THEORY OF ANISOTROPIC THIN-WALLED CLOSED-SECTION BEAMS WITH HYGROTHERMAL EFFECTS

Ashraf M. Badir* School of Aerospace Engineering Georgia Institute of Technology Atlanta, Georgia

Abstract A variationally consistent theory is developed in order to derive the governing equations of anisotropic thin-walled beams with closed sections subjected to axial load, torsion, bending and hygrothermal loads. The theory is based on an asymptotical analysis of twc-dimensional shell theory. Closed-form expressions for the beam stiffness coefficients, equivalent hygrotherma1 loads, stress and displacement fields are provided. The influence of material anisotropy and hygrothermal coefficients on the displacement field is identified. The analysis is applied to predict the response of composite beams under thermomechanical loading. Introduction Thermal and moisture stresses are induced in anisotropic materials as a result of the mismatch in the thermal expansion and moisture swelling coefficients of their constituents. In composite materials residual thermal stresses result from the curing cycle. These stresses can have a significant effect on their strength and stiffness.

Figure 1: Cartesian Coordinate System due to the chan es in moisture content or temperature[4]*[51*[ 1. This hygrothermally induced warping affects the dimensional stability of the composite. The objective of this work is t o develop a theory for anisotropic thin-walled beams which enables the prediction of their response under thermomechanical loading. The theory is an extension of the work of ~ a d i r l and ~l Berdichevsky et al.i81,and considers thin-walled closed section beams with circumferentially varying thickness and stiffness. It is based on an asymptotical expansion of two dimensional shell theory.

Matrix cracking has been observed in ceramic matrix composite due t o the residual thermal A detailed derivation of the theory is presented stresses[lI . In laminated composite plates made first. This is followed by a summary of the govof graphite/epoxy material, residual thermal erning equations. Illustrative examples are p r e stresses increase the interlaminar s t r e s ~ e s [ ~ ] * vided 1 ~ ] in order to isolate the effect of thermal and and consequently precipitate failure. On the moisture stresses on the deformation. other hand, the same studies show that residual moisture stresses tend to alleviate this effect by decreasing the interlaminar stresses. Development of the Analytical Model From a stiffness stand point, unsymmetric laminated composites warp, bend or twist
'Graduate Research Assistant, member of AIAA, ASCE. Copyright 0 1 9 9 2 by Ashraf M. Badir. Published by the American Institute of Aeronautics and Astronautics, Inc. with permission.

Coordinate Systems Consider the slender thin-walled elastic cylindrical shell shown in Fig. 1. The length of the shell is denoted by L, its thickness by h, the radius of curvature of the middle surface by R and

Equations y = y(s) and z = z(s) define the closed contour r in the y, z plane. The normal vector to the middle surface 6 has two nonzero components 6 = n,(s)i, n,(s)i' (2)

The position vector I? of an arbitrary material point can be written in the form

Figure 2: Curvilinear Coordinate System

Equations (2) and (3) establish the relations between the cartesian coordinates x, y, t and the curvilinear coordinates x, s, J. The coordinate J lies within the limits

the maximum cross sectional dimension by d. I t is assumed that The shell thickness varies along the circumferencial direction and is denoted by h(s). The tangent vector the normal vecto56 and the projection of the position vector ?on t and n' The shell is loaded by external forces applied are expressed in terms of the cartesian and curvito the lateral surfaces and at the ends. The shell linear coordinates as undergoes changes in temperature and moisture content as well. It is assumed that the variation of the external forces and material properties over distances of order d in the axial direction and over distances of order h in the circumferencial direction, is small. The material is anisotropic and its properties can vary circumferencially and in the normal direction to the middle surface as well. It is convenient to consider simultaneously two coordinate systems for the description of the state of stress in thin-walled beams. The first one is the cartesian system x, y and z shown in Fig. 1. The axial coordinate is x while y and z are associated with the beam cross section. The second coordinate system, is the curvilinear system x, s and E shown in Fig. 2. The circumferencial coordinate s is measured along the tangent to the middle surface in a counter-clockwise direction whereas J is measured along the normal to the middle surface. A number of relationships have a simpler form when expressed in terms of curvilinear coordinates. A relationship between the two coordinate systems can be established as follows. Define the position vector F of the shell middle surface as

<

An asymptotical analysis is used to model the slender thin-walled shell as a beam with effective stiffnesses. The method follows an iterative process. The displacement function corresponding to the zeroth-order approximation is obtained first by keeping the leading order terms in the energy functional. A set of successive corrections is added and the associated energy functional is determined. The process is terminated when the new cycle does not generate any additional terms. Shell Energv Functional

Consider in a three dimensional space the prismatic shell shown in Fig. 2. A curvilinear frame x, s, and 5 is associated with the undeformed shell configuration. Values 1, 2 and 3 denoting z, s, and J, respectively are assigned to the curvi, are unit vectors associated with linear frame. Throughout this work, Latin suwhere iz, i the cartesian coordinate system x, y and z . perscripts (or subscripts) run from 1 to 3, while

c,

Greek superscripts (or subscripts) run from 1 to 2, unless otherwise stated. The energy density of a 3 0 elastic body is a quadratic form of the strains
U=-E 2
i j k l (m) (m) E . . E~~
13

PI2 =

avl - 3-) av2 as ax a 2 ~ a v2 P22 = as2 - -(-) as R

asax + -(4~
1

aZv

where v l , v2 and v represent the displacements in the axial, tangential and normal directions, repectively as shown in Fig. 2. These are related The material properties are expressed by the to the displacement components in cartesian coHooke's tensor Eijkl. The mechanical strain ordinates by v1 = 211 is the difference between the total strain ~ i and j the nonmechanical strain E ~ T ~ i.e. ),

E!?)

yij and pij represent the total strains and the total change of curvatures, respectively. The therThe 3 0 energy is first minimized with respect mal coefficients are denoted by cvij in Eq. (6), while the swelling coefficients by Pij: The change to E,!?). This is equivalent to satisfying Eq. (7). in temparature between the ambient and the The result is stress-free temperature is denoted by AT. The percentage moisture weight gain is represented by A H .

The energy density of the 2 0 elastic body is obtained in terms of yap and pap by the following procedure.

where Dapy6 represent the component of the 2 0 Following classical shell formulation [9],[10],[11] modulii. The expressions for D@y6 are given in the through-the-thickness stress components ui3 ~ " 8 in 7 the ~ Appendix. terms of are considerably smaller than the remaining - components uaP therefore The strain dm)from Eq. (8) is substituted into u'3 0 "Smtenration of the result over the (7) Eq. (11). After thickness 5 one obtains the energy of the shell cf, The mechanical strains can be written as per unit middle surface area
3

where yap and pap represent the total in-plane strain components and the total change in the shell middle surface curvatures , respectively. For a cylindrical shell these are related to the dis- where placement variables by
711

+-h3 ~;@76 Pappy6 12

cap76 = -

+ +(AT, A H )

(12)

=-

av1

ax

c;h6

1 <~ a p 7 6 > h 2 <D ~ P Y ~ =h2

E>

and a function of 5, say a(E), between pointed brackets is defined as

<a>=

(14)

722

The function \k(AT, A H ) in Eq. (12) is inessential in the derivation of the beam rigidities since By keeping the leading order terms in the it is independent of the strains. For an applied strain-displacement relationships, Eq. (9) can be external loading Pi, the displacement field u, dewritten as termining the deformed state are the stationary points of the energy functional

Since 7 1 1 and pll are much smaller than 712, and pl2, p22, respectively, their contribution to the elastic energy is neglected.

I =

J adzds - J piuidzds

(15)

Asymptotical Analysis Zeroth-Order A ~ ~ r o x i m a t i o n The order of magnitude of the shell energy per Let A and E be the order of displacements and unit area and the work done by external forces is stiffness coefficients C a P y 6 , respectively. Assume that the order of the external forces is EAh P - O ( ~ )

This assumption is shown later to be consis tent with the equilibrium equations. An alternative would be to assume the order of the external Since Piui << a, the contribution of external force as some quantity P and derive the order of forces is neglected. Therefore the energy functhe displacements as P L 2 / ~ from h an asymptot- tional takes the form ical analysis of the energy functional. The nonmechanical strain is assumed to be of the same order of the total in-plane strain, i.e.

For a thin-walled slender beam whose dimensions satisfy Eq. (1) the rate of change of the displacements along the axial direction is much smaller than their rate of change along the circumferencial direction. That is, for each displacement component

Using Eqs. (9) and assuming that d is of the The integrand in Eq. (17) is a positive same order or smaller than R, the order of magquadratic form, therefore the minimum of the nitude of the in-plane strains and curvatures is functional is reached by functions v, vl, and v2 for which 712 = 722 = ,712 = p22 = 0. Fkom Eq. (16) this corresponds to

(18)

where U2 and U3 are arbitrary functions of x. Substitute from Eq. (26)into Eq. (19) to get

Equations (22), (26) and (27) represent the curvilinear displacement field that minimizes the zeroth-order approximation of the shell energy. Using Eq. (10) the curvilinear displacement field The function v in Eqs. (19)and (20)should be is written in cartesian coordinates as single valued, i. e.

u3 = U3(x) Y(P(X) The bar in Eq. (21)and in the subsequent derivation denotes averaging along the closed contour The cross section translation is denoted by U1 (x) ,U 2 (2) and U 3 (x) while p(x) represents r whose length is denoted by I . the cross-section rotation normally referred to in beam theory as the torsional rotation. This disEquation (18) implies that vl is a function of placement field corresponds to the zeroth-order x only, i.e. approximation and does not include bending beVl = U l k ) (22) havior and hygrothermal contribution. Integrate Eq. (20)t o get First-Order Approximation A first-order approximation can be constructed where p(x) is an arbitrary function which is by rewritting the displacement field in Eqs. (22), shown later to represent the cross sectional ro- (26) and (27) in the form tation. From Eq. (21) and (23),one obtains the relation between p(x) and 212.

Substitute v from Eq. (19) into Eq. (23),to get the following second order differential equation where wl, w 2 and w can be regarded as correcfor v7 tion functions to be determined based on their contributions to the energy functional. To solve this equation, one has to recall the relaSubstitute Eq. (28) into (9) to obtain the tions between the radius of curvature R and the components y(s) and z(s) of the position vector strains and curvatures in terms of the displacement corrections associated with contour I '

It follows from Eq. (25) that and are solutions of the homogeneous form of Eq. (24) and v2 = cp(x)rn is its particular solution. The general solution is therefore given by

If the rigid body motion is suppressed the solution is unique. The terms biz, h z are essential to the uniqueness of the solution, however, their where 7 O a P and poaP are the strains and curvatures corresponding to the zeroth-order approxi- contribution to the energy is of order ~ h g mation. These are expressed as and is consequently dropped. This aspect is discussed by Berdichevsky and ~ i s i u r a [ ' ~with I regard t o the accuracy of classical shell theory. The shell energy can therefore be represented by

(4))

It is worth noting that the bending contributioi does not appear in Eq. (31). That is, to the firstorder approximation the shell energy corresponds to a membrane state. The first variation of the energy functional is

+ 6am
8722

(-+- w,) ) dsdx

(32)

The prime in Eqs. (30) denotes differentiation = Nz2, Eq. with respect to rc. The order of w, is as Recall that m = NI2 and shown in Ref. [7]. An alternative approach is to (32) takes the form and verify this asassume the order of wi as sumption, as shown later, once wi is determined. Among the new terms introduced by the function wi the leading ones are denoted by superscript in Eqs. (29). By keeping their contribution over -theother terms, the energy functional can be represented by Set the first variation of the energy to zero, to obtain the following

(y)

(y)

where terms of order

(g) or smaller such as

h;12,712 , h;12%2, h2i12b12, h2h21j22 are neglected in comparison with the following which result in terms
0

711j12

711,722 , 712,712 , 7 1 2 j 2 2

N12 = constant

(33)

(34) Nz2 = 0 of order Similarly, the contribution of the work done by external forces, Piwi, is neglected This is similar to the classical solution of constant since its order is ~hs in comparison with shear flow and.vanishing hoop stress. By Setting NZ2 to zero the energy density is expressed in the order of the remaining terms in the energy terms of 7 1 1 and 7 1 2 only functional Eh-p . Therefore in order to determine the functions wi one has to minimize the functional

(g).
( ( "')

and

(t))

The variables A(s), B(s) and C(s) represent where the axial, coupling and shear stiffnesses, respectively, while the nonmechanical contribution to the energy density is represented by A("")(s), d n r n ) ( s ) ,and \k2(AT, A H ) . They are defined in terms of the 2 0 shell modulii in the Appendix. The function \k2(AT, A H ) in Eq. (35) is inessential in the derivation of the beam rigidities. Equation (35) indicates that, to the first order, the energy density function is independent of functions w2 and w. That is the in-plane warping contribution to the shell energy is negligible. The The area enclosed by contour r is denoted by A, function wl however, can be determined from in Eqs. (39). Eq. (33) and (35) and by enforcing the condition on wl to be single valued as follows The displacement field corresponding to the first correction is obtained by substituting Eq. (38) into Eq. (28) and dropping w2 and w since their contribution to the shell energy is neg1 -dnrn) (s) = constant (36) ligible compared t o wl. The result is 4

Substitute the leading terms from Eqs. (29) and (30) into (36) to get

+ ] :a + -c("")(s)

Displacement Field In deriving Eq. (37) the term B% has been neglected in comparison with This is The displacement field corresponding to the possible, if IBI is less or of the same order of next correction is found in the same way, and magnitude as C. For the case when IBI >> C is given by [7]: additional investigation is needed. Since the e l a s tic energy is positive definite, B2 5 AC, and B could be greater than C only if A >> C . In practical laminated composite designs JBI < C as the shear stiffness is greater than the extension-shear coupling.

= constant

(37)

$c%.

Equation (37) is a first order ordinary differential equation in wl. The value of the constant in the right hand side of Eq. (37) can be found where from the single value condition of function w l :

The solution of Eq. (37) is determined within an arbitrary function of x. The result is the followA third cycle is carried out, however no addiing simple analytical solution of Eq. (37) tional terms result from the third cycle17] and the displacement field converges to the expression given in Eq. (40).

The underlined terms in Eq. (40) correspond to the classical theory of extension, bending and torsion of beams. The additional terms gl ( s ) U ; , g2(s)Ug and g3(s)U; in Eq. (40) r e p resent warping due to axial strain and bending. These new terms emerge naturally in addition to the classical torsional related warping G(s)cpl.They are strongly influenced by the material anisotropy, and vanish for materials that are either orthotropic or whose properties are antisymmetric relative to the shell middle surface. For these layups the coupling parameter b(s) defined in Eqs. (39) and (58) vanishes. The significance of the axial and bending-related warping terms and their effect on the accuracy, is shown in the applications provided in this paper. .-,. The contribution of shear deformation is found1'I to be implicitly accounted for through the out-ofplane warping due to bending. The last term t ( s ) represents the contribution of thermal and moisture stresses. Strain Field The strain field is obtained by substituting Eq. (40) into Eq. ( 9 ) and neglecting terms of smaller order in the shell The result is

where

Explicit expressions for the stiffness coefficients Cij (i, j = 1, 4 ) and equivalent hygrothermal loads ~(""1, MP"), Men), and M,("") are given in the Appendix. The constitutive relationships can be written in terms of stress resultants and kinematic variables by differentiating Eq. (44) with respect to the associated kinematic variable or by relating the traction T, torsional moment M,, and bending Moments My and M, t o the shear flow and axial stress as follows

The shear flow N l z is derived from the energy density in Eq. (36) and the axial stress resultant Nll is given by

The associated axial and shear stresses are uniform through the wall thickness.
722

=0

(42)

It is worth noting that the vanishing of hoop

stress resultant in Eq. (34) and hoop strain in Eq. (42) should be interpreted as negligible contribution relative to the other parameters. Constitutive Relationships Substitute Eq. (42) in the energy functional and integrate over s to get the energy of 1D beam theory

Substitute Eq. (42) into Eqs. (36) and (46) and use Eq. (45) to get

different non-classical behaviors. These behaviors are influenced by the out-of-plane warping The equilibrium equations can be derived by due to extension and bending as outlined in the substituting the displacement field in Eq. (40) following. into the energy functional in Eq. (15) and using CUS Configuration the principle of minimum total potential energy to get This configuration produces both extensionT ' + fp,ds = 0 twist and bending-transverse shear couplings. The axial, coupling and in-plane stiffnesses A, B, and C given in Eq. (58) are constant throughout the cross section and hence the name circumferentially uniform stiffness (CUS) adopted in Ref. 1151, 1161, [17] and [20]. Such a configuration is manufactured by wrapping the composite lay-up using winding technique. For a box-beam, the ply lay-ups on opposite sides are of reversed orientation, and hence the name antisymmetric conwhere P,, Py and P, are surface tractions along figuration adopted in Refs. [14], (181, and [19]. the x, y and z directions, respectively. Equilibrium Equations Since A, B, and C are constants, the stiffness One of the member of each of the following four matrix in Eq. (61), for a centroidal coordinate pairs must be prescribed at the beam ends system, reduces t o

T o r U1, M, or p, My or U$,and M, or Ui

Summarv of Governing Eauations The nonzero stiffness coefficients are given by The development presented in this work encompasses five equations. The first, is the displacement field given in Eq. (40). Its functional form was determined based on an asyrnptotical expansion of shell energy. The associated strain field is given in Eq. (42) and the stress resultants in Eq. (36), (45) and (46). The fourth, are the constitutive relationships in Eq. (47) with the stiffness coefficients expressed as integral of material properties and cross sectional geometry in Eq. (61) of the Appendix. Finally the equilibrium equations and boundary conditions are given in Eq. (48) and (49), respectively.

CI2= BA,

where 1 denotes the length of the closed contour r. For such a case the out-of-plane warping due In the present development the determination to axial strain vanishes and gl does not affect the of the displacement field is essential in obtaining response. The influence of the out-of-plane warpaccurate expressions for the beam stiffnesses and ing due to bending in the x-z and x-y planes are equivalent temperature and moisture loads. expressed by the underlined terms in the expressions of C 3 3 and (744, respectively. These terms are significant in predicting the deflection of antisymmetric configurations. Applications For CUS configurations, temperature and moisture change induce equivalent axial and torsional moment while no equivalent bending moments are generated. This is derived from Eq. (62) of the Appendix.

Two special layups: The circumferencially uniform stiffness (CUS) and circumferencially Asymmetric stiffness (CAS) have been considered in Refs. [14]-[20]. They are associated with

Following the cure cycle, the drop in tempera- The nonzero stiffness coefficients are expressed ture causes the CUS configuration to twist. This by pretwist of the beam p(x) can be predicted by B2 C I 1= A1 - 2 " d considering the mechanically applied load to be Ct zero in Eq. (47) while using the stiffness and applied equivalent loads given in Eq. (51) and (63), Ct c 2 2= A : respectively. For the case where the through-thethickness properties are neglected in Eq. (58), the [d+a result is

-(g)]

where, the stiffnesses K i j are

CAS Configuration This configuration produces both bendingtwist and extension-transverse shear couplings. The stiffness A is constant throughout the cross section. For a box beam, the coupling stiffness, B, vanishes for the vertical members, while its values in the top and bottom members are of opposite signs

Subscripts t and v denote top and vertical members, respectively. The box width and height are represented by d and a, respectively. Equations (57) are derived by substituting Eqs. (54) and (55) into Eq. (61) and considering A to be constant. The underlined term in the expression of the axial stiffness C l l represents the extension contribution to the out-of-plane warping. The bending contributions to the out-of-plane warping are represented by the underlined terms in the expressions of C3 and C44. For the CAS configuration, bending about the y-axis is coupled with torsion while extension and bending about the z-axis are decoupled.

For CAS configurations, temperature and moisture change induce equivalent axial load only while no equivalent torsional and bending moments are generated. This is derived from Buertical members = 0 (54) Eq. (62) of the Appendix. and hence the name circumferentially asymmetric stiffness (CAS) adopted in Ref. [15], [16], [17] The bending slope variation along the beam and [20]. For a box-beam, the ply lay-ups on span for antisymmetric and symmetric canopposite sides are mirror images, and hence the tilevers under a 1 lb transverse tip load is shown name symmetric configuration adopted in Ref. in Figs. 3 and 4, respectively. The beam geome[14],[19], and [18]. The stiffness C along the hor- try and its material properties are given in Table izontal and vertical members are equal and ex- 1. The experimental results are reported in Refs. pressed by [14], [18], and [19]. The influence of the out-ofplane warping due to bending is isolated in these C t o p = Cbottom figures. The bending related out-of-plane w a r p (55) ing, gzU{ and g3Uj' terms in Eq. (40), results Cvertical l e f t = Cuertical right in a 91 and 20 % increase in the bending slope The stiffness matrix, for a centroidal system of for the antisymmetric and symmetric configurations, respectively. A comparison of the response axes, reduces to with other models is provided in Figs. 5 and 6. The analytical predictions reported in Refs. [14], [18], and [19] together with results obtained on the basis of the analyses in Refs. (151, (161 and the present theory are combined in these figures.

Table 1: Cantilever Geometry and Properties

-

Length = 30 in. Width = 0.953 in. Depth = 0.53 in. Ply thickness = 0.005 in. Ell = 20.59 Msi &2 = & = 1.42 Msi G12= GI3 = 0.87 Msi G23 = 0.7 Msi 1~12 =~ 1 = 3 0.42 4 3 = 0.5

Results show that the present theory is in Tip good agreement with the test data and the End closest when compared to the other analytical approaches which include shear deformation, Refs. [14] and [18], and shear deformation corrections, Refs. [15] and [16]. Figure 3: Significance of out-of-plane bending related warping on the bending slope of an antiThe bending slope in Figs. 3-6 is defined in symmetric [I516 cantilever under 1 lb transverse terms of the cross section rotation for theories tip Load including shear deformation. For the geometry and material properties considered, this effect is negligible as shown in Figs. 5 and 6 where the spanwise slope at the fixed end from theories with shear deformation, is indistinguishable from zero. The nonzero value shown by the test data may be due to the experimental set up used to achieve clamped end conditions.
Fixed Spanwise Coordinate,inches

Figure 7 shows the tip twist for an antisymmetric box cantilevered beam made of a [el6 due to a drop in temperature, AT, of 280F. The dirnensions and properties of the beam are taken from Ref. [23]. However, the values for the swelling coefficients pll is zero and ,Bz2is 5560 p&/% weight. The tip twist tends t o increase with 8 and reaches its maximum when the plies are oriented a t 45 degrees. Thermal pre-twist may reach 23" for a [4516 lay-up, while moisture content tends to alleviate thermal effects as illustrated in Fig. 7. Tip Fixed Spanwise Coordinate x, inch The specific moisture content for total alleviEnd ation from the thermal effect is about 0.8 %. This moisture content value is similar to the one corresponding to the alleviation of interlaminar stresses for laminated plates 131. Figure 4: Significance of out-of-plane bending related warping on the bending slope of a symrnetric [3016 cantilever under 1 lb transverse tip Load

25

- . - - - - - * - . - . - . - . . A T=-280F

--

AT+AH

-5

-10
2b

- A ~ ~ l . ~ %# -

/ /

-lo;

30

40

50

60

70

so

90

Ply Angle (degrees)

d 0

Fwd
End

10 20 Spanwir Coordinate. inches

30

Figure 7: Tip twist for an Antisymmetric [BIG layup beam Conclusion

r v

A closed section thin-walled theory has been Figure 5: Bending Slope of an Anti-Symmetric developed based on an asymptotical analysis of [I516 Cantilever Under 1 lb Transverse Tip Load the shell energy functional. The displacement field is not assumed apriori and emerges as a result of the analysis. In addition t o the classical torsional related warping, three new contributions are identified namely, axial strain, bending and hygrothermal related out-of-plane w a r p ing. The influence of hygrothermal loading on the governing equations is isolated. Explicit closed form expressions for the beam stiffness coefficients, stress resultants and displacement field are provided. Acknowledement This work was supported by the NASA Langley Research Center under grant NAG-1-637. This support is gratefully acknowledged. The author wishes to express his thanks to Dr. E. Armanios and Dr. V. Berdichevsky for their effective guidance throughout this research.

References
Figure 6: Bending Slope of a Symmetric Cantilever Under 1 lb Transverse Tip Load Bischoff, E., M. Ruhle, 0. Sbaizero and A. G. Evans, "Microstructural Studies of the Interfacial Zone of a Sic-Fiber-Reinforced Lithium Aluminum Silicate Glass-Ceramic, J. Am. Cerm.Soc., Vol. 72, 1986, pi'4l.

[2] O'Brien, T . K., Raju, I. S., and Garber, D. P., "Residual Thermal and Moisture Influences on the Strain Energy Release Rate Analysis of Edge Delamination," Journal of Composite Technology k 3 Research, Vol. 8, No. 2, Summer 1986, pp. 37-47. [3] Armanios, E. A., and Badir, A. M.,"Hygrothermal Influence on Mode I Edge Delamination in Composites," Composite Structures, Vol. 15, No. 4, 1990, pp. 323342. [4] Hyer, M. W.,"Some observations on the Cured Shape of Thin Unsymmetric Larninates," Journal of Composite Materials, Vol. 15, March 1981, pp. 175194. [5] Hyer, M. W., "Calculation of RoomTemperature Shapes of Unsymmetric Laminates," Composite Technology Review, July 1981, pp. 296-310. [6] Halm, H. T., " Warping of Unsymmetric Cross-Ply Graphite/Epoxy Laminates," Composite Technology Review, Fall 1981.

[14] Smith, Edward C., and Chopra, Inderjit, "Formulation and Evaluation of an Analytical Model for Composite BoxBeams,"Proceedings of the 31st Structures, Structural Dynamics and Materials Conference, long Beach, California, April 2-4, 1990, pp. 759-782. [15] Rehfield, L.W., and Atilgan, A.R.,"Shear Center and Elastic Axis and Their Usefulness for Composite Thin-Walled Beams," Proceedings of the American Society For Composites,Fourth Technical Conference, Blacksburg, Virginia, October 3-5, 1989. [16] Atilgan, Ali Rana, "Towards A Unified Analysis Methodology For Composite Rotor Blades," Ph. D. Dissertation, School of Aerospace Engineering, Georgia Institute of Technology, August 1989. [17] Rehfield, L. W., Atilgan, A. R., and Hodges, D. H., "Nonclassical Behavior of Thin-Walled Composite Beams with Closed Cross Sections." Journal of the American Helicopter Society, Vol. 35, (2), April 1990, pp. 42-50.

[7] Badir, Ashraf M., "Analysis of Advanced Thin-Walled Composite Structures," Ph. D. thesis, Georgia Institute of Technology, [18] Smith, Edward C., and Chopra, Inderjit, February 1992. "Formulation and Evaluation of an Analyti[8] Berdichevsky Victor, Armanios Erian, and cal Model for Composite Box-Beams," JourBadir Ashraf, "Theory of Anisotropic Thinnal of the American Helicopter Society, July Walled Beams, I- Linear Theory of Closed 91, pp.23-35. Cross-Section Beams," to appear in Composite Engineering, 1992. [19] Chandra, R., Stemple, A. D., and Chopra, I., "Thin-walled Composite Beams un[9] Koiter, W. T., "A Consistent First Apder Bending, Torsional, and Extensional proximation in the Theory of Thin Elastic Loads," Journal of Aircraft, Vol. 27, No. 7, Shells," Proc. IUTAM Symp., Delft, 1959. July 1990, pp. 614626. [lo] Sanders, J. L., "An Improved FirstApproximation Theory for Thin Shells," [20] Hodges, D. H., Atilgan A. R., Fulton M. V., and Rehfield L. W., "Dynamic CharNASA TR-R24, 1959. acteristics of Composite Beam Structures," Proceedings of the AHS National Special[ll] Koiter, W. T., "A Consistent First Approxists' Meeting on Rotorcmft Dynamics, Fort imation in the General Theory of Thin ElasWorth, Texas, Nov. 1314, 1989. tic Shells," Proc. IUTAM Symp on the Theory of Thin Shells, Delft, August 1969, 1233, North-Holland Publ. Amsterdam, 1960, [21] Vinson, J. R., and Sierakowski, R. L., The Behavior of Structures Composed of ComEdited by W. T. Koiter. posite Materials, Martinus Nijhoff Publish[12] Berdichevsky, V. L., and Misiura, V., "Efers, 1987. fect of Accuracy Loss in Classical Shell Theory," Journal of Applied Mechanics, 1991 (to [22] Jones, R. M., "Mechanics of Composite Materials," McGraw Hill Book Co., New York, appear). 1975. (131 Rehfield, L. W., " Design Analysis Methodology for Composite Rotor Blades," Proceed- [23] Tracy, Anita L., and Chopra, Inderjit, " An Analytical Model for Composite Box-Beams ings of the Seventh DoD/NASA Conference Including Thermal Effects," Proceedings of on Fibrous Composites in Structural Design, the 32nd Structures, Structural Dynamics AFWAL-TR-85-3094, June 1985, pp. (V(a)and Materials Conference, April 8-10, 1991, 1)-(V(a)-15).

Baltimore, MD, AIAA Paper No. 91-1018, and pp. 770-780.

where the pointed brackets denote integration over the thickness as defined in Eq. (14). In this appendix explicit expressions for some of the relevant variables used in the development Expressions for the stiffness coefficients Cij as well as the stiffnesses Cij (i, j = 1, 4) in (i, j = 1, 4) in terms of the cross section geomeEq. (47) are provided. try and materials properties are as follows The three stiffness parameters A, B and C in Eq. (35) ,are expressed in terms of the Hooke's tensor ~ ~ as 3 follows ~ '

The hygrothermal contribution to the energy density in Eq. (35) is given by

~ ( " " 1 (s) = 4

< > (< ~ [ < D2222 > + < ~~~~~p~~ > A H ) - < ~ -<D ~ > AH] ~

~
~

> AT ~
> AT ~

~
~

~
~

a
a

~
~

6
6

~ (59)

The 2 0 Young's modulus DaPr6 are given by

where

and H F Aare components of the inverse of the 2 0 E ' ~ ~ ~ E ~ ~ ~ matrix ~ p 3 A 3 -

Combining Eq. (39) and (58) the variables b and c can be written as

The equivalent non-mechanical loads in Eq. (47) are found to be

The Aij and Q i j (i, j = 1, 2, 6) in Eqs. (64) and (65) are the in-plane axial stiffnesses and reduced stiffnes co cients of Classical Lamination Theory B1lli'g. The indices adopted in Eqs. (64) and(65) follow the convention of Ref. [21]. The bars over the reduced stiffness coefficients Q i j indicate that these quantities are to be obtained through appropriate coordinate transformations. The summation in Eqs. (65) is performed over the n plies of the laminate. For a CAS configuration, the equivalent nonvanishing hygrothermal axial load, for the case where the through-the-thickness properties are neglected in Eq. (58) , become

For a CUS configuration, the equivalent nonvanishing hygrothermal loads in Eqs. (62) become
1 T("") = --'4(nm)l 2

and the beam undergoes only axial displacement. This is a result of Eqs. (51) and (47). The rate of change of the axial displacement is found to be

If the through-the-thickness properties are ne) glected in Eq. (58), the variables A ( " ~ and dnm) in Eq. (64) take the following form

where