ML17258A520

Forwards Info Re Seismic Capacity of Reactor Coolant Pump. Action Proposed to Close Item

ML17258A520
Person / Time
Site:	Ginna
Issue date:	09/11/1981
From:	NELSON T A LAWRENCE LIVERMORE NATIONAL LABORATORY
To:	RUSSELL W T Office of Nuclear Reactor Regulation
Shared Package
ML17258A195	List: ML17258A194 ML17258A518 ML17258A520
References
CON-FIN-A-0415, CON-FIN-A-415 SM-81-247, NUDOCS 8202090379
Download: ML17258A520 (19)

Text

Lawrence Livermore National Laboratory September ll, 1981 SM 81-247 Docket 850-244 FIN A0415 Mr.William T.Russell, Branch Chief Systematic Evaluation Program Branch Division of Licensing Office of Nuclear Reactor Reg.Washington, D.C.20555

Dear Bill:

I have enclosed information regarding the seismic capacity of a reactor coolant pump at the Ginna plant.While this open item has not been completely resolved, a rather simple action is proposed which should close this item.Sincerely, TAN/mg 0131m Enclosure Thomas A.Nelson Structural Mechanics Group Nuclear Test Engineering Division sooaoqoo79 saoiis PDR ADOCK 05000244 I PDR An Endleaf~tnffy El~NiVeetyef Caifemia PQ BaX 808 Li mmes Caifemia 94550~TelephOne(415)422-1100

~TWX9f0-38&8339 UCLLL LWR i i 11 IE P P.'I STRUCTURAL IECHAfllCS ASSOC l ATES A Collf.Coro.5160 Birch Street, Newport Beach, Caiif.S2660 (714)833-7552 SMA 12205.20 August 27, 1981 Mr.Thomas A.Nelson (L-9Q}Lawrence Livermore Laboratory Nuclear Test Engine'ering Division P.0.Box08 Livermore, California 94550 P

Subject:

Resolution of Open Items on Ginna Equipment l

Dear Tom:

!n my June 22 letter, there were still two open items for Ginna equipment which were: 1.Control Rod Drives and.Supports 2.Primary Coolant Pump This letter is to advise you of the current status of these items..George Wrobel of Rochester Gas&Electric (RGE}Company arranged for me to talk directly to Robert Kelly of'Westinghouse, who was responsible for providing much of the Ginna equipment seismic documentation.

Regarding loading documentation on the CRD assemblies, R.Kelly will send his analysis of CRD housing and seismic support loading for'a 0.8g static coefficient loading.A quick review of the analysis should resolve all outstanding issues.on the CRD system.A legible copy of the primary coolant pump report was transmitted to me from RGE.Conclusions from the review of that report are contained in the attachment, Tom Cheng suggested that I be present at an upcoming open items meeting for Ginna in September.

In view of the current status of open items for which I have been responsible, I don't feel it necessary for me to attend;the=

meeting.After my conversation with R.Kelly of Westinghouse, I feel that the CRD and primary coolant pump issues will be.resolved very'quickly and can be handled without a meeting.It is.therefore recommended that these'tems not be a topic of the meeting.r Yery truly yours, STRUCTURAL MECHANICS ASSOCIATES, INC.rZW Robert D.Campbell Project Manager RDC:lca cc: T.Cheng{NRC)

~~ATTACHMENT REACTOR COOLANT PUMP SEISMIC DESIGN REVIEM The original stress report submitted for review, was illegible.

A more legib1e copy was submitted to SMA by RGE on August 17, 1981, and the following conclusions can be reached from review of that submittal.

I~1.The pump was analyzed for a 0.8g horizontal static coef-ficient and a 0.54g.vertical static coefficient.

As reported in.NUREG/CR-1821, the 7X damped peak spectral acceleration for both horizontal directions is 0.55 g's resulting in a vector sum of 0.78 g's.Thus, the equiva-I lent static coefficient used in the original analysis is conservative by a small margin.2.'ll stresses calculated fdr the 0.8g H and 0,54g V static coefficients ai e within allowables.

designated for the original design basis.3.Analytical methods used in the design analysis are reason-able except in the ca'se of pump nozzles.4.The pump nozzles analysis is unrealistic and inadequate for the followi.ng reasons: a.Only the straight pipe portion of the nozzles were evaluated.

Local membrane stresses in the pump casing were not computed.

ATTACHHENT Continued b.The derived pump nozzle loads have no resemblance to actual achievable loads.The pump was assumed to be supported by the piping for purposes of deriving nozzle loads.This is probably highly conservative but not necessarily so.Actual piping reactions are available and should be used in an evaluation of the pump case.A conversation between R.Kelly of Mestinghouse and R.Campbell of SMA revealed that:.i 1.The San Onofre pumps are very similar to the Ginna pumps and that a detailed finite element analysis was conducted for the San Onofre Units for a specified set of nozzle loads.2.A loop analysis of Ginna was conducted by Mestinghouse for seismic loading.Actual pump nozzle loads are obtainable from the analysis.Reconmended Actions RGE should have Vestingfiouse make a comparison of Ginna vs San Onofre pump casing geometry and pump nozzle loads and scale resulting stresses from the San Onofre pump analysis for Ginna nozzle loading conditions.

The load/stress comparison and a comparison of nozzle and casing geometry should be sufficient to demonstrate seismic capability of the Ginna primary coolant pumps.

8 1 p~~r-STRUCTURAL lTIECHAAICS ASSOCIATES A Calif.Corp.5160 Erch Street, Newport Beach, CaN.92BBO (714)833-7552.SMA 12205.20 June 22, 1981 Mr.Thomas A.Nelson (L-90)Lawrence Livermore Laboratory Nuclear Test Engineering Division P.0.Box 808 Livermore, California 94550

Dear Tom:

In the Reference 1 submittal regarding review of open items on Robert E.Ginna Nuclear Power Plant, there were three (3)open items still remaining.

1.Control Rod Drive Mechanism 2.Reactor Coolant Pump 3.Steam Generator Tube Supports I Tom Cheng of the USNRC contacted me on June 16, to inquire if use of a, site specific spectra for Ginna, in lieu of the regulatory guide spectrum anchored to 0.2g, would eliminate the oustanding items.'indicated that for two of the three outstanding items, lack of information, not marginal stress conditions, was the principal concern.'s a result of the conversa-tion.I have reexamined the thr ee open items and can eliminate the steam generator tube support overstress problem by using Level 0 Service (faulted condition) allowable stresses and scaling up the original analysis results to applicable floor spectral accelerations.

The control rod drive and primary coolant pump items still remain open but should be easily resolved with submittal of the necessary information.

The remaining action items and resolution of the steam generator tube support stress problem are sum-marized in the attachment.

Very truly yours, STRUCTURAL MECHANICS ASSOCIATES, INC.Robert D.Campbell Project Manager RDC:lca Attachment cc: J.Stevenson ATTACHMENT

SUMMARY

OF OPEN ITEMS FOR ROBERT E.GINNA NUCLEAR POMR PLANT AS PART OF THE SEP PROGRAM CONTROL ROD DRIVE MECHANISM The submittals provided by the licensee, References 2 and 3, do not contain a correlation between loading used in the analyses and accel-erations at the RPV support.A conclusion regarding seismic resistance of the CRD system cannot be reached without such a correlation.

Mesting-house should be able to supply the necessary information to Rochester Gas and Electric.REACTOR COOLANT PUMPS The Reactor Coolant Pump Stress Report submitted for review, Reference 4, is illegible due to poor reproduction quality.A legible report needs to be submitted for review.STEAM GENERATOR TUBE SUPPORTS Section 16 of Reference 5 documents the stress analysis of Series 44 steam generator internals for seismic loading.A conservative analysis of the tube supports resulted in a primary membrane stress of.7900 psi for an equivalent static coefficient of 0;19g horizontal.

The stress was calculated in the ligaments between the tube holes and circu-lating holes in a local area adjacent to a wrapper channel.The analysis'conservatively ignored a redundant load path from the.edge of the tube"support to the tube holes.Reference 6, submitted for review, is an update analysis of the Series 44 steam generator but does not address the~1 tube support for horizontal loading.Neither Reference 5 or 6 describe the dynamic characteristics of the steam generator and its internals nor identify the material of construction for the tube support plates.Reference 7, obtained during the SSMRP program, documents a generic dynamic analysis of the Series 51 steam generator for varying support stiffnesses and locations.

The Series 44 steam generator is similar, but smaller.The fundamental frequency of the Series 51 steam generator ranges from 4.8 to 9.6 Hz depending upon the support stiffness and location.The fundamental mode is predominantly translation and rocking of the steam generator shell.A generic response spectrum that is flat through most of this frequency range was used to compute response accelerations and loadfngs in the steam generator.

A review was conducted of Reference 7 to determine the degree of coupling between the steam generator shell and the internal structures and to establish the validity of considering the steam generator as a SDOF system for estimating an appropriate equivalent static coefficient for evaluation of the tube support plates..It was concluded from review of Reference 7 that at the most critical support location, as determined in Reference 5, that the tubes and tube supports would accelerate as rigid bodies with the shell;thus, using the spectral acceleration from the Ginna spectrum for a funda-mental frequency of about.5 Hz is a reasonable approximation of an equiva-lent static coefficient, to use for evaluation of the tube supports.From the response spectra in Reference 8, for 7X damping, the maximum spectral accelerations at 5 Hz are 0.58 g's in each of two ortho-gonal directions.

Combining the two directional accelerations, the resulting maximum vector is 0.82g.Using this value and scaling the ligament stress computed in Reference 5, the resulting ligament stress is 34,095 psi.Reference 7 indicates that Series 51 steam generator internals are constructed of SA 285-Grade C carbon steel.A compar ison of the allowable stress for this material at the design.temperature of 556 F to the allowable stress quoted for the Series 44 steam generator tube supports in Reference 5, indicates that the Series 44 tube supports are also constructed of SA 285-Grade C or equivalent.

This material and a design temperature of'556 F are used as a basis for establishing allowable stresses for the Safe Shutdown Earthquake.

The tube supports are considered to be Class 1 plate and shell type component supports and the allowable primary membrane stress is computed for Level 0 Service from Appendix F of the ASME Code, Reference 9.The allowable stress is the greater of 1.5 Sm and 1.2 Sy.For SA 285-Grade C material, 1.2 Sy governs and the allowable stress is 27840 psi.Note that the original design criteria limited the tube support stress to S.Comparison of the calculated and allowable stress for Level D Service-results in a 22%overstress condition.

If site specific spectra anchored to 0.172g are considered in lieu of regulatory guide spectra anchored to 0.2g, the calculated.

stress decreases.

Decreasing the calcu-lated stress by the ratios of the site specific peak ground acceleration divided by the 0.2g peak ground acceleration used to generate floor spectra, the resulting ligament stress is 29,320 psi.This is still.about.5'.3X over the Level D Service allowable of 27,840 psi.In consideration of the conservatism inherent in obtaining the calculated stress, the computed 5.3X overstress condition is considered acceptable for several reasons.1.The site specific spectrum envelope has lower spectral accelerations in the frequency range of the containment structure than the regulatory guide spectrum if both are anchored to the same peak ground acceleration.

Conse-quently, the in-structure response spectra will be lower than those in Reference 8.'.The static analysis from Reference 5 did not account for'he redundant load path between the outside diameter of the tube support plate and the outer row of tube holes.'The degree of conservatism could not be evaluated since perti-nent dimensions are not provided in Reference 5.The'egree of conservatism is certainly greater than 5X though.

3.The evaluation considered the applicable acceleration to be the vector.sum of the two orthogonal directional acceler-ations.This assumes that both directional responses are in phase and that the resulting vector is aligned in the worst direction.

4-The in-structure response spectra were peak broadened+15%and smoothed so that the resulting spectra are essentially flat from 2-1/2 to 9 Hz, which covers the range of funda-mental frequency for the steam generators.

Items 3 and 4 are consistent with current regulatory criteria and are prudent conservatisms to cover many of the uncertainties in the simplified treatment of the tube support.Recommended acceptance is, therefore, based on the conservatism of Items 1 and 2 being sufficient to overcome the estimated 5.3f overstress condition.

Further analyses or submittals from the licensee are not considered necessar y.

REFERENCES 1.SMA letter, R.D.Campbell to T.R.Nelson, Review of Open Items Resulting from Seismic Review of the Robert E.Ginna Nuclear Power Plant as Part of the SEP Program, 4 May 1981.2.High Speed Control Rod Drive Stress Analysis June 26, 1968.3.Control Rod Drive Mechanism Seismic Frame Calculations, August 13, 1968.4.Static Seismic Load Stress Analysis, Model RGE Pump (V-11001-BI), July 30, 1968.5.Vertical Steam Generator Stress Report, Mestinghouse Electric Corporation, Tampa Division, April, 1969.6.MTO-SM-75-028, 44 Series Steam Generator Stress Report, External Load Analysis Update, May, 1975.J 7.Stress Report;51 Series Steam Generator, Generic Seismic Analysis, westinghouse Electric Corporation, Tampa Division, December, 1974..8.NUREG/CR-1821, Seismic Review of the Robert E.Ginna Nuclear Power Plant as Part of the Systematic Evaluation Program, 15 November 1980.9.ASME Boiler and Pressure Vessel Code,Section III, Nuclear Power Plant Components, Appendices, 1980.

STRUCTURAL ICCHAAICS ASSOCIATES A C~Ilt.Coro.5160 Birch Street, Newport Beach, Calif: 82660 (714)833-7552 SMA 12205.20 May 4, 1981 Mr.Thomas A.Nelson (L-90)Lawrence Livermore Laboratory Nuclear Test Engineering Division P.O.Box 808 Livermore,California 94550

Dear Tom:

SMA has reviewed the package of documents transmitted with your April 15 letter addressing open items on Ginna.Our comments and r ecommended action are contained in the enclosure.

Very truly'yours, STRUCTURAL MECHANICS ASSOCIATES, INC.pi Robert D.Campbell Project Manager RDC.mw Encl osur e cc: J.Stevenson w/encl.

REVIEW OF OPEN ITEMS RESULTING FROM SEISMIC REYIEM OF THE ROSERT E.GINNA NUCLEAR POWER PLANT AS PART OF THE SEP PROGRAM Reference 1 documents a review.conducted by the Lawrence Livermore Laboratory and their consultants on the seismic adequacy of the Robert E.Ginna Nuclear Power Plant.Conclusions of the adequacy of the several items in the NSSS system were based upon summary information provided;however, the sources of the summary informa-tion were not available for independent review'.Those items listed in Section 5.4 of the report as components for which seismic design analyses.have not been independently verified are: Reactor Control Rod Drive Reactor Vessel'upports Steam Generator Reactor Coolant Pumps-Pressurizer and its Supports References 2 through 7 were provided to SMA in response to the above identified open items.-The following summary and conclusions resulted from SMA's review'of the submittals-.

1)Control Rod D~ive Mechanism In reference 7, the allowable bending'oment in the CRDM due to seismic loads is developed.

This report does not, however, provide a correlation between bending moments and acceleration levels, Thus, no conclusion can be reached on the basis of the submittal.

Reference 2 contains a stress analysis of the control rod drive support structure.

The analysis provides stresses as'function of a static load"P" in pounds.There is no corre-lation between this static load and acceleration level.Therefore, a conclusion on the seismic capability of the support structure cannot be reached based upon the submittal.

ls 5~~SMA's experience with the SSMRP reference plant, which uses 106Afull length control rod drive.mechanisms, indicated that a large margin of safety exists for a 1.15g spectral accel-eration loading condition and we would not anticipate a seismic problem with the Ginna CROM.Recomme'rided action-CROM loads documentation applicable to Ginna need to be supplied to SMA for final resolution.

2J Reactor'Vessel Su orts Docume'ntation verifying the seismic adequacy of the reactor l vessel;supports was not submitted.

Based upon SMA's SSMRP experience for nozzle supported RPY's the seismic induced stresses in the nozzles and adjacent shell are very small and the gover ning element for RPV support is the concrete,.shield wall.The shield wall was considered in Ref.1 to be adequate to withstand the 0.2g SSE.3l Steam Generator Reference 6 contains a 1969 static analysis of the Series 44 steam geneator.The most critical area due to seismic load-ing identified in this report is the tube support baffle ligaments which are stressed past yield for a 0.38g horizon-tal static load.The SEP revised spectra'result in an SRSS response of 0.85g horizontal which will increase the stresses by a factor of 2.24.Thus, based on the submittal, the tube support baffles are overstressed for the 0.2g SSE..The static analysis at these tube support baffles was.done quite conser-vatively, however, and a more rigorous'~nalysis will most likely result in lower state of stress.In 1975, an update on the series 44 steam generator was con-ducted (Reference 4)but the tube support area was only evaluated for vertical seismic loading and not for the PI horizontal seismic loading.Thus, the results of Reference 4 cannot be used to update the results of Reference 6 in the area of concern.Recommended Action-Documentation evaluating the tube support baffle ligaments for horizontal seismic loading should be submitted for review, 4)Reactor Coolant Pum s Reference 3 summarizes the stresses induced in the reactor coolant pump by a 0.8g'orizontal and a 0.54g vertical load-ing.The reported stresses are below the ASME Code allowables, but SMA is unable to'evaluate the model or the ana1ysis due to poor reproduction quality of Reference 3.If the static analysis of the pump can be shown to be valid, then the stresses due to the revised Ginna spectra loading will be less than those contained within Reference 3," and thus accept-able.Recommended Action--A legible copy of Reference.3 should be provided to SMA for review.5).Pressurizer Reference 5 contains a 1969 stress report of an 1800 cubic foot pressurizer.

All pressurizers from 800 to 1800 cubic'eet with cast and'fabricated heads utilize the same support skir ts, thus conservative generic analysis was conducted for.the heavier 1800 cubic foot models.Based on this report, the loads resulting from the new Ginna spectra will.cause an over-stressed condition within the support f1ange, This is a very conservative conclusion, however, since the Ginna pressurizer is a smaller 800 cubic foot model plus, the flange analysis itself was very conservatively conducted using a beam bending model in lieu of a more rigorous finite element model.

The 1973 pressurizer report referred to in Reference 1 and the pr'essurizer summary stress report (reference 45 of Refer-ence 1)were not supplied to SMA for review;thus, the ade-quate capacity for a 6.7g equivalent static load portrayed in reference 45 of Reference 1 cannot be verified.Later analysis (References 8 and 9)obtained in the SSHRP program,'for the Series 51 1800 cubic foot cast and fabricated head pressurizers showed the supports to be acceptable for a 0.96g.horizontal and 0.64g vertical loading condition.

This later analysis utilized a finite element model of the skirt and flange as opposed to the conservative beam theory.used in reference 5.Recommended Action-No action required.The pressurizer stress analysis within references 8 and 9 and the statements in References 5, 8 and 9 that the support skirts are identical for the 1800 cubic foot and the 800 cubic foot designs sub-=stantiate the adequacy of the Ginna pressurizer for a 0.2g.SSE.