ML20092P795

Summary of Testimony & Testimony of Ak Singh on Contention 1 Re Reinsp Program,Work Quality.Related Correspondence

ML20092P795
Person / Time
Site:	Byron
Issue date:	07/02/1984
From:	Singh A SARGENT & LUNDY, INC.
To:
Shared Package
ML20092P775	List: ML20092P774 ML20092P778 ML20092P781 ML20092P785 ML20092P787 ML20092P789 ML20092P792 ML20092P794 ML20092P795
References
OL, NUDOCS 8407090382
Download: ML20092P795 (9)

Text

, -_

a t-NEUulb (

~"~~~ COMMONWEALTil EDISON COMPANY Date: July 2, 1984 UNITED STATES OF AMERICA *%,J E7[p NUCLEAR n6GULATORY COM" 9CICh L 3h@

BEFORE Tile ATOMIC SAFETY AND LICENSI'.qq BOgD In The Matter Of ) .

r

) i , '-

COMMONWEALTil EDISON COMPANY ) Docket Nos. 50-454-OL

) 50-455-OL (Byron Nuclear Power Station, )

• Units 1 & 2) )

SUMMARY

OF Tile TESTIMONY OF ANAND K. SINGli ON CONTENTION 1 (REINSPECTION PROGRAM - WORK QUALITY)

I. Anand K. Singh is the Assistant Head of the Structural Analytical Division of Sargent & Lundy.

II. Mr. Singh was involved in preparing the portion of the Byron Reinspection Program Report which dealt with the inferences of work quality from the Program.

III. Mr. Singh has applied principles of statistics and probability theory to the results of the engineering evaluations discussed in the testimony of Messrs.

McLaughlin, Leone and French. lie concludes with a 95% confidence level that, in general, the work per-formed by flatfield and !!unter meets the original design basis with a greater than 99% reliability.

8407090382 840702 PDR ADOCK 05000454 T PDR ,

l

, o l UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION T', FORE THE ATOMIC SAFETY AND LICENSING BOARD In the Matter of )

)

COMMONWEALTH EDISON COMPANY ) Docket Nos. 50-454-OL

) 50-455-OL (Byron Station, Units 1 and 2) )

TESTIMONY OF ANAND K. SINGH Q.1. Please state your full name and place of employment for the record.

A.I. Anand K. Singh, Sargent & Lundy, 55 East Monroe Street, Chicago, Illinois.

Q.2. Please describe your job responsibilities.

A.2. I am Assistant Head of the Structural Analytical Divi-sion. In this capacity, I supervise and coordinate the work of the Stress and Probabilistic Analysis and the Dynamic Analysis Sections in preparation of analytical studies, special problem analyses, and com-puter program development.

Q.3. Please describe your educational background and work experience.

A.3. I have a Doctor in Philosophy and a Master of Science degree in Structural Engineering from the University w _-_ -_ _ - - - - _ _ _ _ _ _ . _ _ _ - _ _ _ _ - _

of Illinois at Champaign-Urbana. These degrees were awarded in 1972 and 1970, respectively. I am a regis-tered professional engineer and a registered struc-tural engineer in the State of Illinois. I am a mem-ber of the American Society of Civil Engineers (ASCE),

and a member of the Seismic Analysis Committee of the ASCE Nuclear Structures and Materials Committee, a member of the Working Group on the Seismic Analysis of Safety of Class Structures of the ASCE Nuclear Stan-dards Committee and a member of the ASCE Committee on Turbine Foundations. I have published numerous tech-nical papers in the area of probabilistic analysis, seismic analysis and dynamic analysis of structures and piping. A list of my publications is attached to my testimony.

I I joined Sargent & Lundy in 1972 as a Senior Engineer-ing Analyst. I was responsible for the development and maintenance of computer programs for seismic and dynamic analyses of structures and piping and for per-forming and/or reviewing seismic analyses of nuclear power plant structures. In 1975, I was promoted to the position of Supervisor of the Dynamic Analysis Section responsible for seismic and dynamic analysis of structures and the development of computer programs for dynamic and seismic analysis. In 1979, I was pro-

~2-L

moted to the position of Assistant Division Head. In that capacity, I supervise and coordinate the work of the Stress and Probabilistic analysis and the Dynamic Analysis Sections in preparation of analytical studies, special problem analyses, and computer pro-gram development. In 1980, I was made an associate of Sargent & Lundy.

Q.4. Are you familiar with the Byron Reinspection Program?

A.4. Yes. The reinspection program was developed by Com-monwealth Edison Company to verify the effectiveness of former inspector certification practices and inspector qualification by re-examining, on a sampling basis, inspections performed by QC inspectors certi-fied prior to 1982.

Q.5. Were you involved in the preparation of the report?

A.S. Yes. I was involved in preparing portions of the

] report dealing with work quality, including the sec-tion on inference of-work quality from the reinspec-tion program.

Q.6. What is the purpose of your testimony?

A.6. The purpose of my testimony is to apply principles of statistics and probability theory to the results of

the engineering evaluations discussed in the testimony of Messrs. McLaughlin, Leone and French.

Q.7. Would you summarize the results of the engineering evaluations to which you are applying your statistical analysis? -

A.7. Yes. Their testimony explains that the results of engineering evaluations performed by Sar9 ant & Lundy demonstrated that none of the 356 Hatfield Electric Company ("Hatfield") weld discrepancies analyzed or ,

any of the 2,311 objective discrepancies analyzed had design significance. Similarly, the engineering eval-uations demonstrated that none of the 109 Hunter Cor-poration (" Hunter") weld discrepancies or 689 Hunter objective discrepancies analyzed had design signifi-cance.

Q.8. Applying a statistical analysis to these results, what conclusions do you reach with respect to the total population of work performed by Hatfield and Hunter?

A.8. From a statistical standpoint, I conclude with a 95%

confidence level that, in general, the work performed by Hatfield and Hunter meets the original design basis with a greater than 99% reliability.

Q.9. Please explain the basis for your conclusions.

A.9. The reliability for a work attribute can be defined as the proportion of work items in the total population of work for that attribute which has no discrepancies with design significance. A generally accepted sta-tistical method for calculating such reliabilities is to compute reliabilities at 95% confidence level from the sampled data. Such a reliability represents a conservative estimate of the true reliability. It is conservative in the sense that there is a 95% chance that the true reliability is greater than the esti-mate. In the case where no discrepant items are observed in a random sample from a large population, the reliability at 95% confidence level can be calcu-lated from the formula R= 1- 2.9955 n

where

~

R = Reliability at 95% confidence level, n = number of inspections in the random sample.

For Hatfield welding, approximately 28,000 welds were reinspected. This resulted in approximately 2,200 observed discrepancies. This shows that 8% of the Hatfield welds do not meet the conservative specifica-l 1

tion requirements. From this population of approxi-mately 2,200 welds, 356 welds were evaluated for the design significance of the observed discrepancies.

This evaluation showed that none of the observed discrepancies had any design significance. By applying the above formula, this sampling evaluation establishes with 95% confidence that greater than 99%

of all obnerved discrepancies do not have any design significance.

For Hatfield objective attributes, approximately 67,000 items were reinspected. This resulted in approximately 2,300 observed discrepancies. All these observed discrepancies were evaluated for design sig-nificance. This evaluation showed that none of the observed discrepancies had any design significance.

By applying the above formula, this sampling evaluation establishes with 95% confidence that, in general, greater than 99% of all Hatfield objective work in the plant meets the design requirements.

The actual reliacility levels for each of the 10 objective attributes reinspected are shown in Table 1. The table shows that for 8 out of 10 attrib-utes the reliability is greater than 99%. For the remaining two attributes, the reliability is computed as 98.9 and 96.3 percent.

l 1

a .

For Hunter welding, a total of 3,725 welds (1,007 AWS and 2,718 ASME) were reinspected. Discrepancies were observed in 60 AWS welds and 49 ASME welds. These observed discrepancies were evaluated for their design significance. None of the observed discreparteies had any design significance. By applying the above for-mula, this sampling evaluation establishes with 95%

cenfide'. ace that more than 99% of all Hunte: welds meet the design requirements.

For Hunter objective work, a total of 34,878 hardware items were reinspected. This resulted in 248 observed discrepancies. All these observed discrepancies were evaluated for their design significance. This evalua-tion showed that none of the observed discrepancies had any design significance. By applying the above formula, this sampling evaluation establishes with 95%

confidence that more than 99% of all Hunter hardware work items meet the design requirements.

~

s

r TABLE 1:

CALCULATED RELIAPILITIES FOR WORK OF HATFIELD No. of Reliability j No. of Discrepancies  % at 95% l Inspected with Design Confidence Items Significance Level Objective Attributes

1. Conduit 2,793 0 99.9 i

2. Terminations 7,784 0 >99.9

3. Equipment setting 778 0 99.6

4. A325 bolting 295 0 98.9 4

5. Equipment modification 1,850 0 99.8

6. Conduits as-built 44,777 0 >99.9

7. Pan hangers 4,776 0 > 99. 9

! 8. Pan 80 0 96.3

9. Conduit support bolting 1,008 0 99.7

10. Concrete expansion anchor 2,840 0 99.9 1

_J