ML18101B128
| ML18101B128 | |
| Person / Time | |
|---|---|
| Site: | Salem  |
| Issue date: | 12/07/1995 |
| From: | Eric Simpson Public Service Enterprise Group |
| To: | NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| References | |
| LR-N95203, NUDOCS 9512130154 | |
| Download: ML18101B128 (9) | |
Text
'*
'1/f.\\\\**
Public Service Electric and Gas Company E. C. Simpson Public Service Electric and Gas CodEC O 7o;ggs236, Hancocks Bridge, NJ 08038 LR-N95203 609-339-1700 Senior Vice President - Nuclear Engineering U.S. Nuclear Regulatory Commission Document Control Desk Washington, DC 20555 Gentlemen:
10 CFR 50.46 ANNUAL REPORT -
ECCS EVALUATION MODEL CHANGES SALEM GENERATING STATION UNIT NOS. 1 AND 2 DOCKET NOS. 50-272 AND 50-311 Pursuant to 10CFR50.46, Public Service Electric and Gas Company (PSE&G) hereby submits our annual report describing the changes to the Emergency Core Cooling System (ECCS) evaluation models for Salem Unit Nos. 1 and 2.
The calculated peak clad temperature (PCT) for the Salem Generating Station small break and large break loss of coolant accident (SB LOCA and LB LOCA) analyses are unchanged from their last reported values:
Analysis SB LOCA LB LOCA PCT 1660°F 2014°F Last Reported December 8, 1994 January 18, 1995 (NLR-N94211)
(LR-N94224)
As previously reported, the large break PCT value is applicable for Unit 2 and will apply to Unit 1 upon restart from the current refueling outage.
The small break and large break PCT margin allocations are summarized in Attachment 1.
The ECCS model changes identified since the last reports are summarized in Attachment 2.
The requirements of 10CFR50.46, including the PCT limit of 2200°F, continue to be met.
If you have any questions about this submittal, please contact us.
l:Dt,. Printed on
~
Recycled Paper
~~
~~.. *~.*-* /!: ':"°* ~~
< * *) ;,_; t.: b....:;.
9512130154 951207 PDR ADOCK 05000272 P
- {\\01)' I
Document Control Desk LR-N95203 ATTACHMENTS (2)
- c Mr. T. T. Martin, Administrator - Region I
- u. s. Nuclear Regulatory Commission 475 Allendale Road King of Prussia, PA 19406 Mr. L. N. Olshan, Licensing Project Manager -
Salem
- u. s. Nuclear Regulatory Commission One White Flirit North 11555 Rockville Pike Rockville, *MD 20852 Mr. c. Marschall (X-24)
USNRC Senior Resident Inspector Mr. K. Tosch, Manager, IV Bureau of Nuclear Engineering 33 Arctic Parkway CN 415 Trenton, NJ 08625 DEC 0 7 1995
ATTACHMENT 1 Small Break LOCA Allocation Description of Issue Analysis of Record Safety Injection in the Broken Leg Improved Condensation Model Drift Flux Regime Errors LUCIFER Error Corrections Boiling Heat Transfer Correlation Error Steam Line Isolation Logic Error Axial Nodalization, RIP Model Revision, and SBLOCTA Peak Clad Temperature
(
- F) :
Allocation (°F)
~R=e~f~e=r~e~n~c~e~~~~~~
1580
+150
-150
-13
-16
-6
+30
+85 1660 NRC letter dated August 25, 1993 NLR-N93173 October 29, 1993 NLR-N93173 October 29, 1993 NLR-N93173 October 29, 1993 NLR-N94126 July 27, 1994 NLR-N94211 December 8, 1994 NLR-N94211 December 8, 1994 NLR-N94211 December 8, 1994 Large Break LOCA Allocation Description of Issue Allocation (. F) Reference Analysis of Record 1978 LR-N94224 January 18, 1995 LUCIFER Error
-6 LR-N94224 Corrections January 18, 1995 e
Vantage + Fuel
+94 LR-N94224 without IFMs January 18, 1995 LOCBART Code Cray to
-52 LR-N94224 Unix Conversion January 18, 1995 Peak Clad Temperature 2014
(. F) :
ATTACHMENT 2 ECCS EVALUATION MODEL CHANGES CODE STREAM IMPROVEMENT
Background
Revisions were made to modeling procedures used to interface the various codes that comprise the entire execution stream for performing a LB LOCA analysis.
Previously, the containment pressure calculated by the coupled WREFLOOD / coco codes was transferred as a boundary condition to the BASH code.
This has been replaced with the direct coupling of the BASH and coco codes.
Now, the same code used to calculate RCS conditions during reflood supplies the boundary conditions for the containment pressure calculation.
In conjunction with this, the portion of the WREFLOOD code which calculated the ref ill phase of the transient has been reprogrammed into a separate but identical code that is also coupled with coco.
This methodology revision was made only as a process improvement for conducting analyses and involved no changes to either the approved physical models or the basic solution techniques governing the solutions provided by the individual codes.
Estimated Effect Due to small perturbations in the boundary conditions resulting from this revised methodology for interfacing the codes, Westinghouse observed small differences in predicted results.
However, the effects were minor with no observed bias.
Since this methodology in a process improvement which is to be implemented on a forward-fit basis, there are no effects on existing licensing analyses, and any small effects on results will be implicitly accounted for in future analyses.
BASH: LOOP/CORE INTERFACE CORRECTIONS
Background
Two corrections were made to the logic for interfacing the loop and BART code models.
One correction prevents an occasional inconsistency in how the core timestep is limited by the loop timestep.
The other corrects the fluid density used in the interface calculation when the inlet flow rate is negative.
Estimated Effect Results from Westinghouse' s sensitivity studies for the corrections demonstrated negligible perturbations in the trends of the system parameters with a very minor net effect on PCT predictions relative to results from the previous version.
Since A!I'TACHMENT 2 this effect is extremely small with no apparent bias, the net change to peak cladding temperature in existing analyses is estimated to be zero degrees.
The change has been implemented on a forward fit basis and will be incorporated implicitly in future analyses.
PELLET POWER RADIAL FLUX DEPRESSION ERROR
Background
A coding error (an incorrect sign) was discovered and corrected in a subroutine that calculates radial distribution power factors in the fuel pellet for the LOCBART code.
Estimated Effect Westinghouse' s sensitivity studies found the error correction to result in less than a +/-O.l°F effect on predicted PCT.
The net effect on existing analyses is therefore zero degrees for margin tracking purposes and will be implicitly included in future recalculations.
IMPROVEMENTS TO FLOODING RATE SMOOTHING
Background
Part of the approved Westinghouse methodology for performing LB LOCA analyses with the BASH evaluation model is the requirement that the core inlet flooding rate calculated by the BASH code be
- linearized in a piecewise manner to remove oscillations prior to use in the hot channel fuel rod calculation.
This operation is termed 11 smoothing."
Guidelines are provided to the analysts describing how to linearize the curve by observing inflections in the overall flooding rate.
To facilitate consistency in performing this operation, the logic has been coded into a program named SMUUTH.
A new version of this has been implemented which incorporates improved logic for determining the inflection points gained through experience initializing the program for a broad range of plant transients.
Estimated Effect There are no changes to the approved evaluation model methodology from this revision.
The SMUUTH program merely represents a convenient way of automation and does not explicitly introduce any effects on the results.
This revision is being reported only as a change to the code stream used for standard analyses.
There are no effects on predicted results from using the new program version.
ATTACHMENT 2 ACCUMULATOR WATER TEMPERATURE
Background
The choice of accumulator water temperature can affect the calculated peak cladding temperature during a LB LOCA analysis.
The early Salem model assumed a generic value of 90°F for the accumulator water temperature based on a conservatively low value of containment air temperature at 100% power.
This supported the Appendix K requirements associated with the calculation of a low containment back-pressure.
Assuming 90°F for the accumulator water temperature with the early Salem model was conservative with respect to the overall effect on LB LOCA PCT.
Newer models have demonstrated that a higher containment air temperature coupled with higher accumulator water temperatures may result in an even more conservative calculation for PCT --
even if containment pressure is slightly higher than calculated with the 90°F assumption.
Sensitivity studies performed with the newer models have shown a small sensitivity to accumulator water temperature.
The effect on PCT was a 1.3°F change in PCT for every l°F change in accumulator water temperature between 90°F and 1'20 °F.
Estimated Effect The value for accumulator temperature assumed in the current Salem analysis is ll0°F.
PSE&G has concluded that this is the maximum temperature the accumulators could reach.
Hence, the existing calculation is bounding and this issue does not affect the Salem model.
CORE NODE ZIRC OXIDE INITIALIZATION ERROR
Background
NOTRUMP models two regions for each core node analogous to the two (mixture and vapor) regions in adjoining fluid nodes.
During a transient, NOTRUMP tracks region-specific quantities for each core node.
Erroneous logic caused incorrect initialization of the region-specific, fuel cladding zirc oxide thickness at times prior to the actual creation of the relevant region during core boiloff.
Estimated Effect Representative plant calculations led to an estimated generic PCT effect of 0°F for this effect.
ATTACHMENT 2 PRESSURE SEARCH CONVERGENCE CRITERIA IN NOTRUMP
Background
The convergence criteria used during the pressure search in NOTRUMP have been found, in some cases, to not ensure a sufficiently accurate value for fluid node pressure under conditions approaching the boundary between subcooled and saturated.
This became more pronounced below pressures normally seen during standard calculations.
The previously hardwired convergence criteria values have been made user input.
Appropriate values have been determined for these, and they will be implemented in all future analyses.
Estimated Effect The nature of this error led to an estimated generic PCT effect of 0°F for existing analyses.
FRICTION VALUE INPUT CORRECTIONS
Background
The SPADES code is used to generate input decks for the SB analysis code, NOTRUMP.
An error was found in the code which involved the values assigned to some of the friction factor input_.
The erroneous values had no impact on transient calculations and were corrected in order to maintain the consistency of the SPADES code with the relevant documentation.
Estimated Effect Representative plant calculations indicate no effect on PCT analyses.
AUTOMATIC CONTAINMENT SPRAY ACTUATION DURING SB LOCA
Background
Historically, the Salem SB LOCA calculation did not assume containment spray to be actuated.
This was because Westinghouse believed that containment pressure would not increase enough to actuate containment spray.
However, Westinghouse has recently discovered that this is not so for some of its customers.
If containment spray occurs and it is early enough in a SB LOCA transient, the spray and SI can drain the RWST before the end of the transient.
When that happens, a switch to cold leg recirculation would occur.
This reconfiguration could reduce or interrupt SI flow.
Additionally, the enthalpy of the reestablished injection flow may be increased.
The effect this would have on PCT is dependent upon how close the switchover A~TACHMENT 2 occurs to the time of PCT.
If they occur closely, the PCT could increase.
However, if switchover occurs sufficiently late (or not at all), PCT would remain unaffected.
Also, Westinghouse discovered a new single failure scenario for some customers that would cause a faster draindown of the RWST, hastening the subsequent switchover to cold leg recirculation.
This causes two competing effects on PCT:
- 1)
A faster draindown means there is more SI flow delivered in a shorter period of time following actuation.
This benefits PCT.
- 2)
A faster draindown means the switchover reduces or interrupts SI flow, earlier.
Similarly, any increase in enthalpy associated with the reestablished flow would occur earlier.
This would represent a penalty on PCT.
Since these are competing effects, site specific considerations must be made to determine if a faster RWST draindown is a benefit or penalty on PCT.
Future SB LOCA analyses will explicitly consider these issues.
Estimated Effect A Salem-specific evaluation indicates that there is no PCT effect due to SI interruption or reductiqn following switchover to cold leg recirculation.
Similarly, Westinghouse has determined a faster draindown of the RWST is an overall benefit for Salem-type plants and compensates for the change in injection flow enthalpy.
Hence, this item does not result in a PCT penalty for Salem.
LOCA AXIAL POWER SHAPE SENSITIVITY MODEL
Background
LB LOCA analyses have been traditionally performed using a symmetric, chopped cosine, core axial power distribution.
Under certain conditions, calculations have shown that there is a potential for top-skewed power distributions to result in PCTs greater than those calculated with chopped cosine axial power distributions.
In 1991, Westinghouse developed a statistical methodology to evaluate and assure that the cosine distribution remains the limiting distribution.
This methodology, Power Shape Sensitivity Model, was submitted to the NRC for review and approval.
It has since been implemented (without approval) on a forward-fit basis as part of the LB LOCA evaluation model in conducting new and reload safety evaluations for Salem.
ATTACHMENT 2 Recently, the NRC has indicated that approval would require extensive modifications to the methodology and a potentially significant PCT penalty to account for uncertainties.
Rather than pursue this further, Westinghouse proposed the use of an alternate, licensed approach for power shape evaluation.
The NRC has concurred that this methodology is sufficient as a replacement.
As a result, Westinghouse is now using Explicit Shape Analysis (ESHAPE) for PCT Effects to evaluate sensitivity of axial power shape instead of the old Power Shape Sensitivity Model.
Estimated Effect Using the ESHAPE methodology, Westinghouse has determined that plants with early PCT times remain cosine shape limited.
As a result, Westinghouse has determined that these plants are not impacted by the change in methodology.
Salem satisfies the criterion as one of these plants.
Hence, there is no change in PCT.