ML20203C184

Safety Evaluation of Topical Rept NEDC-32721P, Application Methodology for General Electric Stacked Disk ECCS Suction Strainer, Part 1.Concluded That Use of GE Hydraulics Design Method Acceptable for All Plants,With One Noted Exception

ML20203C184
Person / Time
Issue date:	02/03/1999
From:	NRC (Affiliation Not Assigned)
To:
Shared Package
ML20203C182	List: ML20203C178 ML20203C184
References
NUDOCS 9902110248
Download: ML20203C184 (3)

Text

.

[f%

g 81 j

UNITED STATES NUCLEAR REGULATORY COMMISSION

~t WASHINGTON, D.C. 20555-0001

%, p$

Attachment SAFETY EVALUATION BY THE OFFICE OF NUCLEAR REACTOR REGULATION RELATED TO TOPICAL REPORT NEDC-32721P, ENTITLED

" APPLICATION METHODOLOGY FOR THE GENERAL ELECTRIC STACKED DISK ECCS SUCTION STRAINER" GENERAL ELECTRIC COMPANY PARTl EVALUATION OF GENERAL ELECTRIC'S METHODOLOGY TO DETERMINE STRAINER HYDRAULIC PERFORMANCE

1.0 INTRODUCTION

By letters dated April 3,1997, and November 21,1997, the General Electric Company (GE)  ;

submitted topical report NEDC-32721P, " Application Methodology for the General Electric l Stacked Disk ECCS Suction Strainer," to the Nuclear Regulatory Commission (NRC, the staff) for review. The methodologies described in this report are being used by multiple boiling-water reactor (BWR) licensees (i.e.,12 plants) as part of their resolution of concerns identified in NRC Bulletin 96-03, " Potential Plugging of Emergency Core Cooling Suction Strainers by Debris in Boiling Water Reactors." By facsimile dated August 18,1997, and letter dated March 10,1998, i the staff transmitted requests for additional information (RAls) to GE. GE responded to the staff RAls in a letters dated November 21,1997, and April 29,1998.

The GE topical report encompasses methodologies for two different types of analyses. The i first methodology is used to determine the head loss across the strainer for estimated debris i loadings. The second methodology is used to determine the structuralloads on the ECCS l penetrations, piping and strainers caused by hydrodynamic forces during an accident. These two methodologies are separate and distinct. Therefore, the staff has broken its review into two parts. This safety evaluation (SE) is Part I of the staff's evaluation of topical report NEDC-32721P. Part I provides the staff's evaluation of the methodology used to determine strainer performance (e.g., head loss across the strainer with assumed debris loadings) only.

Part ll of this safety evaluation will address the methodology to determine the structural loadings due to hydrodynamic forces during a accident. The hydrodynamic load methodology is still undergoing staff review.

2.0 DISCUSSION l

GE's topical report and RAI responses were evaluated by the staff's contractor, Los Alamos National Laboratory (LANL) and LANL's findings relative to GE's methodology for determining the head loss across GE stacked disk strainers are documented in a proprietary Technical Evaluation Report (TER) entitled " Technical Review of GE LTR NEDC-32721P: Application Methodology for GE Stacked-Disk ECCS Suction Strainer," dated December 23,1998. This SE does not contain proprietary information; however, LANL's TER (attached) does. LANL's TER is accordingly being withheld from the public document room.

9902110248 990203 PDR TOPRP EMVGENE C PDR

l <

\ -

j L

2

3.0 CONCLUSION

S Based on the staff's review of all relevant information including LANL's TER and GE's submittals, the staff has concluded that the testing program used by GE for verifying the hydraulic performance of the prototype GE strainer and validating GE's head loss correlation is acceptable. The staff notes that some of the test procedures are not considered to be prototypical because they do not exactly replicate post-LOCA conditions; however, the overall l- impact of these procedures on the measured head loss is considered to be insignificant. The l staff concludes that GE has taken adequate measures to ensure that test results are conservative. The test program provided valuable data that can be used to correlate the effect of flow velocity, strainer shape, debris loading and particulate-to-fiber ratio on head loss. The i only drawback of the GE test program noted by the staff is that it did not envelope the l parametric range that covers all the plant applications. In particular, much of the test data were obtained for low fiber loadings where the gaps between the strainer disks were not completely filled. This comment is applicable to all insulation types, especially when tested in conjunction l with sludge debris. This, in itself, is not a deficiency, and in fact, is often the case for many test ;

l programs where practical considerations influence the parametric range selected for testing. j Extending the test results over a narrow parametric range outside the test range is reasonable.

However, the staff believes that for one case (a plant with a sludge-to-fiber mass ratio in excess of those tested), the application of GE test data (or the correlation) would be inappropriate without additional testing. GE is conducting an additional test to address this concern.

GE adopted an empirical means for correlating the test data. GE's choice was to correlate l head loss in terms of superficial parameters (such as circumscribed velocity) that are easy to determine in plant applications. This approach raised questions regarding the generic applicability of the GE correlation, especit.by application of the correlation beyond the test l range. However, the staff believes that GE introduced margin which is sufficient to compensate for any deficiencies noted in the correlation. For example, GE recommended use of 1.0 (the maximum value it can reach) for the compression factor (f,) irrespective of flow rate. According to the test data, at a flow rate of 10,000 GPM the actual compression factor is approximately 0.75, which is 25% lower than the design value of 1.0. The staff concluded that this margin would allow GE to apply its correlation within a narrow range beyond the range for which the test data obtained. LANL also conducted independent analyses to evaluate the applicability of GE methodology to each of the plant applications cited in GE's submittals. These analyses used an updated version of the NUREG/CR-6224 head loss correlation which was modified to

! better predict the characteristics of the GE strainer geometry. The modified NUREG/CR-6224 L correlation was then applied to each plant specified in the GE submittals. Based on the results l of these calculations, the staff concluded that the use of GE's hydraulics design method is l acceptable for all the plants, with the one exception noted above. For the one exception, the i staff identified specific concerns relative to the application of the GE correlation. The staff is

! concerned about the applicability of the GE correlation to this plant because:

(1) neither the GE correlation nor the NUREG/CR-6224 correlations were tested to sludge-to-fiber ratios approaching the value for this plant (i.e., thin-bed effects), and

'(2) the controlling insulation in this case may be a different type of fibrous insulation for which no head loss data have previously been obtained. The staff concludes, therefore, that i

l l

ll

. - . . - - _ _ . . - -- - - . - - . - - . - ~ . - -. -

g 9

GE's approach of validating its hydraulics methodology using head loss data from an additional test is the most prudent approach.

The staff also reached the following conclusions:

-(1) GE's use of bump-up factors, consistent with the guidance of NEDO-32686A " Utility Resolution Guidance for ECCS Suction Strainer Blockage" (URG), to account for miscellaneous debris is acceptable. l (2) GE's approach to estimate head loss contribution from reflective metallic insulation (RMI) i debris appears reasonatste; however, the staff notes that GE should ensure that NRC 1 comments provided Appendix K to the staff's safety evaluation report on the URG are properly reflected in any GE plant-specific analyses. Specifically, GE should not neglect the potential contribution of RMI debris to strainer head loss without supporting analyses i establishing .that RMI contribution is negligible. The staff was unable to verify the contribution to strainer head loss from RMI debris because GE did not provide information relative to the assumed RMI loadings for any of the plants using GE strainers.

Principal contributors: Rob Elliott, NRC D.V. Rao, LANL l l

I i

I t

F i I

. -_. _