Geometrically Nonlinear Analysis of Composite Laminates J . N. Reddy and C. F. i i u NASA Contractor Report 4056 A Higher-Order Theory for Geometrically Nonlinear Analysis of Composite Laminates J. N. Reddy and C. F. Liu Virginia Polytechnic Institute altd State University BZacksburg, Virginia Prepared for Langley Research Center under Grant NAG1-459 National Aeronautics and Space Administration Scientific and Techni cal Inf ormat i on Branch 1987 Table of Contents Page ABSTRACT ......................................................... 1 . I NTRODUCTI ON ................................................. 1.1 Background .............................................. 1.2 Review of Li terature .................................... 1.3 Present Study ........................................... 2 . FORMULATION OF THE NEW THEORY ................................ 2.1 Kinematics .............................................. 2.2 Displacement F i el d ...................................... 2.3 Strain-Displacement Relations ........................... 2.4 Constitutive Relations .................................. 2.5 Equations of Motion ..................................... 3 . THE NAVI ER SOLUTIONS ......................................... 3.1 I ntroduction ............................................ 3.2 The Navier Solutions .................................... 4 . MI XE D VARIATIONAL PRINCIPLES ................................. 4.1 I ntroduction ............................................ 4.2 Variational P ri nci pl es .................................. 5 . F I NI TE ELEMENT MODEL ......................................... 5.1 I ntroduction ............................................ 5.2 Finite-Element Model .................................... 5.3 Solution Procedure ...................................... 6 . SAMPLE APPLICATIONS .......................................... 6.1 I ntroducti on ............................................ 6.2 Exact Solutions ......................................... 6.3 Approximate (Finite-Element) Solutions .................. 6.3.1 Bending Analysis ................................. 6.3.2 Vi brati on Analysis ............................... 7 . SUMMARY AND RECOMMENDATIONS .................................. 7.2 Some Comments on Mixed Models ........................... 7.3 Recommendations ......................................... 7.1 Summary and Conclusions ................................. REFERENCES ....................................................... APPENDIX A: Coefficients of the Navier Solution ................. APPENDIX B: Sti ffness Coefficients of the Mixed Model ........... i v 10 10 12 16 17 19 24 24 26 28 28 29 34 34 34 40 46 46 46 57 57 68 73 73 74 74 75 82 85 iii A HI GHER- ORDER THEORY FOR GEOMETRI CALLY NONLI NEAR ANALYSI S OF COMPOSI TE LAMI NATES J . N. Reddy and C. F. Li u Depar t ment of Engi neer i ng Sci ence and Mechani cs B1 acksbur g, Vi r gi ni a Vi r gi ni a Pol yt echni c I nst i t ut e and St at e Uni ver si t y ABSTRACT A r ef i ned, t hi r d- or der l ami nat e t heory t hat account s f or t he t r ansver se shear st r ai ns i s devel oped, t he Navi er sol ut i ons ar e der i ved, and i t s f i ni t e el ement model s ar e devel oped. The t heor y al l ows par abol i c descr i pt i on of t he t r ansver se shear st r esses, and t her ef or e t he shear cor r ect i on f act or s of t he usual shear def or mat i on t heor y ar e not r equi r ed i n t he pr esent t heory. The t heor y al so account s f or smal l st r ai ns but moder at el y l ar ge di spl acement s (i . e. , von Kar man strai ns). Cl osed- f or msol ut i ons of t he l i near t heor y f or cer t ai n cr oss- pl y and angl e- pl y pl at es and cr oss- pl y shel l s are der i ved. The f i ni t e- el ement model i s based on i ndependent appr oxi mat i ons of t he di spl acement s and bendi ng moment s ( i . e. , mi xed f ormul at i on) , and t her ef or e onl y C"- appr oxi mat i ons ar e r equi r ed. Fur t her , t he mi xed var i at i onal f or mul at i ons devel oped her ei n suggest t hat t he bendi ng moment s can be i nt er pol at ed usi ng di scont i nuous appr oxi mat i ons ( acr oss i nt er el ement boundari es). The f i ni t e el ement i s used t o anal yze cr oss- pl y and angl e- pl y l ami nat ed pl at es and shel l s f or bendi ng and nat ur al vi br at i on. Many of t he numer i cal r esul t s pr esent ed here f or l ami nat ed shel l s shoul d ser ve as r ef er ences f or f ut ur e i nvest i gat i ons. V 1. I NTRODUCTI ON 1.1 Backgr ound The anal yses of composi t e l ami nat es i n t he past have been based on one of t he f ol l owi ng t wo cl asses of t heori es: ( i ) Thr ee- di mensi onal el ast i ci t y t heor y ( i i ) Lami nat ed pl at e t heor i es I n t hr ee- di mensi onal el ast i ci t y t heor y, each l ayer i s t r eat ed as an el ast i c cont i nuumwi t h possi bl y di st i nct mat er i al pr oper t i es i n adj acent l ayers. Thus t he number of gover ni ng di f f er ent i al equat i ons wi l l be 3N, wher e N i s t he number of l ayer s i n t he l ami nat e. At t he i nt er f ace of t wo l ayer s, t he cont i nui t y of di spl acement s and st r esses gi ve addi t i onal r el at i ons. Sol ut i on of t he equat i ons becomes i nt r act abl e as t he number of l ayer s becomes l arge. I n a ' l ami nat ed pl at e t heor y' , t he l ami nat e i s assumed t o be i n a st at e of pl ane st r ess, t he i ndi vi dual l ami nae ar e assumed t o be el ast i c, and per f ect bondi ng bet ween l ayer s i s assumed. The l ami nat e pr oper t i es ( i . e. st i f f nesses) ar e obt ai ned by i nt egr at i ng t he l ami na pr oper t i es t hr ough t he t hi ckness. Thus, l ami nat e pl at e t heor i es ar e equi val ent si ngl e- l ayer t heor i es. I n t he ' cl assi cal l ami nat e pl at e t heor y' (CLPT), whi ch i s an ext ensi on of t he cl assi cal pl at e t heor y ( CPT) t o l ami nat ed pl at es, t he t r ansver se st r ess component s ar e i gnor ed. l ami nat e pl at e t heor y i s adequat e f or many engi neer i ng pr obl ems. However , l ami nat ed pl at es made of advanced f i l ament ar y composi t e mat er i al s, whose el ast i c t o shear modul us r at i os ar e very l ar ge, ar e suscept i bl e t o t hi ckness ef f ect s because t hei r ef f ect i ve t r ansver se shear modul i ar e si gni f i cant l y smal l er t han t he ef f ect i ve el ast i c modul i The cl assi cal 1 2 dsn along fiber directions. render the classical laminated plate theory inadequate for the analysis of thick composite plates. These high ratios of elastic to shear moduli The first, stress-based, shear deformation plate theory is due to Reissner [l-31. The theory is based on a linear distribution of the , n = 0,1,2, ... s=O inplane normal and shear stresses through the thickness, where (al ,u ) and a6 are the normal and shear are the associated bending moments (which are 2 stresses, (M1,M2) and M6 functions of the inplane coordinates x and y), z is the thickness coordinate and h is the total thickness of the plate. The distribution of the transverse normal and and u ) is determined from the equilibrium shear stresses (a3, a4 equations of the three-dimensional elasticity theory. 5 The differential equations and the boundary conditions of the theory were obtained using Castiglianos theorem of least work. The origin of displacement-based theories is attributed to Basset [ 4 ] . Basset assumed that the displacement components in a shell can be expanded in a series of powers of the thickness coordinate 5. For example, the displacement component u 1 along the c1 coordinate in the surface of the shell can be written in the form (1.2a) where c1 and c2 are the curvilinear coordinates in the middle surface of the shell, and uin) have the meaning (1.2b) 3 Bassets work did not receive as much attention as it deserves. In a 1949 NACA technical note, Hi ldebrand, Rei ssner and Thomas [ 51 presented a displacement-based shear deformation theory for shells (also see Hencky [6]). They assumed the following displacement field, principle of minimum total potential energy. This approach results in t five differential equations in the five displacement functions, u, v, Y. w, bX, and 4 The shear deformation theory based on the displacement field in Eq. (1.3) for plates is often referred to as the Mindlin plate theory. Mindlin [7] presented a complete dynamic theory of isotropic plates based on the displacement field (1.3) taken from Hencky [6]. We shall refer to the shear deformation theory based on the displacement field (1.3) as the first-order shear deformation theory. A generalization of the first-order shear deformation plate theory for homogeneous isotropic plates to arbitrarily laminated anisotropic plates is due to Yang, Norris, and Stavsky [8] and Whitney and Pagano [9]. Extensions of the classical von Karman plate theory [lo] to shear deformation theories can be found in the works of Reissner [11,12] Medwadowski [13], Schmidt [14] and Reddy [15]. Higher-order, displacement-based, shear deformation theories have been investigated by Librescu [16] and Lo, Christensen and Wu [171. In 4 these higher-order theories, wi th each addi ti onal power of the thickness coordinate an addi ti onal dependent unknown i s introduced i nto the theory. Levinson [ 181 and Murthy [ 191 presented thi rd- order theori es that assume transverse i nextensi bi l i ty. were reduced t o f i ve by requi ri ng that the transverse shear stresses vanish on the bounding planes of the plate. the equilibrium equations of the fi rst- order theory i n thei r analysis. As a resul t, the higher-order terms of the displacement f i el d are accounted f or i n the cal cul ati on of the strai ns but not i n the governing di fferenti al equations or i n the boundary conditions. These theori es can be shown (see Librescu and Reddy [ZO]) t o be the same as those described by Reissner 11-31. Recently, Reddy [21-231 developed a new third-order pl ate theory, which i s extended i n the present study t o laminated shells. The nine displacement functi ons However, both authors used 1.2 Review of Li terature Shell structures are abundant on the earth and i n space. Use of shel l structures dates back t o ancient Rome, where the roofs of the Pantheon can be cl assi fi ed today as thi ck shells. Shell structures f or cal and the a long time have been bui l t by experience and i ntui ti on. No l og sci enti fi c study had been conducted on the design of shel l s unti eighteenth century. The earl i est need for design cri teri on f or shel l structures probably came wi th the development of the steam engine and i t s attendant accessories. However, it was not unti l 1888, by Love [24], that the f i r s t general theory was presented. Subsequent theoreti cal ef f orts have 5 been di r ect ed t owar ds i mpr ovement s of Love' s f or mul at i on and t he sol ut i ons of t he associ at ed di f f er ent i al equat i ons. An i deal t heor y of shel l s r equi r e a r eal i st i c model i ng of t he act ual geomet r y and mat er i al pr oper t i es, and an appr opr i at e descr i pt i on of t he def or mat i on. Even i f we can t ake car e of t hese r equi r ement s and der i ve t he gover ni ng equat i ons, anal yt i cal sol ut i ons t o most shel l pr obl ems ar e never t hel ess l i mi t ed i n scope, and i n gener al do not appl y t o ar bi t r ar y shapes, l oad di st r i but i ons, and boundar y condi t i ons. Consequent l y, numer i cal appr oxi mat i on met hods must be used t o pr edi ct t he actual behavi or . Many of t he cl assi cal t heor i es were devel oped or i gi nal l y f or t hi n el ast i c shel l s, and ar e based on t he Love- Ki r chhof f assumpt i ons ( or t he f i r st appr oxi mat i on theory): undef or med mi ddl e sur f ace r emai n pl ane and nor mal t o t he def or med mi ddl e sur f ace, (2) t he nor mal st r esses per pendi cul ar t o t he mi ddl e sur f ace can be negl ect ed i n t he st r ess- st r ai n r el at i ons, and ( 3 ) t he t r ansver se di spl acement i s i ndependent of t he t hi ckness coor di nat e. The f i r st assumpt i on l eads t o t he negl ect of t he t r ansver se shear st r ai ns. Sur veys of var i ous cl assi cal shel l t heor i es can be f ound i n t he wor ks of Naghdi [25] and Ber t [26]. These t heor i es, known as t he Love' s f i r st appr oxi mat i on t heor i es ( see Love 1241) ar e expect ed t o yi el d suf f i ci ent l y accur at e r esul t s when ( i ) t he l at er al di mensi on- t o- t hi ckness r at i o (a/ h) i s l ar ge; ( i i ) the dynami c exci t at i ons ar e wi t hi n t he l ow- f r equency r ange; ( i i i ) t he mat eri al ani sot r opy i s not sever e. However , appl i cat i on of such t heor i es t o l ayer ed ani sot r opi c composi t e (1) pl ane sect i ons nor mal t o t he 6 shells could lead to as much as 30% or more errors in deflections, stresses, and frequencies. As noted by Koiter [27] , refinements to Love's first approximation theory o f thin elastic shells are meaningless, unless the effects of transverse shear and normal stresses are taken into account in the refined theory. The transverse normal stress is, in general, of order h/R (thickness to radius ratio) times the bending stresses, whereas the transverse shear stresses, obtained from equilibrium conditions, are of order h/a (thickness to the length of long side of the panel) times the bending stresses. Therefore, for a/R < 10, the transverse normal stress is negligible compared to the transverse shear stresses. Ambartsumyan [28,29] was considered to be the first to analyze laminates that incorporated the bending-stretching coupling due to material anisotropy. The laminates that Ambartsumyan analyzed are now known as laminated orthotropic shells because the individual orthotropic layers were oriented such that the principal axes of material symmetry coincided with the principal coordinates of the shell reference surface. In 1962, Dong, Pister and Taylor [30] formulated a theory of thin shells laminated of anisotropic shells. presented an analysis of laminated anisotropic cylindrical shells using Flugge's shell theory [32]. A first approximation theory for the unsymmetric deformation of nonhomogeneous, anisotropic, elastic cylindrical shells was derived by Widera and his colleagues [33,34] by means of asymptotic integration of the elasticity equation. homogeneous, isotropic material, the theory reduced to the Donnell's Cheng and Ho [31] For a 7 equation. An exposition of various shell theories can be found in the article by Bert [261 and monograph by Librescu [35]. The effect of transverse shear deformation and transverse isotropy as well as thermal expansion through the shell thickness were considered by Gulati and Essenberg [361 and Zukas and Vinson [37]. [38] presented a theory applicable to layered, orthotropic cylindrical shells. theory. displacements in the surface of the shell are expanded as linear functions of the thickness coordinate and the transverse displacement is expanded as a quadratic function of the thickness coordinate. Recently, Reddy [40] presented a shear deformation version of the Sanders shell theory for laminated composite shells. constant transverse shear stresses through thickness, and therefore require a correction to the transverse shear stiffness. Dong and Tso Whitney and Sun [391 developed a higher-order shear deformation This theory is based on a displacement field in which the Such theories account for As far as the finite element analysis of shells i s concerned, the early works can be attributed to those of Dong [41], Dong and Selna [ 42) , Wilson and Parsons [43] , and Schmit and Monforton [ 441. The studies in [41-441 are confined to the analysis of orthotropic shells of revolution. composite shells include the works of Panda and Natarajan [45], Shivakumar and Krishna Murty [46], Rao [471, Seide and Chang [481, Venkatesh and Rao [49], Reddy and his colleagues [15,50-511, and Noor and his col leagues [ 52-541. Other finite element analyses of laminated anisotropic 8 1.3 Present St udy Whi l e t he t hr ee- di mensi onal t heor i es [ 55-60] gi ve mor e accur at e r esul t s t han t he l ami nat i on ( cl assi cal or shear def or mat i on) t heor i es, t hey are i nt r act abl e. For exampl e, t he ' l ocal ' t heor y of Pagano [60] r esul t s i n a mat hemat i cal model consi st i ng of 23N part i al di f f er ent i al equat i ons i n t he l ami nat e' s mi dpl ane coor di nat es and 7N edge boundar y condi t i ons, wher e N i s t he number of l ayer s i n t he l ami nat e. The comput at i onal cost s, especi al l y f or geomet r i cal l y nonl i near pr obl ems or t r ansi ent anal ysi s usi ng t he f i ni t e el ement met hod, pr ecl ude t he use of such a t heor y. 631, the r ef i ned pl at e t heor y pr ovi des i mpr oved gl obal r esponse est i mat es f or def l ect i ons, vi br at i on f r equenci es and buckl i ng l oads f or l ami nat ed composi t e pl at es. f i ndi ngs, deal s wi t h t he ext ent i on of t he t hi r d- or der pl at e t heor y of Reddy [21-231 t o l ami nat ed composi t e shel l s. The t heor y al so account s f or the von Kar man st r ai ns. The r esul t i ng t heor y cont ai ns, as speci al cases, t he cl assi cal and f i r st - or der t heor i es of pl at es and shel l s. Mi xed var i at i onal f or mul at i ons and associ at ed f i ni t e- el ement model s ar e devel oped i n t hi s st udy. The si gni f i cant and novel cont r i but i ons of t he r esear ch conduct ed ( i n addi t i on t o t hose r epor t ed i n t he f i r st year ' s r epor t [23]) are: As demonst r at ed by Reddy [21-231 and hi s col l eagues [61- The pr esent st udy, mot i vat ed by t he above 1. The f or mul at i on of a new t hi r d- or der t heor y of l ami nat ed shel l s t hat account s f or a par abol i c di st r i but i on of t he t r ansver se shear st r esses and t he von Kar man st r ai ns. 2. The der i vat i on of t he exact sol ut i ons of t he new t heor y f or cer t ai n si mpl y suppor t ed l ami nat ed composi t e shel l s. 9 3. The development of a mixed variational principle for the new theory of shells, which yields as special cases those of the classical (e.g., Love-Kirchhoff) and the first-order theory. 4. The development of a mixed, Co-finite-element and its application to the bending and vibration analysis of laminated composite shells. 2. FORMULATION OF THE NEW THEORY 2.1 Kinematics Let ( c 1, &2, s ) denote the orthogonal curvilinear coordinates (or shell coordinates) such that the curvature on the midsurface s=O, perpendicular to the surface s=O the lines of principal curvature el - and c2-curves are lines of and 5-curves are straight lines . For cylindrical and spherical shells coincide with the coordinate lines. The values of the principal radii of curvature of the middle surface are denoted by R1 and R2. The position vector of a point on the middle surface is denoted by r, and the position of a point at distance 5 from the middle surface is denoted by R. .., The distance ds between points ( E , ~ , ~ ~ , O ) and (cl+dcl, c2+dc2,0) is determined by (see Fig. 2.1) .., (ds)' = dr I dr .., 2 2 = +a2(dc2) ar I where dr = sldcl +g2dc2, the vectors g1 and g2 (gi =-) are tangent I . . , I aci to the t1 and c2 coordinate lines, and a1 and a2 are the surface metrics .., .., I I The distance dS between points ( c1,c2,c) and ((l+dc1,c2+dc2,s+dr) is given by (dS)2 = dR dR I 11 Figure 2. 1 Geometry and stress resultants o f a shel l 12 aR aR - ac1 at 2 2 as - aR - wher e dR =- de1 +- dg + coef f i ci ent s ds, and L1, L2 and L3 ar e t he Lame aR aR . . . I t shoul d be not ed t hat t he vect or s s1 E and G E ar e par al l el -2 ac2 t o t he vect or s g1 and g2, r espect i vel y. - -, 2.2 Di spl acement Fi el d The shel l under consi der at i on i s composed of a f i ni t e number of or t hot r opi c l ayer s of uni f or mt hi ckness ( see Fi g. 2.2). Let N denot e t he number of l ayer s i n t he shel l , and ck and s ~ - ~ be t he t op and bot t om 5- coor di nat es of t he k- t h l ayer. Bef or e we pr oceed, a set of si mpl i f yi ng assumpt i ons t hat pr ovi des a r easonabl e descr i pt i on of t he behavi or s ar e as f ol l ows: 1, t hi ckness t o r adi us and ot her di mensi ons of shel l ar e smal l . 2 . t r ansver se normal st r ess i s negl i gi bl e 3. st r ai ns ar e smal l , yet di spl acement s can be moder at el y l ar ge compar ed t o t hi ckness Fol l owi ng t he pr ocedur e si mi l ar t o t hat pr esent ed i n [ 21] f or f l at pl at es, we begi n wi t h t he f ol l owi ng di spl acement f i el d: wher e t i s t i me, (U,V,.) ar e t he di spl acement s al ong t he ( g1, c2, c) 13 f U a, > 0 7 0 -0 ii 14 coordinates, (u,v,w) are the displacements of a poi nt on the middle surface and $1 and $2 are the rotati ons at 5 = 0 of normals t o the midsurface wi th respect t o the c2- and cl-axes, respectively. A l l of (u,v,w,~l,~2,bl,b2,el,e2) are functions of el and 6 only. The 2 parti cul ar choice of the displacement f i el d i n Eq. (2.5) i s di ctated by the desire to represent the transverse shear strai ns by quadratic functions of the thickness coordinate, 5, and by the requirement that the transverse normal s trai n be zero. The kinematics of deformation of a transverse normal i n various theories i s shown i n Fig. 2.3. The functions bi and ei w i l l be determined using the condi ti on that the transverse shear stresses, ul3 = and bottom surfaces of the shell: and aZ3 = vanish on the top u4 These conditions are equivalent to, f or shel l s laminated of orthotropi c layers, the requirement that the corresponding strai ns be zero on these surfaces. radi i of curvature are given by The transverse shear strai ns of a shel l wi th two pri nci pal a i 1 a i u =- +- - - - 5 as a1 a t 1 R~ - - U 1 aw u - +41 +2sal +3s 2el +- - - - R1 OL1a% R1 (2.7) 1 aw v V - - - +b2 +2cb2 +3s2e2 +-- - - R2 O2 a% R2 Setting E5(c1,c2,k 7 h ,t) and ~q( tl rS 29 k 7 h ,t) t o zero, we obtai n 15 1 UNDEFORMED DEFORMED IN CLASSICAL (KIRCHHOFF) THEORY DEFORMED IN THE THIRD ORDER (PRESENT) THEORY Figure 2.3 Assumed deformation patterns o f the transverse normals i n various displacement-based theories. 16 - 4 1 aw - 4 1 aw e 1 - - 2 ( 41 +- -) e2 - - 2 (@2 +(r -) 3h O 1 a c l 3h 2 ac2 Substituting Eq. (2.8) into Eq. (2.5), we obtain 5 - v = (1 +-)v R, c This displacement fie 3 4 1 aw 3 4 1 aw 3h (2.9) d is used to compute the strains and stresses, and then the equations of motion are obtained using the dynamic analog of the principle of virtual wor k. 2.3 Strain-Displacement Relations Substituting Eq. (2.9) into the strai n-di spl acement re1 at ions referred to an orthogonal curvilinear coordinate system, we obtain where (2.10) 17 0 - aw 1- 4 aw h a x2 K4 - - 2 (Cb2 +-) 4 - @2 + 0 - aw 1 4 aw h axl K5 = - 2 (e1 +-) 5 - @1 + q a@2 a@l +- O a v +- +- - au aw aw 0 6 = a x2 axl ax2 K6 = ax2 (2.11) Here xi denote the Cartesian coordinates (dxi = aidcis i = 1,Z). 2.4 Constitutive Relations The stress-strain relations f or the k-th lamina in the lamina coordinates are given by 0 0 Q66 - ( k ) 0 0 0 - (k) Q44 1 6 4 5 (k) (2.12) where vi ( k ) are the plane stress reduced stiffnesses of the k-th lamina 18 in the lamina coordinate system. in terms of the engineering constants of a lamina: The coefficients qij can be expressed (2.13) - - - 44 = 23 Q55 = 13 Q66 = 12 To determine the laminate constitutive equations, Eq. (2.12) should be transformed to the laminate coordinates. We obtain r4) = 1;: r4} (2.14) 5 ( k ) 55 ( k ) 5 ( k ) where 4 2 2 4 - Q,, = Qll cos e +2(Q12 +2066)~in e cos e +gz2 sin e - 4066)sin 2 e cos 2 e +Q12(sin 4 e +cos 4 e) Q12 = (Q11 +Q22 - 4 2 2 4 Q , , = Qll sin e +2(Q1, +2Q66)~in e cos e +Q,, cos e Q,, = (Qll +qz2 - 2Q12 - ~ T j ~ ~ ) s i n 2 e cos 2 e +Q66(~in 4 e +cos 4 e) 2 2 - Q4, = Q4, cos e +V, , sin e 19 2 2 - = Q , , cos e t q44 s i n e Q55 (2.15) 2.5 Equations of Motion The dynamic version of the pri nci pl e of vi rtual work f or the present case yi el ds dxldx2}dc - q6wdxldx2 - 6(J h/2 { $ k ) ~ [ (t)2+(t)Z+(~)2]dxldx2}dc) ]dt D -h/2 Q 1617 ( a 2" 2" - - - - +9) + IlW)6W av 2 9h4 axl ax2 +(' 3 % +' 5 (2.16) where q i s the di stri buted transverse load, Niy Mi y etc. are the resu 1 tants, 20 The inert? and (2.17) s Ti and Ti (i = 1,2,3,4,5) are defined by th, equations, T1 = I 1 +li;; 2 I 2 , 4 4 - - 1 - I = I +- 2 2 R 1 '3 3h 4 4 2 I 5 - - 1 - 5 = I2 + 5 I3 3h2 I4 - l6 I 8 3h - I 4 = I3 - - Z 1 5 +g h 4 7 l6 I 4 I =- - 3h 2 1 5 - 2 7 ' l6 I 4 1' =- - 3h 2 1 5 - 2 7 (2.18a) I 21 The governing equations of motion can be derived from Eq. (2.16) by integrating the displacement gradients by parts and setting the coefficients of 6u,6v,6w and 6 ' i (i = 1,2) to zero separately: a N 1 +- = aN6 aw axl ax2 6u: - 6v: - 2 2 a '6 ) a p 2 +2 axlax2 a91 aQ2 aK1 aK2 4 a2P1 +2 2 ( 2 ax2 3h axl ax2 axl ax2 h2 axl ( +- ) +- - +- - - - 6w: - - - - - N1 N2 +N(w) =- I3 .; au - - a 4 ~ av +Ti 3 a42 +I1w I 5 +% 3 R1 R2 1617 ,aZ; + a2; 7) - q - - - 9h4 axl 2 ax2 aw aP1 aP6 - - 2 ' + '4'1 - '5 +- - +- K 4 - - - ( +-)=I " Q1 h2 1 3h2 axl ax2 641: - aM1 aM6 axl ax2 (2.19) where aw =) (2.20) a aw aw a (N1 + N6 + - (N6 + N2 ax2 N(w) =- axl ax2 The essential (i .e., geometric) and natural boundary conditions of the theory are given by: 22 Un Uns W an aw as aw @n @ns (essential ) _ _ _ - I Nn Nns* Qn, Pn Ps Mn Mns (natural) (2.21) where 2 2 =n N +n N +2 n n N n x 1 y 2 X Y 6 Nns = (N2 - N )n n +N6(nx 2 2 - ny) 1 X Y = nxP1 2 +n 2 P n Y 2 X Y 6 +2n n P = (P2 - P )n n +P6(nx 2 2 - ny) = (M2 - M )n n +M6(nx 2 2 - ny) ns 1 X Y = nxM1 2 +n 2 M Mn Y 2 X Y 6 Mns 1 X Y +2n n M aPn aPns 4 K - - aw aw 4 Qn - Nn an +Nns +3 (K +-) +Qn - - h2 as =n K +n K Kn x 1 y 2 P n - - - - 4 P . . , 3h2 - - - - - 4 P ns 3h2 - Mn = Mn - - 4 P 3h2 - - - 4 P ns = Mns 3h2 ns (2. 22) 23 and nx and ny are the di recti on cosines of the uni t normal on the boundary of the laminate. The resul tants can be expressed i n terms of the s trai n components I using Eqs. (2.10) and (2.12) i n Eq. (2.14). We get Ni - - Ai j ~ g +6. .IC. 0 +E. .K 2 1 J J 1 J j Mi - - B i j ~j 0 +D O + F i j'j P. = E 1 2 +F..K jKj 1~ j jKj 1~ j 2 O +H..K (i,j = 1,2,6) (2.23) 0 1 5 j j K1 = D5cj +F K where Aij, Elij, etc. are the laminate stiffnesses, ( A . B. . ,Di ,Ei j,Fi j,H. .) ij' ij 'J f o r i,j = 1,2,4,5,6. (2. 25) 3. THE NAVIER SOLUTIONS 3.1 Introduction Exact solutions of the partial differential equations (2.19) on arbitrary domains and for general conditions is not possible. However, for simply supported shells whose projection in the xlx2-plane is a rectangle, the linear version of Eq. (2.19) can be solved exactly, provided the lamination scheme is of antisymmetric cross-ply [Oo/900/ 0/90 O . . . ] or symmetric cross-ply [0/900...]s type. solution exists if the following stiffness coefficients are zero 1211: The Navier Fi6 = Hi6 = 0 , (i = 1,2) ( 3. 1) Ai6 = Bi6 - - Di6 - - Ei6 - - = o A45 = O45 = F45 The boundary conditions are assumed to be of the form, u(xl,O) = u(xl,b) = v(0,x2) = v(a,x2) = 0 w(xl,O) = w(xl,b) = w(0,x2) = w(a,x2) = 0 aw (0,x2) =- (a,x2) = 0 aw aw aw axl axl ax2 ax2 - ( ~ 1 ~ 0 ) =- (xl,b) =- N2(x1,0) = N (x ,b) = Nl(0,x2) = Nl(a,x2) = 0 2 1 where a and b denote the lengths along the xl- and x2-directions, respectively (see Fig. 3.1). 24 25 / Figure3.1 The geometry and the coordinate system f o r a Projected area of -she1 1 element 26 3.2 The Navier Solution Following the Naviw solution procedure (see Reddy [21]), we assume the following solution form that satisfies the boundary conditions in Eq. (3.2): m m where f (x ,x ) = cosax1sin~x2, f (x ,x ) = sinax C O S B X ~ f3(x1,x2) = sinaxlsin~x2, a = mn/a, B = nn/b 1 1 2 2 1 2 1 (3.4) Substituting Eq. (3.3) into Eq. (2.19), we obtain where Qmn are the coefficients in the double Fourier expansion of the transverse load, 27 and the coefficients Mi j and Ci j(i,j = 1,Z ,... ,5) are given in Appendix ~ A. Equation (3.5) can be solved for Umn,Vmn, etc., for each m and n, and then the solution is given by Eq. (3.3). are evaluated using a finite number of terms in the series. vibration analysis, Eq. (3.5) can be expressed as an eigenvalue equation, The series in Eq, (3.3) For free i I ( [ CI - UZIMl){A} = (0) (3.7) where {A} = {Umn,Vmn,Wmn,m,!,n,m~n}T, and w is the frequency of natural vibration. For static bending analysis, Eq. (3.5) becomes 4. MI XED VARI ATI ONAL PRI NCI PLES 4. 1 I nt r oduct i on The var i at i onal f or mul at i ons used f or devel opi ng pl at e and shel l el ement s can be cl assi f i ed i nt o t hr ee maj or cat egor i es: ( i ) f or mul at i ons based on t he pr i nci pl e o f vi rt ual di spl acement s ( or t he pr i nci pl e of t ot al pot ent i al energy) , ( i i ) f or mul at i ons based on t he pr i nci pl e of vi rt ual f or ces ( or t he pr i nci pl e of compl ement ar y ener gy) and ( i i i ) f or mul at i ons based on mi xed var i at i onal pr i nci pl es( Hu- Washi zu- Rei ssner ' s pri nci pl es). f or mul at i ons ar e cal l ed, r espect i vel y, t he di spl acement model s, equi l i br i ummodel s and mi xed model s. The f i ni t e- el ement model s based on t hese I n t he pr i nci pl e of vi r t ual di spl acement s, one assumes t hat t he ki nemat i c r el at i ons ( i .e. , st r ai n- di spl acement r el at i ons and geomet r i c boundary condi t i ons) ar e sat i sf i ed exact l y ( i . e. , poi nt - wi se) by t he di spl acement f i el d, and t he equi l i br i umequat i ons and f or ce boundar y condi t i ons ar e der i ved as t he Eul er equat i ons. No a- pr i or i assumpt i on concer ni ng t he const i t ut i ve behavi or ( i . e. , st r ess- st r ai n r el at i ons) i s necessar y i n usi ng t he pr i nci pl e. The pr i nci pl e of ( t he mi ni mum) t ot al pot ent i al ener gy i s a speci al case of t he pr i nci pl e of vi r t ual di spl acement s appl i ed t o sol i d bodi es t hat ar e char act er i zed by t he st r ai n ener gy f unct i on U such t hat aU( E . .) i j i j --J -J -=. aE wher e E . . and ui j ar e t he component s of st r ai n and st r ess t ensor s, r espect i vel y. 1 J When t he pr i nci pl e of t ot al pot ent i al ener gy i s used t o 29 devel op a f i ni t e- el ement model of a shel l t heor y, t he ki nemat i c r el at i ons ar e sat i sf i ed poi nt - wi se but t he equi l i br i umequat i ons ar e met onl y i n an i nt egr al ( or var i at i onal ) sense. The pr i nci pl e of compl ement ar y pot ent i al ener gy, a speci al case of t he pr i nci pl e of vi r t ual f or ces, can be used t o devel op equi l i br i ummodel s t hat sat i sf y t he equi l i br i umequat i ons poi nt - wi se but meet t he st r ai n compat i bi l i t y onl y i n a var i at i onal sense. Ther e ar e a number of mi xed vari at i onal pr i nci pl es i n el ast i ci t y ( see [64-711). The phr ase ' mi xed' i s used t o i mpl y t he f act t hat bot h t he di spl acement ( or pr i mal ) var i abl es and ( some of t he) f or ce ( or dual ) var i abl es ar e gi ven equal i mpor t ance i n t he var i at i onal f or mul at i ons. The associ at ed f i ni t e- el ement model s use i ndependent appr oxi mat i ons of dependent var i abl es appear i ng i n t he var i at i onal f or mul at i on. Ther e ar e i n whi ch bot h pri mal and dual y i n t he i nt er i or and on t he t wo ki nds of mi xed model s: ( i ) model s var i abl es ar e i nt er pol at ed i ndependent boundar y of an el ement , ( i i ) model s i n whi ch t he var i abl es ar e i nt er pol at ed i nsi de t he el ement and t hei r val ues on t he boundar y ar e i nt er pol at ed by t he boundar y val ues of t he i nt er pol at i on. The f i r st ki nd ar e of t en t er med hybri d model s, and t he second ki nd ar e known si mpl y as t he mi xed model s. I n t he pr esent st udy t he second ki nd (i . e. , mi xed model ) wi l l be di scussed. 4.2 Var i at i onal Pr i nci pl es To devel op a mi xed var i at i onal st at ement of t he t hi r d- or der l ami nat e t heor y, we r ewr i t e Eq. (2. 23) i n mat r i x not at i on as f ol l ows: { N} = [ A * ~ { E ~ } + [B*I{M} + IE*I{P} 30 t o -{K} = [B*1 { E } - [D*l{M} - [F*l{P} -{K} = [E*] { E } - [F*lt{M} - [H*]{P} t o where {KC} = - w 1 a w 2 ay R2 2 ay a v +- +- (-) au av aw aw 1 ay ax ax ay - +- +- - A c1 =2 and c =- 4 3h h2 [B*] = [B][D*] + [E][F*] [E*] = [B][F*] + [E][H*] (not symmetric) (not symmetric) - - [Dl [FI [F*I [H*l [FI [HI (4.2) Note that in this part the notation x = x1 and y = x2 is used, and [ I t denotes the transpose o f a vector or matrix. To develop a mixed variational statement of the third-order laminate theory, the generalized displacements (u, v, w, 41, generalized moments (MI, M2, Me, PI, P2, P6) are treated as the dependent variables. and 31 1 o t = s a [ y { E 1 + {EO}~([B*]{M} +[E*]{p}) A A A - J- (Nnun +Nsus +Qnw +Mn+,, +M,$,)ds r (4.3) where the quanti ti es wi th a hat over them denote speci fi ed values, and Qn = Qlnx +Q2ny un = unx +vny , us = -uny +vnx a - n - - n - a an x ax y ay ' as x ay y ax a a _ - a - n - a + n - - - 2 2 2 2 Nn = N n +2N n n +N2ny , NS = (N2 - N )n n +N6(nx - ny) l x 6 X Y 1 X Y wi th (4.4) (4.5) 32 Note that the bending moments (MI, M2, M6) do not enter the set of essential boundary conditions, and the resultants (PI, P2, P6) do not enter the set of natural boundary conditions also. The (mixed) variational principle corresponding to the functional in Eq. (4.3) can be stated as follows. laminate can assume, the one that renders the functional nH1 stational-y also satisfies the equations of equilibrium (2.19) and the kinematic- constitutive equations [i.e., the last two equations of Eq. (4.3)]. Of all possible configurations a We note that the variational statement in Eq. (4.3) contains the second-order derivatives of the transverse deflection w (through {K}). To relax the continuity requirements on w, we integrate the term {P}t{~} by parts to trade the differentiation to {P}. We obtain A A A A aw en an 9 e~ as - - - aw =- 33 I We note that the stress resultants Pn and Ps (hence PI , P2 and P6) enter the set of essential boundary conditions in the case of the second mixed formulation of the higer-order theory. As special cases, the mixed variational statements for the classical and the first-order theories can be obtained from Eq. (4.6) by setting appropriate terms to zero. 5. FINITE-ELEMENT MODEL 5.1 Introduction The mixed variational principles presented in Chapter 4 can be used to develop mixed finite-element models, which contain the bending moments as the primary variables along with the generalized displacements. Historically, at least with regard to bending of plates, the first mixed formulation is attributed to Herrmann [72,73] and Hellan [741 Since the bending moments do not enter the set of essential boundary conditions, they are not required to be continuous across interelement boundaries. discontinuous (between elements) bi 1 inear or higher-order approximation of the bending moments. The present section deals with the development of a mixed model based on nH2. Thus, one can develop mixed models with 5.2 Finite-Element Model Let n, represented where = n the domain (i.e., the midplane) of the laminate, be by a collection of finite elements (5.1) - n = u n , ne n af = empty for e # f e= 1 ur is the closure of the open domain a and r is its boundary. Over a typical element ne, each of the variables u, v, w, etc. are interpolated by expressions of the form N N N u = c U.$ , v = c V.$ , w = W.$ (5.2) j=l J j j= 1 J j j=l ~j 34 35 etc. , wher e { s, .} denot e t he set of admi ssi bl e i nt er pol at i on f unct i ons and j denot es t he number of f unct i ons ( see Reddy and Put cha [64]). Al t hough t he same set of i nt er pol at i on f unct i ons i s used, f or si mpl i ci t y, t o i nt er pol at e each of t he dependent var i abl es, i n gener al , di f f er ent set s of i nt er pol at i on can be used f or (u, v), w, (Q Q ), (M1, M2, M6) and (Pl , PZ, P6). Fr oman exami nat i on of t he var i at i onal st at ement i n Eq. (4.6), i t i s cl ear t hat t he l i near , quadr at i c, etc. , i nt er pol at i on f unct i ons of t he Lagr ange t ype ar e admi ssi bl e. r esul t i ng el ement i s cal l ed a Co- el ement , because no der i vat i ve of t he dependent var i abl e i s r equi r ed t o be cont i nuous acr oss i nt er el ement boundar i es. J x y Y The To devel op t he f i ni t e- el ement equat i ons of a t ypi cal el ement , we f i r st comput e t he st r ai ns, r ot at i ons, and r esul t ant s i n t er ms of t hei r f i ni t e el ement appr oxi mat i ons. We have { EO} = { EOL } + { EON} ( 5- 3) = [ HL] {nl } +[H"]{A*} (5. 4) 36 aw aw 2- - a x a y N 2 ,: $ [H ] { A } (5.5) { &.ON} = [ HN] {SA'} 37 {Q} = 4 4 [H ]{A } (5.11) (5.12) (5.13) aw , f =- and aw Y aY Here, fx =- ax {$i} = $2 0 . . $N I 9 E$i,x} = {$1,x $2,x * e * $N,xl, etc .( 5.14) and {u}, for example, denotes the column of the nodal values of u. The finite-element model for t h e refined shear deformation theory is derived from the variational statement in Eq. (4.6) over an element. element Q~ is given by The first variation of the functional in (4.6) for a typical 0 = 5 [{~E~}~([A*~{C'} + [ B* l { M} + [E*l {P}) +- {sM}t({~s} + [B*] {E } - [ D* l { M} - [F*l{P}) ne t o 38 +{6P }t(- C1{Ks} + [E*]{E'} - [F*l{M} - [H*l{P}) - (Nn6un +N 6u +Qn6w +Mn6$,, +Ms6$, s s re +c e SP +cles6Ps)ds I n n Substitution of Eqs. (5.2)-(5.14) Eq. (5.15), we obtain [Kl1] [K12] [K13] [K14] [K15] [K21] [kZ2] [K23] [K24] [KZ5] [K31] [K3'] [K33] [K34] [K35] [K41] [K4'] [K43] [K44] [K45] [K51] [K52] [K53] [K54] [K55] where (5.15) nto the vari ati onal statement i n [K"] = [HLIt[A*][HL]dA , IF1} = 8 [H2I t ds ae re [K l2 1'78 [HLIt[A*]([HN] +2[Hol)dA ne (5.16) [K14] = [HLlt[B*I[H3]dA = [K 41 1 t ne 39 [ K15] = [ [HLI t[E*][H3]dA = [ K 51 ] t ae [i Z21 = s (Hol + [ HN t 1) [A*l ([Hol +7 1 N [ H ])dA +s [ ae ae 1 [KZ41 = (([HOI + [ HN t I ) [B*][H3])dA ne = .f ($ [H31t[B*l t([HN] +2[Ho]))dA ne N t I t L t ([Hol + [ H I ) [E*l [H31 +cl [ H I [ H I )dA [KZ5] = ne [K5'1 = ([H3l t[E*l ([Ho1+ 7 1 N [ H I ) +cl [HL1[H1J )dA ae [ K -3 3 ] = .f [ H2 t - ] [A][H2]dA, [i cZ3l =S [H I t A ] [A][H2]dA = [ K -32 ] t ae ne L t 3 43 t = S [H I [H I dA = [ K ] ae [K441 = - [H3l t[D*1[H31dA, Qe [K45] = - J [H3I t[F*][H 3 ]dA = [ K 54 ] t ae \ 40 [ K 55 1 = - I [H3It[H*1[H3]dA ne IF5} = 1 [H3It[Tlt ds ne (5.17) Clearly, the element stiffness matrix is pJ symmetric. case (i.e., [HN] = [O]), the element stiffness matrix is symmetric. For the linear The theories can be form of the finite-element models for the classical and first-order obtained as special cases from Eq. (5.16). An explicit coefficients of the stiffness matrix in Eq. (5.16) is g B. ven in Appendix 5.3 Solution Procedure The assembled finite-element equations are of the form tKO(A)l{A) = {F) (5.18) where [K O] is the assembled (direct) stiffness matrix, unsymmetric in general, and {A} is the global solution vector. matrix i s a nonlinear function of the unknown solution, Eq. (5.18) should be solved iteratively. quite commonly used in the finite-element analysis of nonlinear problems: iteration methods. nonlinear equations. Because the stiffness Two iterative methods of analysis are the Picard-type direction iteration and the Newton-Raphson Here the Newton-Raphson method is used to solve the 41 Suppose that the solution is required for a load of P. The load is divided into a sequence of load steps AP1, A P ~ , ..., A P ~ such that P = c A P ~ . At any load step i, Eq. (5.18) is solved iteratively to obtain the solution. At the end of the r-th iteration the solution for n i=l the next iteration is obtained solving the following equation (5.19) [K T r (A )]{6Ar+'} = - {R} 3 -[K D r (A )]{Ar} + {F} for the increment of the solution {sar+'}, and the total solution is computed from {Ar+'} = {Ar} + {&A r+l where [KT] is the tangent stiffness matrix, (5.20) (5.21) A geometrical explanation of the Newton-Raphson iteration is given in Fig. 5.1. For the finite-element models developed here, the tangent stiffness matrix is symmetric. model in Eq. (5.16) for plates (1/R1 = 1/R2 = 0) is given by For example, the tangent stiffness matrix for the [K21T] [KZzT] symm. [ K~~~ I. 5 1T [K ] .... ........ where [K1lT] = [K" [KZzT] = [kz2 (5.22) 42 F Load, F 0 Displacement, U Figure 5.1 Geometric i nterpretati on o f the Newton-Raphson i terati on f or the sol uti on o f one-parameter probl ems. 43 [ K31T] = [O], ae 44 [K44T] = [ K t [ A*I I HNl + ["?[A*] [$l )dA{A2} t[B*][H3]dA){A4} ae 32 [ K32T] = [ K 1, 42T 24 , [ K 1 = [ K 53 5 4T 54 [ K53T] = [ K 1, [ K 1 = [ K -N [H 1 = [ K33T] = [ K33] K25] , K55] (5. 23) For a1 1 coupl ed ( i . e. bendi ng- st r et chi ng coupl i ng) pr obl ems, t he mi xed model s of cl assi cal , f i r st - or der and t hi r d- or der t heor i es have r espect i vel y. I f t he , t he el ement degr ees si x, ei ght and el even degr ees of f r eedomper node, bendi ng moment s ar e el i mi nat ed at t he el ement l eve of f r eedomcan be r educed by 3 ( see Fi g. 5.2). The f i ni t e- el ement model (5. 16) f or t he dynam f or m, [ K I M + [ M I m = {F} c case i s of t he (5. 24) wher e [ MI i s t he mass mat r i x ( see Appendi x 6). (5. 24) can be wr i t t en as For f r ee vi br at i on, Eq. 44 w. / // 2 u1 (a) Displacement model / J t* 'a 1 1 " 1 M1 ' pl ( b ) Hixed model Figure5.2 The displacement and mixed f i ni t e elements f or the thi rd- order shear deformation theory 45 [ K ] { A } = A [ Ml { A } ( 5. 25) where ;x i s the square of the frequency. The evaluation of the element matrices requires numerical integration. coeffi ci ents associated wi th the shear energy terms. More speci fi cal l y, the 2 x 2 Gauss rul e i s used f or shear terms and the standard 3 x 3 Gauss r ul e i s used f or the bending terms when the nine node quadratic isoparametric element i s considered. Reduced i ntegrati on i s used t o evaluate the sti ffness 6. SAMPLE AP P LI CATI ONS 6.1 I nt r oduct i on A number of r epr esent at i ve pr obl ems ar e anal yzed usi ng t he hi gher - or der t heor y devel oped i n t he pr esent st udy. i ncl uded her e i l l ust r at e t he accur acy of t he pr esent t heory. These pr obl ems ar e sol ved usi ng t he cl osed- f or msol ut i ons pr esent ed i n Sect i on 3. t he mi xed f i ni t e- el ement model descr i bed i n Sect i on 5. The f i r st f ew pr obl ems Then pr obl ems t hat do not al l owcl osed- f or msol ut i ons ar e sol ved by The geomet r i es of t ypi cal cyl i ndr i cal and spher i cal shel l panel s ar e shown i n Fi g. 6.1. Of cour se, pl at es ar e der i ved as speci al cases f r omcyl i ndr i cal or spher i cal shel l panel s. 6. 2 Exact Sol ut i ons I t i s wel l known t hat t he ser i es sol ut i on i n Eq. ( 3. 3) conver ges f ast er f or uni f or ml oad t han f or a poi nt l oad. di st r i but i on of t he t r ansver se l oad, t he ser i es r educes t o a si ngl e t erm. For a si nusoi dal 1. Four - l ayer , cr oss- pl y (0/90/90/0) squar e l ami nat ed f l at pl at e under Thi s exampl e demonst r at es t he r el at i ve accur acy of t he pr esent si nusoi dal l oad hi gher - or der t heor y when compar ed t o t hr ee- di mensi onal el ast i ci t y t heor y and t o t he f i r st - or der t heor y. Squar e, cr oss- pl y l ami nat es under si mpl y- suppor t ed boundar y condi t i ons [ see Eq. (3.2)] and a si nusoi dal di st r i but i on of t he t r ansver se l oad ar e st udi ed f or def l ect i ons. The l ami na pr oper t i es ar e assumed t o be 46 47 Fi gure 6.1 Geometry of a typi cal cylindrical and spherical shell 48 Plots of the nondimensional center deflection [ w= wE2h3)/qoa4] versus the side to thickness ratio (a/h) obtained using various theories are shown in Figure 6.2. The present third-order theory gives the closest solution to the three-dimensional elasticity solution [58] than either the first-order theory or the classical theory. 2. An isotropic spherical shell segment under point load at the center. The problem data are (see Fig. 6.1 for the geometry) R1 = R2 = 96.0 in., a = b = 32.0 in., h = 0.1 in., 7 El = E2 = 10 psi, v = 0.3, intensity of load = 100 lbs. A comparison of the center transverse deflection of the present theory (HSDT) with that obtained using the f irst-order shear deformation theory (FSDT) and classical shell theory (CST) for various terms in the series is presented in Table 1 (for simply supported boundary conditions). It should be noted that Vlasov [76] did not consider transverse shearing strains in his study. The difference between the values predicted by HSDT and FSOT is not significant for the thin isotropic shell problem considered here. 3. Cross-ply spherical shell segments under sinusoidal, uniform, and point loads. The geometric parameters used are the same as those used in Problem 2, and the material parameters used are the same as those used in Problem 1. The shell segments are assumed to be simply supported. Nondimensionalized center deflection of various cross-ply shells under sinusoidal, uniform, and point loads are presented in Tables 2 through 49 z t;; 4 T u U Y w M 2 0 w I I- 0 si 9 0 M 0 m 0 rl 0 0 50 Table 1. Center defl ecti on ( - w x 103) of a simply supported spherical shel l segment under poi nt load a t the center (see Fig. 6.1) Number of terms i n the series h Theory N =9 N = 49 N = 99 N = 149 N = 199 CST Vlasov [79] - - 0.1 FSDT [40] 32.594 39.469 HSDT 32.584 39.458 0.32 FSDT 3.664 3.902 HSDT 3.661 3.899 1.6 FSDT 0.165 0.171 HSDT 0.164 0.170 3.2 FSDT 0.035 0.038 HSDT 0.035 0.037 6.4 FSDT 0.007 0.008 HSDT 0.007 0.007 39.591 39.560 39.724 39.714 3.919 3.916 0.174 0.172 0.039 0.037 0.009 0.007 - 39.786 39.775 3.927 3.923 0.175 0.172 0.039 0.037 0.009 0.007 39.647 39.814 39.803 3.932 3.927 0.176 0.172 0.040 0.037 0.009 0.007 - 51 4, r espect i vel y. hi gher - or der t heor y and t he f i r st - or der t heor y i ncr eases wi t h i ncr easi ng val ues of R/ a. For a/ h = 10, t he di f f er ence bet ween t he def l ect i ons gi ven by FSDT and HSDT i s l ar ger t han t hose f or a/ h = 100. unsymmet r i c l ami nat es (0/90) , FSDT yi el ds hi gher def l ect i ons t han HSDT, wher eas f or symmet r i c l ami nat es HSDT yi el ds hi gher def l ect i ons t han FSDT f or val ues of a/ h = 10. di f f er ence bet ween t he sol ut i on predi ct ed by t he f i r st - or der t heor y and t he hi gher - or der t heor y i s mor e si gni f i cant , wi t h FSDT r esul t s hi gher t han HSDT, especi al l y f or ant i symmet r i c cr oss- pl y l ami nat es. The di f f er ence bet ween t he sol ut i ons pr edi ct ed by t he For Not e t hat f or poi nt - l oaded shel l s t he 4. Nat ural vi br at i on of cr oss- pl y cyl i ndr i cal shel l segment s Nondi mensi onal i zed f undament al f r equenci es of cr oss- pl y cyl i ndr i cal [Oo/9O0], shel l s ar e pr esent ed i n Tabl e 5 f or t hr ee l ami nat i on schemes: [O0/9Oo/O0] , and [0/900/900/00]. For t hi n ant i symmet r i c cr oss- pl y shel l s, t he f i r st - or der t heor y under pr edi ct s t he nat ur al f r equenci es when compar ed t o t he hi gher - or der theory. However , f or symmet r i c cr oss- pl y shel l s, t he t r end r ever ses. 5. Nat ural vi br at i on of cr oss- pl y spheri cal shel l segment s Nondi mensi onal i zed nat ural f r equenci es obt ai ned usi ng t he f i r st - and hi gher - or der t heor i es ar e present ed i n Tabl e 6 f or var i ous cr oss- pl y spher i cal shel l segment s. Anal ogous to cyl i ndr i cal she I s, t he f i r st - or der t heor y under pr edi ct s f undament al nat ur al f r equenc es of ant i symmet r i c cr oss- pl y shel l s; f or symmet r i c t hi ck she 1s and symmet r i c shal l owt hi n shel l s t he t rend reverses. 52 Table 2. Nondimensionalized center deflections, w = (-wh3E /q a4)103, of cross-ply laminated spherical she1 1 segments uGdeP sinusoidally distributed load (a/b = 1, R 1 = R2 = R, q, = 100) - W oo/900 0 O /goo /oo oo/900/900/oo R/a Theory a/h=100 a/h=lO a/h=100 a/h=lO a/h=100 a/h=lO 5 FSDT 1.1948 HSDT 1.1937 10 FSDT 3.5760 HSDT 3.5733 20 FSDT 7.1270 HSDT 7.1236 50 FSDT 9.8717 FSDT 9.8692 100 FSDT 10.4460 HSDT 10.4440 Plate FSDT 10.6530 R/a=m FSDT 10.6510 11.429 11.166 12.123 11.896 12.309 12.094 12.362 12.150 12.370 12.158 12.373 12.161 1.0337 1.0321 2.4109 2.4099 3.6150 3.6170 4.2027 4.2071 4.3026 4.3074 4.3370 4.3420 6.4253 6.7688 6.6247 7.0325 6.6756 7.1016 6.6902 7.1212 6.6923 7.1240 6.6939 7.1250 1.0279 1.0264 2.4030 2.4024 3.6104 3.6133 4.2015 4.2071 4.3021 4.3082 4.3368 4.3430 6.3623 6.7865 6.5595 7.0536 6.6099 7.1237 6.6244 7.1436 6.6264 7.1464 6.6280 7.1474 53 Table 3. Nondimensionalized center deflections, w = (-WE h3/q a4) x 103, of cross-ply laminated spherical shel l s2gmen?s under uniformly di stri buted load 0"/90" 0" /90 O /oo oo/900/900/oo R - Theory a/h=100 a/h=lO a/h=100 a/h=lO a/h=100 a/h=lO a 5 FSDT HSDT 10 FSDT HSDT 20 FSDT HSDT 50 FSDT HSDT 100 FSDT HSDT P late FSDT m 1.7535 1.7519 5.5428 5.5388 11.273 11.268 15.714 15.711 16.645 16.642 16.980 16.977 19.944 17.566 19.065 18.744 19.365 19.064 19.452 19.155 19.464 19.168 19.469 19.172 1.5118 1.5092 3.6445 3.6426 5.5473 5.5503 6.4827 6.4895 6.6421 6.6496 6.6970 6.7047 9.794 10.332 10.110 10.752 10.191 10.862 10.214 10.893 10.218 10.898 10.220 10.899 - 1.5358 1.5332 3.7208 3.7195 5.6618 5.666 6.6148 6.6234 6.7772 6.7866 6.8331 6.8427 9.825 10.476 10.141 10.904 10.222 11.017 10.245 11.049 10.294 11.053 10.251 11.055 54 Table 4. Nondimensionalized center deflection of cross-ply spherical shel l segments under poi nt load at the center - wh3E2 w = - (- )102, a/b = 1, a/h = 10 Pan Center deflection, w R/a Theory 0"/90" 0" /90 " / O " oo/900/900/oo 5 FSDT HSDT 10 FSDT HSDT 20 FSDT HSDT 50 FSDT HSDT 100 FSDT HSDT P1 ate FSDT m HSDT 7.1015 5.8953 7.3836 6.1913 7.4692 6.2714 7.4909 6.2943 7.4940 6.2976 7.4853 6.2987 5.1410 4.4340 5.2273 4.5470 5.2594 4.5765 5.2657 4.5849 5.2666 4.5861 5.2572 4.5865 4.9360 4.3574 5.0186 4.4690 5.0496 4.4982 5.0557 4.5065 5.0565 4.5077 5.0472 4.5081 55 Table 5 Nondimensionalized fundamental frequencies of cross-ply cyl i ndri cal shel l panels (see Fig. 6.1 f or geometry). 0"/9O0 oo/900/oo oo/900/900/oo R/a Theory a/h=100 a/h=lO a/h=100 a/h=lO a/h=100 a/h=lO 5 FSDT HSDT 10 FSDT HSDT 20 FSDT HSDT 50 FSDT HSDT 100 FSDT HSDT P late FSDT m HSDT 16.668 16.690 11.831 11.840 10.265 10.270 9.7816 9.7830 9.7108 9.7120 9.6873 9.6880 8.9082 9.0230 8.8879 8.9790 8.8900 8.9720 8.8951 8.9730 8.8974 8.9750 8.8998 8.9760 20.332 20.330 16.625 16.620 15.556 15.550 15.244 15.240 15.198 15.190 15.183 15.170 12.207 11.850 12.173 11.800 12.166 11.791 12.163 11.790 12.163 11.790 12.162 11.790 20.361 20.360 16.634 16.630 15.559 15.550 15.245 15.230 15.199 15.190 15.184 15.170 12.267 11.830 12.236 11.790 12.230 11.780 12.228 11.780 12.227 11.780 12.226 11.780 56 Table 6. Nondimensionalized fundamental frequencies of cross-ply laminated spherical shel l segments a2 - h - w = w - J p/E2 0"/90" 0" /goo /oo oo/900/900/oo R/a Theory a/h=100 a/h=lO a/h=100 a/h=lO a/h=100 a/h=lO 5 FSDT HSDT 10 FSDT HSDT 20 FSDT HSDT 50 FSDT HSDT 100 FSDT HSDT P late FSDT HSDT 28.825 28.840 16.706 16.710 11.841 11.84 10.063 10.06 9.7826 9.784 9.6873 9.6880 9.2309 9.3370 8.9841 9.0680 8.9212 8.999 8.9034 8.980 8.9009 8.977 8.8998 8.9760 30.993 31.020 20.347 20.350 16.627 16.62 15.424 15.42 15.244 15.24 15.183 15.170 ~~ 12.372 12.060 12.215 11.860 12.176 11.81 12.165 11.79 12.163 11.79 12.162 11.790 31.079 31.100 20.380 20.380 16.638 16.63 15.426 15.42 15.245 15.23 15.184 15.170 12.437 12.040 12.280 11.840 12.240 11.79 12.229 11.78 12.228 11.78 12.226 11.780 57 6.3 Appr oxi mat e ( Fi ni t e- El ement ) Sol ut i ons 6.3.1 Bendi ng Anal ysi s 1. Li near anal ysi s of a r ect angul ar pl at e under uni f or ml y di st r i but ed 1 oad . The geomet r y and boundar y condi t i ons f or t he pr obl emar e shown i n - Fi g. 6.3. The pl at e i s assumed t o be made of st eel ( E = 30 x lo6 psi and v = 0.3). The pr obl emwas al so sol ved by Ti moshenko [77], Her r mann [72] and Pr at o [78]. di spl acement and bendi ng moment s obt ai ned by var i ous i nvest i gat or s ( usi ng t he l i near theory). and ot her s i s very good, ver i f yi ng the accur acy of t he t heor y and t he f i ni t e el ement f or mul at i on. Fi gur es 6.3-6.5 cont ai n pl ot s of t he t r ansver se The agr eement bet ween t he pr esent sol ut i on 2. Nonl i near anal ysi s of r ect angul ar l ami nat es under uni f or ml y di st r i but ed l oad Fi gur e 6.6 cont ai ns l oad- def l ect i on cur ves f or a si mpl y suppor t ed squar e or t hot r opi c pl at e under uni f or ml y di st r i but ed l oad [23]. The f ol l owi ng geomet r i c and mat er i al pr oper t i es are used: a = b = 12 i n. , h = 0.138 i n. 6 0.37 x 10 psi El = 3 x 10 psi , E2 = 1.28 x 10 psi , GI2 = 613 = 623 = 6 6 The exper i ment al r esul t s and cl assi cal sol ut i ons ar e t aken f r omt he paper by Zaghl oul and Kennedy [79]. The agr eement bet ween t he pr esent sol ut i on and t he exper i ment al sol ut i on i s ext r emel y good. I t i s cl ear t hat , even f or t hi n pl at es, t he shear def or mat i on ef f ect i s si gni f i cant i n t he nonl i near range. 58 4 , ~~ 0 L - PRESENT SOLuTIOr4 (LINEAR ANALYSIS) HERRMA a PRATO A TIWSHENKO 0 / / C W E D / SS =SIfVLY SUPPORTED THICKNESS =0,O~ 0 I I I 0 0 2 o14 0,6 018 DISTANCE ALONG X i (IN.) Figure 6.3 Comparison of the transverse defl ecti on o f a clamped- simply-supported-free i sotropi c rectangular pl ate under uniformly di stri buted transverse load,, 59 0 b PRESENT SOLUTION HERRW" - 0 PRATO A TIMOSHENKO 002 0,4 0,6 008 DISTANCE ALONG X2 (in.) Figure 6.4 Comparison o f the bending moment along the line x1=0.4" for the problem of Figure 6.3. 60 6 8 4 X 4 2 W z c. f 0 DISTANCE ALONG X i (I N. ) Figure 6.5 Comparison o f the bending moment M6 along the line x2=0.4" f or the problem i n Figure 6.3. 61 EXPERIMENTAL - - PRESEM' --- 0,4 n z - v" 0,3 2 0 - G k 012 4 * 0 8 1 111 oc W w 0 0 0, 4 018 182 1'6 280 LOAD INTENSITY (PSI) Figure 6.6 Center defl ecti on versus load i ntensi ty for a simply 'supported square orthotropi c pl ate under uniformly di stri buted transverse l oad. The following Simply supported boundary condi- t ions were used : V = W = $y = Idx = Px = 0 on si de x = .a U = W = $ = M = P =O o n s i d ey =b X Y Y 62 Fi gure 6. 7 cont ai ns l oad- def l ect i on cur ves f or a cl amped bi di r ect i onal 1 ami nat e [ 0/ 90/ 90/ 0 I under uni f or ml y di st r i but ed l oad. The geomet r i c par amet er s and l ayer pr oper t i es ar e gi ven by a = b = 12 i n. , h = 0. 096 i n. , 6 6 El = 1. 8282 x 10 psi , E2 = 1. 8315 x 10 psi , 6 G12 = 613 = G23 = 3. 125 x 10 psi , v12 = v13 = '23 = 0. 2395 . The exper i ment al r esul t s and t he cl assi cal l ami nat e sol ut i ons ar e t aken f r om[ 79] . The pr esent sol ut i on i s i n good agr eement wi t h t he exper i ment al r esul t s, and t he di f f er ence i s at t r i but ed t o possi bl e er r or s i n t he si mul at i on of t he mat er i al pr oper t i es and boundar y condi t i ons. 3. Or t hot r opi c cyl i nder subj ect ed t o i nt er nal pr essur e. Consi der a cl amped or t hot r opi c cyl i nder wi t h t he f ol l owi ng geomet r i c and mat er i al pr oper t i es ( see Fi g. 6.8) R1 = 20 i n. R2 = - 6 6 El = 2 x 10 psi , E2 = 7.5 x 10 psi 1. 25 x 10 psi 6 G12 = 6 613 = 623 = 0. 625 x 10 psi , h = 1 i n. = 0. 25 - '12 = '13 - "23 a = 10 i n. , P = 6. 41/ ~ psi Thi s pr obl emhas an anal yt i cal sol ut i on ( see 1771) f or t he l i near case, and Rao [ 47] used t he f i ni t e- el ement met hod t o sol ve t he same pr obl em. Bot h sol ut i ons ar e based on t he cl assi cal t heor y. The cent er def l ect i ons f r om[ 771 and [ 471 ar e 0. 000367 i n. and 0. 000366 i n. , r espect i vel y. Chao and Reddy 1501 obt ai ned 0. 0003764 i n. and 0. 0003739 63 INTENSITY OF TRANSVERSE LOAD (PSI) Figure6.7 Center defl ecti on versus load i ntensi ty for a clamped (CC-1) under uniform transverse l oad (von Karman theory). square laminate [0 /90 /90 / O ] cc-1: lJ = \I = \J = $ - - qY = 0 on al l f our cld;?oed edges 64 i n. usi ng t he f i ni t e- el ement model based on t he f i r st - or der shear def or mat i on t heor y and 3-D degener at e el ement , r espect i vel y. The cur r ent r esul t i s 0.0003761 i n. , whi ch i s cl oser t o Chao and Reddy [501, as expect ed. For t he nonl i near anal ysi s of t he same pr obl em, t he pr esent r esul t s ar e compar ed t o t hose of Chang and Sawami phakdi [80] and Chao and Reddy [50] i n Fi g. 6.9, whi ch cont ai ns pl ot s of t he cent er def l ect i on ver sus t he l oad obt ai ned by var i ous i nvest i gat or s. The agr eement bet ween t he var i ous r esul t s i s very good. 4. Ni ne- l ayer cr oss- pl y spher i cal shel l segment subj ect ed t o uni f or m 1 oadi ng . Consi der a ni ne- l ayer [Oo/900/Oo.. . / Oo] cr oss- pl y l ami nat ed spheri cal shel l segment wi t h t he f ol l owi ng mat er i al and geomet r i c dat a: , a = b = 100 i n. R1 = R2 = 1000 i n. 6 h = 1 i n. , El = 40 x 10 psi 6 0. 5 x 10 psi E2 = 10 psi , G12 = 0.6 x 10 psi , 613 = 623 = 6 6 = 0.25 '12 = '13 = '23 The pr esent r esul t s ar e compar ed wi t h t hose obt ai ned by Noor and Har t l ey [52] and Chao and Reddy [50] i n Fi g. 6.10. Noor and Har t l ey used mi xed i sopar amet r i c el ement s wi t h 13 degr ee- of - f r eedomper node whi ch i s based on a shear def or mat i on shel l t heor y. agr ee wi th bot h i nvest i gat i ons. The pr esent r esul t s 65 Figure 6.8 Geometry and boundary conditions for the octant o f the clamped cylindrical shell. 66 10 8 6' 4 2 - 0 A PRESENT SOLUTION CHANG AND SAWAMIPHAKDI CHAO AND REDDY & A / VERTICAL DEFLECTION (I N,) Fi gure 6.9 Center deflection versus l oad f or the clamped cylinder with internal pressure. 67 0- PRESENT SOLUTION 0 NOOR AND HARTLEY A CHAOAND REDDY - 1 I 1 1 2 3 4 " DI ENSI WL I ZED CENTER DEFLECTION, W/h Figure 6.10 Nonlinear bending of a nine-layer cross-ply spherical shell (0"/90"/0"/90"/ ... /o"). 68 6.3.2 Vibration Analysis Equation (5.24) can be expressed i n the al ternati ve form as where For free vi brati on analysis, we wish t o eliminate { h 2 } as follows. Eq. (6.1) we have From [K l l I {Al } + IK121EA21 = - I M 1 1 1 {q [K2l l {Al 1 + [K221{"21 = 0 (6.3) (6.4) From (6.4) we have, {A21 = -[K221-1[K211{A11 (6.5) Substi tuti ng Eq. (6.5) i nto Eq. (6.3), we obtain, ([K111 - [Kl2I [K221- 11Kz~1) {A11 = - [MI11 {A11 (K1 - u21M1) I A11 = 0 (6.6) For the free vi brati on case, Eq. (6.6) reduces t o (6.7) where w i s the natural frequencies of the system. 1. Natural frequencies of a two-layer [Oo/900] laminated pl ate. The geometric and material properti es used are a = b = 100 in. , h = 0.1 in. El = 40 x 10 psi , E2 = 10 psi . 6 6 Vi3 = V23 = 0.25 G12 = G13 = GZ3 = 0.5 x 10 psi , u12 - - 6 2 4 p = 1 lb-sec / i n 69 The boundar y condi t i ons ar e shown i n Fi g. 6.11, whi ch al so cont ai ns t he pl ot s of t he r at i os wNL/ wL versus wo/ h. nonl i near and l i near nat ur al f r equenci es, and wo i s t he nor mal i zed cent er def l ect i on of t he f i r st node. The r esul t s ar e compar ed wi t h t hose of Chi a and Pr abhakar a [81], and Reddy and Chao [82]. The pr esent r esul t s ar e sl i ght l y hi gher compar ed t o t hose obt ai ned by t he cl assi cal Her e wNL and wL denot e t he and f i r st - or der shear def or mat i on t heori es. 2. Nat ur al f r equenci es of t wo- l ayer [45"/-45'1 angl e- pl y squar e pl at e. The mat er i al and geomet r i c par amet er s used ar e El = 10 x 10 psi , E2 = 10 psi 6 p = 1 l b- sec. / i n , a = b = 100 i n. , h = 0.1 i n. 6 6 = 0.3 G12 =613 = 623 = 0.3333 x 10 psi , w12 = w13 = ~ 2 3 2 4 The boundar y condi t i ons used are shown i n Fi g. 6.12, whi ch al so cont ai ns pl ot s of uNL/uL ver sus wo/ h. avai l abl e i n t he l i t er at ur e f or compari son. For t hi s case, no r esul t s ar e 3. Nat ural vi br at i on of t wo- l ayer [ Oo/9O0] cr oss- pl y pl at es. Consi der a t wo- l ayer cr oss- pl y pl at e wi t h t he f ol l owi ng geomet r i c and mat er i al propert i es: 6 6 El = 7.07 x 10 psi , E2 = 3.58 x 10 psi - ~ 2 3 = 0.3 6 G12 = 623 = 613 = 1.41 x 10 psi , w12 = vi3 - 2 p = 1 l b- sec. / i n. 4 , a/ h = 1000 , a = 100 i n. The r esul t s of uNL/uL ver sus wo/ h ar e shown i n Fi g. 6.13. Compar ed t o t he r esul t s of Reddy [82] and Chandr a and Raj u (831, t he r esul t s of t he pr esent st udy ar e i n general a l i t t l e hi gher , pr esent r esul t s f or nat ur al f r equenci es ar e hi gher t han t hose pr edi ct ed by t he f i r st - or der t heor y i ndi cat es t hat t he addi t i onal i ner t i a t er ms The f act t hat t he cont r i but e t o t he i ncr ease of natural f r equenci es. 70 - PRESENT SOLUTION 0 CHIA AND Pw\f3HAw\RA REDDY f Xl w=+/ =O u=o +,= 0 W =@2= - 0 Figure 6.11 Fundamental frequencies of a two-layer cross-ply ( 0' 190" ) square laminate under simply-supported boundary conditions. 71 0 086 182 1,'s 2 8 - 4 3 8 0 AFIPLITUDE TO THICKNESS RATIO, Wo / h Figure 6.12 Ratio o f nonlinear to linear frequency versus amplitude to thickness ratio for two-layer 'angle-ply square plate ( 4 5 O / - 4 5 O ) . 72 BC 1 boundary cond i t i ons : u = w = $ l = o ; MZ= P2 = O al ong x2 = b/2 v = w = ~ ~ ' 0 ; M1 =P 1 =O al ong x = a/2 u=o1=0; 1'1 = P = O al ong x = 0 v = 02=0; M =P = O al ong x = 0 1 6 6 1 6 6 2 Figure 6.13 Ratio of the nonlinear to l i near frequency versus the amplitude to thickness rati o of a two-layer cross-ply ( oo/ 900) square laminate. 7. SUMMARY AND RECOMMENDATI ONS 7.1 Summar y and Concl usi ons The pr esent st udy deal t wi t h t he f ol l owi ng maj or t opi cs: The devel opment of a var i at i onal l y- consi st ent , t hi r d- or der shear def or mat i on t heor y of l ami nat ed composi t e doubl y- cur ved shel l s. The t heor y account s f or (a) t he par abol i c var i at i on of t he t r ansver se shear st r ai ns, and ( b ) t he von Kar man st r ai ns. coef f i ci ent s. The devel opment of t he cl osed- f or msol ut i ons ( f or t he l i near t heor y) f or t he si mpl y suppor t ed cr oss- pl y l ami nat es. These sol ut i ons ar e used as a check f or t he numer i cal anal ysi s of she1 1s. The const r uct i on of a mi xed var i at i onal pr i nci pl e f or t he I t does not r equi r e t he use of t he shear cor r ect i on t hi r d- or der t heor y t hat i ncl udes t he c f i r st - or der t heor y as speci al cases. The devel opment and appl i cat i on of t he of t he t hi r d- or der t heor y f or l ami nat e assi cal t heor y and t he f i ni t e- el ement model composi t e shel l s, account i ng f or t he geomet r i c nonl i near i t y i n t he sense of von Kar man ( moder at e rotati ons). The i ncr eased accur acy of t he pr esent t hi r d- or der t heor y ( f or t hi n as wel l as t hi ck l ami nat es) over t he cl assi cal or f i r st - or der shear def or mat i on t heor y i s demonst r at ed vi a exampl es t hat have ei t her t he t hr ee- di mensi onal el ast i ci t y sol ut i on or exper i ment al r esul t s. Many of t he ot her r esul t s on bendi ng and vi br at i on anal ysi s i ncl uded her e can ser ve as r ef er ences f or f ut ur e i nvest i gat i ons. 73 74 7.2 Some Comment s on Mi xed Model s The di spl acement model of t he cl assi cal l ami nat e t heor y r equi r es 1 t he use of C el ement s, whi ch ar e al gebr ai cal l y compl ex and comput at i onal l y expensi ve. si mpl e and al l ows t he di r ect comput at i on of t he bendi ng moment s at t he nodes. The i ncl usi on of t he bendi ng moment s as nodal degr ees of f r eedom not onl y r esul t s i n i ncr eased accur acy of t he aver age st r ess compar ed t o t hat det er mi ned f r omt he di spl acement model but i t al l ows us t o det er mi ne st r esses at t he nodes. Thi s f eat ur e i s qui t e at t r act i ve i n cont act pr obl ems and si ngul ar pr obl ems i n gener al . t hat t he bendi ng moment s ar e not r equi r ed t o be cont i nuous acr oss i nt er el ement boundar i es, as was shown i n Chapt er 4. based on t he shear def or mat i on t heor i es al so have t he same advant ages, except t hat t he di spl acement model of t he f i r st - or der shear def or mat i on t heor y i s a so a Co el ement . pr ogr ammi ng ef f or t s ar e l ess wi t h t he mi xed el ement s. The Co mi xed el ement s ar e al gebr ai cal l y I t shoul d be not ed The mi xed model s I n gener al , t he f or mul at i ve and 7 . 3 Recommendat i ons The t heor y pr esent ed her e can be ext ended t o a mor e gener al t heor y; f or exampl e, t he devel opment of t he t heor y i n gener al cur vi l i near coor di nat es, and f or mor e gener al shel l s ( t han t he doubl y- cur ved shel l s consi der ed here). Ext ensi on of t he pr esent t heor y to i ncl ude nonl i near mat er i al model s i s awai t i ng. Of cour se, t he i ncl usi on of t her mal l oads and dampi ng i n t he pr esent t heor y i s st r ai ght f or war d. AcknowZedgements The aut hor s ar e gr at ef ul t o Dr. Nor man Kni ght , J r. f or hi s suppor t , encour agement and const r uct i ve cr i t i ci smof t he r esul t s r epor t ed here. Our si ncer e t hanks t o Mrs. Vanessa McCoy f or t ypi ng. REFERENCES Rei ssner , E.: "On t he t heor y of bendi ng of el ast i c pl at es, " J . Math. Phy. , Vol . 23, pp. 184-191, 1944. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. Rei ssner , E.: "The ef f ect of t r ansver se shear def or mat i on on t he bendi ng of , el ast i c pl at es, " J . Appl . Mech. , Vol . 12, No. 1, pp. A- 69 t o A-77, 1945. Rei ssner , E.: "On bendi ng on el ast i c pl at es, " 9. Appl . Math. , Vol . 5, pp. 55-68, 1947. Basset . A. 6.: "On t he ext ensi on and f l exur e of cvl i ndr i cal and spher i cal t hi n el ast i c shel l s, " Phi l . Tr ans. Roya? SOC. , ( London) , Ser. A, Vol . 181 , No. 6, pp. 433-480, 1890. Hi l debr and, F. B.; Rei ssner , E. ; and Thomas, G. B.: "Not e on t he f oundat i ons of t he t heor y of smal l di spl acement s of or t hot r opi c shel l s, " NACA Techni cal Not e No. 1833, Mar ch 1949. Hencky, A.: "Uber di e ber ucksi cht i gung der schubver zer r ung i n ebenen pl at t en, " I ng. Archi v. , Vol . 16, pp. 72-76, 1947. Mi ndl i n, R. D.: "I nf l uence of r ot at or y i ner t i a and shear on f l exur al mot i ons of i sot r opi c, el ast i c pl at es, ' ' J . Appl . Mech. , Val . 18, pp. 31-38, 1951. Yang, P. C. ; Nor r i s, C. H. ; and St avsky, Y.: "Pl ast i c wave pr opagat i on i n het er ogeneous pl at es, " I nt l . J . Sol i ds & St r uct ur es, Vol . 2, pp. 665-684, 1966. Whi t ney, J . M.; and Pagano. N. J.: "Shear def or mat i on i n het er ogeneous ani sot r opi c pl at es, " J . Appl . Mech. , Vol . 37, pp. 1031-1036, 1970. Von Kar man, Th.: "Fest i qkei t spr obl eme i n Maschi nenbau, " Encykl opadi e der mat hemat i schen Wi ssenschaf t en, Vol . 4, Art . 27, pp. 348-351, 1970. Rei ssner , E.: "Fi ni t e def l ect i ons of sandwi ch pl at es, " - J . Aer onaut i cal Sci ences, Vol . 15, pp. 435-440, 1948. Rei ssner , E.: "On var i at i onal t heoremf or f i ni t e el ast i c def or mat i on, " J . Mat hemat i cs & Physi cs, Vol . 32, pp. 129-135, 1953. Medwadowski , S. J .: " A r ef i ned t heory of el ast i c, or t hot r opi c pl ates, I ' J . Appl . Mech. , Vol . 25, pp. 437-443, 1958. Schmi dt , R.: "A r ef i ned nonl i near t heor y of pl at es wi t h t r ansver se shear def or mat i on, " J . I ndust ri al Mat hemat i cs Soci et y, Vol . 27, No. 1, pp. 23-38, 1977. 75 76 15. Reddy, J . N.: "Bendi ng of l ami nat ed ani sot r opi c shel l s by a shear def or mabl e f i ni t e el ement , " Fi ber Sci ence & Tech. , Vol . 17, No. 1, pp. 9-24, 1982. 16. Li br escu, L.: El ast ost at i cs and Ki net i cs of Ani sot r opi c and Het er ogeneous Shel l - t ype St r uct ur es, Noor dhof f , The Net her l ands, 1975. 17. Lo, K. H.; Chr i st ensen, R. M. ; and Wu, E. M.: " A hi gher - or der t heory of pl at e def or mat i on, " Par t s 1 and 2, J . Appl . Mech. , Vol . 99, pp. 663-676, 1977. 18. Levi nson, M.: "An' accur at e si mpl e t heor y of t he st at i cs and dynami cs of el ast i c pl at es, " Mech. Res. Commun. , Vol . 7, pp. 343- 350, 1980. 19. Mur t hy, M. V. V.: "An i mproved t r ansver se shear def or mat i on t heor y f or l ami nat ed ani sot r opi c pl at es, " NASA Techni cal Paper 1903, November 1981. 20. Li br escu, L.; and Reddy, J . N.: " A cr i t i cal eval uat i on and gener al i zat i on of t he t heor y of ani sot r opi c l ami nat ed composi t e panel s, " Proc. Amer i can Soci et y f or Composi t es, Oct ober 7-9, 1986, Dayt on, OH, pp. 472-489. 21. Reddy, J . N.: " A r ef i ned nonl i near t heor y of pl at es wi t h t r ansver se shear def or mat i on, " I nt l . J . Sol i ds St r uct . , Vol . 23, No. 3, pp. 319-330, 1984. 22. Reddy, J . N.: " A si mpl e hi gher - or der t heor y f or l ami nat ed composi t e 23. Reddy, J . N.: " A Ref i ned Shear Def or mat i on Theor y f or t he Anal ysi s pl at es, " J . Appl . Mech. , Vol . 51, pp. 745-752, 1984. of Lami nat ed Pl at es, " NASA Cont r act or Repor t 3955, J anuar y 1986. 24. Love. A. E. H.: "On t he smal l f r ee vi br at i ons and def or mat i ons of t he- i l ast i c shel l s, " Phi l . Trans. Roy. SOC. ( London) , Ser. A, Vol . 17, pp. 491-546, 1888. 25. Naghdi , P. M.: " A sur vey of r ecent pr ogr ess i n t he t heor y of el ast i c shel l s, " Appl . Mech. Rev. , Vol . 9, No. 9, pp. 365-368, 1956. 26. Bert , C. W. : "Anal ysi s of shel l s, " i n Anal ysi s & Per f or mance of Composi t es, L. J . Br out man (ed. ), Wi l ey, New Yor k, pp. 207-258, 1980. 27. Koi t er , W. T.: " A consi st ent f i r st appr oxi mat i on i n t he gener al t heory of t hi n el ast i c shel l s, " Pr oceedi ngs of t he Symposi umon The Theory of Thi n El ast i c Shel l s. Del f t , pp. 12-33, August 24-28, 1959. 77 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. Ambar t sumyan, S. A.: "Cal cul at i on of l ami nat ed ani sot r opi c shel l s, " I zvest i a Akademi i a Nauk Armenskoi SSR, Ser. Fi z. Mat . Est. Tekh. Nauk. , Vol . 6, No. 3, p. 15, 1953. Ambar t sumyan, S. A.: Theor y of ani sot r opi c shel l s, Moscow, 1961; Engl i sh Tr ansl at i on, NASA Ti F-118, May 1964. Dong, S. B.; Pi st er , K. S.; and Tayl or , R. L.: "On t he t heor y of l ami nat e ani sot r opi c shel l s and pl at es, " J . of Aero. Sci . , Vol . 29, pp. 969-9753 1962. Cheng, S; and Ho, B. P. C.: "St abi l i t y of het er ogeneous ael ot r opi c cyl i ndr i cal shel l s under combi ned l oadi ng, " AI AA J . , Vol . 1, No. 4, pp. 892-898, 1963. Fl ugge, W.: St r esses i n shel l s, J ul i us Spr i nger , Ber l i n, Ger many, 1960. Wi der a, G. E. 0. ; and Chung, S. W. : " A t heor y f or nonhomogeneous ani sot r opi c cyl i ndr i cal shel l s, " J . Appl . Mat h. & Phy. , Vol . 21, pp. 378-399, 1970. Wi der a, G. E. 0.; and Logan, D. L.: "Ref i ned t heor i es f or nonhomogeneous ani sot r opi c cyl i ndr i cal shel l s: par t I - der i vat i on, " J . Eng. Mech. Di v. , ASCE, Vol . 106, No. EM6, pp. 1053- 1074, 1980. Li br escu, L.: "Subst ant i at i on of a t heor y of el ast i c and ani sot r opi c shel l s and pl at es, " St. Cerc. Mech. Appl , Vol . 1, pp. 105-128, 1968. Gul at i , S. T.; and Essenber q, F.: "Ef f ect s of ani sot r opy i n axi symmet r i c cyl i ndr i cal shel l s, " J . Appl . Mech. , Vol . ' %l , pp. 659- 666, 1967. Zukas, J . A.; and Vi nson, J . R.: "Lami nat ed t r ansver sel y i sot r opi c cyl i ndr i cal shel l s, " J . Appl . Mech. , Vol . 38, pp. 400-407, 1971. Dong, S. B.; and TSO, F. K. W.: "On a l ami nat ed or t hot r opi c shel l t heor y i ncl udi ng t r ansver se shear def or mat i on, " J . Appl . Mech. , Vol . 39, pp. 1091-1096, 1972. Whi t ney, J . M.; and Sun, C. T.: " A hi gher or der t heor y f or ext ensi onal mot i on of l ami nat ed ani sot r opi c shel l s and pl at es, " J. Sound Vi b. , Vol . 30, pp. 85-97, 1973. Reddy, J . N.: "Exact sol ut i ons of moder at el y t hi ck l ami nat ed shel l s, " J our nal of Engi neer i ng Mechani cs, ASCE, Vol . 110, pp. 794- 809, 1984. 78 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. Dong, S . B.: "Anal ysi s of l ami nat ed shel l s of J . Eng. Mech. Di v. , Vol . 92, EM6, pp. 135-141, 1966. Dong, S. B.; and Sel na, L. G.: "Nat ur al vi br at i ons of l ami nat ed or t hot r opi c shel l s o f r evol ut i on, " J . Composi t e Mat. , Vol . 4, No. 1, pp. 2-19, 1970. Wi l son, E. A; and Par sons, B. : "The f i ni t e el ement anal ysi s o f f i l ament - r ei nf or ced axi symmet r i c bodi es, " Fi br e Sci ence- & Tech. , Vol . 2, p. 155, 1969. Schmi t , L. A. ; and Monf or t on, G. R.: "Fi ni t e el ement anal ysi s of sandwi ch pl at e and l ami nat e shel l s wi t h l ami nat ed f aces, " AI AA J . , Vol . 8, pp. 1454-1461, 1970. Panda, S. C.; and Nat ar aj an, R.: "Fi ni t e el ement anal ysi s of l ami nat ed shel l s of r evoi ut i on, " Comput er s & St r uct ur es, Vol . 6, pp. 61-64, 1976. Shi vakumar , K. N.; and Kr i shna Mur t y, A. V.: " A hi gh pr eci si on r i ng el ement f or vi br at i ons of l ami nat ed shel l s, " J . Sound & Vi b. , Vol . 58, NO. 3, pp. 311-318, 1978. Rao, K. P.: "A r ect angul ar ani sot r opi c shal l owt hi n shel l f i ni t e el ement , " Comp. Met h. Appl . Mech. Eng. , Vol . 15, pp. 13-33, 1978. Sei de, P. ; and Chang, P. H. H.: "Fi ni t e el ement anal ysi s of l ami nat ed pl at es and shel l s, " NASA CR-157106, 1978. Venkat esh, A.; and Rao, K. P.: " A doubl y cur ved quadr i l at er al f i ni t e el ement f or t he anal ysi s of l ami nat ed ani sot r opi c t hi n shel l of r evol ut i on, " Comput er s & St r uct ur es, Vol . 12, pp. 825-832, 1980. Chao, W. C. ; and Reddy, J . N.: "Geomet r i cal l y nonl i near anal ysi s of l ayered composi t e pl at es and shel l s, " NASA CR-168182, Depar t ment of Engi neeri ng Sci ence and Mechani cs, Vi r gi ni a Pol yt echni c I nst i t ut e & St at e Uni v. , Bl acksbur g, VA, Feb. 1983. Reddy, J . N.; and Chandr ashekhar a, K. : "Geomet r i cal l y nonl i near t r ansi ent anal ysi s of l ami nat ed doubl e cur ved shel l s, " I nt l . J . Nonl i n. Mech. , Vol . 20, No. 2, pp. 79-90, 1985. Noor, A. K. ; and Har t l ey, S . J.: "Nonl i near shel l anal ysi s vi a mi xed i sopar amet r i c el ement s, " Comp. & St ruct . , Vol . 7, pp. 615- 626, 1977. Noor, A. K. ; and Ander sen, C. M.: "Mi xed i sopar amet r i c f i ni t e el ement model s of l ami nat ed composi t e shel l s, I l Comp. Met h. Appl . Mech. Eng. , Vol . 11, pp. 255-280, 1977. 79 54. Noor , A. K.; and Mat her s, M. 0.: "Shear f l exi bl e f i ni t e el ement model of l ami nat ed composi t e pl at es and shel l s, " NASA TN 0-8044, December 1975. 55. Sei de, P.: Smal l El ast i c Def or mat i ons of Thi n Shel l s, Noor dhof f I ntl . Publ . , Leydon, The Net her l ands, 1975. 56. Sr i ni vas, S.; and Rao, A. K.: " Bendi ng, vi br at i on and buckl i ng of si mpl y suppor t ed t hi ck or t hot r opi c r ect angul ar pl at es and l ami nat es, ' ' I ntl . J . Sol i ds St r uct ur es, Vol . 6, pp. 1463-1481, 1970. 57. Sr i ni vas, S.; J oga Rao, C. V.; and Rao, A. K.: " An exact anal ysi s f or vi br at i on of si mpl y- suppor t ed and l ami nat ed t hi ck r ect agul ar pl at es, " J . Sound Vi b. , Vol . 12, No. 2, pp. 187-199, 1970. 58. Pagano, N. J.: "Exact sol ut i ons f or r ect angul ar bi di r ect i onal composi t es and sandwi ch pl at es, " J . Composi t e Mat er i al s, Vol . 4, pp. 20-34, 1970. 59. Pagano, N. J . ; and Hat f i el d, S . J.: "El ast i c behavi or of mul t i l ayer bi - di r ect i onal composi t es, " AI AA J . , Vol . 10, pp. 931-933, 1972. 60. Pagano, N. J.: "St r ess f i el ds i n composi t e l ami nat es, " I nt l . J . 61. Reddy, J . N. ; and Phan, N. 0.: "St abi l i t y and vi br at i on of Sol i ds & St r uct ur es, Vol . 14, pp. 385-400, 1978. i sot r opi c, or t hot r opi c- and l ami nat ed pl at es usi ng a hi gher - or der shear def or mat i on t heor y, " J . Sound & Vi br at i on, Vol . 98, No. 2, pp. 157-1703 1985. 62. Put cha, N. S. ; and Reddy, J . N.: " A r ef i ned mi xed shear f l exi bl e f i ni t e el ement f or t he nonl i near anal ysi s of l ami nat ed pl at es, " Comput er s & St r uct ur es, Vol . 22, No. 4, pp. 529-538, 1986. 63. Reddy, J . N.; and Put cha, N. S.: "St abi l i t y and vi br at i on anal ysi s of l ami nat ed pl at es by usi ng a mi xed el ement based on a r ef i ned pl at e 1986. J . Sound & Vi br at i on, Vol . 104, No. 2, pp. 285-300, 64. Hel l i nger , E.: "Di e al l egemei nen Ansat ze der Mechani k der Kont i nua, " Ar t 30, i n Encykl opadi e der Mat hemat i schen Wi ssenchaf t en, Vol . I V/ 4, 654-655, F. Kl ei n and C. Mul l er (eds. ), Lei pzi g, Teubner , 1914. 65. Rei ssner , E.: "On a var i at i onal t heor emi n el ast i ci t y, " J . Mat h. 66. Hu. H.: "On some var i at i onal pr i nci pl es i n t he t heor y of el ast i ci t y PhyS. , Vol . 29, pp. 90-95, 1950. and t he t heor y of pl ast i ci t y, " Sci ent i a Si ni c, Vol . 4, pp. 33-54, 1955. 80 67. Rei ssner , E.: "On var i at i onal pr i nci pl es of el asti ci ty, I l Proc. 68. Washi zu, K.: "Var i at i onal pr i nci pl es i n cont i nuummechani cs, ' ' Dept . Symp. Appl . Mat h. , Vol . 8, pp. 1- 6, 1958. of Aer onaut i cal Engi neer i ng, Repor t 62- 2, Uni v. of Washi ngt on, 1962. . 69. Tont i , E.: "Var i at i onal pr i nci pl es of el ast ost at i cs, " Mechani ca, Vol . 2, pp. 201- 208, 1967. 70. Oden, J . T.; and Reddy, J . N.: "On dual compl ement ar y var i at i onal pr i nci pl es i n mat hemat i cal physi cs, " I ntl . J . Engng. - Sci . , Vol . 12, pp. 1- 29, 1974. 71. Oden, J . T.; and Reddy, J . N.: Var i at i onal Met hods i n Theor et i cal Mechani cs, Second Edi t i on, Spr i nger - Ver l ag, Ber l i n, 1983. 72. Her r mann, L. R.: " A bendi ng anal ysi s of pl at es, " Proc. 1st Conf . Mat ri x Mech. St r uct . Mech. , Wr i ght - Pat t er son AFB, AFFDL- TR- 66- 80, Dayt on, OH, pp. 577- 604, 1966. 73. Her r mann, L. R.: "Fi ni t e el ement bendi ng anal ysi s f or pl at es, " J . Enqng. Mech. Di v. , ASCE, Vol . 93, EM5, pp. 13- 26, 1967. 74. Hel l an, K.: "Anal ysi s of el ast i c pl at es i n f l exur e by a si mpl i f i ed f i ni t e el ement met hod, " Act a Pol yt ech. Scand. Ci vi l Engng. , Ser i es No. 46, Tr ondhei m, 1967. 75. Reddy, J . N.: An I nt r oduct i on t o t he Fi ni t e El ement Met hod, McGr aw- Hi l l , New Yor k, 1984. 76. Vl asov, V. Z.: Gener al Theor y of Shel l s & I t s Appl i cat i ons i n Engi neer i ng, ( Tr ansl at i on of Obshchaya t eor i ya obol ocheck i yeye pr i l ozheni ya v t ekhni ke) , NASA TT F- 99, NASA, 1964. 77. Ti moshenko, S. ; and Woi nowsky- Kr i eger , S.: Theor y of Pl at es and Shel l s, 2nd ed. , McGr aw- Hi l l , New Yor k, 1959. 78. Pr at o, C. A.: "Shel l f i ni t e el ement met hod vi a Rei ssner ' s pri nci pl e, " I nt l . J . Sol i ds St ruct . , Vol . 5, p. 11- 19, 1969. 79. Zaqhl oul , S. A. ; and Kennedy, J . B.: "Nonl i near behavi or of symmet r i cal l y l ami nat ed pl at es, " J . Appl . Mech. , Vol . 42, pp. 234- 236, 1975. 80. Chang, T. Y.; and Sawami phakdi , K.: "Lar ge def l ect i on anal ysi s of l ami nat ed shel l s by f i ni t e el ement met hod, " Comp. & St r uct . , Vol . 13, pp. 331- 340, 1981. 81 81. 82. 83. Chi a, C. Y.; and Pr abhakar a, M. K.: " A gener al mode appr oach t o nonl i near f l exur al vi br at i ons of l ami nat ed r ect angul ar pl at es, " ASME, J . Appl . Mech. , Vol . 45, No. 3, pp. 623-628, 1978. Reddy, J . N.; and Chao, W. C.: "Nonl i near osci l l at i ons of l ami nat ed ani sot r opi c r ect angul ar pl at es, " Trans. of ASME, Vol . 49, pp. 396- 402, 1982. Chandr a, R.; and Raj u, B. Basava: "Lar qe arnDl i tude f l exur al vi br at i on of cr oss- pl y l ami ant ed compo; i te pl at es, " Fi b. Sci . Tech. , Vol . 8, pp. 243-263, 1975. APPENDIX A COEFFICIENTS OF THE NAVI E R SOLUTION 4E66 +-1 4E 12 c(1,5) = a0(-B12 - 866 +- 3h2 3h2 C(2,3) = 0(- A12 +R) A22 +a 2 "(-7- 4E12 R1 2 3h 2 2 4 16H22 32a B - 9h4 (H12 + 2H66) c(3,3) = a 4 (- z) 16Hll + 5 (- 8E12 8E22 +2) -) 2 8D55 16F55 8Ell 2 8D44 16F44 8E 12 - - - - - 2 2 +a (-A55 +- h2 h4 3h R1 3h R2 + 6 (-A44 +7 - - - - 2 h4 3h R1 3h R2 82 83 - - All 2A12 A22 2 R1 -m-G- 2 16aB (H12 +2F66) - - 3 4a 3h 9h4 C(3,4) =- B1l B12 4 E12 +2ti66) +a [ r +- - - +T I 1 - A55a 1 R2 3h2 (T 2 8aD55 16aF55 h2 h4 +-- 8BD44 168Fqq h2 h4 +-- 86' 16 2 2 2 2 +H66B ) - Dlla - 0668 2 +- 8a 3h 2 F66 - 2 (H1l a C(4,4) =- 2 F1l 3h 2 8F66 2 8F22 16H - :6) + 0 (-DZ2 +- - +3h2 9h 3h2 c(595) = a (-066 84 M(1,2) = 0 4 4 M(1,3) = (7 I4 +- 3h 2 R1 I 5l a 3h 4 -- '3 4 2 I 51 M(1,4) = - (I 2 +- - - R1 3h2 I4 3h R1 M(1,5) = 0 M(292) = -(I1 +-) 212 R2 4 4 M(2,3) = (7 I4 +- 3h 3h R2 2 '5)B M(2,4) = 0 I 3 4 4 M(295) = - (I 2 +- - 7 I4 - - 2 I 5) R2 3h 3h R2 16 2 16 2 M(3,3) = - 7 17a - 7 1 7 ~ - I1 9h 9h 16 3h l6 I ) 8 3h M(4,4) = -(I3 - 7 I5 +2 M(4,5) = 0 8 16 M(5Y5) = 413 - 2 I 5 +gh4 17) 3h APPENDIX B STIFFNESS COEFFICIENTS FOR THE MIXED MODEL The element equation (5.16) can be written in the form 12 22 [K I ...... [K ] ...... [ K(ll)l] [ K(ll)'] . . . 1 aw 1 aw 1 aw + - A 2 12 [axp '12' + 2 A13 '5 '12' + 2 A13 [$ 'ill + 7 1 A32 '3 aw '22' + 7 'A 33 [% axl '22' + 7 A 33 [x ax2 SZ1] 14 15 [K l = [ K l = O a5 86 23 A3 1 1 aw A32 [ K 1 3 - R1 [slO1 + 2 A 3 1 ' T '11' + ['lo1 1 aw 1 aw 1 aw + 2 A32 '3 '12' + 2 A33 '12' + 2 A33 [axp '11' 1 aw 2 22 ax2 +- A [- SZ2 24 25 [ K ] = [ K 1 - 0 A22 s211 + RE [s201 87 A12 A13 aw aw ['0Z1 + R + A 1 2 [ q '12' + A 1 3 [ q '11' =- R1 1 A 1 13 [aw '02' + 2 7 5 '01' A 1 13 [aw A 1 12 [aw 2 R1 +- - 3 '02' + 7 aw 1 aw 2 A 1 ~ [aw - sl ol +- 1- '101 +7 A l l I(-) S I 11 +- R1 aX1 axl R2 axl + - A 1 [-- aw aw aw 2 2 12 ax1 ax2 '12' ' $ A13 '(q) '12' s ] +- A21 + - A 1 [-- aw aw 2 13 axl ax2 11 RI RZ A A22 1 A22 aw 1 23 [aw +2 [Soo1 +- - [- SO2] +- - - 2 R2 ax2 2 R2 axl '02' R2 1 2 3 [ ~ A21 - S20] + - A 1 [-- aw aw [ aw A22 [aw 1 aw 2 1 aw aw 1 aw 2 A31 [ aw 1 aw 2 2 23 ax2 A +- - 2 R2 ax2 '01' + 2 21 axl ax2 '211 ax2 +- - s201 +2 A22[($ s221 +- 2 A 23 [-- axl - ax2 '22' R2 ax2 +- A [(-) S211 +- - '201 +2 A31[($) 5211 R1 axl aw 2 A I(-) S221 1 aw aw 2 32 ax1 ax2 '22' + 7 33 axl A32 [E S20] +- A [-- +- R2 axl 1 aw aw A31 aw 1 aw aw 2 33 axl ax2 '21' + [- ax2 '10' + 7 A 31[ q 3 '11' + - A [-- 1 aw 2 1 aw aw '10' + 7 A 32' ( q) '12' + 2 A 33' q 3 '12' +- + x45 36 51 aw B2 1 R1 R2 [ K I = - [ So , ] +B l l [ ~ Sl o] +- aw + B 3 1 ' 3 '20' + '31[% '10' '001 + 824% s20 89 +B [aw S ~ ~ I +~32[$ aw slOl 32 axl [K3'] = [ K3 11] 41 [K I = 42 [K ] = O [ K ~ ~ I =R ~ ~ [ s ~ ~ I +R ~ ~ [ s ~ ~ I 90 IK41 = CPISol] , [K4,lo] = 0 , [K4,11] = CP[So2] [K ] = [ K ] = 0 51 52 63 Bll aw aw [K ] =- R1 91 67 [K I = -D12[Sool [K681 = -Dl3[Soo1 69 [K 1 = -F11[SooI [K6lo1 = -Fl2[So0] [K611] = -F [S ] 13 00 [K7lI = Bl2[So1I +B32[So21 [K7I = B22[S021 +B32[Soll aw 1 aw +- B [- so11 1 +- 2 32 [- axl 02 2 32 ax2 74 75 [K 1 = 0 [K 1 = [So21 76 [K 1 = -D21[Sool [K I = -D22[Soo1 [K 1 = -D23[Sool [K 1 = -F21[So0l 77 78 79 [K7l0] = -F [S ] [K711] = -Fz3[So0] [K 22 00 [ S 1 I =Bl3[soll +B33 o2 81 92 83 '13 1 aw '23 1 aw [ K ] = - R1 [sOO1 + 2 B13[$ '01' + [sOO1 + 7 B23[$ '02' 93 +- 1 E [- aw So21 +- 1 E [- aw Sol] +CPISll] 2 31 axl 2 31 ax2 94 95 [K I = CP[Slo1 [K I = 0 [ K ~ ~ I = - F ~ ~ [ s ~ ~ I [Kg71 = - F21[Sool [Kg8I = - F31[Sool 99 [K I = - H1l[Sool [K910] = - H [S ] [K9y111 = - H13[Soo] [K1Oll = El2[Sol1 +E32[So21 12 00 [K1Oy2I = E22[So21 +E32 [S 01 1 +- 1 E [- aw So2] +- 1 E [- aw Sol ] +CP[S22] 2 32 axl 2 32 ax2 [K1Os41 = 0 , [K10*5] = CP[Sz0] [K1096] = - F [ S ] 12 00 [K10971 = - F22[Soo] 94 [ K1OY8] = - F32[Soo] [ K109 ] = - HZ1[Soo] [ KIO 9 l o ] = - H22[Sool [K1o ll = - H23[Soo] [ K1191] = E 13 [S 01 .] + E33 [ I d2 [ K11921 = E23[So21 +E33 [S o1 I [ K1194] = CP[ Sz0] [ K 1 l y 5 1 = CP[ Sl 0] [ K 1 l Y8 ] = - F33[ Soo] [K1llo] = - H32[Soo] [ K11911] = - H33[Soo] Mas s M a t r i x [ MI f or t h e Mi x e d Mo d e l 95 .... .......................... - - - - I [S ] [M31] 3hZ 4 01 = I [S ] - - u s I 2 00 3h2 4 00 - - - u s I 3h2 4 10 41 - 1 3hZ 4 00 [ M 1 = '2~sool - - u s 1 - - - - I [ S ] [ M ] = - - 3hZ 4 02 3hZ 4 20 32 = (I2 - - 42 I ~ ) [ s ~ ~ I , [M5'] = [MZ5] 3h 44 = [ M 1 al l others are zero, and Ii are defined i n Eq. (2.18b) 96 - 4 2 A45 = A45 - 2E45 h2 +G45(7) - 4 4 2 = A55 - 2E55 7 + G55 $7) 4 CP =- 3h2 aJli a xO with - - - 6i For the tangent stiffness matrix, the coefficients are given by [ (KT) 13 ] = 2[ K13] = [ K31] [ ( KT) 23] = 2[ KZ3] = [ K 32 ] [ ( KT) 331 = [ K 331 au au au + A 1 3 [ 3 '11l + A 2 3 [ 3 '22' + A 3 3 [ 3 ('21 + '12)' 97 av av av + A 13[ q '11' + A 2 3 [ q '22' + A 33[ q ('21 + '12)' aw aw aw 2 s211 + [(-I '2211 axl 1 2 21 axl ax2 + - A [[-- 1 aw aw aw 2 2 23 axl ax2 1 aw aw aw 2 2 31 axl ax2 + - A [[-- '221 + [(-) '211 [[-- 'ill + [(-) '121 ax2 axl + - A 1 aw 2 aw aw +7 A33[[(-) '111 + [ax ax '1211 + I(-) '2111 a x2 1 2 aw aw aw 2 1 + - A 2 32 [[-- axl ax2 '22l ax2 1 aw aw + 7 q iA1l ["X1 '011 + A12 1% '021 98 99 [ ( K T ) ~ ~ ~ = [K371 [ ( K T ) ~ ~ ~ = [K381 [(KT)931 = IK31 [ (K,)103] = [K3] [ ( K T ) l l y3] = [K311] A l l others are same as those i n [K]. Standard Bibliographic Page 17. Key Words (Suggested by Authors(s)) klgl e-ply , composite laminates, cross-ply, closed- form sol uti on, f i ni te element analysis, Report No. NASA CR-4056 18. Distribution Statement Unclassified - Unlimited , Titie and Subtitle A Higher-Order Theory f or Geometri cal l y Nonl i near Anal ysi s of Composite Laminates - . Author(s) J . N. Reddy and C. F. L i u - , Uncl ass i f i ed . Performing .Organization Name and Address Virginia Polytechnic I nsti tute and State University Department of Engineering Science and Mechanics Blacksburg, VA 24061 Uncl assi f ied I 105 I A06 2. Sponsoring Agency Name and Address National Aeronautics and Space Administration Washington, DC 20546 3. Recipient's Catalog No. 5. Report Date MARCH 1987 6 . Performing Organization Code 8. Performing Organization Report No. VPI- E-86.21 10. Work Unit No. 11. Contract or Grant No. NAG 1 - 459 13. Type of Report and Period Covered Contractor Report _ _ 14. Sponsoring Agency Code 505-63- 31-02 5. Supplementary Notes NASA Langley Research Center Technical Monitor: Dr. Norman K ni ght, J r. Final Report hi gher-order theory , pl ates and she1 1 s , nonlinear anal ysi s, shear deformation theory, vibration Subject Category 39 I 19. Securitv Classif.(of this report) 1 20. Security Classif.(of this page) I 21. No. of Pages I 22. Price For sale by the National Technical Information Service, Springfield, Virginia 22161 NASA-Langley, 1987