ML19263E701

Submits ACRS Consultant 790416 Rept on Containment Shell Buckling Criteria

ML19263E701
Person / Time
Site:	Atlantic Nuclear Power Plant
Issue date:	05/23/1979
From:	Baer R Office of Nuclear Reactor Regulation
To:	Collier A OFFSHORE POWER SYSTEMS (SUBS. OF WESTINGHOUSE ELECTRI
References
NUDOCS 7906250063
Download: ML19263E701 (12)

Text

.

s -

/ a aro,o 1]4.~1, h n

u UNITED STATES E j ) "3, c ' %

NUCLEAR RECULATORY COMMISSION C7 5.%N 1

W ASHING TO N, O. C. 20555 4~!

s

• ..../

MAY 2 41979

'V Docket No. STN S0-437 Mr. A. R. Collier, President Offshore Power Systems P. O. Box 8000 8000 Arlington Expressway Jacksonville, Florida 32211

SUBJECT:

BUCKLIllG CRITERIA FOR FREE STANDItiG CONTAINMENT SHELLS OF NUCLEAR POWER PLANTS We have received a copy of an ACRS consultant report dated April 16, 1979, from Zenons Zudans, Franklin Research Center, on the subject of containment shell buckling criteria. We are forwarding a copy of this report for your information and use.

Sincerely,

[Me[ g. Am >7 Robert L. Baer, Chief Light Water Reactors Branch No. 2 Division of Project Management

Enclosure:

Letter Report, Zenons Zudans, Franklin Research Center,

" Buckling Criteria for Free Standing Containment Shells of fluclear Power Plants,"

April 16,1979 22l0 }'O ces w/ enclosure:

See next pages 7906250063

y' Mr. A. R. Collier

~

President Of fshore Power Systems P. O. Box 8000 8003 Arlington Expressway Jacksonville, Florida 32211

~-

cc: Vincent W. Campbell, Esq.

Vice President & General Counsel Offshore Power Systems P. 0. Box 8000 8000 Arlington Expressway Jacksonville, Florida 32211 Thomas M. Daugherty, Esq.

Offshore Power Systems P. O. Box 8000 8000 Arlington Expressway Jacksonville, F1orida 32211 Barton Z. Cowan, Esq.

Eckert, Seamans, Cherin & Mellott

~*

600 Grant Street, 42.:d Floor

(~

Pittsburgh, Pennsyl\\ania 15219 is Franklin H. Berry, Jr., Esq.

County Counsel County of Ocean

,L P. O. Box 757 i,

34 Washington Street Toms River, Ney.'ersey 0,8753 Anthony Z. Roi sman, Esq.

m Natural Resources Defense Council W shi g o,

C.

2b05 Dr. Glenn L. Paulson U

Assistant Commissioner

(-

State of New Jersey

]j@ j@

Department of Environmental Protection Labor and Industry Building John Fitch Plaza Trenton, New Jersey 08625 L'-

Keith A. Onsdorff, Esq.

Mirian N. Span, Esq.

Assistant Deputy Public Advocates 520 East State Street Post Office Box 141 Trenton, New Jersey 08625

\\

y.

s

• .m 9

Mr. A. R. Collier g

3g cc: Mark L. First, Esq.

~

New Jersey State Deputy Attorney General l

State House Annex Trenton, New Jersey 08625 L<.'l Valore, Jr., Esq.

Va. ore, McAllister, DeBrier, A on & Westmoreland Mai.iland Professional Plaza 535 filton Road P. O. Box 152 Northfield, New Jersey 08225 y.

Mr. Haroid P. Abrams 9100 Amheest Avenue Margate, f.ew Jersey 08402 Dr. Willard W. Rosenberg 8 North Rtcison Avenue Margate, f.ew J6rsey 08402 Mr. John Williamson' b.

211 Forest Drive Linwood, New Jersey 08221 Harold P. Green, Esq.

Fried, Frank, Harris, Shriver and Kampelman Suite 1000, The Watergate 600 600 New Hampshire Avenue, N. W.

Washington, D. C.

20037

[

Nuclear Coordinator Office of Merchant Marine Safety S:

~

M Conmandant (GMMT-4) 2218 170 c.

U. S. Coast Guard Washington, D. C.

20590 r,.

Mr. Mitchell Attalla kU-4028 Ponce DeLeon Avenue Jacksonville, Florida 32217

.J,'

Mr. Reg Crower j-;

Jacksonv'11e Journal r-P. O. Box 1949 E-Jacksonville, Florida 32201

i..-

M

,e a.

.g a %

.e g g

"r. A. P. Collier WWW 2 3 G79 cc: Sheldon J. Wolfe, Esq., Chairman Atomic Safety and licensing Board U. S. Nuclear Regulatory Commission Washington, D. C.

20555 Dr. David R. Schink Department of Oceanography Texas A. & M. University College Station, Texas 77840 Mr. Lester Kornblith, Jr.

Atomic Safety and Licensing Board V. S. Nuclear Regulatory Commission i

Washington, D. C.

20555 Richard S. Sal zman, Esq., Chairman Atomic Safety and Licensing Appeal Board V. S. Nuclear Regulatory Commission Washington, D. C.

20555 Dr. John H. Buck h2 Atomic Safety and licensing Appeal Board U. S. Nuclear Regulatory Commission Washington, D. C.

20555 Mr. Michael C. Farrar Atomic Safety and Licensing Appeal Board U. S. Nuclear Regulatory Commission Washington, D. C.

20555

..aw ms 2218 171 Iw.

m~

=

• e I

Amh y.

ENCLOSURE Ah

_ Franklin Research Center A Division of The Franklin Instina CY / /O C f

April 16, 1979 KttilVED AD\\150D COM 41fTEL UN REAC10R N EO N U Dr. Paul G. Shev=on Professor and Chairman of

,, gj g

Metallurgical Engineering Dept.

p /

34 [40 I

Ohio State University gygCtUtUll i'-1 n,

1 !

Columbus, Ohio.

T Re: Buckling Criteria for Free Standing Contain=ent Shells of Nuclear Power Plants

Dear Professor Shev=on:

In the ecciosure, Ihave discussed in detail the current situation in evaluating containnent adequacy to resisc the buckling code of fairure.

As indicated, there are no precise criteria available; hence, decisions made during the licensing review process are not necessarily unique or consistently the same from case to case. To resolve this proble= h7C should undertake research to:

1)

Evaluate existing expericental results to deter =ine actual buckling loads for configuration similar to contain=ent buckling.

This is equivalent to determining the factor CA which indicates by how much the theoretical linear bifurcation buckling load cust be reduced to account for deviations from perfect gec=etry and nonlinearities as they cay exist.

2) Encourage (or sponsor) develop =ent of shell of revolution co:puter program with capability to consider the nonsy==ctric prebuckling states as well as mutliple Fourier su==ation for buckling = odes.

As indicated by Dr. Hafi, Ites 1 above might be partly resolved in the research sponsored at International Structure Engineers. I have seen a Draft Report, dated October 1973, uhich presu= ably does not represent the total output anticipated under this contract.

The coe=ents I made at the Sequoyah Fall Cocaittee : ecting en 6 April 1979 are brought into better perspective if one looks at the results in the light of the discussion of the enclosure to this letter.

The applicant presented in the buckling analysis report two sets of results which were claimed to support cach other. One set of results was based on linear bifurcation analysis which came up with a buckling load factor 3 = 4.6 (CB1 analysis).

If this result is interpreted in the context of C

NE-3222.1 (a) (2) discussed in the enclosure, the acceptance criteria would require application of a factor of safety Cs = 1/3 and possibly C3 = 1.2 for service type.

2218 172

Dr. Paul C. Shermon Ohio S.ta te University April 16,19 79 s

The criteria then would read CCC A

3CS C

(4. 6) ()1.2>1 1.82 C 2* 1 C 2;0.54 which = cans that the imperfections and other unaccounted-for effects should not cause the actual structure to buckle at less than a half of the coeputed load.

I believe r' ls is adequate deconstration of the Sequoyah's centain=cnt capability to resist gross buckling failure code.

What I consider inappropriate is the statenant in FSAR. page 3.S.2.-3, which essentially says that in lieu of originally defined buckling criteria (which could not be satisfied), the applicant used Anscet Dynamic Analysis.

This is an implication that the analysis referred to represents an exact analysis in the spirit of NE-3222.1 (a) (1).

If such had been the case, the results by this analysis should have indicated in tability at a considerably lower load than that by C3I analysis.

Instead, both analyses seeted to have confir=atory results. My concern in this context is that I consider the computer program used for the above analysis not adequately validated to put such a trust in it.

The local panel buckling analysis using STACS programs' casa up with C although it is true that C 3 = 2.5, A may be close to 1 for this case, the result is uncomfortably close to the limit and it might indicate some additional conservatism in the codel not specifically identified.

Very truly yours,

!g4~;cs) fY

&cnons 7.udans ces Senior Vice President, Engineering

encls, cc:

Prof. M. Plesset,' Cal. Inst. of Technology Mr. R. Savio, ACRS Mr. A. Bates, ACRS D * * ]O

]D

' S T M

~

2218 173 6 dju eJu,1.kfru.a

APPE:; DIX DISCUS 5tJN OF EUCKLING CRITERIA FOR STEEL CONTAIN'CNT VESSELS

~

Subsection NE of the ASME Boiler and Pressure Vessel Code,Section III, Division 1 establishes rules for design and certification of cetal contain-ment systems.

Subsection NE-3133 of this code provides rules for determining the thickness under external pressure loading in spherical sh,cils and cylindrical shells with or without stiffening rings.

In most practical applications, it turns out that cylindrical contain-ment vessels reinforced only with circumferential rings does not satisfy the buckling criteria of NE-3133.

In su,ch cases, the actual contain=ent vessels is provided with meridional stiffeners (stringers) having sub-stantial cross-section in addition to the vessel wall and circu=ferential rings. There are no criteria for this type of design, neither are there criteria for the design of rather co plex configurations with cut-outs, curvature transitions and other detail consistent with functional require-Instead of such criteria, NE-3200 allows design by analysis.

ments.

In particular, NE-3222 describes the buckling stress allowables as quoted below.

2218 174 D""

'T I

WW, 3

1 a

me

s NE-3222 BUCKLING STRESS VALUES NE-3222.1 Basic Compressive Allownble Stresc.

The maximum buckling stress values to be used for the evaluation of instability shall be either of the following:

(a)

One-third the value of critical buckling stress deter =ined by one of the methods given below:

(1)

Rigorous analysis which considers the effects of gross and local buckling, geometric i= perfections, nonlinearities, large deformations, and inertial forces (dynamic loads only).

(2) Classical (linear) analysis reduced by margir.s which refice:

the difference between theoretical and actual load capacities.

(3) Tests of physical models Onder conditions of restraint and loading the same as those to which the configuration is expected to be subjected.

t>

(b) The value determined by the applicable rules of NE-3133.

While the intent of NE-3222 appears to be clear the following discussion demonstrates the difficulties the designer fa'ces:

NE-3222.1 (a) (1) sounds nice, but is obviously i= practical to apply. First there are no analysis tools which can be generally accepted as " rigorous." Most analyses claimed today as having the " rigorous" characteristics, suffers from symptoms which to say the least make that claim unreal. While purely technically it is possible to define methodology qualifying as rigorous the lack of precise information of i= perfections of the containment "as built" preclude this approach anywhere.

There have been statistical imperfection definitions used in connection with the reduction of buckling capability esiculation.

Such efforts, however, are essentially well enough developed for publication of a paper and not for actual design work.

In conclusion, I do not believe SE-3222.1 (a) (1) provide guidance, it is subject to misuse by claims of rigorous capability which in reality does not exist.

2218 175 ~

$$ 23]@.

2 enume g

p

• e r e e

=

EC_}222.1 (2) (2) describes a method which has a good chance of unambiguous it.plementation. To sum arize what this paragraph says let's us walk through analysis steps performed to comply vitii these criteria.

s Suppose ve have been given a steel containment structure such as that Sequoyah Nuclear Power Plant. With today's existing capability we can compute stresses in such a structure due to all applied heads (such as ice condenser originated nonsymmetric pressures). We call this "prebuckling state of stress" and denote the corresponding load P.

Linear bifurcation buckling analysis will yield a minimum buckling factor, C. This factor, D

times the applied load, P, vill yield the minir.us linear, bifurcation buckling load for a perfect structure Phif " BL Paragraph NE-3222.1 (a) (2) requires that this classical linear bifurcation buckling load, Pgg, be reduced by a factor, call it C, to determine the g

actual buckling load capacity, P as it =ay exist for a real structure, g,

crit " A bif " ABL The value of C depends on the geometry and the nature of the i=perfec-tions in the real structure as =anufactured as well as the =aterial's non-linearities should they occur.

In the next step the allowable buckling load, P,, is determined by applying a safety factor, C, and a service condition factor, C,

e the 3

C critical load,

=C

=CCC a

C S crit g3CSL The safety factor, C, is given as 1/2 by NRC Reg. Cuide 1.57, and as 1/3 by 3

NE-3222.1.

2218 176 3

s hT-3222.2 (1977) gives C a 1 f r n raa p r ting s a cc, 1.2 for C

emergency service, and,1.5 for faulted condition.

By this approach design will be acceptabic if the allowable buckling load exceeds the actual load s

P

?

11 or

<1 which is equivalent to 1

l ABCS CCC g3CS where C :

Factor reducing linear bifurcation load of a perfect structurc A

to actual buckling load of a real structure.

~

C-Factor multiplying applaed loid to yield linear bifurcation 3

buckling load of a perfect st;ructure.

C: Factor correcting the load allowable for the load service C

condition.

C: Factor of safety applfs1 to the critical buckling load of a 3

real structure.

Are there any problems associated with the implementation of the approach based on NE-3222.1 (a) (2)? We answer this question by analyzing various multipliers needed to obtain the allowable buckling load P,.

1) C actor yielding linear bifurcation load with ? being the pre-B buckling load. When shell of revolution analysis methodology is used in currently popular form there may be some question as to the accuracy with which C can be computed. This arises from the fact that applied loadings 3

are not axisymmetric (such as compartment pressures for the ice condensor containment building, SRV discharge lead in MARK III Sk2 and seismic loads in all Lb2's).

Shell of revolution based computer programs would produce nonsymmetric prebuckling state. However most such computer programs cannot use the nonsy= metric prebuckling state as the basis for bifurcation buckling analysis. This is due to the fact that for non-symmetric prebuckling load pattern the rajor advantage of shell of revolution approach is losr-the Fourier expansions of leads and responses no longer uncouple and one is forced 4

2218 177

to analyze aZZ har~monics simultaneously. Com=only used approach is to scicct the highest coepressive stressed meridian and use it as an axisymmetric prebuckling state.

It is then claimed that the C t b.u s 3

computed is conservative, an assertion intuitively plausible, but yet to be proven for all structures to which'it is applied.

If one censiders non-symmetric prebuckling load say such that only one quarter of the shell is under compression in meridional direction it is not unreasonable to postulate that the governing code of buckling may not be periodic in circumferential direction as it will result with axisyemetric prebuckling state, but may consist of a number of Fourier terms acting si=ultaneously such that only the vicinity of compressive regions will determine the buckling code. All shell of revolution computer programs'used today look at buckling codes in the form of Fourier terms one by one and select the Fourier term producing smallest C as the bifurcation buckling mode. The -

g obvious result is that the co=puter buckling code circumferentially is of a single Fourier ters type. This situation can be resolved by deve, loping a shell of revolution program capable of using non-symmetric prebuckling state and also examining C for various combinations of several Fourier 3

terms. Alternative way of accomplishing the sama objective is to use currently existing surface type (two di=ensional) finite ele =ent pregra=s to compute C.

If ne is able to cope with the resulting econo:ic penalty B

this approach can be used. Additional refinc=ents in the for= of refined local modelling can be used to improve the resolution and the econocies.

2)

C - fa t r reducing linear bifurcation load to that of real A

s t ruc ture. This factor, C, represents the major unkncwn in connection with the methodology of NE-3222.1 (a) (2). The only way credible C can 3

be produced is by comparing analytical results to actual test results.

Significant amount of testing already has been done which would allow to determine C with fair degree of confidence for a number of configurations.

g What is needed is a systematic and comprehensive review of all available data and comparison of these to linear bifurcation analysis to define C 3

with an acceptable level of confidence. To quote but a few NASA SP-8007, [1}*,

NASA CR-912, [2], represent a significant contribution in this direction.

Similarly a recent paper by Hiller [3] provides a ce=prehensive study on

• Number in brackets, please refer to list of references.

e 5

2218 178

axially compressed cylinders most recently final draft report " Suckling Criteria Application of Criteria to Design of Steel Containment Shell" [4]

s attempts to address this subject in greater detail.*

At the time of this report I have not studied the conclusions of the above re, art in detail.

Based on oral information conveyed to me by Dr. Hafi: it appears that authors of [4] have come up with conclusions similar to those expressed here.

3) C'S ere is no d W culty in assigning appropriate.a h es C

for these factors.

NE-3222.1 (a) (3) allows experimental method to be used to determine the buckling load.

1 believe this is totally impractical for two reasons:

a) can not test full scale containment building, b) there is no vay to =edel containment building to represent it realistically.

In conclusion it is my opinion that a method similar to that described here under NE-3222.1 (a) (2) should be considered as the basis for bucklin; criteria of containment buildings.

Research should be conducted to determine C f r t pical structures and loads.

Factors C can be determined by A

3 finite (Le=ent (2-D) methodology.

Improve shc11 of' revolution analysis should made as discussed under 1) C above.

B REFERENCES l.- NASA SP-8007 Buckling of Thin-Walled Circular Cylinders, Sept. Ic65, Rev. August 1968.

2.

NASA CR-912, Shell Analysis Manual, by E. H. Baker, A. P. Cappelli, L. Kovalevsky, F. L. Rish, and R. M. Verette, prepared by North American Aviation, Inc., April 1968.

3.

C. D. Miller " Buckling of Axially Compressed Cylinders," ASCE, Journal of the Structural Division, ST3, pp. 698-721, March 1977.

4.

Buckling Criteria and Application of Criteria to Oesign of Steel Containment Shell, International Struc tural Engineers, Clendale, CA 91206, October 1973, prepared for Structural Engineerin;; tranch, Division of Systems Safety, U.S.N.R.C., under Contract NRC-03-77-131.

• Dr. A. Hafi:, USNRC technical monitor.

2218 179 6