ML19352A980

Revised Final Deficiency Rept,Originaly Reported on 800321, Re Improper Documentation & Utilization of Computer Program Inertia.Inertia & Bigdyna Properly Utilized & Only Seismic Analysis for Control Bldg Affected

ML19352A980
Person / Time
Site:	Yellow Creek
Issue date:	05/28/1981
From:	Mills L TENNESSEE VALLEY AUTHORITY
To:	James O'Reilly NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION II)
References
10CFR-050.55E, 10CFR-50.55E, NCR-YCNCEB-8004, NUDOCS 8106020490
Download: ML19352A980 (4)

Text

"

400 Chestnut Street Tower II Hay 28, 1981

1. [ P(' N%

[.[ 'j ( \.

m g,:;w '-}F 7-Mr. James P. O'Reilly, Director ~

- ej O 1 ES Office of Inspection and Enforcement 2 nc^^*" '

U.S. Nuclear Regulatory Comission Region II - Suite 3100 C["N h ' # # '

/

101 Marietta Street Atlanta, Georgia 30303

/^,VMY x

'WI'#

'/9'

Dear Mr. O'Reilly:

YELLOW CREEK NUCLEAR PLANT UNITS 1 AND 2 - IMPROPER DOCUMENTATION AND UTILIZATION OF COMPUTER PROGRAM INERTIA - NCR YCN CED 8004 - REVISED ,

FINAL REPORT The subject deficiency was initially reported to NRC-0IE Inspector R. W. Wright on March 21, 1980, in accordance with 10 CFR 50.55(e).

Interim rwports were submitted on April 21, May 5, and August 1, 1980, and our final report was submitted October 14, 1980. As discussed with J. Lenahan and J. Harris en April 28, 1981, enclosed is our revised final report.

If you have any questions concerning this matter, please get in touch with D. L. Lambert at FTS 857-2581.

Very truly yours, TENNESSEE VALLEY AUTHORITY L'. M. Mills, Manager N'iclear Regulation and Safety Enclosure cc: Mr. Victor Stello, Jr., Director (Enclosure)

Office of Inspectial and Enforcement U.S. Nuclear Regulatory Comission Washington, DC 20555 QOl9 s

/'

8106020Y96

ENCLOSURE YELLOW CREEK NUCLEAR PLANT UNITS i AND 2 IMPROPER DOCUMENTATION AND UTILIZATION OF COMPUTER PROGRAM INERTIA NCR YCN CEB 8004 10 CFR 50.55(e)

EEVISED FINAL REPORT

References:

1. John A. Blume, Nathan M. Newmark, and Leo H. Corning,

" Design of Multistory Reinforced Concrete Building for Earthquake Motions," Portland Cement Association,1961.

2. N. C. Chokshi and J. P. Lee, " Shear Coefficients and Shear Force Distribution in Nuclear Power Plant Structures Due to Seismic Loadings," International Symposium on Earthquake Structural Engineering, St. Louis, Missouri, August 1976.

3 " Analysis of Small Reinforced Concrete Buildings for Earthquake Forces," Portland Cement Association,1955.

Description of De_ficiency Improper docu=entation and utilisation of the TVA computer program, INERTIA, could have resulted in incorrect shear-related cross sectional properties being used in seismic analyses. INERTIA documentation did not include verification of the program for structures with open cross sections (such as a channel section). Structures with closed sections (cells) connected by a wall may also have response charaateristics similar to an open section. The Control Building at Yellow Creek Nuclear Plant has cross sections of this latter type. Incorrect shear-related cross sectional properties will affect the natural frequencies of the structure and, consequently, the structural response and instructure response spectra.

In our investigation of the applicability of this problem to the Watts Bar, Bellefonte, Hartsville, and Phipps Bend Nuclear Plants, it was discovered that INERTIA is used as a subroutine in a computer program known as BIGDYNA. BIGDYNA is a structural analysis program that calculates the modal frequencies and mode shapes ot a structure and determines the structural response to dynamic loads by the modal superposition method.

Safety Implications A preliminary seismic analysis was performed in September 1977 for the original Control Building geometry using normal seismic analysis methods followed by the detailed design in accordance with standard TVA design procedures. Later, during the period from mid-1979 to late 1979, a second ceismic analysis was performed taking into account structural changes in the building that had occurred to that date. In the second analysis the computer program SASA was used to calculate the shear-related cross section properties. The results of the second analysis were 20 to 40 percent higher than the original analyses. An evaluation of why the two seismic results differed showed that several (shear) walls were changed, thus creating, in effect, an "open section" in the structure cross section.

4 l

This changed the natural frequencies of the structure and the seismic loads. Furthermore, SASA used " beam theory" to determine the shear-related cross sectional properties, whereas TVA's computer program INERTIA used in the original analyses used the rigidity method. Both beam theory and rigidity analytical methods are technically correct for analyzing closed sec tions. Both computer programs had been checked and verified except that the TVA program INERTIA had not been documented as being technically correct for use in analyzing open sections such as those found in the Control Building. A nonconformance report was filed because of the uncertainty involved in the preliminary seismic analyses in accordance with regulations stipulated in 10 CFR Part 50.55(e).

In that INERTIA uses the rigidity method which is known to be a technically acceptable method, nonconformance report (NCR) YCN CEB 8004 as reported on March 17, 1980, was not a safety concern. It is now clear that the method of analysis addressed by this NCR was in fact not a nonconforming condition. The nonconforming condition was the lack of documentation on the aspect ratio limits of the program for open structures.

Corrective Action The computer program INERTIA uses the rigidity method described in Reference 1 to calculate the shear-related properties for use in seismic analysis. Other methods are beam theory and finite element method. The documentation for using the rigidity method in the seismic analysis of structures with open sections was determined by comparing the beam theory and rigidity methods with the finite element meth:4 Beam theory and rigidity methods give essentially the same results for closed sections.

Comparison of the two methods does not agree as closely for all open sections and closed sections connected by a single wall, but each is preferred in certain applications. Theoretically, finite element methods give "ccrrect" resulta and, therefore, were used in Reference 2 to compare with the rigidity and beam theory methods. A comparison of the these methods is given below.

1. Literature Survey - The literature showed that the rigidity method was a standard method used in seismic design of structures (for example, References 1 and 3). In one study directly applicable to this NCR, the rigidity method was found to be more accurate for open structures than beam theory. The rigidity method and' beam theory give essentially the same aesults for closed sections (Reference 2).

2. Induatry Survey - An informal survey of utilities and architect-engineers angaged in nuclete power plant construction showed that the rigidity method is comsonly used for seismic analyses of structures with both open and closed sections.

3 Study by TVI - Fol:owing filing of NCR YCN CEB 8004, TVA conducted a study to accccmine the relative merits of the rigidity method and beam theory. Analyses were performed on structures with four different open cross sections and aspect ratios of 0.75, 0.50, 1.0, and 2.0 using beam theory, the rigidity method, and finite element method. It was found that the rigidity method was consistently more accurate than the beam theory when compared with results from the finite element method.

1 t

1 t

. The extent to which a model predicts how the shear forces are distributed  !

across the cross section of the structure determines the accuracy of the l natural frequencies, mode shapes, and, consequently, the structural responses. While both the rigidity method and beam theory are valid approaches, neither can fully account for the interactions (shear flow) I between walls and walls and slabs in complex "open" structures. But until work now being done by TVA, NRC, and others to better define how shear forces from earthquakes are distributed is completed, open structures will l be conservatively analyzed on a case-by-case basis using one or more of the  !

three methods to calculate shear-related properties. The programs INERTIA and BIGDYNA will continue to be used.

A check has been made of all other TVA nuclear plants, and it has been ,

determined that INERTIA and BIGDYNA were properly utilized and only the I seismic analysis for the Control Building at Yellow Creek was affected. It has now been established and documented that the rigidity method is acceptable for open structures with aspect ratios up to two, which covers the open structure found in the Yellow Creek Control Building. Some designs and design drawings were changed because of the increased seismic loading calculated by the computer program SASA in the second analysis.

However, because of the number of design enanges that occurred up to the l time of the reanalysis TVA is not able to determine drawing changes that occurred as a direct result of tne analysis. The structural changes were more significant than those that resulted from the reanalysis.

Furthermore, still other structural changes occurred before the Centrol Building was redesigned to incorporate the second seismic analysis results, and a third seis=1c analysis was performed. The results of the third

• seismic analysis have also been incorporated into the structure design. ,

l

- _ _ _ _ ._ , . , - . , . . _ , - . _ . . _ . . . . _ , , , . - , _ . , , _ , - , . . . _ . _ , , . . _ _ - - , , - - . . - - -