ML20216E569
| ML20216E569 | |
| Person / Time | |
|---|---|
| Issue date: | 04/09/1998 |
| From: | Skay D NRC (Affiliation Not Assigned) |
| To: | NRC (Affiliation Not Assigned) |
| References | |
| NUDOCS 9804160244 | |
| Download: ML20216E569 (8) | |
Text
..
e n
pn atc
\\
UNITED STATES 3"
NUCLEAR REGULATORY COMMISSION WASHINGTON, D.C. 30s86 e001 o
'.....,d April 9, 1998
SUMMARY
OF MEETING WITH THE GENERAL ELECTRIC COMPANY (GE)
SUBJECT:
DISCUSS ISSUES RELATED TO GE'S METHODOLOGY FOR COMPUTING HYDRODYNAMIC LCADS FOR THE GE EMERGENCY CORE COOLING SYSTEM SUCTION 0 TRAINERS On January 27,1998, the NRC staff met with tne General Electric Company. The purpose of the meeting was to discuss GE's proposed approach to compute the hydrodynamic loads for the GE stacked disc suction strainer. Upon resolution of all staff issues, it is the aim of GE to obtain NRC approval of this approach for all plant specific applications. Enclosed is a list of meeting attendees.
In November 1997, GE submitted a licensing topical report (LTR), NEDC-32721P, " Application Methodology for the General Electric Stackoo Disk ECCS Suction Strainer," for NRC review ine NRC staff provided seven proposed discussion topics to GE prior to the January 27,1998, meeting, all of which were addressed by GE during the meeting. The following discussion contains a portion of the staffs evaluation of the LTR.
BACKGROUND The LTR describes the GE methodology for the calculation of drag loads imposed upon the stacked disk strainer when subjected to a hydrodynamic event within the suppression pool.
These hydrodynamic events are the result of either a postulated loss-of-coolant accident (LOCA) or the actuation of one or more safety relief valves (SRV). Either event can cause a significant movement of the water mass within the suppression pool which, in tum, imposes drag forces on submerged structures within the suppression pool.
The need to perform these calculations was first recognized during the original pool dynamic program undertaken by GE during the design of the Mark lli containment. During this prcgram, calculations of hydrodynamic forces were found to be necessary. Documentation of the hydrodynamic loads as applicable to Mark ! cad il containment designs (the Mark I containments were the operating plants at the time while the Mark ll containment plants were in the construction phase) can be found in various topical reports generated by GE and others as part of the Containm6nt improvement Program (CIP) and is described in NUREG-0661, " Safety Evsk ation Report for the Mark l Containment Program," July 1980. The Mark I original plant licent;ing basis did no: include hWiver,amic loads and, therefore, the plant design basis was supp!omented by the GF. Report NEDO-21888,
- Mark l Program Load Definition Report (LDR),"
\\
\\
November 1981 and doceented in each plant's Plant Unique Analysis Report (PUAR). The 5
LDR and NUREG-0661 fomi part of the basis for establishing the hydrodynamic load methodology and provide a path via attached references to cil other documents which describe the analytical procedures that had been previously accepted b the NRC. A similar process was
/0 i
used for the Mark 11 containment riesign plants. For these planta, GE documented the y
hydrodynamic criteria in a topical report and plant specific respor,ses to this report were included in the plants' Final Safety Analysis Report (FSAR). NRC review of the hydrodynamic I ad basi g
~
40%-3 UCL 9804150244 980409 PDR TOPRP EMVGENE Yh'7[
NRC FU CEiTER fl0P1
s.
2-for the specific plants was accomplished during the review of the plants' FSAR prior to issuance I
of the plants' operating license.
DISCUSSION GE stated at the January 27,1998, meeting that, for the most part, the calculation of loads on the disc strainer will be identical to the previously approved Mark I and 11 programs disused above. GE also stated that the calculation of the acceleration drag loads would contain some differences from the previous calculations used for the Mark I and ll programs, but that the methodology will remain the,same. The difference will be in the values selected for the hydrodynamic mass coemcient and acceleration drag volume. The original calculations used a bounding value of 2.0 for the hydrodynamic mass coefficient (Cm). This value is bounding since it is for an infinitely long cylinder with solid surfaces. For the new disk stra!ner, GE proposed to conduct a test and measure parameters which would allow the hydrodynamic mass coefficient to i
be calculated for the tested strainer. The test assumed that the straineris both finite in length as well as perforated on all exposed surfaces.
)
The staff was interested in the value of the mass coefficient and in the manner in which GE was determining the forces acting on submerged structures like the strainer. Previously, the methodology identified specific events such as LOCA and SRV air bubble and the submerged structure loads were calculated for each event. The staff believed that a similar approach should be selected for the new disc strainers.
GE responded with a description of its method of calculating the forces. GE stated that the calculations necessary for the determination of the drag forces acting on the stacked disk strainer will use the same methodology as was used in the original program. Therefore, the generation of the fluid velocities as a function of radius within the pool will be identical to the previously accepted values. These velocity profiles around the new stacked disk strainer will then be used i
in the calculation of the drag forces. GE concluded and the staff agreed that this process is the same approach used for the previously accepted calculations. Other forces caused by fall-back and fluid structure interaction would also be calculated using the same methods. In other words, the approach will be identical to the approach used in the original design.
In response to an NRC question regarding the change to the input value for the hydrodynamic mass coefficient for the new stacked disk strainer, GE stated that the new strainer is very different from the original strainers. Because of the increased length and diameter of the new strainer, the methodology previously established for calculating the drag load would yield unreasonably high drag forces acting on the new stacked disk strainer. Therefore, GE reviewed the previous approach to determine where excess margin existed. GE found that the drag forces acting on the stacked disk strainer are the most important source of excess margin. Further, the most important parameter needed to calculate the drag forces acting on the stacked disk strainer is the hydrodynamic mass coefficient.
Previous calculations used a very bounding value of 2.0 for this coeff cient. To continue to use this same value for the much larger stacked disk strainer would impose very large forces on the new strainer that are not realistic. Testing could demonstrate that a coefreient of 2.0 was an overly conservative value. GE indicated that there are two ways in which margin could be greatly L
+
3 reduced: (1) the presence of the perforated plate rather than solid plate, and (2) the consideration of the finite rather than the infinitely long solid cylinder. To better account for these affects, GE established a scaled testing program.
The scaled testing program was performed to establish the imped of perforated plates versus solid plates on simple right circular cylinders. A simple geometry was selected so that an analytical model could be developed to compare with the test results. This comparison showed that close agreement could be demonstrated between test and analysis and that the effect of the perforations had a profound effect on the results. Unfortunaiety, the analysis could not be developed for the more complex geometry of a stacked disk strainer under evaluation. However, this testing program was sufficient to conclude that significant reductions in the mass coefficient were possible by the affect of perforations alone.
GE then developed an analytical model which was capable of modeling the complex configuration of the stacked disk strainer. As stated earlier, this analytical modeling could only be accomplished for solid surfaces and not the perforsted surface of the strainer. To accomplish this objective, a commercially available computer program was used for the study of the stacked disk geometry of the strainer. Since the passing fluid could be either air or water, it was adapted to compute the necessary parameters to allow the calculation of the hydrodynamic mass coefficient.
Parametric results of the computer based analysis of the stacked disk geometry showed the affect of each parameter (strainer hydraulic length and strainer diameter) on the value of the mass coefficient. It was noted that the results are only valid for solid surfaces, however, GE assumed that the same affect would be realized for the stacked disk strainer with perforated plates.
i GE acknowledged that the available technology would not allow the direct analytical modeling of the complex configuration of the stacked disk geometry with perforated plates. Therefore, a mixture of a test and analysis was selected as the preferred approach. A single test was i
conducted by GE using a prototype strainer to determine the mass coefficient value. This test was then used to estaolish the scale positions for the calculation of hydrodynamic mass of -
different stacked disk strainers with varying hydraulic lengths and diameters. This approach assumes that the analytical results would be the same for both the solid and perforated surfaces.
The staff concems are based on the fact that only one size strainer nas been tested for submerged structure drag forces. From a single strainer, drag forces were extrapolated for all other size strainers. The basis for the extrapolation as described above was briefly presented in the GE LTR and was discussed at the January 27,1998, meeting.
During the meeting, GE explained that the testing described in the GE LTR should be sufficient to establish a data base which GE believes ghe=s confidence in its analytical methods to predict how drag forces change with respect to a change in strainer geometry. However, without benefit of test corrections, the analytical tools can not preCict stacked disk strainer drag forces. GE stated that the reason the analytical method does n'st accurately predict the experimental results is the inability of the model to property evaluate the presence of holes in the perforated plate of the strainer and, therefore, the mass coefficient valt a would not be correctly calculated. in this case, hydrodynamic mass, as calculated by the corr puter code, is much larger than the measured hydrodynamic mass of the prototype strainer by about a factor of four. Normally, this
4 large factor between analysis and test would be a cause for concem with the analysis. It should be noted, however, that GE uses its analytical procedure only to evaluate changes in stacked disk strainer geometry ; length and diameter) and then estimates the resultant submerged fluid drag forces acting on die strainer.
NRC STAFF CONCLUSIONS
' The new stacked disk straber design is significantly larger than the original plant design strainer.
This increased size increases the loads that are imparted to the stainer assembly and eventually to the torus penetration. Recognizing that the computed loads could not continue to use boundmg assumptions, GE embarked on a comprehensive program to address and refine these larger drag loads. The program consists of both testing and analytical methods to compute drag forces and both appear to be well-founded.
The staff strongly supports the GE program as being a rational approach to establishing reasonable loads on the stacked disk strainer. However, because of the elaborate geometry of the stacked disk strainer, a purely analytical approach to estimating drag forces acting on the strainer can not be accepted by the staff. The staff has concluded that the analytical effort needs to be sufficiently supported by appropriate testing. The staff has not concluded that the analysis is flawed, but rather, that the analysis is incomplete and must be adequately supported by additional test results.
The analysis described to date by GE is supported by the testing of one prototype strainer. The analytical model used to extrapolate the prototype design to other strainers is based on a computer based method intended to model solid plates as an air foil. The perforated plate of the strainerwhich also contributes to the reduced drag of the strainer has been modeled by GE for this strainer configuration in separate calculations. It has only been modeled for a very simple right cylindrical geometry. Therefore, the approach relies completely on testing to determine the affect of perforated ar.d stacked disk geometry rather than solid plates. Both analysis and testing are necessary for evaluating strainer drag forces.
While the staff agrees with the concepts, it also acknowledges soms limitations of the modeling methods chosen. The computer based analytical model used by GE to evaluate different stacked disk strainer geometries has predicted higher coefficients for the prototype stralaer than the actual strainer achieved during testing. GE stated that it believes that the competer.wse.d analytical method accuracy is sufficient to use this analytical method to predict the drag *>rce on different strainer geometries. GE also stated that the fact that the computer based analytical method does not calculate the correct absolute drag coefficient should not invalidate its use to predict changes in drag coefficient. GE stated that changes in strainer geometry should be considered only as relative changes and, therefore, the absolute value of the calculated drag coefficient does not need to be considered. GE believed that testing one prototype was sufficient to valedate the use ofits analytical method. The staff can not agree that testing one prototype strainer is sufficient to qualify the above described analytical methods. However, the staff has not conduded that a separate test for each unique strainer is necessary.
Added testing is necessary to validate the current assumption that strainer behavior relative to the drag coefficient will be the same for both perforated and solid surfaces. To accomplish this objective, the staff believes that a strainer design with hydraulic length and diameter significantly different from the prototype should be selected for testing.
w.
~ Based on the staff's identified need for an additional test, it is important to understand the l.
program status until the tests confirm the adequacy of the entire approach. The prototype test j
result is approximately a factor of two below the GE recommended design value for the stacked disk strainer acceleration drag volume.
i In light of the known improvement in the strainer's ability to accommodate large amounts of l
debris, the staff believes that the installation of the new strainers is a major step toward improved j
safety. Although the staff believes that additional testing is needed to demonstrate the accuracy l
of the analytiel r athods, the anticipated design margins which GE has specified are sufficient i
for the short-term. However, without a better understanding of the design margin available for l'
the drag load, the staff has a concem relative to exceeding the design loads of the torus l-penetration. On this basis, the staff believes that the additional testing can be considered confirmatory.
l ORIGINAL SIGNED BY:
Donna M. Skay, Project Manager l
Project Directorate 111-2 l
Division of Reactor Projects j-Office of Nuclear Reactor Regulation l
Enclosure:
List of Attendees
' cc w/ encl: See next page i
i DISTRIBUTION-Hard Copy E-Mail Central Files SCollins/FMiraglia (SJC1/FJM)
JKudrick MJordan, Rill PUBLIC -
BBoger (RPZ)
RElliot MRing, Rlli l
PDill-2 r/f EAdensam -
MMarshall DHills, Rlll L
D. Skay SRichards AD'Angelo OGC CMoors
. CBerlinger L
~ACRS.
TMartin (SLM3)
THiltz (TGH)
DOCUMENT NAME: G:\\CMNTSPtSKAY\\127MTG. SUM To HECEIVE A COPY oF THis DoCUMER WDICJ iTE W THE aoX: "C" = copy WITHouT ENCLOSURES '*E"
- COPY WITH ENCLOSURES "N" OFFICE PM:PD3-2 O
LAD L2 R
NRR:SCSk h D:PD3-2.
C-DSkay:sph dM' CBerlingeh $ SRichardsM NAME' QATE 4/ 4
/98 4/
/g8 4/ P'
/08[ 'k 4/
9 /98
. OFFICIAL RECORD CQ$ '
f I
Based on the staffs identified need for an additional test, it is important to understand the program status until the tests confirm the adequacy of the entire approach. The prototype test result is approximately a factor of two below the GE recommended design value for the stacked disk strainer acceleration drag volume.
In light of the known improvement in the strainer's ability to accommodate large amounts of debris, the staff believes that the installation of the new strainers is a major step toward improved safety. Although the staff believes that additional testing is needed to demonstrate the accuracy of the analytical methods, the anticipated design margins which GE has specified are sufreient for the short-term. However, without a better understanding of the design margin available for the drag load, the staff has a concem relative to exceeding the design loads of the torus penetration. On this basis, the staff believes that the additional testing can 1:e considered confirmatory.
A
/
Donna M. Skay, Project Man er Project Directorate Ill-2 Division of Reactor Projects Office of Nuclear Reactor Regulation
Enclosure:
List of Attendees cc w/ encl: See next page
(
1 l
l*
MEETING ATTENDANCE LIST l
JANUARY 27,1998 l
NAME AFFILIATION l
Rob Elliot NRC Michael Marshall NRC Jack Kudrick NRC Tony D'Angelo NRC I
Donna Skay NRC l
Martin Torres GE Nuclear l
Joe Quirk GE Nuclear i
John Lynch GE Nuclear Alan Bilanin Continuum Dynamics i
Paul Bunker PP&L Kevin Brinckman PP&L Richard Pistolas PP&L R.R. Sgarro PP&L Thomas Oldenhare PP&L ATTACHMENT
~
,?.
Rocky Sgarro, Chairrnan BWR Owners' Group ECCS Suction Strainer Committee c/o Pennsylvania Power & Ught 2 North Ninth Street, Mail Code GENA 61 Allentown, PA Thomas A. Green, Technical Project Manager General Electric Company 175 Curtner Avenue, Mail Code 182 San Jose, CA fr5125 Thomas J. Rausch, Chairman 8WR Owners' Group c/o Commonwealth Edison Nuclear Fuel Services 1400 Opus Place Downem Grove, IL 60515 Joseph Quirk General Electric Company 175 Curtner Avenue, Mail Code 182 San Jose, CA 95125 l
f