--e

,----- :.,.
-h
..
--+-
-.-
*
..
&
s
. . --. -- -. -.......-- .. . .. .... ,. . . .. .. .. . -
L. --i.:
-, ---
-.
.. W. AD?-lSORY C(NIXiHTiL2
~..,-
FORAEROPNJrIICS
OAR
TECHNICAL NOTE 3781
TECHNI CAL LI B
AFL2291
=~- I- BUCKIJNG OF FLAT PLATES
q -e -~d andHerbertBecker
Xew YarkUniversity
. .
?*.
.
Wmngtoil
my 1957
.
.,
.
.
..
. . . --
G
---i- - i
----
MY
I
.
I
.- -1
TECHLIBRARY KAFB, NM
llllllllll]llll[lkll~ll! E
IIOLLL+
S!!m.lw.-. -.a . . . . . . . . . . . . . . . .. . . . . . . . . 1
,~
1N3%C)IXJCTIOH... . . . . . . . .
G .
S!mKKIJJ. .. . . . . . . . . . . .
.
*..
13ASZ2.PRIIJC?XL =....... . .
G
GeneralRemarks. . . . . . . . .
F@.ilLbriun Di.t<=z~alEquation
*_~lntegrak . . . . . . . .
dilutions . . . . . . . . . . . .
. . . . . . ----- . . . . . 1
. . ...* .: ---- -----
3
. . . . . . ------ ----
7
. . . . . . .--. * . ---- 7
. . . . . . . s.-.....- -..--
i
. . . . . . . . .. . . . G *.. !3
. . . . . . .- .?... . ..- 10
~coxmmcxs . . . . . . . . . . . . . . . . . . . . . ..*I2
lkthe=timl~sis . . . . . . . . . . . . . . . . . . . .. -12
Jhticlastic e?lzz-ture . . . . . . . . . . . . . . . . . . . . ..l~
SZEESS-STRA12T ~IGlii311 YIELDRBHON . . . . . . . . . . . . . . s
5hree-Parameter kcriptionofStress-Strain Cum=5=5. . . . . . . K
inelastic lkdti. . . . . . . . . . . . . . . . . . . . . .*. 17
me3.astic Poi=$=*sRatioG . . . . . . . . . . . . . . . . ...17
~fx!rY-Rxmnz%aFxxmS . . . ..*...-.. . . . . . ..+19
2Mastic-EwI&Lqj-Stress Equation. . . . . . . . . . . . . . - 19
J3mrparisoa of.55kories and&pertiintal &ta . . . . . . . . . . 2Q
Assumptions of ZEelastic-BuckMng Theories. . . . . . . . - . . 20
Inelastic-Bu_ !X%eo5es . . . . . . . . . . . . . . . . ...23
FactorsUsedimC!mqwtations. . . . . . . . . . . . . . . . ..24
CcLnstmction of 3hndizensionlQ Euckling-s . . . . . . . . . 25
~m-GREMscTJEE EwmRs. . . . . . . . . . . . . . . . . . ..-
33esic Principles. . . . . . . . . . . . . . . . . . . . . .. --
lkrivation ofW Stress-Strain Curve . . . . . . . . . ..-.=
#2cmpzmL5un ofmeQrysmd Eqerhent . . . . . . - . -......27
kivations Of~fied c~g Reduction ~CtC2ZZZ * a e - a.- 28
I!sse 1. IOng~Iy supporte~ pletesincompr~~~ . . . . - 26
Csse2. Pkteaums . . . . . . . . . . . . . . . . . . ..2g
Case3. Longs&k@y supporte~ platesinshear - . . . . . . . 2g
~ QFmAT E=2!2K?I=ULAR -Pup UN!ERccMmEssm= X&ms . . . . 31
3fstorical E.se~ . . . . . . . . . . . . . . . . . . . . . .31
~cal Values=f C=pressive-13uckMng Coeffic =- =ur
Ehtes . . . . . . . . . . . . . . . . . . . . . . . . . . . . -
. -- -- --- .
... ,
.
.,
1.
r.-
......%1.6;-,
n
. .-:.
l?.-:.:
Sup~rtedPlate,E~cesElasiicQ>-?==T==.-==== &=illst
Rotation. . . . . . . . . ------------- -- =-
32
PlatesWithUnequalFklge RoLLL ZZ-L ~ . --- - . - . . - . . . - 33 ......
Supported klanges WithElastic:~-====..%===-=int. . . . . . . 33
EffectofIateral Restminton ~~? --~- - - . . . . . . . . . . 3%
EUCKLIWJ OFFiiTRECTAXXJLAR PIJ== ===== ==== WADS G . = . - . - ~~
Historical Ihckground . . . . ------------ ---- =35
Symmetric andAntisyrmetric M&-= . . - . . . . - . . . . - - - . 3%5
Eumerical ValuesofShear-~cUq 2==--=:- - - - - - - = - -
EffectofPlateLengthonBuckltLK,i&kfEZL*2s?-. . . . . . . . . 36
KXKGINGOFFIATFWTANGULMPLATESEZZ=T LSXJ=XG 3JMDS.. - - - - 36
Historical l?ackgroun d... ----------- .-- ~----%
Numsrical Vah~esofBending-Ewxz%l$-~ _Z~
;&s*- e-.--*. -
36
HJCHJN3OFPhil RECTMGULAR PrA!i55 Q3KK==. --* ~--?mms------ 37
~ne~~~c~o~a.... . . . . . . . .. . .. . --.....37
Biaxial .Co-=pression . . . . ------------ ------ t--3T
Shearan&HoxmalS+xess.. . . . . . . . . . . . . . . . . . . . %
BendingandNormalStress.----------- -..--..-39
Bending,and ShearStress.. -------------- -.=---$x
Ber~ing, Shear,andTransverse
-.***. ---- kl
Longittiinal Benin, Longitud+= S.-==-, and
Transverse Coxzpression. . . . . . . . . . . . . . . . . ...42
Combined ~elasticStresses . ---------- . . . . . ...&
EETK?l? OFPRESSURE ONKHKDG = E%====== .-IT-T= - - - . - &3
?kmgeofPublished Results.. . . . . . . . . . . . . . . ...~~
Longitudiral%f Compressed Iaw .~!=
~Plates . . . . .
Lmgitudini CoqressedLcmz~ +.iz.%~. . . . . . . . ..*
.
SPECIALC AS:3 . . . . . ...* -,-----------*- -- -*
UseofElastic-Eucklin6-%ress ~
%*..=-.Z-. -. . . ... &
AxiallyCoqressedPlateWithW=-------.-~~ and!IMclmess. . #
AxiallyCompressed P:.lte Withm-i=+- 2 =3 Cond=nt
l%ickne.= . . . ..-. ---------- --------o-
Par&Llelogmm Panels4LnCmp~-- ------------- .-
:;
ParaUelogram Plates.m. . ------------ .--- =-~
Triangular Plates. . . . . . -- a --------- -----.==
APPEWDIX A -APPlzT4TIoNsEtmmN --. --- m.---. . .. . . . . ...=
Intr&.uction =...... . - j----------- --------~
Physical Properties ofMaterial=. .... -.- . . . - . . . . . . . @
Compressive Eucklinc. . . . . . ---------< -..---~~
-,
.
. ----- . ..-- . .---. ... .
.-. .
.,
.. . .. ..-
PIRteE. . . ...
FlsrAes. . . . .
PlateColuzms. .
9xw.rB~ckli ng . .
*
%in~ Buckling.
_
m
C!cmbir~& Lading .
RL3RJCH
.
TEtiAEs*.
FKXZELZS . .
.
.
.
. .
. .
. .
. .
.-
. .
.
.
.
.
.
.
.
.
s
.
.
.

.
.
.
.
.
.
.
.
.
.
.
-
.
-
.
--
.--, .- .,
.
.
.
.
-,
.
.
,.
.
.
.
-.. .
-.. .
-.. .
-.. .
.-.
----
-.. .
*-. .
. -*-
.
.
.
G

. . . .
-.. .
. . . .
. . . .
. ..-
. ..-
----
----
**. -
.
.
.
.
.
.
-
.
.
.
.
.
. .
--..:-
-=
.-
&-
--
1
1
I
1
.
.-
.
.. . . t. -. .. . , ..... - .-
A
. ,. ..;.. ,
.. . . . .... . . .. . . . ....
.,-, .1.- . . :. -. .: :.,. ., ., ,,, ..:c-:r - -
,.
-. .--
T~A:~[!i ~~~ ~{~~ ~fij~
HN21PC)OK OFoSINJCZIJLIL SYYLKCLIiY
PART1 - PLJCKLIRG OFFIATPL4TSS
~ GeorgzGerarda&.EerLert Becker
SUMMA.RY
Thevarious factors governinz bucklirg offlatplatesarecritically
retie-~ed andtheresults aresumarizedina coqrehensive series ofcharts
acd+&bles.Numerical valuesarepresented forbuckling coefficients of
flatplateswithvariousboundary conditions andapplied loadings.The
effects ofplasticity areinco~ratedinnondimensional buckling charts
utilizing thethree-parazeter description ofstress-strain curves.
INTRODUCTION
His Ikmdbook ofStmcturalStability presents a rathercoqreken-
siveretiewandco.~ilation oftheories &ail experi=iental datarelsting t-o
thebuckling arid failure ofplateelements encountered intheairf~.
~ =et theanticipateii ceedsofthosewhowouldusethisrew.e~:aadcorri-
pilation, itappeared best. tQadopta handbook styleofpresentation.
Thersterial isnotintended asa textbook inwhichtF&ez@ELsls isoften
onther~themtical develowent ofdif~erent typesofrelated problenz. r
Neither isitintended to>orpetetiththef~ar aircraft-c&pany
$
structures mcualswhichgenerdJypresent designinfo?nr~tion, empirical
.-
7
data,andmetkais ofextending results beyondthescopeoftheoriginal
report.
This handbook attezpts tocoverthegenerally neglected areabetween
thetext-k andthestxwctures ranual.Noattempt istie topresentan
exhaustive coverage ofmathematical techniques whichareofgreatimpor-
tanceInthesolution ofbuckling prablers. This=terialhaBbeenwell
presented inseveral excellent booksandpaperswhichareincluded inthe
reference list.Thesubject ofcolunns is comprehensively treatedin
several booksand,therefore, theinclusion ofsuchmaterial inthis
reviewdidnotappeartobewarranted.
Thispresentation primrilyconstitutes a critical retiewofdevel-
opwsntsconcerning buckling andfailure of,plateele=ntssincethe
early1940s.Thisdatehnebeenselected sincethelastcomprehensive
reviewof thisnature(ref.1)appeared attlwttime.
.

----- -
. _
G
Inthe
thegeneral
raintextofthi~report, thevarious factors appearing in
buckling-stress equation
ucr(cm7==) = qij
ha
(
12
(
))
b
(1;
121-
~e2
arecritically examined fromthes~.dpoint oftheirtheoretical develop-
mentanQtheagreement of-theory withtestdaa.
Intbesection entitled EasicPrinciples a briefreviewofthe
Lasicxmthematical principles involved insolution ofbuckling problem
lagiven.Theprbaryobjecti--e inpresenting thismaterial is to
acquahtthereaderwiththeapproximate methods usedinordertobe
abletaindicate theaccuracy oftkeresults cfparticular solutions
discussed insubsequent sections.
Tnthesection entitled %uudaryConditions theinilu=.ce ofthe
geometric &nm&ry conditio~s upontkebuckling stresstsdis~sed et
somelength.Itisindicated thattheuseofa freeunloaded edgeina
plateinvolves Poissons ratiointhecompressive b??ckling coefficient.
Assanexample, thebuckling coefiicients forplatecolu!ms, flanges, and
simplysupported platesaredetermice5 fromtheorytodezmnstrate the
effectafvarious boundary conditions uponthebehatior ofsuchelements.
Also,theth&e-parar&er methodofrmthe-~tically describing stress-
strainrelations ispresented inanintroductory mannerinthesection
entitled Stress-Strain Relations in.theYieldRegion.Useofthis
methdaffords a considerable simplification inthepresentation of
results ofineLsticbucking theories.
Theeffects ofexceeding theproportional limitofa rmterial are
incorporated ina plasticity-reduction factorq. Because ofthenri-
ouatheories thathavekeenrecently advanced together withthefact
thatnoonepublication hasreviewed theconflicting assumptions of .
!
*
.
.- -... I
-..-
-.
,,. .,-:p . . . . .
... ..- f:.
J
k;
..
.
Theeffectofcladdin& u~n theL.ickling stressof flatplaheshas
keentreate~ hyanextension ofinelastic-buckling theory.Intl:e sec-
tionentitled Cladding Reduction Factors a simplified treat..mnt of
buckling ofcladplatesispresented iriwhichvaluesforthecladding
Correction factor~ arederived.
Thebaclc~~?dfordeterringtheelastic-kuckllng coefficient k
haskeenwelldocumented. Tkerefore, thelastsections ercconcerfied
withthebuckling coefficients fora largenumberof cases.Thepresen-
tationconsi6ts, fortt.e ?m6tpart,ofa straightforward cataloging of
results intheformofbuckling-coefficient charts.
Theappendix hasbeenorganized fcrunimpeded useinanalysis and
designandforthisreasonROreferences appearinthisportion ofthe
repOrt.Ther e fe r e n c e s are examined indetailinthepertinent. part of
themaintext.Theliterature isreviewed anddiscussed bothas tocon-
teutandapplication totheparticular problem.Experimental evidence
Ispresented whereittend6toSubstantiate onethecryamng several
whichmayhavebeenadvanced ona particular phaseofthebuckling prob-
lem;plasticity-reduction factors areperhapsthemastconspicuous e~m-
pleof this.Thus,therecozzzendation fora particular theoqyisgen-
e- supported byexperimental data.
The=& textalso contains somenewmterialdeveloped duringthe
courseofthiscompilation. Although suchmaterial isimportant tothe
unification of priorresults, ithasnotbeenconsidered ofsufficient
consequence tomeritseparate publication. Therefore, wkensuchmate-
rialdoesappearinthishandbook itisina detailed form.
Thissurveywasconducted u n de r t h e sponsorship andwiththefinan-
cialassistance oftheEational Adviso~Committee forAeronautics.
-IS
.
Ar areaofribcrosssectioa, sqin.
a longdimension ofplate,usually unloaded edgefnuniaxial
compression, in.
.. . .
b shortdimension ofplate,usuallyloadededgeinuniaxial
compression, in.
#
... -
()
. .
..-.
I
]
.
G
D
cl. ..c4
D
D1
E
platecross-cectioa rigidity,Et3/12(1 - V2),lb-h.
plastic platecross-section rigidity,
/
E6t39,lb-in.
Youngcdulus, psi
secentmdulus, O/c
tangent mdulus,du/de
secantandtangent rzodulus forcladplates, respectively
ratiooftotalclaading thickness tototalplatethickness
shearnmdulus
cmuient ofinertia
3 = (%/E) (1- Vez)(1- Vq
K
k
L
M
N
n.
P
P
mdifiedbuckling coefficient, kx2@(l - V2)
buckling coefficie~t
lengthofplate,in.
bending momentapplied inplaneofplate,in-lb
axialload,lb/in.
numberoflongitudinal halfwavesinbuckled plate;also,
shapeparameter forstress-strain curve
.
normalloadappliei inpla n e ofplate,atb,lb
nornulpressure, psi
4
!
b
i
i
;*
e
d
.
}
.. - ... . .
. --- -. . .-J. .....
-,-T-
. .
, . .
!
i-,
,.
. .
-;
i;. G-- veo+\)~
~
shesrlouding, lb/in.
6=
E2+ ve(flb/A)2
R strezs ratio
t thickness ofplate,W.
u=
(
- ks
)/(
%+ -
+ ks
%+ -
)
w
w
X,y,z
y.l+3pf
a
potential energy,in-lb
displacement nomal toplaneofplate,in.
coordinates
edgeangle,deg;also,12M/(Pb + 6M)
& =fi(b/A)L/2
*
$
also,loadingratioforplatewithvarying axialbad,
Maxhmlload/14imim71 lead
ratioofcla~ding yieldstresstocore stress,
cl/acore$
Y
shearstrain
normalstrain; also,ratioofrotational rigidity
e~gesti?fener toro~=tional rigidity ofplate
..
plasticity-reduction factor
cladding reduction factor
total-reduction factor,qfi
bucklehalfwavele n gt h , in.
ofplate
v inelastic Poissons l%tiOj V = Vp-
~p-Je)(%/E) for
orthotropic solids

.
-.

.
.- .
,--

,
,.. ,
........ ....
~o.7JJo.85
T.
~
Subscripts :
A,B
av
b
c
c1
cr
e
P
pl
r
B
Sm
X,y
+.
stressintensity,
(
~x2+
stressatsecantmdulus,
shearstress> psi
angleofdiagonal support
~ 1/2
~y2-
)
axay+ 3T, ,psi
o.p.=d 0.85E,
taplatewidth,
respectively, psi
radians or deg
valuesatstation A andstation B;seefig.30
average
bending
compression
cladding proportional limit
critical orbuckling
elastic
plastic
proportional limit
intraverse ribofcompressed plate
shear
shearoninfinitely longplate
directions ofloading
loadings producing tension
loadings producing compression
.
G
--- ~u:,.:if:c,::s:
. ....
c c h !!. pd
F free
.
G

.+-
.
Ss s iqiy suppurtei (hinged)
Insketches acccnyanying figures, supported edgeswithelustic rota-
tiofial restraint areshwn shaded.Unshaded loadededgesaresimply
sup~orted. Unshaded unloaded edgesarefree.
BASICPRINCIPLES
General Renarks
Thetheoretical buckling stressofa flatstructural ele~~nt isthe
stressatwhichanexchange d stableequilibrium configurations occurs
tet-~en thestraight and.the slightly bentform.Itmarkstheregionin
whichcontinued application ofloadresults inaccelerated growthof
deflections perpendicular totheplaneofthe~:ate.Itsimportance lies
inthefactthatkuckling initiates thephysi~-~.l processes whichleadto
eventual failure oftheplate.
74<
*..
Themathe.=tical solution ofparticular kuckling problems reqtires
thatequilibrium andboundary conditions besatisfied. Thiscanbe
accozplishe~ byintegration oftheequilibrium partial differential equa-
tionoftheflatplateorbyuseofmathematical n&thods whichmaynot
completely satis&theboundary orequilibrium conditions. Theformer
solutions areexactwhereas themethods basedgenerally onecer~inte-
gralsareapproximate although usually veryaccurate.Theneedfor
approximate rithods arisesfromthefactthatexactsolutions canbe
foundforonlyalimited numberofbuckling problems ofpractical
importance.
.
/
t
In thissection, a briefoutline ofthemethods ofanalysis of
bu c k li n g problems ispresented. Forextensive discussions ofthevari-
ousmethods ofanalysis andtheirapplication toa widevariety ofprob-
lems,reference tothelooksofTimshenko, Sokolrdkoff, andDleich
(refs. 2tok) issuggested.
Equilibrium Differential Equation
Thegenemlformofthedifferential equation describing theslightly
bentequilibrium conflgumtion ofaninitially flatplatewasderived by
StQvell inthefollowing form(ref.5]:
.
. . --. ,--.
,..
.,. ..
I
.
.

.
.
.
,
I
*
inwhichtheconstants aredefined as:
cl= 1- (3/4) (a+=,)~ -..(%/s]
( / 2) ~- ( %, s]
C2= 3u~Tai
C=j=l- (3/4)
py-)k
( I)p(a
C4= 3ay7ai
(w ]
-
s

(2)
(3)
t
Thesedefinitions oftheconstmtsare&sed ontheassumption thatno
elastic unloading occursduringthebuckling process.
Furthermore, a
valueofMissonsratioeqwl to112was assured forboththeelastic
andinelastic rangee.
In the elastic range,%.~% = 1,and,therefore, fora32loadings
C1.C3=C5=I and C2= C4= O,andequation (2)reduces tothe
familiar equilibria equation fortheelastic case:
&=&+2&+3
ax4
.
.
(4)
. ..._. ____
-
_. ._. _ .,.
~. ~ .. . .. .- -. .---- ..
t
.=
.
d
.:.
G
.-
J
.
8 krdingrigidity of D= l?L3/9, whereasthe elastic valueis
D = Et3/12(1 - Ve2).
Thesolution ofindividual buckllng prohle=s canbemostreadily
handled byselection ofappropriate solutions ofequation (2),insertion
of properboundary conditions, andminimization to obtainthebuckling
stress.h thisconnection, thebuckling stresses fors@ly supported
pletecolumns, compressed flanges, andplatesareconsidered insome
detailinthesection entitled Boundary Conditions toillustrate the
differences inb~ckling behatior oftkse structural elezients.
KnergyIntegrals
Sinceexactsolutions toequations (2)and(4)canbefoundfor
onlya llmited numberofbucklidproblems ofpractical importance
approximate solutions genemllyutilizing ener~integrals havefound
tideapplication.
Thepotential energy of theplateanditsloading systemisrepre-
sentedbythedifference oftwointegrals. Thefirstintegr~ofequa-
tion(5)represents theincrease instrainenergyduetabending and
tuisting oftheplatedurir~thebuckling process, whereas thesecond
integral represents energyassociated with rrembranestresses resulting
fmm lateral deflection. Ifthepkte edgesarefixedduringbuckling,
thelatterrermesents themembrane ener~. Iftheedgesexperience a
relati.~ shif~,thesecondintegral rep~esents thewo=k
loading system. t
Thegeneral energyintegral forpkteswithsimply
wasderived byStowell(ref.5)fortheinelastic case:
oftheexternal
. .
supported edges
1
+
(5)
.
,. ..~
-.. .. ---
. . . . ..
~--- --
#--
,
,
I
restraints ofmagnitudec
enerminthesercstrnintm
theform
along t!.o edgesofth!plate, tlien tt~e strnin
isaddedtoequation (5).Thesetermshave
oJ \ QY/y=yo
wherey. Istheedgecoordinate.
Fortheelastic case,equation (5)c-k simplified to
dxdy-
(6)
Soluttons
h principle, ofaU thedeflection functions satisfying thegeo-
metricboundary conditions of theproblem, thepotential ener~ AM will
be zeroforthatfunction whichalsosatisfies theequilibrium differen-
tialequation.Thisfunction wouldbeanexactsolution oftheproblem.
Sinceexactsolutions cank-efoundinonlya lfmited numberofcases,
theener~integrals areofgreatusefulness infinding approximate solu-
tionswhichsatisfy thegeometric bcmndsry conditions exactly andthe
differential equation approximately. Thus,oftheseveral ~.ctions
satisfying thegeometric boundary cotiitions butnotnecessarily thedif-
ferential equation, thefunction forwhichtheener~integral Isa mini-
xmnnconstitutes tkebeetapproximate solution ofthedifferential equation.
Probably thebestlmmunenergymethodfordetermining thebuckling
stressofthinplatesistheRayleigh-Ritz procedure. Themethodcon-
sistsofthefo~ouingsteps:
. (1]Thedeflection surface ofthebuckled plateisexpressed in
expended fom asthesumofaninfinite setoffunctions havingundeter-
minedcoefficients. Ingeneral, eachtermoftheexpmsionmst satisfy
thegeometrical boundary conditions oftheproblem.
,- .- - . - -
.
.
.
z
~
G
,
i
f
(3)Thisminimizing procedure leadsto
equations intheundetcr.i~c2 coefficients.
vanishing solutions onlyifthedeteralr%nt
Thevanishing ofthisstability determintint
rnybesolvedfortheb!!ckling stress.
a set oflineu.r horzogeneous
Theseequations hRvenoll-
oftheircoefficient vanishes.
provides theequation that
. .
2
.
4
.
Whentkesetoffuctionsuse~isa complete setcapabls of repre-
senting thedeflection, slope,arid curvature ofanypossible platedefor-
mation,thesolution obtained is,inprinciple, exact.Since,however,
theexact. stability determlnant isusuRlly infinite, a finitedeterminant
yielding approximate resultsisusedinstead.
Thebuckling stresses obtained bytheapproximate methodarealways
higherthantkeexactsolution although theynaybeveryaccurate.This
is8 resultofthefactthatthedeflection function approximates the
truebuckleshapeandtherefore the~tentialeneraresulting fromuse
oftheapprox~tingfunction isgreater thanzero.Ifthedeflection
fumctian isthetrueone,thenanexactsolution tothedifferential
equation isobtained.
Ifa deflection function ischosenwhichsatisfies thegeometrical
boundary conditions approxtitely, itispossible toobtaintuckling
stresses whichapproach theexactsolution fromthelowerside.This
canbeaccomplished bya revision oftheRayleigh-Ritz procedure known
astheLagrangian multiplier method.
TheIagmngianmultiplier methodfollows thegeneral procedure out-
linedfortheRayleigh-Ritz methodwithbutonesignificant cluznge. The
restriction instep(1)thatthekaundary conditions besatisfied by
everytermoftheexpmsionisdiscafied andisreplaced bythecondition,
thattheexpansion asa wholesatisfies theboundsrj conditions. This
condition ismathematically satisfied instep(2),duringtheminimization
process, bytheuseofIagrangian .titipllers.
Theadvantage of theI.agrangian multiplier methodliesin thefact
that,withtherejection of thenecessity ofthefulfillment ofboundnv
conditions termbyterm,thechoiceofanexpansion ismuchlessrestricted-
Forexample, intheclamped-plate compression problem, a simpleFourier
expansion naykeusedinstead ofthecomplicated functions usuallyassured
in theRayleigh-Ritz analyses ofthisproblem.Ftihermore, theorthogo-
nalityproperties ofthesimpleFourier expansion leadtoenergyexpres-
sionsofa simplicity that1sinstnment.al inpermitting accurate
computations. c
.

.-~ -r-~---- ..
,.. .,
... ,,
,
.
-.
.
d
.
ThLslWi!IOi &Itdi.tS fi~~IliCZLiO:l iOSpectticprO*L~VIG isU,:..: rit;wl
byh.dimlsky WA HU (Wr.6). l%cyhivetreatei~kckgranl~immul. tl. -
plierr=tlxxl ina namerinwhichitispossible toobtuinupproxIz&Le
solutions foriothupperandloxerbourds.Asdewrmin:~nts oflli~!ler
orderarz usedtoo!.:.ain betterapproxlmitioas, taththeup>erandl~wcr
boundsapproach thetme buckling strcsc.Thu6,theIagrangian mlti-
pliermethodraybeusedtoohtsd.n results tithir. anydesired degreeof
accuracy.
Inaddition totheaboveprocedures whichsrebasedonenergyir.tc-
grals,othermethods of obtaining approximate solutions ofbuckling prob-
lemshavebeenuse~whichinvolve theequilibrium differential equation.
~CtiOnSwhich&&:l sfythegeorr~trical tiundary conditions exactly are
usedtosat~sfy t~.e governing differential equation approxkately by
processes thatleadtointegration ofthesefucctions. Galerklns =thod,
finite-difference e.~uations, relaxation techniques, anditeration aresome
ofthenumerical nethods thatcanbeused.
KUNDARYCONDITIONS
Thenatureofthebucklerattern ina relate dependsnotonlyupon
thetypeofapplied loading butalsouponthe=%er in whichtheeds=s
aresupported. Thisisillustrated infigure1 inwhichthes&me-al
compressive loading isseen to generate threetypesofbucklepatterns
ona longrectangular platewithdifferent geo~trical boucdary condi-
tions.Thesinglewaveisrepresentative of c olu zn behatior, thetwisted
waveis representative offlangebehatior, andtkemultiple-buckle pattern
58representative ofplatetehatior.
N indicate them=nne=in whichtine geometric boundary condltio-=
mathe-tically influence thebuckling behavior an~alsotodamnstrate
thesolution oftheequilibrium differential equation (eq.(4))forsome
particular cases,theplatesshowninfigure1 cueanalyzed.I!o-..f
conditions vhichc~cterize simplysqrporte~ vi~ecolumns, ~ges,
andplatesareconsidered.
G
Wthemtical&ialysis
Theequilibrium differential equation forelastic buckling ofa
unlaxially compressed platecanbecbtained fromequation (4)inthe
form
(7)
-.. -.. . . . .. .. . . .- .. . --- -. - . .- . -. .

.
. .
.
.
.
*
. .
;
I
-,, ,,. .J. ;
....!
T!, isES.5UT.*2,3
.&Ild ~h~refore
w=
(
c1
uhere
.
(9)
(lo)
(m
Thecoefficients cl to C4 areto be dete.nzined bythegeometrical
bmndaryconditions alongthe&nloadei edgesoftheplate.
Fortiewidecolumn, tkeunloaded edgeslocated at y = tb/2 are
free,andconsequently theedgeEmQent6 andre-iuced she?~s must be zero.
Therefore,
.
Ii
A
~ + 2 (1
1
-Ve)zz- *O
ax~ytib/2
(1)
Fortheflange, theunloaded edgeat y = O isassumed tohesimply
supported andthatat y G b isfree:
.
.
. . . .
. .. . . .
.
. .
.
.
.
d
. .
A:-
(W)y=() =o
(=+veG),=o,b=o
&+2~-e)-]y=b=0
.,.,..),;:;;,:.
.
Theplateisassumed tobesiMPlysuppmtedalongtheunloaded ed~es

located ai y = ~b/2:
Incorporation of
byequation (8)leads
(w)y~b/2 =o
()
a%
+,,22
* & ~2 ~*~/2 = 0
theseboundary conditions
tothefollowing implicit
.
intothesolution given
expressions for kc.
Forthecolumn,
1
. for the flange,
1
#p sinh&cosp - ~%coshtisln~= O (16)
andfortheplate
.
[it a n h {~/ 2 ) +j3 tan(5/2~-1=O (17)
where
#-
5=&2
- ve(xb/A)2
.
,-. .-. . _
----- .-._
!.,
~lebuckling coefficient fora sixplysupported flangewasderived
~,yt,w@@standStowell
(ref.8)intheform
{ - e)i[fib/A)2/6]}
kc= (6/Lr2) (1

(18)
kc=0.83 -0.93ve+L34(A/fib)2+ 0.10(xb/A)2
(19)
Forthe s*W supported plate
kc= [m+(b/AjJ2
(a
Anticlastic-titure
ASmy beseenfromthesolutions inthepreceding section, the
,,U:kling coefficient for
thesimplysupported platedepends upononly
~d 1sindependent ofRissontsratio,whilethecoefficients for
J/Adecolumnandflmgearefunctions ofboth Ve
the@
and b/L ThiS
isnotlimited totbecaseof6fmplesupport alonebutper-
~5tufition
~tl,vtoa~ degreeofrotational restraint alongtheW-A ed=s of
Theinfluence of v= upon ~
istraceable tothereduced-
~ ~lfitec
atthefreeedgesofflanges andcolumns.
~~lemr te-
Boundary condi-
til)rl~
suchassimplesupport donotImposetherequirement ofzero
~ettuc~ she= ~ong theunloaded edges
, whicheliminates the Ve
influ-
,ntofromtherelationship for kc.
me valueofthecompressive buckling coefficient foranelement
unloaded edgedepends uponthedegreeofanticlastlc
~mtdinin6a free
developed. Fora verynarrowelement suchasa be&m,complete
~~lrV/LtUL~
.
,---
.
I
.
.
.-
.
.
m
. -
enbiclastlc curutureoccurs anithelending rigiditiy iuslc.ply EI. ?01
a relatively vitie Btrip,t?ma:iticlaztic CU-JatIJre is suppressed sot?.t~t
.

thecrosssectiofi reraizs relntlve>y flatexceptfora highlylaculize;

Curling atthefreeeigesvkre the~tressdistribution rermrapges itself
tosatisfy thegeometrical bow-simy comiibions. Therestraint ofant.i-
cksticcurntureres-dts inm increase inbending stiffIMss.

Fora
verywideele~ent, tkek.::.<irg stiff~ess approaches EI/(1- v2);this
limiting condition is~.r-~:: ascylindrical kending.
case
This
tion
Platecolumns andflanges my oft-en herelatively narrow, inwhich
thebending stiffness liesLctween thelimiting valuesdiscbssed.
effectcanteacccunted forbyuseoffigure2.

STRESS-STRAIN R!21ATIONS INYIELDRIGION

Three-tineter Description ofStress-Stmin Curves
Stress-strain cumesareoffundanental iqcmtanceinthecosrputa-
ofinelastic buckling stresses. Thenumberofdesirechartsrewired
forthemny materials a~-ilable andthevarious allowabie stresses ~or
thesemterialsatnomalandelevate~ temperatures canbetrewmdous;ly
reduced byuseofa nondimensional mathematical description ofstress-
strainrelations.
Ramberg andOsgod(ref.9)haveproposed a three-pszaneter repre-
sentation ofstress-strain relations intheyieldregionwhichFRSfound
wideapplication. TY.eir e~uation specifies thestress-strain curveky
theuseofthreepar==ters:ThemduluEofelasticity E,thesecant
fieldstressCO-7 corres~nding totheintersection ofthestress-
stralncurveanda secantof0.7E,andtheshapeparameter n which
describes thecwxa:uzze ofthekneeofthestress-strain curve.The
ebapepa.raceter isa function of
Jo.7~d ~0.85Jthelatterstress
corresponding toa secantof0.55Easshowninfigure3(a).Theshape
~ter n ispresentti infigure3(b)asa function oftheratio
/
0.70.85
Thethree-paraceter methodisbase~ontheexperimen~l observation
thatformny mterialsa simple ~wer lawdescribes therelation between
#
theplastic andelastic components ofstrain.Eyuseofthis fact,tke
following nondimensional equation canbederived:
Ee=
()
n
=0.7
+2A
e 7U0.7
(a)
.
.
i
!
,
Thequalities
-/
M l?~.y
/
an: u C(-J,7 tire nond~rwmio::al andccmsc~uefiLly
tk nondimensional stress-struin curvesshorninfigureh canLeplottei.
Therefore, thestress-strain curwxofranyffi%erials m y be foundwith
theaidoffigy?eh providing E, n,
cificmterials.
Inelastic
Forinelastic-tickling problems,
and Et/Esappear.Theseratioscan
byuseofequation (21).SinceES=
tion(21)that
ahd
=0.7
areknownforthespe-
Moduli
tile modulus ratiosEe/E,Et/E,
beco~~ted innondimensional form
u/c,itfollows directly fromequa-
~~ = 1+ (3/7) @o.7)n-1
Since~ = du)de,differentiation ofequation (21)leadstothe
expression *
E~ =
.
Fronequations (22)and(23)
Et/Es=
Thesequantities
inelastic buckling.
it follows that
( %%) / ( %)
1+ (3m\u/%.7)n-l
~+ (3/7) n(u/uo.~)n-1
areusedinsubsequent seetions
(22)
(2 3 )
(24)
concem,ed with
Inelastic Poissons Ratio
Poissens ratioforengineering mterialsusually hasa valuein
theelastic regionofbetween 1/4and1/3and,ontheassumption ofa
plastically irlcompressible i$otropic solid,assures a valueof1/2in
theplastic region.Thetransition fromtheelastic totheplastic value
isnxxtpronounced intheyieldregionofthestress-strain curve.Since
G
.
.---. .- .- ..
.
.
.,
-
*
.-
*
isofsorei:uport:mce inir.elastic-iucklit:~ proLlc%s.
GerardandWildhorn, w..;::K ot~.ers, havestudieJ thisproblem on
several alu~dtuum alloysandhuveshornthatPoissonc ratioisseriously
affected bymisotmpyofthematerial (ref.10). ForrmterklsUMCII
cankeconsidered toteorthotropic (e.g., hatingthesmc properties
alongthey- .?U_d z-~e6iflotied810ngtkle x-is) the rolbW@ rehtion
describes thetransition Intheyieldregion:
v
= P - @sIE)~p -e)
(&j)
ti thisrelation,
P
isthefullyplastic valueofPoissons ratio.
For isotropic rmterialsVp= 1 / 2 , u k r e a s fororthotropic materialsVp
isgenerally different froma valueof1/2.
Itisevident from thebuckling stressexpression thattworaterials
whichdifferonlyintheirvaluesofPoissons ratioshouldhavedifferent
buckling stresses. Asa rule,however, thevalueof v= istirtually
constant fora material whoseproperties m3ychangeas a resultofheat
treatment, details ofcomposition, orarcou.nt ofcold-work.
The USUEL1 mge of ye for EOst technically inportant structural
mterialsisI#sween 0 . 2 5 and0.35.~Lere are exceptions, however.One
ofthemostextrer~ ~terialsIsberyllium, forwhichUdy,Shav,8nd
Wulgerreportatiue of0.02(ref.n).
Intheinelastic range,presmblybecause ofanisotropy, numerical
valuesof v havebeenfoundwhichme considerably inexcessofthe
theoretical upperlimitof0.5,whichisderived on tke assumption of
incompressibility of- isotropic =terial.Forexa@e, Gerardand
Wildhorn obtained valuesof v asImge as0.70forseveral high-strength
aluminum alloys (ref.10),vhile~n andRussell reported a value
of0.77forcanzercially puretitanium sheetand0.62forFS-lh=gnesium
alloy(ref.12). Stang,Greenspan, andNewcanalsoobtained dataatvar-
iancewiththetheoretical valueof0.5forplastic strains(ref.13).
Thesethreer epor t s cover a large variety ofalloys, deformed byvarious
total. strains inbothbarandsheetstock,andshouldbeconsulted for
morecomplete data.
.
0-
.
~-
- . . ._
-
-...-- . . . .,__ ___ .
.-. ..-- ----
. - .. .- --- ~. . ~.
.-
!
I
,
I
I
I
.
t
,
.
,.
.
. .
I
/
0-
Theelastic
expressed inthe
PLASTICITY-RZHKTICW FACTURS
Inelastic-Buckling-St .-ess EquatSon
buckltng stressofa flatrectangular
form
platecanbe
~2E ~ 2
Cre=
)()
12(1 2 b - Ve
Whenthebuckling stressexceeds theproportional limitoftheplate
(26)
material, thetermsinequation (26)whichareinfluenced are
~, E,
and v. Thebuckling coefficient k depends uponthetypeofloading,
thebucklewavelengthasaffected bythegeometrical features ofbound-
aryconditions andaspectratio,thestresslevel,andPoissons ratio
inthecaseofplateswithfreeedges.Theelastic MUIUS E iS altered
by thereduction intending stiffness associated withinelastic behatior.
Poissons ratiointheyieldregionexhibits a gradual transition from
theelastic valueVe to8 valueof1/2for8 plastically incompressible
isotropic material.
G
Forsimplicity ofcalculation alleffec ts ofexceeding thepropor-
tionallimit-are .&nera12y incorporated in
toastheplasticity-reduction factorq.
~ = aq~cre
a singlecoefficient referred
Bydefinition
(27)
,Substituting equation (27)intoequation (26),
(28)
Sinceq =1 Inthe elastic range,equation (28)isperfectly general
anditisnotnecesssxy todistinguish between elastic andplastic
buckling . !Ihe valuesof k @ ve arealwaystheelastic values
sincethecoefficient ~ contains allchanges inthosetermsresulting
frominelastic behatior.
.
I
._ ___ _ . -
-
3J
* ,,. .\ f.,..
IL... i.J:, ;. :
.-
,
.
.-
.
s
G
-
.
, ? .
Thetheoretical andexperimental determinations ofthevnluesof q
appropriate tov.trious typesofloadings andtiund[lry conditioner lIQVC
reaulte~ i3extensive literatu~e. Theassumptions underlying thev%riau$
theories differwithrespect toplaatlclty laws,Btress-strain relaticws,
and.buckling mdelsused. Inordert~avoidpossible confusion indis-
cussing thevarioustheories, itappears desirable toresorttotke
expedient ofcomparing theories withtestdatafirst.
Ratherpreciseexperimental dataexistforplastic buckling of
colurms, sinplysu~ortedflanges andplatesundercompressive loads,
andelastically sup~rt.ed platesundershearloads.Forpractical
alundnum-a~oy columsundercompression, itisa well-known facttk=t
theexperimental failing stressisclosely approximated bytheEuler
formula withthetangent nmdulus substituted fortheelastic modulus.
Infigure5,testdataforbuckling ofsl.mply supported flanges
undercompression areshownincomparison withthetheoretical values
asderived byStowell(ref.14)according tothemethodofGerard
(ref.15). Excellent agreent isobtained.
Infigure6,test dataofPrideandHeimerl(ref.16]andPeters
(ref.17)forplasticbuckling ofsinplysupported platesundercovres-
sionareshowninco,xparison withthetheories ofBijleard (ref.18),
Handelman andPrager(ref.19),H.yushin (ref.20),andStowell(ref.5),
andthe~thodofGerard(ref.15). Pooragreement isobtained betweea
thetestdataandtheflowtheoryofHandehn andPrager, whereasrelat-
ivelygoodagree~ent isobtained forthedeformation theories oftke
othqrswithStowells theoryinbestagreer~ct.
Infigure7,testdataforplastic tickling ofelastically supported
platesundersheerareshownincomparison withthetheories ofBi~laard
(ref.18),Gerard(ref.21),andStowell(ref.5). Itcanb.cobs_.ed
thatthemethodofGerard, whichisbasedonthemaximum-shear plasticity
lawtotransform anaxial. stress-strain curveintia shearstress-strain
curve,isingoodagreement withtestdataonaludnum alloys.
Onthe-basis oftheagreercnt withtestdata,thevaluesof q
recommended forusewithequation (28)appearintheappendix.Also,
nondimensional buckling chartsderivedthrough theuseofthesereduc-
~tionfactors appearinfi~es 8,9,and10foraxially .compressed
flanges a n d plates and for shear-loaded plates.
Assumptions ofInelastic-Buc?Xng Theories
Thestateofknowledge upto1936concerning inelastic buckling of
platesandshellshasbeens~rfzed byTiimmkenko (ref.2). The=:Q
,
I
i
ei?or<s rc?crt.cl tl:cv-ei~ Jere ca[~c.~~ccd vlthat l~iii~)t . s t -o modifyLk!vari-
ousL-erldiu~-rrfirest tcrl:s of
td::~ e~:l~litriu~, dlfrcrentltil e.~u~tionE byt!.c!
useofsuitak,lc plasticity coefficients determiried frome~.eri.rental date
oncolumns.Alt?.oIIgh suchsenier~frical efforts rfit witha reasonet~le
degreeofsuccess, thetkoretical deterntinatlon ofplasticity-re~uction
factors forflatplateshasbeenechleved withinrecentyearsasthe
resultofthedevelopzxmt ofa Satisfactory inelestic-kuck~ing theory.
Because suchdeveloprxmts arerecentandbecauqe thevarious theories Iave
notbeen,asyet,adequately treated intextbooks,thefollowing dis-
CUS6LOn concerning the.assu~ptions andreSLiltS oftheVEir iOU6 theories is
presented insouedetail.
Mathematical theories ofplasticity arephenomsnological innature
sincesuchtheories generally proceed fromtheexperirr~ntally determined
stress-strain relations forsimpleuniaxial loadings.Intheelastic
range,stressW strainarelinearly related bytheelastic mdulus.
At strains beyondtheproportional lfm~t,a finitestress-strain rela-
tioncanbeusedinthefora
oranincremental relation canbeused
h eitherrelation thesec~tmodulusEs orthetangent tiulus %
varieswithstressandapplies aslongastheloading continues to
increase. Unloading usual.ly occursalonganelastic we parallel to
ttie initial e-tic portion ofthestress-strain curve.
Inthebuckling process, forexample, thestressstateisc&sidera-
blynrmecomplex thansimpleuniaxial loading.Therefore, formulation of
suitable stress-strain lawsforthree-dimensional stressstatesbeyond
theproFortiorsl limitformsoneofthebasicassumptions C*thevarious
phsticitytheories.Ease3ongeneralizations ofequation (3) which
involve finiterelations, deformation t~s ofstress-strain lawshave
beenad-ted. Similar generalizations ofequation (30)involving incre-
mentalrelations arereferred toasflow-type theories.h boththeories,
unloading occurselastically.
Theuseofthevarious plasticity theories isgreatly facilitated
bytheintroduction ofrotationa~y invariant fimctions todefinethe
three-dimensional stressandstrainstates; suchfunctions aretermed
stressandstrainIntensities. Theassumption thatthestressintensity
isa uniquely defined, single-valued function ofthestrainintensity

.. . . . .. . . ... . ___ _. ._. .. .. . . .___

- -
. . .. .-
fora @ven rzterial wt.en thesLressintensity incre%ses (loadin~) ur.i
iselastic Wtenitdecre-fies (unloading) is a secondof thefunda%entsl
hypotkesefi ofplasticity theory.
Thedefinitions .of,the stressw.dstrainintensities tkeoretlcali.v
canLechosenfro.n a f~all%y ofrotationally invariant fi,anctior.s. Two
suchfunctions referre~ toas diemQximLK-shear lawend-octahedral-skar
Iawhavebeenfoundtobeofconsiderable usefuheseforcorrelaLin3
stress dataonductile zterials.Thus,troth of theselaws.h.ave teen
assumedtoapplyinvarious solutions forinelastic buckling.
Inordertoobtainsolutions tovarious plasticity problerrs, adiii-
tionalassurzptions are~-nerally er@oyed.Theseordinarily include the
assun@ionthatthepri~cipal axesofstressandstraincoincide andthe
assunrption ofpksticTsotropy.Wtherrmre,thevariation ofPoissons
ratiofromtheelastic valuetothevalueof0.5fora plastically incom-
pressible, isotropic solidisnmstpronounced intheyieldregion.Some
solutions account forthe instantaneous valueof Poissons ratiowhereas
othersassumea valueof0.5forboththeehsticardplastic region.
The latterissumptioa servestosirplify theanalysis considerably.
Corrections fortheuseofthefullypl.astic value ofPoissons ratio
cangeneraUybeincorporated inthefinalresults.
Alltheforegoing assumptions formthebasisforsolution ofplas-
ticityproblems ingeneti. Forthespecific problea ofinelastic
buckling, Itisnecessary tamakeanadditional assumption concerning
thestressdistribu~idn attheinstant ofbuckling.
Fromthestandpoint ofclassical stability theory, ttte buckling load
istheloadatvhichanexchang@ ofstableequilihriurr configurations
occursbetween thestraight formandthebentform.Sti.ce theloadrezains
conskntduringthisexc.Wnge, a strainreversal mustoccurontheco~vex
sideand,therefore, tkebuckling nndelleading tothereduced-mdulus
consept forcolumns iscorrect theoretically.
%icticalplatesarxi COIUMRS invariabwcontain initial @erect-
ions ofsomesort,and,therefore, axialloading andbending proceed
simultaneously. Inthisczse,thebentformistheonlystableconfi~-
uration.Sincein thepresence ofrelatively largeaxialcompressive
stresses thetending stresses aresmll,no strainreversal occursand
theincremental bending stresses intheinelastic rangearegivenby
equation (30).
G
Sincefailing loadsobtained from testson ahuninum-alloy columns
areclosely approxinuited bytheEulerbuckling equation withthetangent
modulussubstituted fortheelastic modulus, certain of theinelastic-
tuckling theories assumetheno-strain-reversal, or ~.gent-nmdulus,
modelas thebasicbuckling process andthenproceedtosolutions byuse
ofequilibrium equations basedonclassical stability concepts.
------ .. - .. .. --- ,- ~ - ------- - -. ..,

G
::;..-;, x.;f;.: 1.
.
.
.-
,
.
.. .
. .
!
.
Inelusiic- ikcklirrg i-~.eori~s
Different invectlgntcrs hnveusedvarious onesof t},ose assuwLions
di.~cussed above,Inordertoindicate themaJorassuimpkions underlyirl~
eachoftl:e theories, ELsunrary ispresent~a intable1.
Historically, EiJlaard appezrs tohavekeenthefirsttoarrive
at6Qtisfactor~ theoretim~ solutions forinelast.ic-buck~ing theories
(ref. 18).Hisworkisthenestcomprehensive ofallthoseconsidered
inthatheconsiders kth incre.~ntal anddeforrztion theories andcon-
cludesthatthedeformation typeiscorrect sinceitleadstolowerine-
lasticbuckling loadBth&nareobtained fromincrerzntal theories.Eis
workwasfirstpublishe~ in1937.!Ihis pnperancl laterpublications
include solutions to.wnyi.@ortnnt inelastic-buckling problem.Hov-
ever,thisworkappears tohaveremained unknoum tonestofthelater
investigators.
Ilyushin briefly referred toBijkards workandthenproceeded to
derivethebasicdifferential equation forinelastic buckling offlat
platesaccording tothestrain-reversal rdel (ref.20). Thederivation
ofthisequation isratherelegant andwasusedbyStowell, who,however,
usedtheno-strain-reversal mdel (ref.5). Thedifferential equation
obtained byBiJlaard reduces tothatderived byStawell bysetting
v . 1/2 intheformer.
Handelrxm andFrager, duringthistime,obtained
solutions toseveral fnelastic-buckl.ing groblems byuseofincremental
theory(ref.19). Testdata,suchasshowninfigure6,indicate that
theresults ofincremental theories, regardless ofthebuckling model,
aredefinitely unconsemtive, whereas defomtion-type theories arein
relatively goodagreerent.
Alltheforegoing theories weretzscdontteuseoftheoctahedral-
shearlaw. Eovever, testdataontheinelastic buckling ofaluminum-alloy
platesinshearindicated thattheresults oftheabovetheories were
unconservative. Gerardusedthemaximum-shear lawinplaceofthe
octahedral-shear lawtotransform axialstress-strain curvestoshear
stressandfoundgoodagreezent withthealuuinum-alloy-plate shear-
buckling data(ref.21.).
Tcsumarize,then,theassuruptions whichleadtothebestagreement
between theoryandtestdataoninelastic buckling ofaluminum-alloy flat
platesundercompression andshearloadings include deformation-type
stre~s-strain laws,stressandstrainintensities defined bytheoctahedral-
shearlaw,andtheno-strain-reversal modelofinelastic buckling.Althos.$
theremaybetheoretical obJectior-is todeformation theories asa classand
theuseofa no-strain-reversal modelinconjunction withclassical sti-
bilityconcepts, testdatadosuggest theuseofresults obtained froma
theorybasedontheseassurrgtions inengineering applications. Thechoice
,oflawstotransform axialsress-strain datatoshearstress-strain data
-. - -.
-
,
-- - .
.
.
.
.
.
.-
-,I
-,}
-:-..,.), :.,.;.i,:l
de~ends upontj:e de~reeofcurrc::,:l G:lObt.::irtczl b::dc%!rl eachorth,JJs L -
lawswlihpolyaxial test, dntufor ir.iivldwd I:LLt F:r lfLIs .
FactorsUsedinCo:qm&it.icms
As already inilicated, thelnelastlc-kuckl~ng stressu.~ybccox-
putedbyuseofplasticity-reductfon fectors appropriate tothebound-
aryandloadin~ conditions. ThefacLar3incoqymate alleffects of
exceeding the proportion92 lizxit upon k,,E,a~d
in.preparing desi~chartsforinelastic buckding,
straincanbeused:
k# tz
Ecr =
()
5
+ 1#)
Frcmequations (28)end(31)
Ucr* V%r
v. Forconveaienee
thecritical elastic
(31)
(32)
Therecommended valuesof q sregivenintable2. Forcowressive
loads,thevaluesof q derived by Stowell forinfinitely longplates
exceptinthecaseofplatecolumns(seerefs.5 and22)havebeencor-
rectedtoaccount forthetistant~leous valueofPoissons ratioaccording
toa methodsuggested by Stowe12. andPride(ref.23). Thusj
,=,8(+)
k- v2)
(33)
whereqs istheoriginal valuegivenby Sti#ell.Equation (33)isthe
formoftheplasticity-reduction factors tkatappears intable2 sndhas
beenuse~toconstmctthenondimensional buckling chartsoffigures 8,
9,andlo.
Forlongsimplysupported platesundercoxzbined axialco~ression
andbending Bijlaard foundtheoretically, bya finite-difference approach
(ref.24),that
G
.
.-.
*.>
.
. . .
~
-.
,.. . ,,7 .,,
.*- . . . . . . :, ..-& ~?
.
.
.
.
.
.
.
t}.e plasticity-red.uction factorfcraxialcompression. Equation (Zl))
r~luces tOtMs valueforaxialloudalofie, since
a = O forthiscase.
Forpurebendingu = 2 andequation (34)isequaltotheplasticity-
reiiuction factorfora hingedfb.nge.
Todetermine theinstanta~eous value ofPoissons ratio,equation (25)
canbeused.Forthenond~nencional buckling chartsthetheoretical fully
plastic value of0.5wasas3um4forPoissonts ratio,aswasassvced by
Stowell inhisdeterminations oftheplasticity-reduction factors.Stowell
andPridereported on computations tie usingequation (34)instead of
v = 0.5 andshowedthattherewaslittledifference betweenthetwocurves
forflanges andsimplysupported pliates (ref.23). Bijlaard tookexception
tothisreport(ref.25);however, thedifferences wereslight, aswns
pointed outbyStowell andRride,anditcanbeassured forpractical pur-
posesthattheplasticity-reduction factors shownintheappend~ix aresat-
isfactory forgeneral desi~andanalysis.
&mStrUCtiOn OfNondimensional BUC~Ilg(%Sl_tS
Thenondimensional buckling-stress chartsoffigures 8,9,and10
wereconstructed fromthebasicnond~sionalstress-strain curvesof
figureh andtheplasticity-reduction factors shownintheappendix,
incorporating themethodofcritical strains asdepicted through equa-
tions(31)and(32).Sincethere islittledifference amongthenwzeri-
calvaluesofthebuckling stresses thatwcxildbe obtained forthe
plasticity-reduction factors applicable toa longckuxped flangeandto
a hug platewithanyamuntofedgerotational restraint, tkesecases
weregrouped intoonee@oying thereduction factorforthesiuplysup-
@rtedplate,whichistheaverage ofthethreefactors.
CIADDIM3 RRXCTIONFACTORS
Eaiic%inciples
Thepresence ofcladding onthefacesofplatesmy haveanappreci-
ableeffectonthebuckling stresssincethecladding raterial, wh i c h
u s u a lly h a s lowe r m e c h a n i c a lstrength thantheplatecore, is located at
theextreme fibersofthephte crosssection (fig.11)wherethebending
strains duringbuckling attaintheirhighest values.
Buchert determined buckling-stress-reduction factors forcladplates
whichinclude plasticity effects aswellasreduction duetocladding
(ref.26). However, itispossible todetezmine a reduction factorfor
.- ---- .-. -- --- - ---- -. .-. -- . -- . .
-. ., .,.,
.. . ..L ..I. L-il ~. ..;
.
,
.
.
G

claddind <a;: :.. -!J..lt lxiy & uciltli~tli~-i L:(tl,einclzstic t)uc}:lic~ stress
t~yiel~a iinnl bucklifi~ strc:sforL1.e clmipl;ite tl~:~t a~reesquite
close~withtketestdata.Thecladdir~g reductio~l iuckrs maytlhc:. kc ..._
usedwiththeexi.stin& inelasti.c-bucKLix~ curvesoffig~es8,~,aT.A 13.

Theformofbuckling eqmtionco~mdy useifor detercinin~ t!:c

.
buckling stressofa bureflatplatewithanytypeofloading andlm~n~~ -
arysuppcrts isgivenas equation (28).Forclaiplatesthisexprec~ian
isusedtofindc nominal buckling stress, wliere thethic!v.ess istt~t
ofthetotl-.l plateandtheraterial properties arethoseofthecore.
Theactual buckling stressofa clad@ate thenmy be foundbyapplying
a simplentczerical multiplier fitothisstress.Thisrultlplier,
termedthecladding reduction factorbecause itreduces theratioo:the
nominal corestresstothebuckling stressoftkecladplate,isa fxnc-
tionoftherelative coreandcladding stresslevelsandtherespective
mduliofthecoreandcladding materials. Tkeclad-plate buckling stress
canbe foundfrom
1
acr
=<ucr
(35)
E tine nominal buckling stressexceeds tk.e proportional limitof
thecore =terial,thenthenominal buckling stressforthecladplate
my be foundbyusingtheappropriate valueof q,theplasticity-
reduction factorofthecoremterial.Valuesof q maybeobtained
fromtheclad-plate stress-strain curveshowninfigure12,thederi-
vationofwhichisdiscussed belo-w.,
Itshouldbenotedthattheplasticity-reduction factordepenti
uponthestresslevelandconsequently requires anestimteofthefinal
buckling stressoftheplatebeforeequation (28)canbeusedtofi~~
ffcr.Thecladding reduction factorhasbeenfourdtobeofsucha -ture,
however, thatlittleerrnrisinmlvedinfirst~inding thenominal buck-
lingstressandthenmultiplying itdirectly bythecladding reductiaa
fact-or tofindtheactualbuckling stressoftlce cladplate.Theprod-
uct qfiis ~T$whichwasdetermined byBuck.ert.
Table3 containsalisting ofthe=riouscladding reduction fac-
torsdeteminedinsubsequent portions ofthissection.Inthetable,
allplatesarelor&andsti,ply supported. R.allcasesforwhichtke
cladding proportional-limit stressUcl exceeds thenoninal buckling
stressUcr thecladding reduction factorisequaltounity.Theymn-
tity p %sdefined as P = aclci
~ cr,ad
f istheratioofthetotal
cladding thickness totheclad-plate totalthi-kness.
-.- .
-.
--.
.
.-
.. . ....-
... .... . . .. . :.,:1
7
DcrfvaLion OrCoreStress-ZkrEln Curve
Tkecorestress-strein curve My be derived froma skress-str8fn
curvefortheentirecladp~%teusshowninfi=ve12. Usingthenoti-
tionoffl~e 11,inuhtcha sectiofi ofa cladplateisshown,thetotal
axialloadactingonthesection isdetermirtible ikon
.
. .
.
.
. .
Dividing thisexpression by tucoreyields
,
5/ocore
=1 -f+pf
where$ =
cllacore
(36)
(37)
!Ums,thecorestress-strain currecanbeconstructed byplotting
tF&corestressdeterrzined fromequation (37)ateachvalueofstrain
forwhichthecorresponding clad-plate stress wasfound.(Seefig.12.)
~.einitial. slopeofthecorecurve,whichisthesarze astheinitial
slopeoftheclad-plate curve,istheelastic mdulustobeusedinthe
noril.n.=1-buckling-stress equation.Sincethebuckling stress refersto
thecoreruiterial., Ucorewasreplaced byitscounterpart Ucr inthe
succeeding derivations.
~ical valuesof f foralcladplateappearintablek forsev-
eralaluninum alloys.Duchert showeda valueof acl= 10,~ psifor
ll(X)-KL4alloy (ref.26).However, thecladding stress tillvarywith
thecladding material, ofwhichdifferent typesareusedondifferent
alloycores.
.
Comparison ofTkeovand@erfm_ent
Thetotal-reduction f8ctor, defined 8stheproduct oftheplasticity-
andcladding-reduction fnctors, hasbeenplot-ted infi,-.re 13asa func-
tiffn ofstressforboththetestdataandthetlfeory ir,thecaseofaxially
compressed plates.Wo mterialsarerepresented, eachwitha different
percentage ofcladding thl.ckness. purthe~ore, thefirst(202h-T@L sheet)
isa sirrply supported platewhereas thesecond(~zk-~ sheet)isa long
colum. Plasticity-reduction factors forthesetwocaseswereobtained
frontable2. Itisinstructive tonoticetheclosecorrelation forthe
colunncase,forwhichthetangent mdulusistheapplicable plasticity-
reduction mdulus.Thisfo~owstheprediction ofthesimplified theo~,
.-- .--- ---
..
I

,,
,/,. ,
,
,.. l\... :.:
4
.
.
.
i
Derivations ofSimplified Cladding Reduction Factors
Buchert derived expressions forthetotal-reduction factorforfist
simplysupported rectangular platessubjected toseveral typesoflc e . iin gs .
In thefollowing sections arepreriented derivations ofsimplified cla.iding
reduction factors thatyieldbuckling .stres6es atallstresslevelsz~rely
bymLtLplying thenGzinal stress(elastic or inelastic) bythecladding
reduction factorattkt stress.Thisisdofie by separating theclad.iing
effectfromtketotal-reduction factorbyusingtherelationship fi= ~/~.
Case1 . LongSs!lySuppo tiedplatesin coqression.- Buchert derived
theevressiori for ~ at a==>upl (ofthecore)(ref.X):
where
~ = (3f%/Es) [1A) + (3/4)(%/%]
Fora bareplatef = O and ~= q,ufiich give
(*)
(39)
.
(cf.table2). Then
3 )
1.+(,&,E=jl +{[l +(mJEs] p,,,+(,,4) (2,,++w]~f
5= 1
l+3r
(Q)
1+11A)+ (3/4) (%/%]12
-. ,.-. ---
. .=
.
.
.
.
((
.
.
.
.-
<=
{
(l+3Pf) +*p+39f)(4+313fp2}
(41)
2(1 + 3f)
vhich JllSy bewritten
IfitiS ass-d
sionisobtained
-(8[+i~-[$vfm + 3,f)l}2)
that 9W/(1+ 3Pf)<<4, thefoUowingsiapleexpres-
forthecladding reduction factor:
*
(42)
(c)Forl.arge stresses,P-O mdtheretore
1
?=
(43)
l+sf
Equations (42) ad (43) m=r infi~ 13~nthefo~of qT= Tfi>
wherethey, maybeseentoagreeclosely withthetotal-reduction factor
andthetestdata.
.,-
Case2 . Platecolumns.- Thederivations of I forshortandlong
platecolumns followthefomnusedincase1 forthesupported plate
uithout anysimpli~ing assumptions.
Theresults areshownintable3.
W COlumncurveisplotted infi~e 13int~~e fo~ qT= vi>whereit
isseentoagreeclosely withthedataandwithF!ucherts theory.
Case~. Langsimplysupported platesinshear.- Buchert (re&.26)
showsthat ~ forshearona longsimplysupported plateis
I
.
. .- . _ -. .-- -.... ..
. ...- --
.-
,
.,
.
.
.
.
.
.-
wherethenodal-li[,e
@licit equation
slopeoftheshearbuckles isobtaine~ from the
The&nLmm-energy state occurs foruncladelsstic plateswhen
I
a-l ~, andthereislittleyeasontoexpecta significantly different
valueforcladplates.Consequently, thisvalueof a isassured in
thefollowing development:
C5T EsE
.wpa+p3T/%)+*iJ (45)
M 4(l*3f)
(a)When acr< ~c13 Ipfi=lq=l.
(b)Theplasticity-reduct~on factorfor a=r>dcl isderivable
fromthetot.al-reductian factorintheform
,= (E8,8E)f +(~ps) +-+]
,.
.
.
.
.
.
G
.
-.. .
-,,.. ,. :
I
hi-(E@)++ - (@.)]
:=
1:3f 4+(qEJ+3/~-(%iq
1
where Y . 1+ 3pf.
(47)
The expression in bracesdeviates about2 percentfromunityfor
f . 0.10 andfor $>0.2,u~ch willbeintheneighborhood of the
proportional limitfortypicalstructural aluminum alloys.Consequently,
ittillnotintroduce anappreciable errortoconsider itequaltounity,
inwhichcaseequation (42)forthecompressed simplysupported plate
F=ldstrue.
(c)Forlargestresses,$-+0,and therefore {= ~.
l+3f
KEKLINGOFFIATILECTMGUIAR PLATESUNDERCOMPRESSIVE LOADS
Inthe preceding sections themathe=tical endphysical background
fortheflat-plate buckling problea hasbeenpresented.Itwasshown
thatbasicequation (1)canbe usedforthesolution ofbuckling problems
pertaining toflatrectan~ platesundervarioustypesofloadings in
theelastic andInelzstic rangesbysuitable choiceofreduction factors
andbuckling coefficients. Considerations thatinfluence thedetermina-
tionof k havebeenanalyzed inthesections entitled Easic.Principles
andBoundary Conditions.! Theplastici@-reduction SM cladding reduction
factors werediscussed inthesections Plasticity-Reduction Fac~ors and
Cladding Reduction Factors.h thissection, andinthosetofollow, t%.e
buckling coefficient k wi~ bediscussed anditsnumerical valuesfor
various loading andboundary conditions ti12bepresented.
Historical Background
_ inves-kigated thebuckling ofa simp~ys-~pported flatrec-tangu-
larplateundertxialloadingintheelastic rangeusingtheenerw
method(ref.27). Heobtained theexplicit formfor ~ forthistype
of l~dingandsupport:
kc = ~a/nb)+ (nb/a]2 (48)
I
.
... . .- ----- . . .. . - . . - - . -
.
.
.
.
.
.
.
+.
.. .
,,,,, ,, :.. j: : :. --
Tiw.. ,sI.w2s tr~tcdt].wnerou~ niiit. iMAI ,-nses oth )l(t i 1 , ;; :. I;. . 1 t ? , UI, ~I{Uy
cOr t d~t iOI1 5 Ut ~liZf:lg bo~h bk e Ii . 2 .-~ u p~r c :t c t : ar,d Llk4: solul-iott 01t~.e
differential equntion (ref.2). liiil constructed Q churtof kc covisY:ng
thecomplete rzn.ge ofpossi~;?: til&X:~;~ conditiol]s fOrf~i[l~]@k~iIl+i:
simplycj~port~~d, clamped orfreeCir;c: onone st.ie, awl simply Guppor!&
orck,peded~~sontheother,withthelomhi C:WS eitherclumped or
simplysupported (ref.28).
Lundquist and%owe~ presented tkefirstunified treatrent oftkc
compressive-buckling problem intheiranalyses, byboththedifferential-
equation andenergyrethods, ofthecasesof supported platesandflanges
withsimplysupported lomltiedgesandwithvarying degrees of elastic
rotational restraint alongtkesupprkd unloadeii edges(refs.8 and2 9 ).
SteinandLibove, inconsidering co~bined longitudiml andtransverse
axialloads,covered theeffects ofc~~ing alongtheunloaded edgesot
rectangular plates
Numerical
(ref.~).
ValuesofCompressive-Buckling Coefficients
*
forPktes
Figure14isa sumarychartdepicting thevariation of ~ asa
~CtiOn Of a/b fOrVSriOUS limithgconditions OfedgeSupport and
rotational restraint ona rectangular fht plate.Itisapparent that
forvaluesof a/b greater thanfcmrtheeffectofrotational restraint
alongtheloadededgesbeccmes negligible andthatthecl~ed plate
wouldbuckleattirha12.y tine samecompressive bad asa platewith
siqpl.y supported loadededges.
.
Supported Plate,~~ws
,,
Against
.
Elastically Restrained
~otation
Thebehavior ofcompressed plateswithvarioua amuntsofelastic
rotational restraint alongtheunloaded edgescanbeunderstood by
exminingtherelation between buckling coefficient andbucklewave
length.Forplatessupported alongbothuriloa.ied edgesthecurves
appearinfigure15forrotational restraint fromfullclamping (e= =)
tohingedSUppO1tS (~= O). Fromthisfigure, which1stakenfromthe
reportbyLundqulst andStoveI-l (ref.29),itispossible toseether~-
ner inwhich~hebucklewavelengthdecreases asrotational restraint
increases, andthevalueof A/b for a minimum val ue of ~ canbe
seentoincrease from2/3fore-lamped edgesto1.00forhingededges.
Thelowerportions ofthesecurvesandtheportions totheleftoftke
mininum~ lineformthefirstarmsofthec u r ve s of ~ a s a function
.
.
.
. .
--- - -. - -. - .. .. --.. ... .- ___ ____ ___
. .
.
.
.
.-
i
.
i
.-
..
!&:;,~,~;il. >~
of a/b,asin ? X&m e16. Forco.~)leieness,
tr~nsttims from1 to2,2 to3, . . .n ta
included infigure15. Theintersections oftheseIities withtl.e curves
of ~ againstA/b correspon~ tothecuspsortthecurvesoffigure16.
PlatesWithUnequalEdgeRotational Restraint
Figure15canalsobeusedwhenthereareunequalrotational
restraints alongtheunloaded edgesofa plate.Thiscanbedoneby
determining the & valuefor$he e oneachunloaded edge.Theeffec-
tivevalueforuseinequation (l)canthenbefoundfro=
(49) -
Theaccuracy ofthismethodhasbeerI demonstrated byLun.iquist andStowell
whocompared results soobtati.ed withthevaluesobtained by solting
directly withtheequations usedbythemforthegeneral caseofroca-
tionalrestraint (ref.29).
-.
Theelastic restraints aremathemticallye quivalenttoa seriesof
unconnected torsional springs.Sincethisdoesnotriecessarily conform
tothebehatior oftheusualedge~r orstiffener ofa flatpanel,
itisnecessary toevaluate theeffective singlespringstiffness ofthe
actual. stiffener inordertouseeitherfigure15orfigure16. However,
itfsnotnecessary todetermine thisstiffne=s toa highdegreeof
accuracy sincetheinfluence of E upon & embraces a large range of
sttffness ratios} asisshowninfigure1?forinfinitely longplates.
Whenthestiffener rotational rigidity hasbeenfound,e naybecon-
putedlzyforming theratioofthisrigidity tatherotatiordrigidiw
ofthe@ate.
RorntestdataGerardwasabletoconstruct a chartof k for
lo~gplatesasa functionofb/t forstrongandweakstiffeners
(-f. 31andfig;18).Aboveb/t=200 itisseentbatnnststiff-
enerswilleffectively clmzptheplateedge.
Supported Flanges WithZlastic Rotational Restraint
i
Therelationships amng ~, A/b,-d e aredepicted forflanges
infigure19. Itshouldbenotedthatthesecurveswereconstructed for
a @issonsratiovalueof0.3,whichalsoapplies tothecurvesof &
asa function of a/b infigure20. Thedetermination of ~ forother
valuesof v isdis$ussed inthesection entitled Boundary Conditions.
.
.
I
..- -,- -. -
-.
.
b.
.
.
. .
.}::
&
. . .. ..... ..
. .-..-/..,4 >,.
Thc~lTUISit iO!l lh ~S fO11 tO 2, ~ tb 3, . . . n ~ n + ~ buckl~:, :LW
shorninfigure19. %Jever,itshouldbenotedthaktherinimuzn lizc
doesnotintersect thecurvefora hir~ged flange(c= O). Forthiscase
.
thereisonlyoccbucklewhichextends thefulllengthoft?.e Ilange.
As inthecaseoftheplate,thetheoretical restrafnt actionon ttLe
unloaded supported eiige oftheflangeisassumed tobea seriesof dis-
connected torsional springs, anditisnecessary inthiscasealsoto
determine theeffective restraint fortheedgestiffener inordertouse
thecurvesoffigures 19and20. However, asfnthecaseofsupported
plates, itisnotnecessary todeterminec tooaccurately, asfigure17
shows,since& isrelatively insensitive tolargevariations in c.
EffectofIateral Restraint onBuckling
h theusualbuckling-stress computations theplateanalyzed is
assumed tobeunrestrained against distortion initsplaneuriler the
external loadsapplied.However, forlongitudinal compressive loadson
a rectangular plate,theedgesparallel totheloadswouldtendto~ve
a~rtasa resultofthePoissons ratioexpansion. Ifthismtion
should be restrained toanyextent, forceswouldbedeveloped transverse
tatheapplied bad whichwouldinfluence thelongitudinal stresstkat
thephte tightWithstand beforeitwouldbuckle.Iftheinteraction
concept isemployed, it5sapparent thatthetransverse compression uculd
loyertheperniissible longittiinal stress byan amuntthatcould.be
foundfrominteraction-curves utillzing stressratios.
Iftheplateedges
~transverseribseach
forcesrequires that
Thedirections of ax,
.&erest~ined byrigidstiffeners heldIn place
witha section area ~, thebalance oftransverse
..
.
(50)
Uy,and Cr r ar e shown infigure2L. Theec@va-
lenceof transverse strainrequires that
G
(5U
r.
assuming thattheribsand platea r e of thesamemterlal.Fromequa-
tions(!%3) and(51),thetransverse stress becomes
, .

~ . ..- -- --.
.
k
.
,-.
..:
s
1
.
.
.
.
.
u
.
.
.
. .
.
Fromthispointitisa sirple rottertodetermtfie tt.e reduced
longitudinal-buckling stress.Thismy keexpressed inte=5ofthe
newvalue ofthebuckllng coefficient ~ esshowninfigure21,which
isa modification ofcurvespresented byArgyris andDunne(ref.32).
BWXZNG OFZIAT RECTANLUUR HATESUNDERSII!WR LOADS
Historical E!ackground
Southwell andSkancomputed thecritical shearlosdfora flatrec-
tangular platewithsimplysupported edgesandwithffiededgesbymeans
oftkebuckling differential equation (ref.33). Tinxmhenko investigated
shearbucklirg also(ref.2);however, heusedtheenergymethodand
obtained a critical loading 6.5 percent. higherthantheexactresultof
Soutbwell andSkan.
Stowel.1 deteqmi.nedshear-buckling coefficients forinfinitely long
supported plateswiththeedgeselastically restrained agair~t rotation
(ref.%). Heutilized thedifferential equation foranexactsolution
andtheenergyintegrals forplotting purposes.StoweHpresented his
results inthenannerofSouthwell andSkan,whoplotted thebuckling
coefficient asa function of A/b forlongplates.Thisisthesame
procedure usedbyI.undquist andStowell forcompressive loadlng onplates
ofanylength(refs.8 and2g).
mtricatihtisptric Modes
Thesolutions obtalned~Southwell andSk- (ref.33)andby
!J !imshenlm(ref.2) pertained toa buckleformtemedthesymetricmode
because ofthe~etxy ofthemodeshapewithrespect toa diagonal
acros6theplateatthenode-line slope.SteinandNeffexminedthe
antisynmetric buckletie forsimplysupported platesandfoti thatit
hasa lowerbuckling stress, withina stall. rangeof a/h values, than
doesthesymmetric-tie (ref.35). Steinaiid Neffalsorepeated
Timoshenkos calculations forgreater precision andobtsined anesti-
=ted errorof1 percent.
Budiansky andConnorinvestigated theshortclamped plateforlmth.
symmetric andantisymetric bucklemalesusingthelagrangian mltiplier
method(ref.36). Exceptfora smallrangeof a/b values, tkesymmetric
tie wasshowntoyieldthel ow er buckling stress.
.
..

.. .-.r.
.
.
.
.
G
Numerkal ValuesofShe4r-Duckllng Coefficient
r
Theplotof k~,asa fmctim of a/b appears infigure22. It
.
maybe seenfromthecurveshozthesymzitric andnntisynmetric modes
alternate withoneanotk.er as a/b increases. Forlongplatesthe
valueof ks my befoundfromfigure
a function of c.
=fectofPlateLengthon
23(a),inwhichksm appears as
Buckling Coefficient
When k~ isplotted asa function of a/b forinfinite andzero
valuesof e (clamped andhi.nged edges)asshowninfigure23(b),it
my be seenthatthereislittledifference between thetwocurves.
Thissuggests a rapidmethodofcomputing theshear-buckling coefficient
foranyvalueof e. Thecoefficient forthespecifiedE isob-ti~
fromthecurw=cf &= asa function of e (fig. 23(a)), whichisa
replotofthe-minimum ks line(n= m)offigure24. Also,theratio
ks~ksm
isfoundfromfigure23(b).Then k~ forthespecifieda/b
and E maybe foundbycomputing theproduct ofthesetwonumbers.
Estimation ofthecorrect valueof ks&=
I
willberelatively freefrom
errorbecause oftheproximity ofthetwolimiting curvesinfi~e 23(b).?
BUCKLIMOFFLAT
#
REC711GUIAR PLATES UNDERBEND~W2 LOADS
Historical Background
.
. .
Ti m m h e n k oinvestigated thebuckling stresses forflatrectangular
platesundercombined longitudinal andbending loadsusingenergyinte-

gralsandobtained valuesfor kb thatagreewe~ withlatercalculations

ofhigherprecision (ref.2). Schuette andMcCulloch analyzed longplates
underpurebending tithsupported edgesandelastic rotational restraint
(ref.37).Johnson andNoelalsoinvesti~ted thebuck31ngQf plates
underlongitudinal axialloadandbending(ref.38),andNoelanalyzed
platesforlongitudinal bending plussxial- loadcombined withtransverse
=ial load(ref.39).
Numerical ValuesofBendin~-Buckling Coefficient
me relations between bucklewavelengthandbuckling coefficient
for=ious valuesofrotational restraint appearinfigure25together
withthewave-length transition lines.Thecurvesof ~ asa function
of ajb areshowninfigure26. Itisofinterest tonotethatthe
,,. .b
.. ___ ... .
-- - .---- . -
.
4
.
.
. .
G

7
valueof kb forinfinite pl.te6isroughly slxtinesasgreatasthe
valueforthesupported plate~ forallvaluesofrotational restraint.
BUCRLING OFFIATRSC3XMWLAR PIATESUNDERCOM31NED LOADS
General Background
Flatrectangular platesfrequently aresubjected tocombinations of
elementary loadings.Ithasbeencommonpractice toconsider elementary
loadings inpairsandtodetermine aninteraction curveorcurvesforthe
combination. Eowever, tworecentpaperstreattriplecombinations ofthe
elementary loads,sothataninteraction surface instressratiosisgen-
erated, andbytakingapp~priate sections (e.g., letting oneofthe-
stkessratiosequalzero)it ispossible toreproduce theinteraction
curvesthatkere derived previously intheliterature. .
Interaction curvesforthecodination ofbending, shear,andtrans-
versecompression onlongplatesweredeveloped byJohnson andlhcbert
(ref.&.O), andNoelConstmcted tbetwo-dtinsiond sections ofthe6ur-
faceforlongitudinal bending, longitudinal compression, andtransverse
compression (ref.39). Thebackgrounds forthevarious combinations of
loadings arediscussed inthefollowing paragaphs.Interaction charts
areshowninfigures 27and28, in whichsections ofthetriplestress-
ratiosurfaces appear.
A sumzaryoftheloading C~ditiOn6discuese~ inthefo~owingpars-
9gmphsappears intable5. Interaction equations whichexistfora few
casesareincluded inthetable.
t
,r
Biaxial Compression -
Tinnshenko derived a relatlon between thelongitudinal andtransverse
edgestresses actingonarec~r plate~tbuckling (ref.2). This
relation wasenluatedforthelowestpossible combination ofstresses by
meansofa chartthatMUStbedrawnforeach a/b mlue underconsidera-
tion.AsoneMsltingcaseofplateproportion andloading, !lEhmshenko
denmnstrated thata squareplateloadedbyequal. biaxial strecses hasa
bu c k lLn gcoefficient of2,orhalfofthatfora uniaxially loadedsquare
plate.
e
LiboveandSteinevaluated buckling underbiaxial loadings by the
energymethodforrectangular platessupport@inseveral different nan-
nersandpresented theresults inchartsof-~ asfWnction6 of a/b
forvarious valuesof kyjwhere
*
... - .--- .
.,,,
- .--!-
.:
. .
.
qf
()
I&D
=kx
b%
()
s*D
~Y=%~
1
(53)
and ax and Uy =e thetwostresses actingontheplateatbucklira
(ref.30).
Nosimpleinteraction expressions existforthestressratiosin
thegeneral casefortheloadings emdsupports investigated byLibove
endStein.However, forsqusrepanels, orforlongpanelsthatbuckle
Insquarewaves,itcanbeshokm,fromTircoshenkos results, that
Rxt~=l
Noelconsidered marecomplicated loading conditions andpresente~
dak fromwhichinteraction curvesuybe constructed forbiaxial loadtis
foranyvalueof a/b (ref.39). Noelscurvesappe=infigure28.
Shear andkimalStress
Bye,pplicationof theener~method, Stowell andSchwartz examined
theconditions underwhichbuckling willoccurona long,flat,rectan-
gularpanelwithedgeselastlca~restrained against rotation underthe
sinmltaneous actionofshearandnozmall. stresses (ref.41). ~ey derived
theinteraction re~tionship between the
.~+F@=
Theyalsoderived anexpression for
buckling throughuse. of thedifferential
stre3sratioshthefo-im
1
(55)
thestressco~bination at
eqyation andtestedtheinter-
actionequation forseveral valuesofrestraint coefficient e. The
agreement withtheinteraction equtionwasfoundtobe excellent, asa
consequence ofwhichtheinteraction equation written aboveraybe
applied tothisloading caseforallvaluesof restraint coefficient
andmy beusedwhentheaxialloadiseithercompression ortensioc,
protided therestraint coefficients arethesameonbothedgesandtie
panelisinfinitely long.
I
1
I
.,.
i
!
....
.
.. ..-

.
..
.
.
.
.
.
.
c
,
.
.
.
.

Theproblem of~etci%inln~ critical loadin~ combinations fa?d.cur

anitransverse normlst r ess w as sol ved by~ltdorf andHouLolt byboth
tt.e energymcttiod ud thedifferenti~l equation (ref.42). Tti~signifi-
cantresultof thisworkisthedemonstration thatroughly halfoftl:e
critical shearstressraybeapplied toa transversely compressed pmel
without lcx~ering itspermissible compressive-buckling .atress.
Thisworkwasdoneoninfinitely longpanelswiththelongedges
supported andelastically restrained against rotation. Therestraint
coefficient wasfoundtoexertanappreciable (although notverylarge)
effectuponthecriticaLloading combination. Theresults forthistype
ofloading, conseque~tly, donotlendthemselves tothewriting ofa
simpleexplicit interaction equation between thestressratios.The
curveswereplotted byBatdorf andHoubo2t forbothcompressive andten-
siletransverse nomalloadings incofiination withshearovertheentire
=nge ofrestraint coefficients.
Thetwopreceding loading ~onditions werereexamined forsimply-
supported platesoffinitea/b byBatdorf andSteinwiththeuseof
theener~equations (ref.43). Theyshowedthattheparabolic inter-
act~onexpression ofStowell andSchwartz (eq.(55))agreeswiththe
interaction curvesforfinitevaluesof afi forshearpluslongitudi-
nalcompression (ortension) (ref.41).However, thecurvederived for
infinitely longpanelsundershearandtranmerse loading requires mxli-
fication for finitevaluesof.a/b. Fora squar e paneltheparabola
agreeswithther!ndified curve,whilethesimple+dge-support caseof
Eatdoti andHoubolt(ref.-k2) maybeusedfor a/b.=k. Thetransition
regionfromthemdifiedcurvestothosefor
a/b=m liesbetween
thesetwotiuesof a/b.
TheMge shearstressthatmaybesuperimposed uponthecritic~l
Compressi- stresswithout lowering thepermissible cwressfvestress
forinfinitely long panels isnotpossible forsqmreplates.= fact,
itappears tobepossible for.infinitely longplatesonly.
1.
Bending andHOrmSlStres;
Tirmshenko determined thecritical combination ofbending andnor-
=1 stresses actingonsimplysupported fl%z.rectangular platesusing
theenergymet?xd(ref.2). Hedetermined thebuckling coefficient as
a function of a forseveral ratiosofmomentlosdlng toaxialloading
forpanelswithvarious valuesof a/b.
Johnson andNoelbroadened thescopeoftheproblemhyincluding
elastic rotational restraint alongtheunloaded compression edge(ref.38).
Theirresults wereplotted as kb versusA/b forallvaluesofrestraizt
.
. . ..
.
--
-
.
.
.
.
.
.
..
.
.
.-

Lo
coefficient. Onecharti6rquiredfOreachofthe
gitudinal loading tomomentloading), of.which four
Theloading ratio.is de~iced by
12M
~.
Pb+6M
Pb=6(2-G)
M al
?iACA V! 5751
loading rntios(lon-
valueswerechosen.
(56)
whereP isthelongitudinal load,M isthebending nxment, and b
Isthepanelwidth.TheyalSOplOt kb asa function of a/b forthe
casesofsimplesupport andclamping oftheunloaded compression edgeof
thepanel.Inaddition, theeffectoffti}~yofthetioadedtension
edgeisdepicted forvarious valuesof u..% a plotof kb vers= ab
inwhichthehinge~andfixedcasesaredrawqonthesanegraph.Itis
apparent thatedgefixitydoesnotbecomeimportant untilu falls
below7/4,whichcorresponds approximately toa Pb/M of1 ormore.
Grossman examined bendingincofiination withtransverse compression
usingtheenergymethod(ref.4f+).Hefoundthatforinfinitea/b tke
bending stressmtio canbe0.9atthesametir&thatthetransnrsecom-
pressive stressratiois1. Healsoprovides a graphofthestress
ratiosforseveral values of 8/b; however, apparently onlytheinfi-
nitelylongplateiscapable ofwithstanding bending stresses without
buckling whilethetransverse stressisatitscritical value.Thisis
similar totheresultfourdbyBatdorf ti Ste~~forshearandtransverse
compression (ref.45).
Noelprotides Interaction c&ve&forsimplysupperted re&angular
platesloadedinlongitudinal bendi~,longitudinal compression, and
transverse compression (ref.39]. Ibrthelimiting caseofnotransv&rse
, loading theyagreewiththeresults ofJohnson andHeel(ref.38),and
whenthelongitudinal compression vanishes theyagreewiththoseof
Grossman (ref.44). Consequently, theirchartscanbeusedforbothof
theseloading combinations. ~.ecurvesappeminfigure28.
ThedataofJohnson andNoelandoflibel wereobtained fromequa-
tionssolvedforinfinite valuesof am andwereapplled tofinite
=luesof 8/b byuseoftheidentity
/
A/b / =amb
(57)
.-
..
.
.-. .. .... --. . -. ----- . -. .- - .
l%i~procedure raybe qtiestioned for62311aspectratios; hc.yever, it
maybeJustifici byco~arisonwitil theworkofTimx:~enko (longitudi-
nalco~ression amlbending) andwithtk.e workofGrossman (transverse
compression andbending), withwhichtheresults ofJohnson and Nuel
andofNoel6howgod agreex.e;.t..
G
.
Berziing andShearStress
r
TLmoshenko reports theresultofanalyzing a rectangular flatplate
todetermine thecritical combination ofbending endshearstresses
(ref.2). Heusedtheefier~mthodand.plotted thebucklirz coeffi-
cient, ofthepanelasa fhctionoftheshearstressratio.Tinecoeffi-
cient,whendivided byth& forthebending loadalone,becomes the
bendingstressratio,andthesetofcurvesprotided byTtishenkofor
various ~ues of a/b becomes anfiteraction chart,fromwhichitnay
be seen that theinteraction
Themnge of a/b forwhich
to1.0. However, thecurves
equation isa unitcircle:
*
R~2+Rb2=l
(*)
Tinm&enko plotted thecurvesisfrom0.5
loopbackO%themselves as a,b increases,
thusindicating thatlargervaluesof a/b wouldyieldcurvesfalling
withintheplot.The~imun variation ofstressratiosabmt thevalues
obtainable frm thecircular interaction equation is7 percent, withthe
equation valuesthehwest (andhencether.stconservative) ofall.
8
Bending, Shear,:and Transverse Compressfbn
JohnsonandBuchert utilized theLagran&ian multiplier rethodto
determine thecritical co~binations ofbendtng, shear,andtransverse
compressive loadson rectangular flatplatesofinfinitea/b (ref.40).
Theresults appearas interaction surfaces inthethreestressratios~,
R6,and ~. Thetwotypesofmpportfortheplatearesimplesupport
alongbothlongedgesandsimplesupport along thetension(dueto
bending) edgewithclamphgalongthecompression (duetobending) edge.
Sections of tliiinteraction-surfaces takenperpendicular toanyof
thethreestress-ratio sxesyieldplanestress-ratio curvesthatagree
withtheresults obtained directly forthesecasesinprevious publica-
tions.Thisistzmeonlyofthesimplysupported plate,ofcourse,since
nothing hasappeared intheliterature forshearplusbending ofplates
withthecompression edgeclamped.Theinteresting resul~ofa shear
St r es s r at io e qu a l. t O l. ~, t it h Rb e qwil t o 0 . 5 , isrevealed (fig.27(b)),
aswellas thecombination of,Rc= 0.94,Rb= O.X$ and Rs= 0.43.
42
.
.
-
.
.
.
.
.
Transverse CGrpress ion
TheworkofIrcel (ref.39)ont~Le pmb~em~flonEl ~udiru~~ ~XOWl {I:{
~on@tudinal compression, andtransverse coripression LQShccrl t~[xcjfi
inthesection onconibinc,i bending andfiorc-al stress.me pertlni~rl~
interaction curvesappearinfigure28.
Combined tielastic Stresses
Stowell utilized theconcept ofaneqtivalent=stress ~tensftfl r
combined stresses appliedinconsbnt ratio during loading in t~l:
I:,,.]nLl-
ticrange(ref.45).
Heexamined theproblemofdete~ni~gtl.c
~rlt[cu~
combination ofshearandlongitudinal compression inelastically sl~]
,,r@
-.
flatrectangular platesbyusingtheenergymethodtodetemnfnc tl;c
bUC~llgstresses.rnQ theseresfit~, stressratioswerep~otte?d
,directly fromthetheoretical results andwerealsocorrected forie
changesineffective madulus.
Wornthis,StowellCODClUd~that~jt]
little@mor thefollowing stress-ratio equation isapplicable:
(ES)
[1
()
E
2
%
s ps
2+ %
() ()
=1
% ai Sai
(ya
i Inequation (591s(Es)p.
isthesecantmodulusat a = acr forWrd ,
compression,
()
Es~~ iSthecorresponding secant~ti= forput?a a~)nnr)~
~ (%)=i
,
isthesecantrmdul.usror the effective stress ofthem- :
bindldngathcW&; (~)ai=
~(~+3T)~3.$+~]2*g
s~~i~y ofthisexpression tothatforthee&tic caseisapPnret]t;
tifact,intheehsticrangetheexpression r~ucestotheeq~tionor
elastic loads.
A recentinvestigation ofpetersonlong ~gwe tu&5 lo~ded f~
t Qr Si OEl and compression (ref.17)fidic~tes t~t a StEss.mtio eqwb~ion
of thefoxq
G
R&+R#=l
(60)
.
,.
4
i-.

. .
.
.
.
.
. .
.
.
.
ag?ec=sli@KLybetterwit.11 thetestIiatu (fig.29)thandoestkemdi~icd
Ff4ratmla ofStowell(ref.45). Actually, tlie data yieldslightly hifjler
stress-ratio cmi-oination.s tli2n do ei~herofinteraction eluations (59)or
(@),tiththediscrepncy i,,crea~ing l~ithdeereasir@ stresslevels.For
stresces whollyintheelastic rangethedataareasmuchaslCQpercent
higher(thatis, Rb is0.4instead of0.2for &- equalto1). The
dataalsoagreeclosely withe theoretical curveobtained byEkx3iansw,
Stein,andGilbertforlongsquaretubesloadedelastically intorsion
I .._
i
andcompression (ref. 46). -
EFFECT OFPRESU~@01{BUCKGIG3 OFRECTANGULAR
RangeofPublished Results
FL4TPLATES
Theeffectofnormalpressure on thelongitudinal compressive-
buckling stressofa rectangular flatphte hasbeeninvestigated for
bothsimplysupported andclanped edges.Levy,Goldenberg, and
Zibritos~ (ref.47)analyzed thesimplysupported plateusingthe
.
largedeflection differential equations ofVonI&r&. Theplatelength.
=
wasfourtimesthewidth,wkiich placesitinthelong-plate category.
Thedatar%feala r i se in longitudinal compressive-buckling stressfor
thisconfi~tionwhichincreases withpressure.However, thisrise
E%yW realizable onlyina plateofsuchproportions endloading because
oftkesignificant difference inwaveformsof thelongplateundercom-
pressive andpressure loadings.Itmaybe intuitively evident thatwhen
thereislittledifference between thesewavefoz%i,suchasfora short
plateundercombtie~ lor~itudinal compression andnorml pressure, there
raybe a reduction intheco~ressive:buc.kling stressoftheplate.No
dataareavailable inthiscase,however.
.
Longitudinally Corcpressed LongSimplySupported Plates
Highnormlpnessure =s foundtoincrease thecompressive-buckling
stress~nsiderably forthelongsimplysupported platetestedbyLevy,
Goldenberg, andZibritosky (ref.47).. Forexample, whenthdpressure
applied toa platewithlengthfourtimesthe@.dthreached24.03Et4/b4,
thebuckling stresswas3.1 timesthatforzeronormalpressure onthe
phte. Levy,@lderiberg, andZibritosky alsoshowedthatmorethanone
equilibrium configuration oftheplatewaspossible Whennorml pressuzze
wasapplied, withtheconfiguration atanyinstant depending uponthe
previous loading history.Theplatecouldbeeitherbuckled orunbuckled
undervariousspecific combinations ofaxial l oad and 110~1 pressure.
G
, ... ..n,
.
.
.
.
.
.
Lcm.gitudinally Compressed LongC1.n.mped Plates
Wmlley,Corrick, smdLevyanulyzed a lox~itudinnlly co.rpressed 10T1c
c~ei plate(ref.43). Forthiscasetheeffectofpressure wasnotso
prcmxncel as forsimplysupported ewes. Theratinmm bueklir.c loadfor
44
a pressure of 37.75tt ~u wasfoundtobe1.3timesthstforno norxul
pressure.Also,forcti~edplatesthebucklepattern wasfo~ti tohe
*
uniqueforanyparticular combination ofpressure andaxiallcmding.
SPECW*CASB
Useof Elastic-Buckling-Stress Expression
It hasbeenshownthattheelastic-buckling stressfora~ flatrec-
tangular plateofconstant thickness canbecomputed usingequation (%)
forvarious loading findboundary conditions. Therearealso flatplates
ofinterest toaeronautical engineers thatareneitherrectar~arnor
of constant thichess.Bysuitable choiceofthebuckling coefficient
anddefinition oftheplatethickness andproportions Itispssibleto
utilize equation (26)tocompute thebuckling stresses fortheseplates
alsoG
AxiallyCompressed PlateWithVariable Loading
andThiclmess.
Pines and Gerard inve~clgated theproportions ofa simplysupported
flatrectangular plateundervsxying axialloadingtodetemiaean effi-
cientthicbessvariation forminimum weight(ref.49). Theplaterigid-
ityvasassmedtobeproportional totheaxialloadinordertosatisfy
equation (26)atanyspanwise station.Theloadvariation alongthe
plateuasassured tobeproduced byshearstresses smallenoughtohave
negligible influence uponthebuckling characteristics oftheplate.
R.rrthernore, theairloading ona typical whg develops a coveraxial
loadir~ thatclosely follovs anexponential variation thatdecaysfrom
therootoutboard.Thiswi~ dictate maximum axialloading ontheco~r
attheroot,whichisdepicted as station A infigure30,inwhicha
sketchofthetapered plateisshowntogether wcfth theloading andplate
thickessvariations thatfollovasa resultoftheassumptions madeby
PinesandGerard.
Results presented intheformofthebuckling coefficient asa ftinc-
tionof a/b forvarious valuesofthelogarithm oftheloading ratio
(Y-_ lotiing,~inimm loting)reveallittleincrease ofbuckling
coefficient untiltheloading ratiobeginstoexceede (thebaseof
naturallogarithms) (fig.30).
.
-.
.
.
.
. .
3
.
.-
Axially Corpresscd Plate
cons-ant
Theproblem ofdetermining the
WithVarieble Loading and
Thickness
buckling stressofanaxially com-
pressed flatrectan~lar platebasinvestiga~ed byLibcve, Ferdmin, and
Peuschfora simplysupported platewith cozstant thickness anda linear
axialloadgrdient(ref.~). Theyplotted theeffective buckling-
stresscoefficient asa fumction ofthe.loading ratiofor various values
of a/b. Forthesakeofuniformity ofpresentation, theircurveshave
beenreplotted hereintheformof ~av asa function of a/b for
various valuesoftheloading ratio,including negative values(tension
atoneedge)aslargeas -3. Tke.se curvesappearinfigureS1.
Thebuckling coefficient -~av applies totheaverage axial
loadingontkeplate,whichisequalto (oA+6B)~ with UA assumed
tobethelargerofthetw endloads.Theaverage plateloadis
(uA/2)~+(1/P)].~iswtit. rapidc~~=isontiththe~c~ing
stressofa platewithconstant axialload,whichisthecurvefor
S = 1 infigure31.
LmR dateswillbuckleattheendatwhichthemaximum loadis
applied,~f=r which~ isequalto4.
parallelo~Panels
Anderson inv&tigated compressive
in Compression
buckling ofa flatsheetsub-
divided intopanelsbynondeflectirg supports thatforma parallelogram
griduork tier thesheet(ref.51). Onesetofsupports (allequally
spsced) runslongitudina~y, andtkLeothermm atanangleq tothe
nomal., ortransverse, direction. Thelongitudiml. spacing ofthediag-
onal.supports is a,andthetransverse spacinq ofthelongitudinal sup-
portsis b. Euckling coefficients wereplotted asfunctions of a/b
fof%othlongitudinal corspresgion andtransverse compressionfor various
vahmSOf theangleQ (figs. 32(a)and32(b)).tiaddition, inter-
actioncurveswereprovided forcombinations ofthesetwoloadings in
theformof,buckling-coefficient combinations forvarious valuesof q
(fig.32(c)):
.
:.r-
!:? Lii m; f,. :
.
.
.
- m
.
-
1.
,:
i
#
Thetransverse-buckling coefficient isnotsoseverely af~eci.ed Iy
V, sincek increases from4 to~ as q increases fromzero to joO.
For q equalto60, k is9at a/b=l.
Parallelo&am Plates
-.
Witt~ick-determined thelxc?di
~ stressofaparallelogran plate
withclazped edgesundertheactionofuniform compression inocedirac-
tion(ref.52). Hiswork differs fromtheworkofAnderson (ref.~)
inthatspecified rotational boundary conditions areapplied totke
plateinthiscase.BothWtttrlck a~dAndersoa employed tkeener~
apgroach inobllq~e coordinates toobtainsolutions. Results arepre-
sentedintheformofcurvescf%kebuckling coefficient k= asa Fazc-
tionof a/b. Wittrick presented dataforedgeangles of 00 (rectar~~-
Iarplate), ~, and47 as shQwninfigure33(a),in vhich thepbte
geomet~isdepicted.
Guest(ref.53) andGuestandSilberstein (ref.~) analyzed sirply
supported paral.lelcE.~m pla~~sundgrlongitudinal compression and,fora
rhoffibic plateof3@ edgesagle,deteru...ed t~~t ~ =5.60.Wittrick
al so anal yzed clarped parallelogram platieifishearandobtained the
results shouninfigure33(b)V(,ref, 55). Hasee%wa analyzed buckling or
clamped rhombic platesinshear(re~. ~)j forwhichbuckling coefficients
appearinthetablebelow.Thegeneral plategeor~tzyof figure33(b)
appliesM thiscase.
G
1
Et , deg. . 0 15 20
w 35
ks G . . . 14.7 21.0 6 L .6
?
40.0 51.0
*
IYiangular Plates
Ike :uckling oftrien.@arplatesundervarious lea@:-and edgescp-
portswasinvestigated byVoinowsky-.Krieger (ref.57),Klit&hieff (ref.7J
Wittrick (refs. 59to 61), andCoxandKlein(ref.62). Woinowsky-YXicgPu
ccrptedthebuckling stress ofa simplysupported equilateral triangular
.-

..
._ , .-
.
4-
.
FkLe unteruniforv. cc,~rczs 3..nsmlfcna,dkw Laequal> when1.Ec base
s?thetri~r.yle istakene;ualLO b inequ2.icn (X,. Klitchieff
:nv~sti~~t edthek.xklin~ ofrig?, t--, -w~e Lsoscelcs tria:>d=llar pla=s
.it.11 pureshearofithe0rVr,0;J2nSl side3upplied soas toproduce cOm-
?ressicn alGP<thealtitu~e up>ntlie kypotenase. Wittrick evuluated the
buckling coefficient forshearapplieiso as toproduceeftkr coqm2s-
sienor tension alorg the alti%lude cadalsoincluded tk.e effects ofnor-
.--1 stresses applied tot!.be;ual le~sof L?.e triangle.CoxandKlein
&nalyzcd buckling inisosceles trlanz~es ofa~yvertexanglefor normal
stress aloneandforshearalone.
Thebuckling coefficients presented intinis section aretobeused
inconjunction titkequation (%). lrLe ~eo~etry ofa triangular plate
issho%~. infigure34. ThedataofcoxandKleinappearinfigure3k(a)
foruniform co~re~aion andinfigwe34(b)forshearalongtheequal
legs.Eothsimplysupporte~ andchwed -es werec=,sidered. The
results ofCO%and~ein agreewiththedataofWi.ttrick forri&ht-an@e
isosceles triangdar plates, ~hich appe=intable6. R,e shearbuc-kling
coefficients ks+ and ~- refertopuzze shearloadings whichproduce
tension andcompression, respectively, alongthealtittie upon.the hypote-
nuseofthetriangle.
For.she=andnormlstressona right-angle isosceles platethe
interaction equatien
applies,inwhich
(
2T
)
2
+U + A(1 - U2] =1
cr+ + cr. acr
u=
(%
- k~
+- )/( )
ks++ k6-.
Research Division, College ofEngineerir!!,
NewYorkUniversity,
NewYork,R.Y.lOctober 29,19.34.
-,.
. .
.
(61)
t
..-.
-.
,
. .
.
.
.
.
o
r .
?
Froccdures forthecomputation oftheelastic andplastic buckli::~;
stresses offlatplatesbasedongeneral plate-bucklir~ equation (1)mre
summarized inthiseectica.me factors appeari~g inthisequation are
briefly discussed andchartsarepresented fromwhichnurerical values
ofthesefactors my beobtained.
Theelemeritary loadings suchascoqressionj shear,andbending
frequently areconsidered inprelizdnary designbyusir~thebucklir~
coefficients forthe13miting cases(infinite valuesof ah, ckping
orhinging oftheplateedses,andsoforth).Forconvenience table7
has beencompiled containing thewaluesoftkebuckling coefficients
thatpertain toso~eoftheselimiting cases,whilefigure14displays
thecurvesfor ~ asa function of a/b fordifferent combinations
ofllmiting edgeconditions.
.-.. ,
.......
Physical Properties ofYsterials
G
The bucklir!!stress of a flat plateisdetermined whentheIoadir&,
plategeometry, andcaterial aresFeclfied. Theloading dictates the
particular charttobeusedtofindthebwkMng coefficient k, andtke
platea/b andedgerestraintlocatethenunerical valueof k tobe
found fromthatchart.Foranuncladplate(fi= 1)whichbuckleselas-
tically(~= 1), Ucr canbe fourid directly fromequation (1)if E is
k now n * Theeffects ofcladding ~~dplasticity dependuponthetypeof
loading andthestresslevelandtherefore require a moredetailed kncul-
edgeofthestress-stmfn characteristics ofthematerial.
.
Thethree-parameter description ofstress-strain da- c-be useif
= a convenient ge~eralized approach inbuckling problems.Withthis .
~tkodfigure3 c an be employed tofindtheshapefactorn. Since2,
uoa7,and n canbereadily determined (seetable8 foraveke values
of n),nondimension~ curvesareavailable fromfigure4. ItiS tobe
notedthat$in~ny cases,plastic,-buckling chartshavebeenpreparei

fromwhichtheplnstic-buckling stressmaybedetermined ifoneknows

E; U0.7) md n.
.
0.1loys . !l%e neudesignat.ions. areusedWm3ughcmt tMf3 reportawl t.ke
I
tableisinclu+ed forusewithLkevarious references. Characteristics
.
1
of theckddicgusedonse;eral structural aluminum alloysareshe-min
~
table4.
I
.- i
I
Poicsons ratiobeyondtheproportional limitc= be calculated
.
Usfng Vpl = 0.5 in th~expression
i-
1
(
ve)(ES/E)
v = Vpl- Vpl
.
Frequently buckling stresses areco~~tecl usingtheequation
ucr= fiq~(tfi)2
(Al)
(A2)
[
whereK = h2/1.2(1 - @). me expression K/k canbefoundasa func-
tionof v infi~e 35.
i
I
Compressive Buckling
1
Plates. -Forplates,kc appears inflgue16Interms-ofafb
and e andinfigure15internsof A/b ad E. Foraninfinitely
lo~gplate,~ rfiy be foundfrcnfigure1~intees of E alone.
Wcen e isnotthesameforbothumloaded edges,thegeclzetric mean
ofthe & valuesforeachedger&ybeused(eq.(49)).
1
Theplastici,m-reduction factorfo$& longplatewithsimplysup-
portededgesis
.
1
w%~lefora longclamped plate
[
Izelastic plate-buckling ~tresses maybecalculated usingthenondiren-
sionnlchartoffigure9.
.-.
.
.
.
.
.
J
.
I
Theeffectoflateral restraint inreducing & maybe detem~ned
fromfigure21forv=~~es of Ar/at)andtk.e effects ofthickness taz-r
andaxialloadvariation maybecalculated with theaidoffigures .V
and31.
Thegaininbuckling stresswithobliquity oftheloadededges:s :
shovn infi~-e33(a)forclaE@ parallelogram plates, vhilefigure32
depicts buckling coefficients fo:largesheetsdivi~e~ intoperallelcgran
.
panels by nondeflectkg supports.Fordataontriangular plates,fig- !
m,34(aj raybeusedtofind ~.
The variation in & with b/t forstiffezedphtes withtors~cnally
weakor strongedgestiffeners &~earsinfigme18. Because of tkesparse
dataa=ilable,norecG~.endatica canbemadecomse.rning theeffectofnor-
malpressure uponbuc-kling.
F
.-Forfhlges, ~ naybefoundinfigure20asa fucction
of.a b acd c andiafigure19asa functionof A/b and E for
v = 0.3. Foraninfinitely lonGflange, fi~xe17contains~ asa
function of e alone.Tkeeffectofvaryingv appears infigureZ.
- Theplasticity-reiuction factor
.
~ = @s/E)(1-
Fora longcla~dfla%ge,
fora long hingedflang~ is
)/
Vez (1- &) (A~)
...
!-J].
t
.
.
.
.
.
.
,-
8
For flange c.lxkling reiuction factors wkn
~cr B upl~
it rayh
permissible touseequation.(&O). Alth,.P@l thisfuctor;asnotco.~ute.J
inthesection e~titled Cl@cling !hxluction Factors, itappearstobe
reasonnb:e bycoxp~riso~ withtl.e iectors forpletescmdcolurms.
Platecol.ums.- Fc5rplatecolums., thebuckingstressraybe deter-
minedusingfigure2(a). -,
Fore sh~rtplatecolun (L/b< 1)tkeplasticity-reduction factor
i.s
Fora squareplatecolwzn(L/b= l),
(A7)
,.
(A8]
,
Fora longplatecolum (L/b~1), theplasticity-reduction factor
is
)/
IIS(%/E)@- V=2 (I- - ~) (A9)
The claddingrduction factorfor shortplatecolurns inwhich
andwhen UcrB ~Pl equaticm (4C)kol~strue
4-
3f)
(Ale)
..
..-
whichisalsoapplicable
to long plate colu!!sat all stresslevelsabove 6=1.
ShearBuckling
The shear-buc.kliri-stress coefficient asa function of a/b is
shcwninfigure22J%I.clmpd Endhingedplateedges.
Forlongplates, ,
. .
..
.1
.
.
.
,.,
,
.
.
.
.
withthe
The
equation
For
whenthe
ai~ OStke nondtrenslocal chart of ?igure 10. -
cladding reduc%ion factorfor
GCI< acre ap~ isgivenby
(42),andfor Ucrs Cpl equation (LO)holdstrue.
clawedoblique platesfiKme32(b)~ beussitofind k~
plateedgeangleis45. Fortriargflar-plnte shear-ticklicg
coefficients figure34(b)~y beused.Inadiition, tile section
entitle,d Special Casesshouldbeconsulted.
Bending Buckling
Thebending-buckling coefficient appears infigure% asa fuznction
of a/b and e endinfi~e 25esa function of A/b. Theplasticity-
reduction factorf or a siKp4supported plateisthesaneas fora hinzed
me-
Littleerrorshouldbe expected inusingelastically restrained
f-e phsticlty-reduction factors for elastically restrained platesin
beading.Forthesecasestheplastic-bucklkg chartoff@re 9 raybe
usedtofind CTcr, whichistheEaximum co.qressive stressontheplate
section.Inorder.to firdtheccmresponding rxxrent itisnecessazy to .
i.
integrate thestressdistribution, forwhichpurposethecurvesoffig- !:.
ure 9 may be used.
#
Corbined badi~g
Interaction equations forvarious co-~tilations o?compression, shear,
andbending appearintable5. ~ieseexpressions arepresented in-phi- ..
calformin figures ~ axial 28for elastic buc~ing.Forlongitudinal com-
pression andshearona long rectangular plate,withkothappliedstresses- ,
in theinelastic range,equation(6o)holdstrue:
~e2 +RE2=1 (~)
The plasticity-reduction factorfora simplysuf~r+.ed platein;om-
blnedcompression andsu?ial loadvariesbetween thatfora hingedfl%-$e
~d thatfOra SiEPIY GUppOrtC+dpht~ under=id COInpZWSiO!I}depemii~~
. .
.
Actually, utilization oftheplastic-buc?ding chartof figu!!e 9 forall
CxcsOfcoxbined kcfidi~ anda.ial~O[fi tOfir?dacr
(aftmr whichthe
plateloadingmy be foundby integr&ti~tk.ecross-section stressdM-
tritution) shouldgiveconservative results.
-.
Onright-angle i~osceles triar+y.rlar platesloadedunder shearand
co.zpressionas shown is the sketches in figures 34(a) ami 3h(b), equa-
tion(61)applies:
.
( -)
2T 2 ~ ~
.,
*U
+ Tcr
( -
+ Ucr
U2) = ~
cr+
(61)
Table6 contains nmericaivaluesOf ~ and ~ fordifferent tws .
ofplateedgesupports.
::
I
,
-,;
*
.
.
.
.
.. .
.
b
1.
2.
39
4.
.*-.
5*
6.
7*
8.
9.
10G
11.
12G
13.
14.
. .. ,- P.,, ,,-:
. ..-. ,i ..1-.f . .
Scchler, E.E.,andKewell, J.S.: Pro~essReport m Wtkods of
Analysis Applicable to Monocoque f.ircraft Structures. Tech.Rep.
No.4313,Air-terielComATi,Ar.~AirForces, Cict. l~\7.
Timskenko, S.: TheoryofElastic Stability. Firstcd.,McGraw-HiU
BookCo.,rnc.,1936.
Sokolnikoff, 1.S.: Mathematical Theoryof~asticity.Firstcd.,
McGraw-Hill BookCo.,~AC.,19460
Bleich, FrieHch: Buckling Strength ofMetalStructures. First
cd.,McGraw-Hill lbokCO.,tic.,19~2.
StoweU,Elbridge Z.: A UnifiedTheoryofPlastic B~ckling ofcol~
andPlates.likcA Rep.898, 19k8. (Supersedes NACATN 15%.)
B@iarx@,Bernmt,andHu,PaiC.: TkeLagrangian HiLtiplfer Mettii
ofFinding Wer andLoverL@d.t6toCritical Stresses ofClamped
Plates . NACARep.848,1946.

Ho~bolt, JchII C.,andStouell, Elbridge Z.: Critical StressofPlate

columns . NACAT?f~63,1959.
Lundquist, Euge~e E.,andStowe~,Elbridge Z.: ~ritical. Compressfi(e~
Stressforoutstanding Flmges. HACAPep.734,1942.
Raniberg, Walter, andOsgood, William R.: Description ofstress-st~in . ....
Curvesby ThreeParameters. NAcATNW2, 1943. -
Gerard, George}andWihihorn, Sorrel:A StudyofPoissons Ratiok .
theYieldRegloa.NACATN2561,Jw. 1952. .
Ud.y, M.C.,Sha*~, H.L.,aril Badger,F.W.: &pertiesofBe@li-2.
Nucleonics, vol.U, no.5, M2y1953,pp. 52-59.
Goodran, Stanley, aiidRussell, Stanton B.: Poissons RatioofAir-
craftSheetI.kterial forLargeStrains.Tech.Rep.53-7, Contracz
PO33(038)51-4061, WADCati Nat.Bur.Standards, Feb.19j3.
Stang,A.H.,Greenspnn, M.,afil;e~, S.B.: Poissons Ratioof
someStzmc.-lAlloysforLargeStrains.Res.Paper RP 17L2, Jo>-.
Res.,Nat.Bur.Standards, vol.37, no. 4, Oct.19%,pp..211-221.
Stowell, Elbridge Z.: Compressive Strength ofFlsnges.NACARep.1229,
1951, (Supersedes NACATN2020. )
:
.
.. ...1 ,f ..i
...
.
.
.
,
.
-.
. .
- /.
, .P
-,.
17.
13.
19.
20.
21.
22.
23.
24.
25.
26G
27.
~~rard, G.: SecaatMalulusM2t!mdforIkteninir;g PlateInainbiiity
_Lbove tlieProFortizmal Limit.J~ur.Aerti. Sci.,vol.13,no.1,
Jan.@6, pp.33-44,48.
PridejRichard A.,andHeimerl, GeorgeJ.: ?lastic Buckling of
SiE@ySupported Co~res~edPlate~.XACAIN181~,1949.
Peters, RogerW:: Buckling of Lou SquareTubesin Cm~incdCoIE-
pressionand Torsionand Comparisonz~ith F1.at-Plat.e Buckding
Theories.NkCATN31@+,1954.
Bijlaard, P.P.: TheoryandTestsonthePlastic Stability ofPlates
andShells.Jour.Aero. Sci.,vol.16,no.9,Sept.1943,
PP*5a-541*
Handelman$ G-H.,andPrager, W.: Plastic Buckling ofa Rectangular
PlateUnderEdgeThrusts.NACARep.946,1949. (Supersedes NACA
TN1530.)
I@xshin,A.A.: TheTheoryofElasto-Plastic Swains and ItsAppli-
cation.Bull.Acad.Sci.,URSS,Sec.Tech.Services, June@@,
pp.769-788.
Gerard, G.: Critical ShearStressofPlatesAbovetheProportionRl
I&nit.Jour:Appl.Mech.,vol.15,no.1,Mar.194-8, pp.7-12.
St6uell, ElhrMgeZ.: Critical ShearStressofanInfinitely Img
i
,PlateinthePlastic Region.NACATN168121948.
i
1
StQwell, E.Z.,andPride,R.A.: TheEffectofCompressibility of
theMaterial onPlastic D~ckUngStresses.Rea.3ers Forun,Jour.
Aero.Sci.,vol.1.8, no. I.lj Nov. 1951,p. 773;andAuthors Reply,
Readers Fonun,Jorr.Aero.Sci.,vol.19,no.7,July1952,p.4*. !
. .
Bijlaard, P.P.: TheoryofPlastic Buckling ofFlat.es S& Applica-
tiontoSi@.ySupported PlatesSubjected toBending orEccentric
Compression inTheirPlaue.- Reprint No.55-A-8,A.S.M.E.,
Rev.1955.
Bi~hard,P.P.: WsingAccountOftkeCorpressi=~Lf.ty oftheMate-
rialinthePlastic Buckling ofPlates.Readers Forum,.70ur.
Aero.Sci.,vol.19,no.7,JuIY1952,PP.493-494.
Mc & t j Kenneth P.: Stibility ofAIcladPlates.IWICA TN1986,1949.
.
Eryan,G.H.: OntheStability ofa PlanePlateUnderThrusts inIts
&n-Plane, WithApplications-to theBuckling oftheSfdesofa
I
Ship.Froc.LmdonMath.Sot.,vol.22,WC. ~, 18x, pp.54-67.
, 3;
,.. F...
. . . . . w. . ; . -
. - *
.
.
.
.
.
.-
28. Hill, E. N.:
C~.r-~t ~Orc~f~.~c~l co~.p~essive strcxs Of Fl?3tR~CkALL-.~-
larPltites. KL:ATN733, l;~!J O.
29. Lurdquist, .EuLwfie E.,andStm-ell) Elkridge, Z.: Critical Compres-
siveStressforFlat~cc:~-=tizr PlzcesS~p~orted AIGzgAllEdge:
andElastically Restrair.ei Agahst iwtution Alor?g theUnloadeti
Edges.NACARep.fij,lS~2.
30.Litmve, Charles, andStein,Manuel:CkartsforCritical Cor.binatLLr.s
ofLar@tudinal andTrans-{erse DirectStressforFlatP.ectanguk-cr
Plates.NACAXlL-22h,@6. (Fo~erlyNLCAARRL&05.)
31.Gerard, George:Effective YidthofT-astlcally Supported FlatP19zes.
Jour. Aero.Sci.,vol.13, no.lo,&t. 1946,pp. 518-724.
t
32. fWYZ-iS,J. H., and Dw-me,P. C.: Structural Principles andData.
HandbookofAero.,no.1,pt.2,X52.
33.Southwell, R.V.,andSk-, SilviaW.: OntkeStability Under
Shearing Forcesofa FlatElastic Strip.fiOC.Roy.%C. (Locdon),
ser.A, vol.105, 1924,pp.582-607.
34. Stwell, Elbrldge Z.: Critical ShearStressofanInfinitely Lor&

FlatPlateWi& EqualElastic Restraints Against Rotation
theParallel SSges.NA2A%RL-476,1943. (Forccrly .X4CA
ARR3K12.)
35.Stein,M3nuel, andNeff,J&-An:Buckling Stre5ses ofSimply
Rectangular FlatPlatesinShear.HACAT?{1222,1947.
X. Budiansky, Eerr.ard, andCo=!rzm, RobertW*: %uckling Stresses of
Clfcped Rectangular FlatPlatesinShear.:iACA IN1559, 1348.
37. Schuette, EvanF.,andW2dloch,JarrsC.; ChartsfortheMini-z-
WeightDesignofMultiweb WingsinBendhg.NACAlTT1323,19ki.
%. Johnson, ~amesE.,Jr.,ani, Noel,RotertG.: Criticti Bending Stress
forFlatRectan@arPlatesotipported AlongAllHgesandZlas-
t~cally RestminedAgairs,t.Rotat+on AlongtheUnloaded CorqressXn
. Wge. Jour.Aero.Sci.,vol.20,r~.8,Aug.1953,pp.555-540.
39.libel, RobertG.: Elastic Stability ofSimplySupporte& FlatRec-
tangukrPlatesUnderCritical Combinations ofLongitudinal BeL5i:g,
Longitudinal Co-wressia., andLateral Compression. Jour.Aero.
Sci.,vol.19,no.12,Tee.1952,Pp.8?9-834.
40.Johnson, AldicE.,Jr.,E@ I@chert, Kenneth P.: Critical Co~bir&-
tionsofBendir&, Shear,and!!!ransverse CoWressive Stresses fcr
Buckling ofInfinitely La~gFlat~tes. wicATN2536,1951.
.:
*
.
.
.
.
-.. .....
.. ..:.i A:;, ;. .
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51G
52.
Stcr.-ell, Flbridge Z.,and
s.~Infinitely Low Flat
.
-i
Schwtirt. z, F~iwcrd Z.: Critical Strc3sfor
PlatehlthElastically Restrained Edges
UnderConbl~zL ShcnrandDirectStress.
(Fornerly RX!AARR3Kt3.)
NAMWRL-3$0,1543.
Batdorf, S.B.,atiHoukolt, JohnC.: Critical Ccr~binution3 ofShear
andTransverse DirzctStressforanInfinitely LongFlctFluteWith
~&es Elastically RestminedA@~st Rotction.RACATR847,1346.
EQtdorf, S.B.,andStein,Manuel:Critical Combinations ofShear
andDirectStressforSimplySupprtedRectangular FlatP@tes.
wcA n{1223,1947. .
-.
Grossran, Uorzz: Elastic Stability ofSimplySupported FlatRec-
tangular PlatesUnderCritical t?cmbinations ofTransverse doEPres-
slonandLcmgitudinal Bending.Jour.Aero.Sci.,vol.16,no.5,
Pay1949,PP.272-276.
Stowell, Elbridge Z.: Plastic Buckling ofa LongFlatPlateUnder
Combined S~&arandLongitudinal Coqression.lLtCA TN19, 1949.
Budiansky, Bernard, Stein,Mknuel, andGilbert, ArthurC.: Buckling
ofa LongSquareTubeInTorsion andCompression. NACATNi751,
1948.
\
Levy,Samel,Golden%rg, Daniel, andZlbrltosky, George:simply
Supported LangRectangular PlateUpderCombir.ed AxialLoadti
Xoml tiessure. NACATN949,194.
. .
Wcalley, RuthF!., Corrick, Josephine N.,endLevy,Sazuel:Clamped
LongRect~alarPlateUnderChibined AxialLaadandNorrBlPres-
sure.NACATN1047,1946.
.
Pines,S.,and&ard, G.: htebilityAnelysis andDeGigy. ofan
Efficiently Tapered PlateUnderCompressive Loading.Jour.Aero.
Sci.,vol.14,no.10,Oct.1947,pp. 594-599.
Libove, Charles, Ferdmn,Saul,arsd Reusch, JohnJ.: Elatiic Buckling
ofa SimplySupported PlateUndera Compressive StressthatVaries
#
Linear~intheDirection ofIoeding.NACATN1831,1949.,
Anderson, RogerA.: CkrtsGivingCritical Compressive StressofCon-
tinuous FlatSheetDivided IntoParallelogram-SkiFed Panels.NACA
TN2392,1951.
Wtttrlck, W.H.: Buckling ofOblique PlatesWithClamped &lges Under
Uniform Compression. Rep.94.182,Aero.Res.Labs.,Dept.SUpply
(Melbowne), k W. 1951. (lUso, Aero.Quart., Vol. 4, pt. 11,
Feb.1953,pp.151-163.)
..
.
.
.
G
.
.-
.
i
,
. ..
.
>3.
24G
55.
56.
>7.
58.
59.
60.
61.
62.
Guest, J.,andSil&rsLeit~, J.P.O.: ;,Hoteor,theBuckling of
SimplySqqmrte3%rallelo~ram Plabes.Str~ctures and. I.titcrinl:
Hote20~,Aero.Res.Labs.,Dept.Supply(Mel&~rie), M.y1953.
.
JitiriCk, ?. H.: l!?~:~~irig ofOblique F:ztesWithClamped ZigesU~~er
UniformSkar. Rep.SM.210,Aero.&s. Labs.,Dept.Supply
(Melbourriej, June1953.
Hase@ra,M.: OnB~skling ofa Clamped Rhonb16 Y%fnPlatein Shear.
Readers Fom.m,Jour.Aero.Sci.,vol.21, no. 10, Oct. 1954,
p.720.
Woinowsky-Krieger, s.: ~er dieBieL~ vonPlatten durchEinzelLns~e&
mit rechteckiger .hfstandsftiche. lr#.-ArcMv, ~l. 21,1953$ . --
pp.331-338.
*
Klitchieff, J.14. : ~ckl.ing ofa Triangular Plateby Shearing Forces.
Quart.Jour.Mech.andAppl.Math.,VO1.~, pt.3, 19Z, PP.257-253.
Wittrick, W.H.: B~ckJirg ofs SimplySupported Triangular Platein
Combtne& Compression & She%r.Rep..SM.lm, Aero.Res.Labs.,
Dept.Supply(Melbourne), July1952.
Vittrick, V.H.: Backling ofa F?i@-.&.gled Isosceles Triangular
PlateinCo=klrA-C oxpressian atiSiear(Perpendicular Ed~es
S~ly Supported, $~tenuseClamped). Rep.@. 220,Aero.Res.
Labs.,Dept.Supply(Melburne), Nov.1973.
Kittrick, W.H.: F!uckling ofa Right-Angled Isosceles Triangular.
PlateinCo~bine~ Co~ression andShear(Perper&cular Si~qS
Clamped, Hypotenuse Simply~w~rted). Rep.SK.211,Aeti;Res.
Iabs.,rfipt.Supp2J@elbowxe), Jwne 1953.
Cox, H.L.,and~ein,B.: Ike BucklfrK ofIsosceles Tria~mlar
Piates.Jo~ur. Aero~SCi.,vol.22,n:.5,MaY1955,pp.~21-325.
63.Cozzone, F.P.,and%lcon,M.A.: Ncniimensionsl Bucklir~ Curves-
TheirDevelopment andApplication. Jour.Aero.Sci.,vol.13,
no.10,tit.196,pp.511-517.
64.Anon.:Chemical Co~sitionLimitsfor Alurdnun Alloys.AlcoaDati
Sheet,Aluminuzn Co.ofLm.,Jan.4,1955. -
,,
-.
.,
1
. .
.
TABLEl.-
Invmtigator
BIJlaard
(ref.18)
Neudelnsm-prager
(ref.19)
Ilyushln
(ref.~)
Stowell (rof 8.~
and34)
I
Btrom-strain law
Incrernentoll anddcfor-
mtion types,
v ins*tmeOus
Incrmentsltyp3,
v instantaneous
Defommtlon type,
v R 0.5
Deforsmtlon @pe,
v= 0.5 --
IImA9m--m THmRIEs
.,
,,, ,
.
....
.k.a a-------.-.---, -.$-- ,
,. ..,. . .
1
PIMticlw law
Octahedral.
oc t ahedr al
Octahedral
octahedral
8hear
shear
shear
shear
SuckliW umkl
Nostrain reverml
Strainreversal
Stmin reversal
No strsin rcverml.
$4.
.-
-..
..
t=..
.,
.
d
.
.
.
,.
.
TABLE2.- PLASMCITY-RZLUCTION FACTORSi
Loadir~ Structure
v/d
Compression Img flange,ofieunlcz:ed ~
edgesimplysupportd
Longfla~e,oneunloaded
elge clsuzped .
0.330+ 0.335~ + (3Z@j 1/2
Longplate,bothunloaded
edgessi.np~y supported
O.~ + 0.250~ + (3~/Es~1/2
Icing plate,bothunloaded
edgesCk7,F~
0.352+ 0.32i~+ (~~,E~~l/2
Shortplateloadedesa
coluxm(L/b<< 1)
[(
0.291 + 3q/Esj
b
Squareplateloadedasa
G
column(i@ = 1)
0.114+ 0.886 (~/Es) .
bng column(L/b>>1)
%/%
Shear
e:g%%l:$;:;r::e:d=es
0.83+0.17@/E~)
.

t,
-., -
.,--.,, ,.
RZIXX!!IOI-f
.
I
TAILS3.- SU?WA.RY GF CLADDIK!2
--i
Ec~> Ispl
1
r
.
Loading
Shortplatecolurms
1 + (3Df/4)
l+sf
.
+-
i
l+3f
1
1
Img plate columns
1
li-3f l+3i?
I
Coxrpression and
I
1 +- 3pf
shearpanels
l+3f
1
l+3f
TAELEA.- FORALCLADP~TES CLADDIKG
.
TRzclcmss
reference
,:.:.
IIIt
6~
,,
ii
.,:. iDatatakenfrom
I L
I
..:
MLterial Cladding Totalplate Tbtalc14dding
designation mterial thickness, in. tlnicknes~, f,in.
,.
I
Alclad2014 6U53
<0G04.0 0.23!
2.040
G
.10 .
Alclad2024 12x <0.064 Old
>*@54
.OS
I
.
.
.
. .
Alclad7075
I
Allthicknesses
I
0.09
.. ..
.
.
.
.
,
.
.
.
s
. -
.
TABLE~.- COXBINiZ3 LQADIHG CONDITIONS FORWHICHINTERACTION CU-fl.% EXIST
nleory
Dsstic
nelastic
Loading corcbir.ation Interaction equation Figure
Bi=ialcompression ~ Forplatestir~t. bucklein 28
squarewaves,Rx+~=l
G
Longitudinal cml- Forlongp~lates,
%+~z=l
~
press ionandshear
Lcmgitudina coE-
1
Hone
I
28
pression andbendfng
Eending andshear Rb2+Rs2=l 27
Bending, shear;and None
I
=7
transverse comres6ion
Longitudinal cowression Hofie
I
28 I
andbending and, trans-
versecompression
.!1 .
hnlgicudinal c%=. --ess ion@+ Rs2.1 ,, 29
andshear
#-
. . ,.
.,
4*
;-
.
--%.-...
.,,. -., .,
. . ..- L.. zI
:
.
.
mm 6.- EUCKIWOCO!IFFICEJ iS FORRIGHT-AHGLE ISOSCEL2S TRIANGULAR
PIATESLO&D 311DEPENDENTLX ~{U?~ORMCOZtPRFSSION,
POSITIVE SHEAR,ANDNZXYiTIYE SHE4.R
-.
Edgesupports
? ..
(a)
G& %+ %.
AU edges simply 10.0 62.0 23.2
supported
Sidessimplysupported, 15.6 70.8 *.O
hypotenuse claqed
Sidesclampe~: hyp&- 18.8 80.0 44.0
enusesiE@ysupported
b
%ypotenuse = b Lnfigure34.
.
-.
.
.
.
.
.
-
.
.
-.
i,:.
Ikadin.g
I
Rtge support Coefficient
Compressim SS on all edges ~ = &.o
hi
}
HACARep.733
C on all e3ges ~ . 6.98
(ref.29)
r
,P
SSony=O, y=a, x=O
a
b
Fonx=b ~ = 0.43
L
~ Cony. =0, y=a, x=O
1
MICA Rep.~b
! 11! L ~ = 1.2, ref8)
Fonx=b
G
-...
Shear

1
n, , .
!
SS on all edges
ks =3.33IL4CATN 1222
~ref. 35)
I
t C onallefiges
i
ks= 8.98~~CA~ Q3
t
(ref.43)
B
I 1
I
.
,.
Bending
D :. -
SS onalledges
I
%= 239&A TN1323
C onalledges
h = 41.8.(ref=37)-
H
.
.
t-, . .-. . . .
. ..-. !. L.i ~ll..i
* . ,
i
-.
.
,
.
.
.
.
G

TABLE~.-VALUESOFSHAPEPAF#&EXER n FORSZVZZLEWGIXTi311K MATI.RIALS -

. .
z. .
[ 1
Data takenfromreference 63
n Material
3
Gne-fourth hardtofullhard 16-8stainless steel,. withgrain
One-fourth hard18-8stainless st,e?l, crossgrain
5
One-half hardandthree -fourth,s hard18-9stiim.less steel,
crossgrain

IMllhard18-8stainless steels cross grain

10 2024-Z=nd7075-T alumin=-a12Gy skeetandext~=ion
2024X-T &urdnu%-&oysheet
2021L-350, 2Q24-~1,& ~2k-T86aluezinum-~llcy sIieet
20-LO25 2024-Tahuninu%-alloy extmsions
SAE4130steelheat-trea=d uptalCO, CXXI psiclti,=te stress
~~b~ XX.*-Taluminu!-alloy extrusions
SAZh130steelheat-tre~ted abovela,~ psiultimate stress
m SAE1025(mild)steel
G
.
.
.-.. _.
.
, .._.. ...=
..-
:
.
.
.
-.
.
.
!,.,.
.
TABLE 9.- DESIG:i4TIOl@ FORWRCCIGHT
.
G
.
,
-A. .
lwMINuM -
Old I
14S, R301 2014
2017
X 2CX24
61s 6061
75s
7075
.
,.
:
.7
I
r.
,..
.

..
t
i.(
.
.
.
..
.-.
.-. .
.
-.

I
z
x
1
.
Figure1.-Transition fromcolumntoplateassupports areaddedalong
unloaded edges.
Note changesinbuckleconfiwatlons .
..
.
..
..,;
. . . . . ..
. .
.,.,
.
.. .-r -.
1
..
4
j
IA?--
kg=l~
. - ...- .
W
.. -. . . ..
.
f
i/ b
(a)Platecohms withhingedloadededges.
Ucr= &n%3/(L/P)2.
2s
m
13

.
\
-
-, -.
..
h
(b)
F5.gurc2.- Compressive-bUcklW co~fficients of platecolumnsami
ritio.
. - ----- -.. ----- .,
,+
kc,4r
Pa
@=m_.
0..P5 0.456
.30 ,426
5.35 .395
-. .---
Y J 4
->..
-
-- .-
,!
If
2 3 4 .5
@b
Hi&d il.atqes.
flangesaafunctionsof Fmlsx4fs
..,
F
s
I

. , -,-
., ... --- -
,. .,
,-
.
.,
~~
e- . -. -. -. ..-
,. .-
)
~
n
~. .
G .
S=.* --l
c
(a)Significant stmm quantities ona typical
atre65-5train curve.
Mgure3:-Char=teri.stica ofstress-utrein curves forstructur~ rdl.oys depicthwqwmti%ies
usedinthree-parameter method.
.
%m%
(b) Ikpendenceof shape factoron
~~t~o ao.~opo.ay
n = 1 + 1%(17/7)/io* (Jo .To/%. c>)
,.
-. ..-. --- -- -.
.,

~-. . . . . . . . . . .
-----
,,
,
.
,..
.
12
-i .
n
40
8.
.
6
!
4
.. .-
0
02 4 6
Fl&rc 4.- Mondiwneional
for
.8 # e M 16 18 20
stress -Btraln curveefor vkricm volues of n
three-par+ter method.
)
%?
,,
. . ,,.
-!
.
.
.
.
.
..
..
0
/
&
E
0
O TBT DATAFRO/U20Z4-T4
CWCFOtW SECTIONS
o
THEORETICAL VALUES
[ref. , 14J
1111 1 1 1 I I
Jv m m 40 50 t
Figure5.- Com~isonoftheoryande~rimentforcmpressedsimply
supported flanges.
.
-.
.
.,
I
i
I
..
.-
-,,.
1 * ,
1
T
.. -
J
8
ad
ZEST DAZA FROM 2024- T6 WARE TUBES *
o PRIM and HEIMERL (rd. 16)
. PETERS hf. 17) F
kl~mretY. -
.. ___
al
Cofiq-mrison
t & ,ad
of theoryamlexperhwnt
mpportedplates.
for
=WN end PRAGI
(M t9)
<+
.
\
\
ILnKW {ref. 20)
mzutn (mf J c 7)
\
J nmELL(rel 5)
comprcmmd
-= ., f
-,
.
1
,
...
,, .,. ,,. .
.
.. .
*
40
.8
.6
9
.4
.2
I
I
!
,
I
o
0
, -.-. ..,-- .- .,-> *._, .--. z ~,, . . .
,.
---
..
Ll-
5 1s
rw, w
30 35 40
Fi@re 7.-comperition of,daear buckling theories andtestdateforpla~:,c
buckling of compressed elmtlcdl.ysupported plates.
\*
-4
ir.
,.
1
, ,
,,
v,
#
u
!
~ - .,. .
I
G
2
I I I I I I
u
I
-J-L
.
.
I
? L6 18 20
,.
Figure8.- Churtofnorulimms ioml comprcmlve buckling stressforlong
(Eufl)(l - Ve2)/(Z- v%).
,.1
L:
hln@-d flumTm. TI=
\.,
,.
-. ....- ____ .. . .._ -. . .. . . .
---. . - .-- ..- . ...-
G


,, ,!
.
..-
. .
/2 -
\
a
# -
i!
.
.8
%&- ,
6
#
2 .
*
I I I I I I I I I
o 2 d & .9 lo Lt M 1.6 1. 2
Fdgure 9.- Chat or nondimmmional compres~ivebuckling atrcss ~or lonC
clamped flanges and for supportedpl@es with edge rotationalrestraint.
~= (%/++0..++(3Et/EJ]/}(l - %2)/(1- ,2).
..:
, II
I
i
.
,,
. .
,,
.
-- b
,
4
,
-.. - ...
.7
.6
.5
,4
V%
.3
.
.2
,1
0
! 1
,
,
1
1 i 1
+ i l-l,-{---l&
1-- -t---r242z2z>xz>x i I
I-+--l #!5jjfz~p-.J.. - !
I
-----
,- . ..- ,.$\\\\\\yl .
/? 6 .-H
El
>
~~:1
I
!$
- .-
-1_
p
:1
.1
~:.,,..v..,,i
1
--.-.
,. _ .-.,.. . - ----
~--r T-
,
I 1 I I I I I I
n
2
3
5
10
20
~50
I
(/ ./ ,.? .3 .4 .5 .6 .7 .8 .9 1.0
.:;;, (#e
::
.-.
~,
.
.-f
Fimme 10.- Chartofnomll mensio]wlshearbuckling stresm(orpnnelswLtll
I
LX@ rotatkmul rmtrui nt.
) / ( )
q G ( E@) ( 1 - L# - v?.
. ,
. . 1
i
.
,
----- .--.. -,. . . . . . . . . . . . . . .-,.,...,-, . . . . . . . . .
I
.
ff/2
1
,
.-.
,,,~
,..:
-.
>
ft/2
FiWe I.lt- Cross section of CICU plute.
.
i . .
..
.
,.
G
..
,.
*,.:
;;. , !.
f
f
q cf. ADmvG
.
.-
i
Figure12.- Stress-str~in cu.wesfurcladding, core,andalcladcombina-
tions G
/score
=l-f+fif; $=
acl-/acore
. . . .. . ... . . .. .-.

,,.
* ~.
. . - .. .. . . .
a
mm
6 -
VT
4 o
z+mrkm _
I
U*, hl
(a) simply supported 2024-TfJ4 platevith
3 percentcledding.Tc~t,c Wre EWJIC
on Z- andchannelsect. ionc.
(b)Long E%224-T3 pletcC.MXUM
M.pmcentcleailng.
Figure13.-
Comparison ofexperlmntml @ theoretlca cladding reduction factor~.
:!
.,
,.h
-.1
1-
)
-!
.
.-. .. .
I
., 7, -..,
. .. 1.,, ! J -:: ,,
--, WADED EVGESCLAMPED
LOADEDEDGES SIMPLY
Su?pot?m

9------
-----
J llll~l!l
I ! 1 i I I I
/ 2 J 4 G
Figure14.- Corcpressive-buckling coefficients forflatrectangular plates.
.
-.. . . . ....-. ..
I
i
.. .
... ,. .-.L
. .. .
. .
.
r
-J
m
m
76
Z2
k
G .
68
64
60
*
fie
.
45.67B9D L1
1.2 1.3 /Y
FISH 12.- Cocqmessive-buckli~~-s tresscoefficient of platesasa
Zliir; Cf A/b forvariousamounts
Oredgerotational restraint.
uretakenfromreference 29.
6
func -
Fig-
1
..[.
..-
. . . :,
. .
..
.
t
-
. .
.
.
/ 5
M
/3
P
//
/0
.
9
8
k=
7
6
5
8
4
3
2 )
/
i
t
---
I
-
-
Ez3
Figure16.-
tionof
.
%
Compressive-buckdlng-st.ress coeffi.ci~~t ofplatesasa fuzc-
a/b forvarious smounts ofedgerotaticmal restraint.
G
v
. . --- -. --- - ..- ... . .
.
- -..*.
.- ---, .... !
/4
.!2
lo
l& ,8
fz4wGE
.6
.2
.
0
. . ..
u ./ #
14
.
$y:o

FLAME
(J-m)
ko.0,400
.4r G0
.
In m IcJ n lot
. , ---
Fun? @
FlgurI? 17.- Compressive-bucklingcoefficientsfor Infinltely long fhnLcs
end pl.ate8ae functionsof edge
restraint. rotational.
I i ?
m
8
4
14X7E
6
4
2
.. -,
I
1
.
.
*
,,
-! *
.Y. . .*.. ..* ----
4 --.. . . .
i
I
i
1
I
!
.
G
-.
7W7SIONALLY
1 -
STRONG S?Wt7M9
(HAT)
6
WEAK SWFENER
.-
5
/
.
4
.I!.o?wEDGESSlhmr
SUPFW?TED&=4.00)
3 :-
- +-4
~DGESCLAMtWD
.
(&496)
I
I
!
I
1 1 1 1 1 1 1 \
! I
I 1 (
s o Iwo i50 zoo 250 300
b/t
.
Figure 18. - Ccqre6slvc-buckl@ coefficient forlonGructan@bmstiC-
ftm!d PIUW1 O no a funeticm of b/t and ntiffencr tor~ioml rl@iiLy.
FiglucLukm lranrc.icruncc 51.
. .
t,
.,
..-
.:
!-,,
______ .
1. ..
I
1,
w
.
.
.
..
.
.
.
4!7l
4
I
4-I-L
.
7
6
5
4
Figure 19.- Compress ive-buckling-stress codficient offlanges asa func-
tionof A/b forvarious .sc ountsof edgerotational. restraint.
G
....
:;,-,\.,., ii.:, ,.-
,.
II
\\
. ..
.. i- t
. . . .
t
/4
/3
/0
.9
.5
!$(
u,
b
azzE
////.///,...////
. .. - m. . ...! -.L.
1
LOUD EOGESCLAM%Z
------ -.-
I
.. .
...
I
!

A! ~ I t
.5 u [5 2025 303 40 4-5 5
Fimre 20.- Compress ive-buckling-stress coefficient offlanges asa func-

tion of a/b forvarious ~zounts ofedgerotationalrestrcint.

1
t .

. ...
. - ------ --. F.-.- v, ,- , ........ .-
----
.. 1.-
*
b,
4
. .
,. : .-~-r--.-=.=-=,
E
i
.
6
.
.
5
3
I
.
F 1
.
I
I
R&we 21.. Compressive-buckling coefficient offlatpla~s
restrained cqsinst lateral e~snsion.
Poissons ratio
equ91S0.3;~~x =
WaM W
.
, ----
.
-. - -- .. . .-.?.. .
~~.., .- ----
I
*
.
.
.
.
-
.
.
.
/5
/J
*
9
7
5,
t
SYMMETRIC
MOE
LAMPkD EDGES
AmfsYHMETRfc SYMM5%7C
MODE
/YVGED EDGES
!
/
I
3 5
0/3
Fig~e22.- Shesr-buckl~--stress coefficient ofplatesasa function
a/b forckupe~sndhingededges.
of

.
. ., -------- ...
-. --. 9. -. -
-
I
i
. .,
.,,..,-. -
.,
i
.
.
.
.
I
.
I
.
I
i k,*898
9 -
ATE~W
i
!
8
k
%
7 -
IQ m
-.-
0 2 4 6 ,8 m
(b) k+s=
.
as a qJllctionof b/a.
~Qu= z3=-
Cu r ve s for cstimtion of sh~=-buckli~ coifficimt of plmtes
with vsrious amnm% of eti&crotatioml restraint.
. .
.
.
.
.
.
t
.
.
.
.
.
-
.
\
J
/25
/20
//..
1[0
.Q5
A2i9
S?5
9.0
k
85
a
z
Zo
6S
6.0
55:
I
so I
\ ~ ~.... ......- ------- .. : }s
#.68 A2 1 2 Mt # 5 bg 20 ez 24
A,
b
Figure24.- Sheer-buckling-stress coefficient forplatesobtatned Cfon
analy. isofinfinitely 10%plateeasa function of A/b forvarhms
aaounts ofcc@.rotational restraivt.
G
..
.
.. -
--
.-. _
. . .
,,. .
.
.
G

.
b
i
.
G

.
G

.,
i
i
J
t
t
.
i
52
51
45
47
45
43
4f
39
k ,
37
35
33
3/
23
27
25
23
i
Figure25.-
1
I
~
I
- ......
a
!\ \YI I\/! M42WW &bAr
,,1 I I I
!\ll.
I I
;: II
1354555 65Z3W
S!!5 In5 / [5 m
A/
b
asa function of A/b
I
restraint.
.-
-
I
.
.
.
.
.
.
.
k b
5 6
I
5 2 -
Izz%
-.- ?----- ---
32-
___- .... -----
24-
1
m-
/6
/2
8 ,
4
L
1
wtii~-bucm~ coefficient ofplatesasa~function of
e
=/b
F@.lre6*-
for}tiiomSr.OuntS
of edgerotat ioml restra~t.
_... -- -3 ._
. ..
----- ------
.,i...- ., I
c-) -,
..
4
.
.
.
.
.
.
0? 468
r o
Ei !i !!!
Ov
l ,
J ? .- .,. . .
i.,
< ---:--,-- - -
R= :!
6
.. ~..~. !. . .
[
> b
8
.
I
, ;...
fo
!
(a)Upperandlo*r edgessir.ply supported.
m
Jo
...-.+ .- a
t
,,
. . . ---.
,,
I
Is
-:. - . y J
i 1
.. >.-
-t- -; ~
I
o
10 8
6/?=4 * 0
3
oz4u*nrz
B
o
~;
2 .-1--;
t
i +
II !.
# . 4 -
f --- --i
~i;
#- +. +. ;
;iO* 1
, ._. L -.f
I ,:t
1 --
Rc
Upperedgessimplysupported, loweredgescls=peL
InteractIon curvesforlongflatplatesuniier variouE cm--
binatlons ofcompression, bending, andshesr.
d
4
.
.!
1.
.. . .. .
e
.
.
0.2.4 -44W
Rr
(a) a/b= 0.8.
(b) a/ ii = LO.
10
s
%
4
2
0
(d) a/b= 1.60. (e) a/b
= 2.0.
RY -
(C) a/b= l.=.
(f) a~b= 3.0.
(g) a/b= =.
Figure28.- Interaction curvesforflatrectangular platesunderconbir.ed
bin.xisl-compres sionamilongitudinal-bend~ng loadings.Curvestaken
fromreference 39.
. ..._
.-. -.. . .. . .. . .
..-. .
.! .,,...
., -,.,, ,, :.1. I
,.
.

G
i
!
i.
. i
TwZM57?CALCWVES
~STIC FLATPLATE
R=
ZEST RESULTS
---- ELAS?7CFLAr PtiTE
.
E~ST1~ SQUARETuBE G PLAST7CRANGE
--
/?=+/?==/
G ELASW RANGE
.
Figure 29.- Cor.perison ofplastic interaction theory and test data for
cozbined axialcompression andshearonflatplates.Testsweremade
with2014.T6 squaretubes.
. .. ... ., . %
. .,-
. ...- .-. - -.
,
------- .-
Eiz3ii
,- ------ --
&l
6 -
5 ~
.
Fill
I
I
% 5 10 Ls Sf2 25 30 Js 4D 45
:
o/b
Figure~.. Compress ive-buckling-stress coeff Icientforn ElmplysupporLed
rectangular flatpl.ute ofminimumweight.Thickness andl.ondl@vary
1
I
(
o
t
.
.. 7.,.
-., ,,: .- l ,l ?
f
6
5
4
3
2
f
,
-.

A
-..
--:-- :
m
~
i
-.. --
0
.
+ *,rj 15 ~~!Z.=
-J
i
!
/
Ss
I.

I
/
t
I
f
I
{
/fcm A7
u/ b =m
au
i
/
I
.
--- -. w. --..
--. -.-~. ---- ... ---- m. -...,--.
-.. .-.
\
I I I
.
G
2 3 4
0/ .
load. C?av=
:
i
I
I
7? -+-i-
.-.,,
.-.--,.. .,
.*
..
,- : ,*
-{s
. ---
,..
,.-
~a >
. .
.,.
-.,
5
-.
. .
db
.
. ..:+
(a) Lmd.lng inx-direction.
e.:;
#
[ I I
I .1
3 I I
Figure
.
uzib
.
(b)Loading iny-direction.
3~m-
Compressive-buckling coefficients forflatsheetonnon-
deflecting supports dlvlded fitopmtielogem-shaped panels.JKIJ
penelsidesnreequal.
..
. .. . .
~
.. +.
J .-
..:
,. L.
. .
%--
. ...
.
*
-. .- . y
. . . ......,.:
.* ..,:-
.. -. -,
,,-
>
.-
-.
---
-< -
,.; . .
.-
-.
.
f
.
,,
-.
,- .-,
/i
. .
. .
4.:
P lQ/&
1-1
.4
F
.,
..
.-.
.
. ,.
..
u
.(
. . ...-
. .
.4 : ..,=---6
. . / 0
.. ;
. . .
. .
-- . .-
.. --.:
-.
d
}.
. . -. ... .
a,
1-:=
..>-.
.
...
.
-.
*
- (c)C=b$nedz@-=mrse los.di&.
i .-----
!, --- .
z.
.-,-.
-=we 32.- coIxckd,d. ~ -
.
.,~ ;;-. . . . .
.....
f
a- .
..- ,-..
... -
,.
4
.-: .
..
,4
.
,-=:- .
,-.
1
.
-. .. .. . . .. .
-. .
- .- .-
.
k
)
.
. F.
i
I
2C
.. - - - - -
. -
/s
4
a
s
4
..
.. .... ,-.
. .
.+ .
.-.
... ----- ..: .<.
.. . . ..- .
- :7? T-+L -
.s -: .- -.- :
w+. . .L. , .-
. .. .
. . .
-., ----- ._.-;.-.. . ..
-:
, i > 4;..
. . 1 . --r.g.--.
. .. ...
%. --= -. -..,. .!
.a#-- - .:%
.... .
. . .
. . .
.-
.,
. .
,.-. . .
{a)-tiwressim lciding. -
. ..-.
., -..
... , ., --,,.
.
,. .
:,
,..,-
,.
,. t.,
.. .
..
.,
. .,
1
;
.,
1
8A. . . .
. .-
(b)Shem1.oa&g. - -..
..
mgure33.-. Buddingcoefficient ofclamped bblique flatplstes.... --
.
.,
.,
.
-.
---
. . ... . ... .
-Q* , .*
, ,
--~ -. . . . . . .
+ --- ----- .* ----- .,.+ -. ..., .._ ... ...,. - .,
I
60
so
m .
30
-,.
20
.
10 -
0
, I 1 I
0
am
(a) W3iform compreEsicm.
.
---4
F
SW%.Y SIJPPOR7D
----
,.,
O/b
(b) Shmr
Figure 34. - Buc k l i ng coefficients for iBo6celes trianguhr plates.
/
d-. .4-
.,.
.,. . . . .- . . .
P
H&we 35, - K/k M u Sunctlon Of Poj.snonsrutlo. K/k = #/12(1 - @) .
-.. .. . . . ._, . . . .. . .
7
6
s
4
K/k
3
2
1
,0