You are on page 1of 4



tial number of eddies before decaying and that the modifica- iors of composite laminates, has demonstrated the importance
tion for the interval preceeding the first multiple jump in the of transverse shear effects for different delamination sizes.
criterion, Eq. (7), is important in preventing unacceptable This Note presents a solution for unsymmetric cross-ply
errors. For the most massive particle considered (dp = 100 jum; laminates, including both large deflection and transverse shear
pp = 2500 kg/m3), c0> At/T and c = c0 for all interactions. effects. A pinned-pinned laminated plate subjected to a uni-
For practical application of the truncation criterion sug- form transverse load is demonstrated. The results from the
gested in this technical note, it will certainly be necessary to nonlinear shear deformation theory (NSDT) are also com-
first validate the criterion for the flow conditions of particular pared with those from the nonlinear classical lamination the-
interest, as demonstrated herein for nearly homogeneous tur- ory (NCLT),1'2 the linear shear deformation theory (LSDT),
bulent flow. Once established, however, the potential gains in and the linear classical lamination theory (LCLT).
efficiency, through elimination of unnecessary and redundant
computational operations, may be significant. Formulation
Consider an unsymmetric cross-ply laminate subjected to a
Acknowledgments uniform transverse load. For cylindrical bending, the govern-
This work was conducted as part of an ongoing spray com- ing equations are assumed to be independent of the y axis.
bustion stability project supported by the University of Ten- Based on the large deflection shear deformation formulation,3
nessee-Calspan Center for Advanced Space Propulsion, UT the laminate displacements u and w are expressed by
Downloaded by Universitats- und Landesbibliothek Dusseldorf on May 22, 2013 | | DOI: 10.2514/3.11082

Space Institute, Tullahoma, TN under NASA Grant NAGW-

1195 and in part by Rocketdyne Division, Rockwell Interna- (1)
tional, Canoga Park, CA with Merlin Schuman as technical
monitor. w(x,z) = (2)
where w° is the midplane displacement and \{/x is the rotation in
^itchford, R. J., and Jeng, S.-M., "Efficient Statistical Transport
Model for Turbulent Particle Dispersion in Sprays," AIAA Journal, the xz plane. The equilibrium equations are given by
Sept. 1991, pp. 1443-1451.
Snyder, W. H., and Lumley, J. L., "Some Measurements of Par- Nx,x = 0 (3)
ticle Velocity Autocorrelation Functions in a Turbulent Flow," Jour-
nal of Fluid Mechanics, Vol. 48, 1971, pp. 41-71. Mx>x - Qx = 0 (4)

Qx,x + NxwtXX + q = 0 (5)

where Nx, MX9 and Qx are membrane force, bending moment,

Cylindrical Bending of Unsymmetric and transverse shear force resultants, respectively, and q is the
Composite Laminates transverse loading. The constitutive relations for the cross-ply
laminates are characterized by

Hsin-Piao Chen* (6)

California State University at Long Beach, Mx = Bn(u°x (7)
Long Beach, California 90840
and Qx = k (8)
Jeffrey C. Shut
The Aerospace Corporation, where An, Dn, Bn, and A55 are extensional, bending, exten-
sion-bending coupling, and transverse shear stiffnesses, re-
El Segundo, California 90245 spectively, and k is the shear correction factor.

Introduction 2000

E ARLY work on unsymmetric laminate analysis was mostly

based on the linear classical lamination theory. Recently,
Sun and Chin1'2 found that the linear lamination theory is 1500
— NSDT f/
inadequate for analysis of unsymmetric laminates, and the — NCLT
nonlinear large deflection theory must be used even for prob- — LSDT //
lems that are normally considered to be in the small deflection 1000 — - - LCLT '
domain. In their studies, however, the transverse shear effect
was not considered. This effect is important for resin matrix
composite laminates because the interlaminar shear moduli of c, 500
composite materials are very small as compared with the in-
plane elastic moduli of reinforced fiber. A recent study by
Chen,3 using a large deflection shear deformation theory on
the delamination buckling, postbuckling, and growth behav-

Presented as Paper 91-0961 at the AIAA/ASME/ASCE/AHS/ -500

ACS 32nd Structures, Structural Dynamics, and Materials Confer-
ence, Baltimore, MD, April 8-10, 1991; received May 9, 1991; revi-
sion received Aug. 13, 1991; accepted for publication Aug. 15, 1991.
Copyright © 1991 by Hsin-Piao Chen and Jeffrey C. Shu. Published -1000
by the American Institute of Aeronautics and Astronautics, Inc., with -4000 -2000 0 2000 4000
permission. q(p»0
* Associate Professor, Department of Aerospace Engineering.
Member AIAA. Fig. 1 In-plane force induced by transverse loading based on linear
tMember of the Technical Staff. and nonlinear theories.

1.0 The general solution of Eq. (12) is

- NSDT w(x) = C2 cosh ax + C4 - -— x2 (14)
-*- NCLT
0.5 - ~ LSDT For pinned-pined boundary conditions,
--- LCLT



-0.5 = 0 (17)

The three unknowns, NX9 C2, and C4, determined from Eqs.
Downloaded by Universitats- und Landesbibliothek Dusseldorf on May 22, 2013 | | DOI: 10.2514/3.11082

(15-17), are given by

-4000 -2000 0 2000 4000
Fig. 2 Maximum deflection due to transverse loading based on linear
and nonlinear theories. (19)
2NX ~ An N2
From Eq. (3), it is concluded that Nx is a constant. Substitu- B\\ q
tion of Eq. (6) into Eq. (7) yields
aa 'A^~NX~ \ 2cta

(9) sinh a«\ q2a2

+ C2 —1 cosh aa - =0 (20)
N "6NJ
Equations (18-20) are three nonlinear algebraic equations cou-
-"11 (10) pled together. They can be solved by the Newton-Raphson
'An iterative method. Note that Eq. (20) contains both q and q2
terms. Therefore, the solutions are different for positive and
By using Eqs. (4), (5), and (7-9), the governing equation for negative loadings.
deflection w is derived as
Compressive In-Plane Force (Nx <0)
Nx For the case with P = - Nx >0, the governing equation (11)
(Nx/kA55)lDc (Nx/kA55)\Dc can be expressed as
Note that the formulation for LSDT can be obtained by
dropping the nonlinear terms in Eqs. (6-8). The formulations (21)
- (P/kA55)}De
for NCLT and LCLT can be obtained from the NSDT and
LSDT formulations, respectively, by neglecting the transverse
shear effects, i.e., by assuming the transverse shear stiffness is
infinite (A55 — oo).
Laminated Plate with Pinned-Pinned Edges
The characteristic of Eq. (11) depends on the coefficient of
the second derivative of the transverse deflection where the 1.E+4 — NSDT
induced in-plane force Nx determines the form of the displace- --- NCLT
ment solution. For a pinned-pinned laminated plate subjected
to a uniform transverse load, it is found that Nx can be 1.E+3
positive (tensile), negative (compressive), or zero decided by
the sign of the coupling coefficient # n , the sign and magni-
tude of the transverse load q, and the slenderness ratio of the
laminate L/h. These different conditions are discussed indi-
vidually in the following paragraphs.
Tensile In-Plane Force (TV* > 0)
For the case with Nx>0, the governing equation (11) can be
expressed as 1.EO
0 (12)
where 0 20 40 60 80 100
(13) Fig. 3 Variation of transition transverse loading.

where shows the variation of the transition transverse load qt, ob-
tained from Eq. (27), for laminates with slenderness ratios L/h
up to 100. For states above this curve, the induced in-plane
(22) forces are in tension. For those below this curve, the induced
[1 - (P/kA55)]Dc
in-plane forces are in compression. Note that qt varies drasti-
The general solution of Eq. (21) is cally with the slenderness ratio. For a thin laminate with rather
large slenderness ratio, qt becomes negligibly small. Since the
transverse shear effect is more important for thick plates, it is
w(x) = C2 cos C4 + -- x2 (23) noted in Fig. 3 that the transverse shear effect on qt becomes
increasingly more pronounced as L/h is less than 20.
The three unknowns, -P, C2, and C4, can be determined from
the boundary conditions, Eqs. (15-17), in a similar way as was References
the case of Nx>0. Sun, C. T., and Chin, H., "Analysis of Asymmetric Composite
Laminates," AIAA Journal, Vol. 26, No. 6, 1988, pp. 714-718.
Transition State with Vanishing In-Plane Force (Nx = 0) 2
Sun, C. T., and Chin, H., "On Large Deflection Effects in Un-
The state corresponding to Nx = 0 is referred to as the symmetric Cross-Ply Composite Laminates," Journal of Composite
"transition state," which implies that the induced in-plane Materials, Vol. 22, No. II, 1988, pp. 1045-1059.
force changes its sign when this state is passed. For this case, Chen, H. P., "Shear Deformation Theory for Compressive De-
Downloaded by Universitats- und Landesbibliothek Dusseldorf on May 22, 2013 | | DOI: 10.2514/3.11082

the governing equation (11) with a uniform transverse load qt lamination Buckling and Growth," AIAA Journal, Vol. 29, No. 5,
can be expressed as 1991, pp. 813-819.
Whitney, J. M., and Sun, C. T., "A Higher Order Theory for
Extensional Motion of Laminated Composites," Journal of Sound
_qL and Vibration, Vol. 30, No. I, 1973, pp. 85-97.
The general solution of Eq. (24) is
Matrix Transformation Method for
= C2x^ 22. +
- ^
C4 - &
24A Updating Dynamic Model
The three unknowns, C2, C4, and qt, are determined from the
boundary conditions Eqs. (15-17). The resulting transverse De-Wen Zhang*
deflection w and transverse load qt are Beijing Institute of Structure and Environment,
Beijing, People's Republic of China
w =• (5a2 - x2)(a2 - x2) (a2 - x2) (26) and
24DC 2kA,
Lingmi Zhangf
Bn\ 2a2 Nanjing Aeronautical Institute,
^? = (27)
5kA55 2(kA55)2 Nanjing, People's Republic of China
It is noted from Eq. (27) that qt must be in opposite sign as the
coupling stiffness Bn. Therefore, a transition state can only Introduction
exist when Bn and q are in opposite signs.
ANY systematic methods1'13 have been developed in
Results M recent years for updating analytical models to predict
modal test data more closely. The methods in Refs. 1-7,
Numerical solutions are obtained based on the following
referred to as matrix-type procedures here, correct the whole
graphite/epoxy composite properties: E\ = 20 x I06 psi, E2 =
mass and stiffness matrices. Correspondingly, the methods in
I Ax I06 psi, G l2 = G l3 = 0.8 x I06 psi, G23 = 0.6 x I06 psi,
v l2 = 0.3. According to Ref. 4, the shear correction factor k is Refs. 8-13, element-type procedures, modify only some
taken as 7r2/l2. The ply thickness is 0.005 in. For the laminate nonzero elements of mass and stiffness matrices. In the former
[904/04], Figs. I and 2 show the comparison between the non- procedures, the connectivity of the original analytical model is
linear solutions and the linear solutions. The induced in-plane not preserved, causing the addition of unwanted load paths.
In the latter, particularly in Refs. 9 and 11, the stiffness matrix
force and maximum deflection predicted by the linear theories
may be identified exactly in certain cases even when some of
(LSDT and LCLT) are proportional to the transverse load.
the test modes are not known.
However, the nonlinear theories (NSDT and NCLT) give quite
different results. As shown in Fig. I, negative q induces posi- The purpose of the work presented in this Note is similar to
that of Ref. 8. This Note proposes a matrix transform method
tive in-plane force Nx with a negative coupling stiffness Bn. As
(MTM), in which the derivations are much simpler than those
discussed earlier, a transition state exists only when Bn and q
are in the opposite signs. As a result, positive q will induce a in Refs. 1-6, which use Lagrange multipliers. In the present
compressive in-plane force initially, which turns into tension MTM, transform matrices for the correction of dynamic mod-
once q passes the transition transverse load qt. els are created. The effect of the transformation is opposite to
Figure 2 shows that the maximum deflections predicted by that of the objective function in the Lagrange multiplier
nonlinear theories are also significantly different from those method (LMM); the former is to reduce the number of un-
predicted by linear theories. In addition, the results from known parameters of the governing equations and the latter is
to increase the number of equations. Not only can MTM
NSDT and NCLT also show that the deflection with positive
q is larger than that with negative q. This is attributed to the reproduce the formulas in Refs. 1-6, but it can also derive
compressive in-plane force initially induced by positive q, some new formulations that are difficult to be formed by
which tends to aggravate the transverse deflection. Figure 2
also shows that the transverse shear effects are less significant Received Oct. 29, 1990; revision received April 22, 1991; accepted
from the nonlinear theories for the present case. for publication May 9, 1991. Copyright © 1992 by the American
Theoretically, the transition state exists for every unsym- Institute of Aeronautics and Astronautics, Inc. All rights reserved.
metric laminate as long as the transverse load q and the *Senior Research Engineer, P.O. Box 9210. Member AIAA.
laminate coupling stiffness Bn have opposite signs. Figure 3 fProfessor of Vibration Engineering, Box 109.
This article has been cited by:

1. Audrey Petit, Habib Eslami, Yi ZhaoNonlinear Forced Vibrations of Unsymmetrically Laminated Plates under Cylindrical
Bending . [Citation] [PDF] [PDF Plus]
2. Marco Di Sciuva, Ugo Icardi. 1995. Large deflection of adaptive multilayered Timoshenko beams. Composite Structures
31:1, 49-60. [CrossRef]
3. Pinhas Bar-Yoseph, Daniel Ben-David. 1995. Free-edge effects in unsymmetrically laminated composite plates. Composite
Structures 30:1, 13-23. [CrossRef]
4. Ugo Icardi. 1994. The nonlinear response of unsymmetric multilayered plates using smeared laminate and layerwise models.
Composite Structures 29:4, 349-364. [CrossRef]
5. ERASMO CARRERA. 1993. Nonlinear response of asymmetrically laminated plates in cylindricalbending. AIAA Journal
31:7, 1353-1357. [Citation] [PDF] [PDF Plus]
6. M. Liu, D. G. Gorman. 1993. Improved Method for Evaluating Damping Ratios of a Vibrating System. AIAA Journal
31:7, 1357-1360. [Citation] [PDF] [PDF Plus]
Downloaded by Universitats- und Landesbibliothek Dusseldorf on May 22, 2013 | | DOI: 10.2514/3.11082