ML19274F019
| ML19274F019 | |
| Person / Time | |
|---|---|
| Site: | Millstone  |
| Issue date: | 04/24/1979 |
| From: | Counsil W NORTHEAST UTILITIES |
| To: | Grier B NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION I) |
| References | |
| NUDOCS 7906080042 | |
| Download: ML19274F019 (8) | |
Text
1 we ts c:an IMHtTHi!AST UTII.ITII!S 4 s,w m v,,, m --. o w.,'
P.O. BOX 270
]
w m.c r... i.#, r r.***'
HARTFORo, CONNECTICUT 06101
- l":::::::::.';?:::"~""
a m> ses-se,s L
L a ::::: :."?;;rlt";'":I""
April 24, 1979 Docket No.' 50-245 Mr. Boyce H. Grier Director, Region I Office of Inspection and Enforcement U. S. Nuclear Regulatory Commission King of Prussia, PA 19406 Gentlemen:
Millstone Nuclear Power Statica, Unit No.1 IE Bulletin 79-07 In response to I&E Bulletin No. 79-07 regarding seismic stress analyses of safety-related piping, the attached information is hereby submitted.
Northeast Nuclear Energy Company (NNECO) wishes to note that the seismic design methods utilized in the analysis of safety-related piping at Millstone Unit No.1 will be reviewed by both the NRC Staff and its consultants within the framework of the Systematic Evaluation Program.
Very truly yours, NORTHEAST NUCLEp ENERGY OMPANY
/
/
/
[M
/
W. G. Counsil Vice President Attachment 7 906 0 806Q.:
e DOCKET No. 50-245 i
MILLSTONE NUCLEAR POWER STATION, UNIT NO.1 RESPONSES TO I&E BULLETIN 79-07 APRIL, 1979
n' Seismic analyses of safety-related piping for the original plant design of Killstone Unit No. 1 were performed by two organizations. John Blume Associates was contracted by General Electric Corporation to perform the analyses of the Recirculation Loop Piping and the Suppression Chamber Saction Header. All other analyses were performed by Ebasco Services Company.
IE Bulletin 79-07 requests information intended to allow the NRC to determine whether algebraic summation techniques were used in the design of safety-related piping. The information requested is paraphrased below in the form of four questions.
Question (1)
Identify which, if any, of the methods specified were employed for the seismic analysis of safety related piping in your plant and provide a list of safety systems affected.
Question (2)
Provide complete computer program listings for the dynamic response analyses portions for the codes which employed the techniques identified in Item 1 above.
Question (3)
Verify that all piping computer programs were checked against either piping benchmark problems or compared to other piping computer programs. You are requested to identify the benchmark problems and/or the computer programs that were used for such verifications or describe in detail how it was determined that these programs yielded appropriate results (i.e., gave results which correspond to the correct performance of their intended methodology).
Question (4)
If any of the methods listed in Item 1 are identified, cubmit a plan of action and so estimated schedule for the re-evaluation of the safety-related piping, supports, and equipment affected by these analysis techniques. Also provide an estimate of the degree to which the capability of the plant to safely withstand a seismic event in the interim is impacted.
Responses to the questions raised by th IE Bulletin follow, categorized by responsible organization:
Responses for Analyses Performed by John Blume & Associates Response to Question (1)
The Suction Header was not analyzed directly for seismic inputs by John Blume &
Associates. The Suction Header report (Reference 1) states: " Response estimates used in the investigation are based on dynamic analysis of the 20-inch suction header for the Monticelle Nuclear Generation Plant performed by H. J. Sexton and Associates, Engineers".
Hence, the results reported in Reference 1 are inferred from another report.
l'ae seismic analysis of the Recirculation Loop pipes described above used the response spectrum modal analysis method. Algebraic summation of codirectional
. spatial components or codirectional intermodal components was not used. The modal inertial forces were calculated using standard response spectrum analysis procedures. The codirectional modal inertial forces were then combined on a squar e-roo t -o f-t he-sum-o f -t he -s qua r e s (SRSS) basis. Resulting displacements, stresses, etc., were then determined from the SRSS inertia forces which were applied as static loads.
Response to Question (2)
Algebraic summntion techniques were not used for the seismic analysis of the recirculation piping or the 20-inch suction header at Monticello.
Response to Question (3)
Two computer codes were used in the dynamic analyses of the Recirculation Loops:
3DFRM and SMIS.
Two computer codes were used for the suction header:
SMIS and FRAMEl.
SMIS was developed at the University of California at Berkeley and was used for matrix manipulation, eigenvalue solutions, and response spectrum dynamic analysis.
The program 3DFRM was used to develop the flexibility matrix for the loops. The program FRAMEl was used to develop the flexibility matrix for the suction header.
SMIS is a public domain program and has been widely used. A description and listing may be found in Reference 1.
A listing of 3DFRM is not readily available.
Verifications for 3DFRM and FRAMEl were performed at the time the programs were written. This documentation is not readily available. As mentioned above, SMIS is a public domain program and has been widely used.
Some aspects of the program are self verifying, sach as orthogonality checks.
In addition, the program was tested and verified by ORS /Blume engineers when it was first used in the 1960's.
However, verification documentation is not presently available.
R_esponse to Question (4) e Since John Blume & Associates did not employ any of these techniques, this question is not applicable.
Reference 1 - Wilson, E. L., SMIS Symbolic Matrix Interpretive System, Department of Civil Engineering, University of California, Berkeley, 1973.
Responses for Analyses Performed by Ebasco Services Company Response to Question (1)
No dynamic seismic analyses were performed by Elasco Services Company in terms of response spectra or time history evaluations. Only static analyses were performed, hence, the question is not applicable.
Response to Question (2)
Since dynamic analysis techniques were not employed, this question does not apply.
~
- Response to Question (3)
The code used by Ebasco for analyses of Millstone Unit No.1 piping systems has been verified against Ebasco's current version of the PIPESTRESS 2010 pro-gram. The PIPESTRESS 2010 program has, in turn, been compared with solutions to sample problems generated by similar, independently written programs in the public domain, namely, ANSYS, PIPESD, AND ADLPIPE. This comparison shows the PIPESTRESS 2010 results to be essentially identical to results generated by the above programs and by hand calculations.
Response to Question (4)
Since Ebasco did not employ any of these techniques, this question is not applicable.
Backfit Projects Involving Safety Related Piping Backfit projects involving modifications to or additions of safety related piping were investigated to determine applicability of the subject IE Bulletin.
Dynamic seismic analyses of safety related piping were performed by Teledyne Engineering Services and Northeast Utilities Service Company.
Responses for Analyses Performed by,Teledyne Engineering Services Response to Question (1)
TES did not use algebraic summation for combining the effects of orthogonal components of earthquake motion. The actual methodology used is delineated in the response to question (3).
Response to Question (2)
TES did not employ algebraic. summation techniques, therefore, this question does not apply.
Response to Question (3)
The computer program used by T2S for seismic analysis of piping at MP1 was TMRSAP.
TES has used a technique of running earthquake directions separately and combining resultant effects (forces, moments, deflections) outside the computer program as follows:
The representative maxinum value of the three moments Mx, M, and Mz at y
any point in the piping system subjected to each of the three independent spatial component response spectra was obtained by taking an SRSS summa-tion of the modal responses for all significant modes of the system.
Mathesatically, this is expressed as follows:
]!
.M
=[
M g
k=1 k
where Mj is the representative maximum value of moment, j is the moment component direction x, y, or 2.
Mjk is the peak value of moment component due to the kth mode, and N is the number of significant modes.
. The combined effect of the three spatial components of earthquake was determined subsequently by the following procedure. The representative maximum values of the codirectional moments (either Mx, My, and Mz) from the two horizontal components of earthquake were combined by the SRSS method and this SRSS value then added absolutely to the representative maximum value of the codirection moment for the vertical component of earthquake. Mathematically, this is expressed as M = [(M ) + (M )
+ (M )Y (2) j j
j j
where Mj is the total seismic moment component Mx, My, or Mz, 04j )X,Y,Z are the representative maximum values of codirectional moments (SRSS values) for each of the X, Y, Z earthquake directions, respectively.
Since all terms are SRSS values, they all possess a positive sign.
THRSAP is a TES proprietary version of SAP IV in which pre-and post-processing options were added by TES. No changes were made to the basic SAP IV program. THRSAP has been completely verified and a verification report is on file in the TES QA Document Control files. A sample of the results from a comparison analysis by TES follows:
Comparison of ADLPIPE and TMRSAP Seismic Stresses for PIPDYN Manual example problem that is also used in SAP IV Manual.
X - Direction Seismic Spectral Loading, B31.1 Stress Sumnary Mode Intensification ADLPIPE TMRSAP Number Component Factor Stress, psi Stress, psi 3
Run 1.00 411 409 3
Elbow 2.80 1122 1146 4
Elbow 2.80 1105 1108 4
Run 1.00 397 395 8
Branch 1.00 896 894 9
Run 1.00 537 537 9
Elbow 2.71 1448 1452 Note that effects of orthogon:1 components of earthquake motion were combined outside the computer programs for both the verification cases and the original analyses performed by TES. The methodology used internally by the computer program for combining effects of orthogonal components of earthquake motion is, therefore, irrelevant for these cases.
Response to Question (4)
Since TES did not employ any of these techniques, this question is not applicable.
Responses for Analyses Performed by Northeast Utilities Service Cocpany Response to Question (1)
The two analyses identified were performed using the computer program ADLPIPE.
Algebraic summation techniques were employed in these two enalyses as previously reported to the NRC Staff via telecon en April 5,1979.
. Response to Question (2)
A computer program listing for the dynamic response portions of ADLPIPE has not been included as a portion of this response.
Northeast Utilities Service Company has been in contact with A. D. Little (owners of ADLPIPE) and have recently received permission from them to provide this requested information.
- However, transmittal of this information is not provided at this time in order to respond to your request in a timely fashion. Transmittal of this information would involve an affidavit pursuant to 10CFR2.790, and is considered super-fluous as it is NNECO's understanding that the intent of your request has been fulfilled by representatives of A. D. Little during meetings with the NRC Staff.
Response to Question (3)
Important steps in the verification of ADLPIPE, as supplied by A. D. Little Company, are as follows:
Verification of ADLPIPE was undertaken in a serf es of fundamental checks.
In important modifications, a supporting document was prepared as an ADLPIPE reference. The verification procedure was as follows.
The thermal and deadweight loadings were checked by a Hovgaard Bend and hand calculated systens given in 'tesign of Piping Systens", M. W. Kellogg, Second Edition,1956, and " Formulas of Stress and Strain", R. J. Roark, McGraw-Hill.
The dynamic analyses were checked by " Response of Structural Systems to Ground Shock", Shock and Structural Response, ASME,1960, in "ADLPIPE Results of Model Given by Youug (ADLPIPE Reference 4), and " Dynamic Behavior of a Foundation 'Like Structure", Mechanical Independence Methods, ASME,1958, in " Experimental Verification of ADLPIPE Mod 1" (ADLPIPE Reference 3).
The time history analysis was checked by a separate analytical solution of the problem given in " Analytical Methods of Vibrations", Page 395, Leonard Meinovitch, "ADLPIPE Time History Response Compared with a Known Solution for a Heavily Lamped System (ADLPIPE Reference 14). A second check was made usins " Pressure Vessel and Piping 1972 Computer Progress Verification",
ACME,1972 (Problem 5).
The thermal transient analysis was verified by a separate analysis, " Transient Thermal Gradient Stresses", E. B. Branch, Peating, Piping and Air Conditioning, Volume 43,1971, Pages 132 - 136, "ADLPIPE Ihermal Transient Analysis (Reference 15).
The computation of intia and inter-modal moment component summation has been verified by a separate computer program for that purpose. A report "ADLPIPE Fbdal Response Combination for Closely Spaced Modes", is available as ADLPIPE Reference 24.
Various calculation procedures required by ASME Section III were verified in ADLPIPE References 10,11, and 18 entitled "ADLPIPE Computation of Bending Stress in Tees and Bran h Connections, ASME Section III, Class 1 Piping",
"ADLPIPE Computation of Resultant Moments for Section III Class 2 and 3 Stresses",
and "ADLPIPE Stress Computation of Piping Components: A Comparison with Hand Calculations for ANSI B31 and ASME Section III".
. In 1978, an independent third party review of ADLPIPE (Section III, Class
- 1) was performed " Verification of ADLPIPE, ASME Section III, Class 1 Piping Stress Program", Teledyne Engineering Services, Report No. TE-2884-1, August 11, 1978.
We understand that further information regarding ADLPIPE verification and methodology have been presented by A. D. Little to the NRC Staff during an April 16,1979 meeting and in a memorandum dated April 19, 1979. Therefore, we have not included this additional information as a part of this response.
Response to Question (4)
The affected piping systems have been identified as:
(1) CRD exhaust piping from the tie-in to the feedwater line to the second isolation valve, 8 feet of 3 inch diameter piping.
(2) 1-CU2 Bypass Line - 1 inch diameter bypass line around a valve in the clean-up water systen supply.
A reanalysis of the affected portions of these systems has been started within Northeast Utilities Service Company.
Stress analyses of both sections of affec^ed piping has been performed and che stresses are within allowable code limits.
The piping restraints for these systems are presently under evaluation.
This evaluation will be completed and any design modifications required will be performed during the spring, 1979 refueling outage.
It is our estimate that the ability of the plant to withstand a seismic event is not adversely impacted in the interim.