ML19261B009

Forwards First Progress Rept Re Evaluation of Buckling Design Criteria for Steel Containment Shells.Study of Static & Dynamic Loading Conditions to Which Containment Shell Is Subjected Will Be Completed Shortly

ML19261B009
Person / Time
Site:	McGuire
Issue date:	01/11/1978
From:	Masri A INTERNATIONAL STRUCTURAL ENGINEERS, INC.
To:	Hafiz A Office of Nuclear Reactor Regulation
Shared Package
ML19261B007	List: ML19261B009
References
RS-77-8, NUDOCS 7902090189
Download: ML19261B009 (8)

Text

.

.%. Tit L' CTL ir A 1,

. *if'

.fi '. . ' ExclNi: :sts. INc.

cm, D,;f;} INTI:leN ATif )Na a.

' t P. o. nus p , n s

,- us.ta i m .r c.u.ir. oi2<,o c.s.r.

J anuary 11, 1978 Dr. A. Ilafiz Division of Systsra Safety Office of Nuclear Reacter Regulation Nuclear Regulatory Ccumission Washi;rJton, D.C. 20555 Subjec'- Buckling Criteria a:xl applicat icn of Criteria to design of steel containnent shell. Number RS-77-8.

Dear Dr. IIafiz:

Our first progress report is encloscd in accordance with the requirenents of our URC centract.

We have starte1 preparing a buckling design criteria deeznent covering the buckling design of steel containnent shells. Je parts of

, als doctment are canpletcd, they wiu be forwarded to you.

's - We are still evaluating the L,tatic and dyncaic-loading conditions

' which the steel containnent shell is subjectcd. This study should b 2 ccrupletcd shortly.

Plcase contact us if you have any questions relato-1 to the pragress r eports.

Sincerly,

, k. Ty i > . ,

A.F. bbsri 79020901M

'g a January 3J973 Progress Report _ f or "isucQin3_ Criteria and Ap g ca n s of Criteria to "tcel Containment Shell" (#RS-77-8) s As stated in our proposal, af ter we received the f.o-ahead f ro:n NRC a detailed literature survey would be carried out to determine the state of the art on the use of buckling criteria on the design of metal containment vessels e static and dynamic loads. The following work has been completed or 51s phase of the contract:

1. Librarv search. Je have conducted a detailed literature search using infot 2ation retrieval systems such as the Engineering Index, .

N SA Publications, U.S. Defnese Department Publications, and the Inter-national Engineering Index.

2. Solicited Information. We have contacted the leading authorities in the buckljng field requesting them to send us any informat;on that m

would help us to establish buckling criteria for steel containment sessels.

Appendix A contains a sample letter and a list of people contacted.

Individual meetings were also held with:

Dr P. Gou (General Electric)

Dr. R. Citerley (Anamet Laboratories)

Dr. C. Babcock (California Institute of Technology) to obtain their views on establishing buckling criteria, cafety f act ar and aSNE Code requirements. Subsequent to the meeting with Dr. Gou we received a summary of the dynamic loads that General Electric uses in the design of their containment structures.

Based on our investigations the following statements can be made about the state of the art to date:

1. Most of the experimental results available in the literature

~

for determining design criteria are baked en model tests and the correspondence betwecu model tents and ful1 ize structures still needr.

to be assessed. Design criteria verified by experiment which censiders effects of imperfections, dynimic loads, anynmetric loadings and non-linear effects is practically nonexistent. To obtain this type of information will not be an easy or inexpensive taux. It appears that our best method of obtaining experimental data for establishing design criteria is through carrying out a large number of carefully planned model tests.

2. A large number of computer programs exist for determining buckling loads of shells of revolution and general shells. Programs which seem to have gained the confidence of engineers developing design criteria are BOSOR 4, STACS, NASTRAN and MARC. Even though many of

\ these programs consider nonlinear effects, very little correlation has been obtained between the results of these computer programs to predict experimental buckling results even when the imperfections of the test models are well known beforehand. For the actual design condition when imperfections and loadings are not well defined, computer prograns can only be used us guidelines or as a first step before knockdown factors arc imposed. It also seems important that the limitations of these computer programs should be well documented and the codes should be casily available to those interested in the buckling characteristics of containment structures.

3. The ASME Section III Buckling Criteria Regulation Guide 1.57 NE-3224 which states that (A) One half tne value of critical buckling stress determined by one of the methods given below

1. Rigot nes analysis which considern gross and local but.kling,,

geome tric 1: aper f ec t ions , nonlinearitics, la rge de f orn.:it ions ,

and inertia forces (dynamic loads only).

2. Classical (linear) analysis reduced by margins which reflect the difference between theoretical and actual load capacities.

3. Tests of physical models under conditions of constraint which reflect the difference between theoretical and actual load capac it ics .

must be changed. The use of these criteria permits designers to select the method uhich yicids a bucklin;; stress which is least con-servative. In fact, even with the use of the one hal f factor it is possible for a shell to buckle at a stress below that predicted by Method 3. Fo :xauple, it is well known that some axial compression

'\ cylinder mo/.el tests yield results for carefully made specimens close to 90 percent of the classical buckling value and others with imper-fections yield results less than 20 percent of the classical value.

The use of Method 3 is valuable in establishing guidelines for buckling criteria but could Se dangerous and yield unconservative buckling stresses if the physical models did not exactiv approximate the loading and imperfections of the full scale operating nodel. Since it is impossibic to know the exact geometric imperfections and static and dynamic loadings of the full scale operating model, Method I which uses rigorous analysis has some of the same problems of Method 3. In cases where these factors were known for test models, rigorous analyscs were not, in most cases, able to accurately predict the experimental buckling values. Most authorities in the fielo agree that !!ethod 2 is the most reliable method and this should be reflected in the AStE Section Ill Regulatory Guide 1.57-

o The other methods should be used in conjunction with Method 2 and only in special cases, determined by 2C, used to establish design criteria.

4. Until more test data is obtained to study the effects of imper-fections, asymmetric loading, load interaction, dynamic and nonlinear effects, a conservative factor of safety such as 3 should be used.

5. A general procedure for dett mining the buckling stress nf a metal containment structure has been developed and is summarized below.

1. The containment structure will be accurately modeled by using a general finite element program such as SAP 6 or NASTRAN.

2. The dynamic and static load con.binations of a) dead loads b) construction loads c) accident design loads (LOCA) d) external pressure c) scismic loads f) penetration loads g) thermal loads it) syrmetric and asymmetric loads will be imposed on the finite element model of the containment structure and a linear static and dynamic analysis using SAP 6 or NASTRAN programs will be performed for all critical laod combina-tions. Maximum stresses will be determined and tabulated.

3. After determining a set of critical maximum stress combina-tions the maximum stress along any meridian will be assumed to bc axisymmetric. This has been shown in the pant to be an accurate and conservative approximation. These critical maximum stress

.. , combinations will then be input to the BOSOR 4 program and the

overall buckling load will be determined. The DOSOP. 4 5.rogram

• considers nonlinear prebuckling deformatior.s and pertorr.s a bifurcat. ion analysis to determine the buckling load. 1;e ;ng, th i s proposed procedure asyn. metric loads , interaction cffccts, dynamle loadings, scismic ef f ects and non1inear prebuckIing deior:.ation can be consideredd.

4< Once the overall bu.-kling stresses are determined, these buckling stresses will be reduced by margins which will reficct the difference between theoretical and actual load capacitics.

The NASA design criteria lower bound curves based on experimenta1 data will be used to determine these reduced margins of safety.

5. Af ter overall buckling is investigated, localized buckling will then be considered based on tne stressen obtained from the linear static and dynamic analysis. Any part af the structure that does not satisfy both the local and overall bucklian requirements will be redesigned until these criteria are satisfied.

At the present time we are

1) evaluatirg the various containment vessel loading conditions which must be considered to determin, the applied static.and dynamic stresses.

2) synthesizing the information that we'have obtained and evalua-ting and recasting this information in the form of a cackling criteria design document.

U NIV liRSITY OF SOUTH ERN C A!J FORNI.

g/

J l.((.s% s '.$ - -

,, 1,,,., .

UNIVD$rlT pal'K

• 1.t >S ANGELES, CAL 1FOut.l A 9. . g u

3.)b .

SCllOOL OF LNGINEETING DEPARTMI.NT or CIVIL F,NGISLEP.ING October 12, 1977

Dear Colleague:

The unde::cigned arc involved in a project which requireo the cor.pilation of information on the buckling of r.helln, including shells of rr_,volutien, un:iar localized and nonsymetrd.c loading.

Uc intend doing a thorouah survey of the open literature as ucll as relying on cuch con.pc trlite.a an the Colu;r.n Reroarch Co:nittee of Japan's llandbook of Structural Stability and Applied Mechanics Reviews. We aro c:ncerned, ho :ver, that much useful infor:ntica will be overlooked becatme of the relative otccurity of the journal in which it in published or its unavailability in journal form.

':'hus, ue would be grateful f er a:.y help uhich you might give un in this ta:,k by taking a few cor.ents to cearch your ine:.ory anc'.

your files for titles and authora of pc.porc and rer.otta on the cebject of buckling un !or nonsyn.utric loading. Ccpicn of hard-to-get itemr. would be appreciated. Your aid will be acknowledged in the final report on the nubject.

Sincerely,

{ E.l'*] 6't -'g

'i- / .,.. _, i /

S.F. Manri p.P. Seif:

V.I. Weingartetr Profcasor Professor Professor and Chaircun Dept. Civil Engineering PS/lrm s

M

~

t B.11udiannky , lia rva rd Univers ity J.W. Ilutchinson, !!arvard Universit'y W.T. Koiter, Technological Univerafty of Delft, the Ietherlands II.J . llo f f , S t:2nford Un. ;crsit y C IL S teel e , Stanford University W. Flugge, Stanford University J. Sinner, Technion-Israel Insti tut e of Techmilogy W. Ilachbar, Univernit y of California at La Jolla Dr. L.ll. Dount 11 Dr. D. Bushnell, Lochheed-Palo Alto Research Laboratories Dr. B.O. Almroth, Lockheed-Palo Alto Research Laboratories D. Brush, University of California at Davis C.D. Babcock, Californ'i ins t itute of Technolor,y E.E. Sechler. California Institute of Teciumlogy M. liaruch, University of Wisconnin G.J. Simitnes, Georgia institute of Technology G. Wempner, Georgia Institute of Technology T . ll . II . Pian. Massachusetts Institute of Technology W.A. Nash, University of Massachusetts, Amherst C.S. Usu, Universi ty of California at Bet heley E . l! . Dill, University of Washington

. _x J. Arbocz, California Institute of Technology

\ Dr. J.ll. Starnes, Jr. , NASA--Langley Research Center E.F. Masur, University of Illinoi:. at Chicago Circle Dr. V. Tvergea rd , Dan inh Center f or Appli cd '!a the: ,t i es ansi Mechanics Dr. F.1. I;iocdson, Danish Center for Applied 'tathematics and Mechanies Dr. M. Esuling,er, Inntitut fur Flugzer ban, Braunschweig, Germany A.C. Wal ker , Univern i t;. College, London J .M.T. Thoinpson , Unive rs i ty Coller,c , London R.M. Evan-lwanowski, Syracuse University D.C. Ashwell , University College, Cardif f , Wales Dr. E.1. Gri;;olyuk, Acadeny of Sciences of the USSR, Moscew Dr. W.F. 'Ibielemann, DVL Inst. fur Feltiglcit, Mul he in-Ruhr , Cernuny W Schell, Technological University, Darnstadt, Germany Dr. C.D. !! iller, Chicago Bridge and tron Company i N