting factors between the affected'and unaffected. l regions in the core are used to conservatively estimate the reactivity .[
feedback. -1 3.2.6 Inter-loop Mixing Model q A nonconducting heat exchanger, modeled via a RETRAN control system, was; ,!
developed to preset the amount of interloop mixing assumed for the upper and- j lower plena in the reactor vessel.
3.2.7 Boron Transport The RETRAN general transport model was used to simulate. boron transport ,
through the RCS. Conservative allowances related to the timing of ~the boron actually reaching the core were assumed; for example, timing of safety .
1 injection (SI) initiation, purge time, and the transport time of boron from. ..
the injection point in the cold legs to the core. ]
j i
i
Before the boron entered into the RCS, the core power had reached a nearly equilibrium plateau at 20.9-percent rated thermal power (RTP) and the rate of its change was less than +0.005-percent RTP per second. As long as auxiliary feedwater (AFW) cc,ntinue to be fed into the SG providing cooling, the gradual increase in core power would not have ceased. TV Electric stated that the power was turned over and minimum DNBR occurred immediately after boron reached the core. However, since this is not a function of the boron transport modeling, TU Electric chose not to qualify the boron transport model.
Because the timing is influenced by the conservative assumptions and the magnitude of minimum DNBR is not influenced by the use of options in the transport moCel, TV Electric's decision not to qualify the boron transport model for use in the SLB analysis is acceptable.
3.2.8 DNBR Analysis A steady-state core thermal hydraulic analysis was performed using the state points provided by the RETRAN-02 analysis as boundary conditions. The state-point boundary conditions correspond to the most limiting set of conditions with respect to DNB at a given point in the transients, such as the conditions at the time of peak heat flux.
The core nodalization used for this analysis represents a full core. On the basis of a sensitivity study, TU Electric determined that the use of the full-core model was more appropriate, realistic, and more conservative than the use of the 1/8 core model. ,
The TUE-1 correlation was used for pressures above 1500 psia and the W-3 correlation for the pressure range between 1500 and 500 psia. The TUE-1 correlation for use with VIPRE-01 was submitted for staff review in a separate topical report (RXE-88-102P) in January 1989 and was approved on June 11, 1993.
3.3 Transient Analysis 3.3.1 Steamline Break at Power The purpose of the evaluation of the SLB event occurring while the reactor is at power is to demonstrate that the plant protection systems perform their intended function to protect the core against overpower conditions before a reactor trip occurs. Potential for a return to power as a result of the continued cool down after scram is bounded by the SLB initiated at hot .
shutdown conditions because there is a greater amount of stored energy in the '
RCS following a reactor trip from power compared to the initial stored energy at hot shutdown cond'tions.
1 Sensitivity analyses performed on the initial power levels of 102 percent, 70 '
percent,and30percentwerpexaminedforaspectrumofbreaksizesranging !
from approximately 0.025 ft per loop to 1.4 ft2 per loop, using input l assumptions typical of FSAR analyses. For large break sizes, the resulting ;
l
k rapid depressurization generates an Engineered Safety Features Actuation Signal (ESFAS) low steamline pressure signal quickly actuating safety injection and a reactor trip. For the intermediate break sizes, a reactor trip is generated by the overpower N-16 function prior to reaching a low steam line pressure signal setpoint due to a larger reactor power increase. The reactor power eventually reaches a higher equilibrium level equal to the steam release rate. The core thermal power remains below the limiting level. In all cases the DNBR remained above the design limit of 1.45.
3.3.2 Steamline Break at Hot Shutdown For the SLB analysis at hot shutdown, a SLB was postulated to occur upstream of an MSIV or it was assumed that a single MSIV failed to close, resulting in a break flow assumed to continue unchecked from one of the SGs. To simulate the asymmetric plant response, the following modifications were introduced into the plant nodalization for the hot shutdown cases:
(1) The reactor vessel model was modified to include a 15 volume model incorporating a split core model to account for the effects 07 the coolant asymmetries during the SLB event. All volumes except the core bypass region were split into two parts, one associated with the ,'
faulted loop representing 25 percent of the total volume and the other with the intact loops representing 75 percent of the total volume, (2) The interloop mixing was assumed to occur in the downcomer, lower plenum, core region, and the upper plenum. TV Electric controlled the degree of mixing in the plena by the use of a nonconducting heat exchanger model, (3) The moderator coefficient was weighted 50 percent /50 percent between the affected and unaffected regions to ensure conservative reactivity modeling.
In addition, the effective steam generator tube bundle height was reduced to delay any possible tube bundle uncovery, and thus, maximize cooldown during the event; for simplicity, the steamlines were not modeled.
Table 14 of Reference 4 summarized the sensitivity studies performed. In all cases presented, minimum DNBR remained higher than the design limit of 1.45.
The table also demonstrated that a representative case using assumptions typical of those used in the FSAR analyses was conservative.
l l
1 l
l
8
4.0 CONCLUSION
S The staff finds that the subject topical report, together with TV Electric's supplement, contains sufficient information to ensure that the use of TV Electric's RETRAN/VIPRE SLB methodology developed for CPSES analysis results in predictions with adequate assurances of conservatism and is, therefore, acceptable subject to the following conditions:
(1) The RETRAN methodology approved in this safety evaluation is applicable only to CPSES analyses.
(2) TU Electric elected not to qualify the boron transport model as required by the RETRAN SER because it has no significant impact on the results of the analysis. Although the power turns over when boron reaches the core, this happens late in the transient when the power has leveled off .
and there is near zero total reactivity. Therefore, the impact of changes in the boron transport model would only change the time at which the peak power occurs with minimum effect on the value of the peak power or the minimum DNBR. In the future, should it become necessary to rely on modeling of this effect, TU Electric must qualify that its modeling is conservative.
(3) For cooldown event analysis, a high value of gap conductance should 1 be used to ensure conservatism.
(4) Use of the critical heat flux correlations with VIPRE should be limited to the approved correlation and within the approved range of aplicability.
(5) Limitations and restrictions cited in the respective SEs for RETRAN and j VIPRE remain valid.
5.0 REFERENCES
- 1. " Methodology for Reactor Core Response to Steamline Break Events," RXE-91-005, May 1991.
- 2. "RETRAN-02-A Program for Transient Thermal-Hydraulic Analysis of Complex ,
Fluid Flow Systems," EPRI NP-1850-CCM, F'evision 2, Electric Power 3 Research Institute, November 1984. I
- 3. " Acceptance for Referencing of Licensing Topical Report VIPRE-01: A l Thermal-Hydraulic Code for Reactor Cores, EPRI NP-2511-CCM Vols.1-4." l May 1, 1986. j 4 Letter from D. R. Woodlan (TU Electric) to NRC, "Comanene Peak Steam Electric Station Reload Analysis Topical Review Transient Analysis Methods for CPSES Licensing Applications," September 3, 1993.
I I
=,--
g 8:
1
- 5. " Transient Analysis Methods for Comanche Peak Steam Electric Station Licensing Applications," RXE-91-001, February 1991.
- 6 "VIPRE-01 Core Thermal-Hydraulic Analysis Methods for Comanche Peak
-Steam Electric Station Licensing Applications," RXE-89-002, June 1989. -
Principal Contributor: L. Lois NRR/SRXB Date: lDEC 3 ( qy.
I a