Journal of

Mechanics of
Materials and Structures

THE EFFECTIVE THICKNESS OF LAMINATED GLASS PLATES

Laura Galuppi and Gianni Royer-Carfagni

Volume 7, No. 4 April 2012

mathematical sciences publishers

JOURNAL OF MECHANICS OF MATERIALS AND STRUCTURES
Vol. 7, No. 4, 2012

msp

THE EFFECTIVE THICKNESS OF LAMINATED GLASS PLATES

L AURA G ALUPPI AND G IANNI ROYER -C ARFAGNI

The flexural performance of laminated glass, a composite of two or more glass plies bonded together by
polymeric interlayers, depends upon shear coupling between the glass components through the polymer.
This effect is usually taken into account, in the design practice, through the definition of the effective
thickness, i.e., the thickness of a monolith with equivalent bending properties in terms of stress and
deflection. The traditional formulas à la Bennison–Wölfel are accurate only when the deformed bending
shape of the plate is cylindrical and the plate response is similar to that of a beam under uniformly
distributed load. Here, assuming approximating shape function for the deformation of laminated plates
variously constrained at the edges, minimization of the corresponding strain energy furnishes new simple
expressions for the effective thickness, which can be readily used in the design. Comparisons with
accurate numerical simulations confirm the accuracy of the proposed simple method for laminated plates.

1. Introduction

Laminated glass is a sandwich structure where two or more glass plies are bonded together by thin
polymeric interlayers with a process at high temperature and pressure in autoclave. Because of the shear
deformability of the polymer, there is not a perfectly coupling between any two consecutive glass plies
[Behr et al. 1993], and the degree of coupling depends upon the shear stiffness of the polymeric interlayer
[Hooper 1973]. Consequently, the flexural response is somehow intermediate between the two borderline
cases [Norville et al. 1998] of layered limit, i.e., frictionless relative sliding of the plies, and monolithic
limit, i.e., perfect bonding of the plies. This problem has close similarities with the case of composite
beams with partial interaction. The most classical contribution, conceived of for a concrete slab and
a steel beam bonded by shear connectors, is associated with the name of Newmark et al. [1951], who
considered a linear and continuous relationship between the relative interface slip and the corresponding
shear stress. More recently Murakami [1984] introduced the usual hypotheses of Timoshenko beam to
model the interlayer in the analysis of composite beams. In a recent paper, Xu and Wu [2007] presented
a very comprehensive approach for static, dynamic and buckling behavior of composite beams with
partial interaction, accounting for the influence of rotary inertia and shear deformation. Approximate
formulations of this kind are particularly important for studying the problem of buckling of composite
columns (e.g., [Le Grognec et al. 2012; Schnabl and Planinc 2011]), applicable to various materials,
including lamellar wood [Cas et al. 2007].
Geometric nonlinearities are usually important because of the slenderness of the laminated panel [Aşik
2003], but are usually negligible when the loads are orthogonal to the panel surface and no in-plane forces
are present. From an analytical point of view, it is often very difficult to obtain a closed-form solution for
the strain and stress field in a laminated glass plate. An analytical approach has been recently proposed
Keywords: structural glass, laminated glass, composite structure, laminated plate, effective thickness, energy minimization.

375
376 LAURA GALUPPI AND GIANNI ROYER-CARFAGNI

by Foraboschi [2012] for the case of rectangular plates made of laminated glass, simply supported on
four sides. The precise calculation of the resulting state of stress and strain is quite difficult and usually
requires numerical analysis, complicated by the fact that response of the polymer is nonlinear, viscoelastic
and temperature dependent [Behr et al. 1993; Bennison et al. 2005; Louter et al. 2010].
This is why simplified methods are becoming more and more popular in the design practice , and much
of current research is directed towards their definition and verification [Foraboschi 2007]. Reference is
made to [Ivanov 2006] for an updated list of the most relevant current literature.
A commonly accepted simplification is to assume that the polymer is linear elastic, with proper secant
elastic moduli that account for environmental temperature and load duration. There are various com-
mercial polymeric films: polyvinyl butyral (PVB), ethylene vinyl acetate (EVA), and sentry glass (SG)
[Bennison et al. 2008; Bennison et al. 2001]. Depending upon the polymer type, temperature T and
characteristic load-duration t0 , the secant shear modulus of the interlayer may vary from 0.01 MPa (PVB
at T = +60◦ C under permanent load) up to 300 MPa (SG at T = 0◦ C and t0 = 1 s). On the other hand,
glass remains linear elastic up to failure (Young’s modulus E ≃ 70 GPa and Poisson ratio ν ≃ 0.2).
A simplified method of very practical value makes use the notion of “effective” thickness. This
method has been introduced starting from the analysis for sandwich beams with linear elastic components
originally developed by Wölfel [1987] and later transferred to the case of laminated glass [Bennison 2009;
Calderone et al. 2009]. To illustrate, consider (as in Figure 1) a beam of length l and width b composed
of two external glass plies of thickness h 1 and h 2 and Young’s modulus E, bonded by a soft polymeric
interlayer of thickness t and elastic shear modulus G. The latter has negligible axial and bending strength,
but nevertheless it can transfer shear coupling stresses between the external layers. Let

1 3 A1 A2
Ai = h i b, Ii = 12 bh i (i = 1, 2), H = t + 12 (h 1 + h 2 ), A∗ = ,
A1 + A2
Itot = I1 + I2 + A∗ H 2 , A = bt, Bs = E A∗ H 2 . (1-1)

Clearly, Itot is the cross sectional inertia of the composing glass layers properly spaced of the interlayer
gaps, associated with the case of perfect bonding of the glass plies as in Bennison et al. [1999] (monolithic
limit). Besides, Bs is the bending stiffness when the external layers have negligible inertia, while the
mid-layer can only bear shear stress. When, as in the case of laminated glass, the external layers have

y
p(x)
h1 E, A1, D1

t H
x G
h2 E, A2, D2
l
b

Figure 1. Beam composed of two glass plies bonded by a polymeric interlayer. Longi-
tudinal and cross sectional view (not in the same scale).
 THE EFFECTIVE THICKNESS OF LAMINATED GLASS PLATES 377

nonnegligible inertia, Wölfel proposed the expression Bs∗ defined as

1
Bs∗ = E I1 + E I2 + Bs , (1-2)
1+K
where the coefficient (1 + K) indicates the degradation of the bending properties of the composite due
to the incomplete interaction between the external layers. Using the principle of virtual work one finds
that the coefficient K is of the form
χ
K = β Bs , (1-3)
Gbtl 2
where, as explained in [Galuppi and Royer-Carfagni 2012], the shear coefficient of the intermediate
layer χ is evaluated as χ = t 2 /H 2 and β is another parameter that depends upon the load condition.
For simply supported beams, the corresponding values of β are recorded in [Wölfel 1987] for various
loadings. Notice from (1-3) that G → ∞ ⇒ K → 0, so that from (1-2) also Bs∗ → E Itot , i.e., the
monolithic limit; moreover, G → 0 ⇒ K → ∞ and Bs∗ → E(I1 + I2 ), i.e., the layered limit.
Bennison [2009] and Calderone et al. [2009] have referred specifically to Wölfel’s approach for the
case of laminated glass. More precisely, they define the nondimensional coefficient Ŵ = 1/(1 + K),
Ŵ ∈ (0, 1), introduce the equivalent moment of inertia of the cross section in the form

A1 A2
Ieq = I1 + I2 + Ŵ H 2, (1-4)
A1 + A2
and consider for Ŵ the expression

1 1
Ŵ= = . (1-5)
χ Bs t Bs
1+β 1 + 9.6
G b t l2 GbH 2l 2
This is equivalent to using in (1-3) the value β = 9.6, which corresponds to Wölfel’s analysis for the
particular case of a simply supported beam under uniformly distributed load. Consequently, recalling
(1-4), for calculating the laminate deflection one can consider a monolithic beam with deflection-effective
thickness h ef;w given by
s
h1h2
h ef;w = 3 h 31 + h 32 + 12Ŵ H 2. (1-6)
h1 + h2

Once the deflection of the laminate is established, one can estimate the degree of connection offered
by the deformable interlayer and, from this, the maximum stress in the glass can be easily estimated. The
result [Bennison 2009; Calderone et al. 2009] is that the maximum bending stress in the i-th glass plies,
i = 1, 2, is the same of that in a fictitious monolithic beam with analogous constraint and load conditions,
with respectively stress-effective thickness
v v
u u
u h 3ef;w u h 3ef;w
h 1;ef;σ = u
t , h 2;ef;σ = u
t . (1-7)
H h2 H h1
h 1 + 2Ŵ h 2 + 2Ŵ
h 1 +h 2 h 1 +h 2
378 LAURA GALUPPI AND GIANNI ROYER-CARFAGNI

It is important to notice that this method relies upon the particular form of Ŵ given by (1-5), which
assumes the coefficient β = 9.6, i.e., the one proposed by Wölfel for the very particular case of sim-
ply supported beams under uniformly distributed loading. Moreover, according to [Wölfel 1987], the
validity of the method is limited because its simplifying assumptions are valid for statically determined
composite beams, where the bending stiffness of the composite plies is considerably small. Nevertheless,
this approach is widely used. In the design practice, the calculations for laminated glass panels are
usually performed on an equivalent monolithic plate whose thickness is assumed to be given by (1-6),
to determine maximum deflection, or by (1-7), where it is the bending stress to be calculated. This
effective thickness is usually adopted in place of the actual thickness in analytic equations and simplified
finite-element analysis; sometimes the method is abused in very delicate conditions, for example when
calculating the stress concentrations around holes and contact regions in a neighborhood of the point-wise
fixing of frameless glazing. In general, no approach based upon the definition of the effective thickness
can be used to evaluate local effects. In any case, the Bennison–Wölfel method may lead to inaccurate
results also when calculating maximum stress and deflection at the center of a laminated plate, especially
when load and boundary conditions are different from that of a rectangular plate simply supported at two
opposite side (cylindrical deformed shape) under uniformly distributed load.
In [Galuppi and Royer-Carfagni 2012] we treated the classical problem of a composite laminated
glass beam under flexure using a variational approach similar in type to that proposed in [Aşik and
Tezcan 2005] for numerical purposes. Using convenient shape functions for the beam deflection, simple
formulas for the effective thickness were obtained which, for the one-dimensional case of beams with
various constraint and load conditions, fitted with numerical experiments much better than the classical
expressions (1-6) and (1-7). Our aim now is to extend this approach to the two-dimensional case of
a rectangular laminated glass plate under uniform pressure, variously supported at the borders. For the
cases considered in [Galuppi and Royer-Carfagni 2012] the problem is certainly much more complicated,
but we show that by assuming again proper shape functions for the plate deflection, simple expressions
of the effective thickness can be found. Comparisons with careful numerical experiments on full three-
dimensional models, show the proposed formulation furnishes results more reliable than those obtainable
with the classical Bennison–Wölfel approach. The method can be readily extended to plates of various
shape, under diverse load conditions.

2. The variational problem

As indicated in Figure 2, with notation analogous to (1-1), consider a laminated plate composed of two
glass layers of thickness h 1 and h 2 with Young’s modulus E and Poisson’s ratio ν, connected by a
polymeric interlayer of thickness t and shear modulus G. Let

Eh i3
Di =
12(1 − ν 2 )
represent the flexural stiffness of the i-th glass plate, i = 1, 2, while H is the distance between their
middle planes. Upon introduction of a reference system as indicated in Figure 2, the plate is identified
by the x − y domain  with border ∂, and is loaded by a pressure per unit area p(x, y), not necessarily
uniformly distributed.
 THE EFFECTIVE THICKNESS OF LAMINATED GLASS PLATES 379

z
y

h1 E, n,D 1
x H
t G
W
h2 E, n,D2

Figure 2. Plate composed of two glass plies bonded by a polymeric interlayer. Isometric
and cross sectional view (not in the same scale).

No slippage is supposed to occur between glass and polymer (case of perfect bonding). Under the
hypotheses that strains are small and rotations moderate, the kinematics is completely described by the
out-of-plane displacement w(x, y) in the z direction that, by neglecting the interlayer strain along z, is
the same for the two glass layers, and the in-plane displacements of the i-th glass layers, i = 1, 2, for
which the x and y components are denoted by u i (x, y) and vi (x, y), respectively (Figure 3).

2.1. The minimization problem. As represented Figure 3, let us denote with u sup (x, y) and vsup (x, y),
u inf (x, y) and vinf (x, y), the x and y displacement components of those faces of the superior and inferior
glass plies, respectively, in contact with the polymer. Then, the shear strain in the interlayer, constant
through its thickness, is characterized by the components
1 1
γ̃zx = [u sup (x, y) − u inf (x, y) + w,x (x, y)t] = [u 1 (x, y) − u 2 (x, y) + w,x (x, y)H ],
t t (2-1)
1 1
γ̃zy = [vsup (x, y) − vinf (x, y) + w, y (x, y)t] = [v1 (x, y) − v2 (x, y) + w, y (x, y)H ],
t t
where subscript commas denote partial differentiation with respect to the indicated variable. The strain
energy E of the laminated glass plate is provided by the flexural and extensional contributions of the

w,x (x,y) w,y (x,y)

v2 u2

vsup usup z z
z vinf uinf
y y
x v1 u1 x

Figure 3. Relevant displacement components and corresponding deformation in the

composite plate.
380 LAURA GALUPPI AND GIANNI ROYER-CARFAGNI

two glass layers, by the part corresponding to the shear deformation of the interlayer, and by the work
associated with the external loads p(x, y). We have, suppressing from the notation the dependence of
w, u i , vi on x, y (see [Timoshenko and Woinowsky-Krieger 1971]):
Z n
1
E[w, u 1 , u 2 , v1 , v2 ] = (D1 + D2 )[(w,x x + w, yy )2 − 2(1 − ν)(w,x x w, yy − w,2x y )]
 2
D1 

+ 12 2 (u 1 ,x + v1 , y )2 − 2(1 − ν) u 1 ,x v1 , y − 41 (u 1 , y + v1 ,x )2
h1
D2 

+ 12 2 (u 2 ,x + v2 , y )2 − 2(1 − ν) u 2 ,x v2 , y − 41 (u 2 , y + v2 ,x )2
h2
G   o
+ (u 1 − u 2 + w,x H )2 + (v1 − v2 + w, y H )2 + 2 p(x, y)w d x dy. (2-2)
t
The analysis of the first variation of the functional with respect to w(x, y), u i (x, y) and vi (x, y), i = 1, 2,
gives respectively the following Euler–Lagrange equations:
GH
(D1 + D2 )△△w − [(u 1 − u 2 + w,x H ),x +(v1 − v2 + w, y H ), y ] − p(x, y) = 0, (2-3)
t
12D1  1−ν 1+ν  G
u 1 , x x + u 1 , yy + v 1 , x y − (u 1 − u 2 + w,x H ) = 0, (2-4)
h 21 2 2 t
12D2  1−ν 1+ν  G
2
u 2 ,x x + u 2 , yy + v2 ,x y + (u 1 − u 2 + w,x H ) = 0, (2-5)
h2 2 2 t
12D1  1−ν 1+ν  G
2
v1 , yy + v1 ,x x + u 1 ,x y − (v1 − v2 + w, y H ) = 0, (2-6)
h1 2 2 t
12D2  1−ν 1+ν  G
2
v2 , yy + v2 ,x x + u 2 ,x y + (v1 − v2 + w, y H ) = 0, (2-7)
h2 2 2 t
with conditions at the boundary ∂
I  
∂ GH  
(D1 + D2 ) △w − (u 1 − u 2 + w,x H ), xn x + (v1 − v2 + w, y H ), yn y δw ds
∂ ∂n It
 
− (D1 + D2 ) (w,x x +νw, yy )n x + w,x y n y δw,x ds
∂
I
 
− (D1 + D2 )2(1 − ν) w,x y n x + (w,x x +νw, yy )n y δw, y ds = 0, (2-8)
I h ∂
i
1−ν
(u i ,x +νvi , y )n x + (u i , y +vi ,x )n y δu i ds = 0, i = 1..2, (2-9)
Iδ h 2
1−ν i
(u i , y +vi ,x )n x + (vi , y +νu i ,x )n y δvi ds = 0, i = 1..2, (2-10)
∂ 2
where δw(x, y), δu i (x, y) and δvi (x, y) are the variations of w(x, y), u i (x, y) and vi (x, y), i = 1, 2,
respectively. As customary in the calculus of variations [Sagan 1969], the geometric constraints at the
border furnish restriction on the possible variations of the displacement components. Then, (2-8), (2-9)
and (2-10) give the boundary conditions for the problem at hand.
 THE EFFECTIVE THICKNESS OF LAMINATED GLASS PLATES 381

2.2. Interpretation of the Euler–Lagrange Equation and boundary conditions. With standard notation
from plate theory [Timoshenko and Woinowsky-Krieger 1971], the x- and y-components of the in-plane
forces-per-unit-length in the i-th glass ply are of the form
12Di 12Di 12Di 1 − ν
Ni x = (u i ,x +νvi , y ), Ni y = (vi , y +νu i ,x ), Ni x y = (u i , y +vi ,x ). (2-11)
h i2 h i2 h 21 2
As shown in Figure 4, imagine cutting the interlayer of the laminated plate with a plane parallel to x
and y at an arbitrary height t ∗ . The tangential stress component on the resulting surfaces are τ̃zx = G γ̃zx
and τ̃zy = G γ̃zy , where the shear strains γ̃zx and γ̃zy have been defined in (2-1). Since γ̃zx and γ̃zy do not
depend on z, τ̃zx and τ̃zy are independent of t ∗ . It is then easily verified that (2-4), (2-5), (2-6) and (2-7)
are the equilibrium equations in the x and y directions of the composing glass plies, that is,
N1x ,x +N1x y , y +τ̃zx = 0, N2x ,x +N2x y , y −τ̃zx = 0,
(2-12)
N1y , y +N1x y ,x +τ̃zy = 0, N2y , y +N2x y ,x −τ̃zy = 0.
Moreover, from the moment-curvature relationships [Timoshenko and Woinowsky-Krieger 1971] for
the i-th glass plate,

Mi x = −Di (w,x x +νw, yy ), Mi y = −Di (w, yy +νw,x x ), Mi x y = Di (1 − ν)(w,x y ), (2-13)

one finds that Equation (2-3) can be rewritten as

(M1x + M2x ),x x +(M1y + M2y ), yy −2 (M1x y + M2x y ),x y −(τ̃zx ,x +τ̃zy , y )H − p = 0. (2-14)

Notice from Figure 4 that the shear stress at the surfaces resulting after the horizontal cut of the
interlayer are statically equivalent, for each one of the component glass plies, to distributed torques per
unit length. In particular, τ̃zx generates m 1zx (x, y) = −τ̃zx ( 21 h 1 +t −t ∗ ) and m 2zx (x, y) = −τ̃zx ( 21 h 2 +t ∗ )
in the upper and lower glass plate, respectively. Then, the overall torque per unit length on the two glass
plates is m zx (x, y) = m 1zx (x, y) + m 2zx (x, y) = −τ̃zx ( 12 (h 1 + h 2 ) + t) = −G γ̃zx H . Similarly, τ̃zy

m1yz
h1
t-t* m1xz Ny
Nxy

tyz Nx
t* txz
y
h2
x
z
y
m2xz
x
m2yz

Figure 4. Equilibrium of portions of the laminated plate.

382 LAURA GALUPPI AND GIANNI ROYER-CARFAGNI

generates m zy (x, y) = −τ̃zy ( 21 (h 1 + h 2 ) + t) = −G γ̃zy H . Henceforth, condition (2-3), or equivalently

(2-14), represents the flexural equilibrium of the package glass + polymer under bending, in the form
(D1 + D2 )△△w − p + m zx ,x +m zy , y = 0.
For what the border is concerned, (2-11) allows to write (2-9) and (2-10) in the form
I I
(Ni x n x + Ni x y n y )δu i ds = 0, (Ni x y n x + Ni y n y )δvi ds = 0 (i = 1, 2), (2-15)
∂ ∂

these being the standard in-plane boundary conditions for the i-th glass layers [Timoshenko and Woinow-
sky-Krieger 1971]. Moreover, using (2-13) and defining Mx = (M1x + M2x ), M y = (M1y + M2y ),
Mx y = (M1x y + M2x y ), condition (2-8) may be rearranged as
I
[(Mx ,x −Mx y , y −H τ̃zx ,x )n x + (M y , y −Mx y ,x −H τ̃zy , y )n y ] δw ds+
∂
I I
[Mx n x + Mx y n y ] δw,x ds + [M y n y + Mx y n x ] δw, y ds = 0. (2-16)
∂ ∂

This is readily interpretable because Q x = Mx ,x −Mx y , y and Q y = M y , y −Mx y ,x represent the sum of
the transversal shearing forces per unit length acting on the two glass plies. As in classical Kirchhoff
plate theory [Timoshenko and Woinowsky-Krieger 1971], (2-16) gives the boundary condition in terms
of bending couples and transversal shear.

2.3. Correlation between the displacements of the external layers. There are noteworthy identities, im-
portant for the forthcoming considerations, that correlate the displacements of the two constituent glass
plies. In fact, we will prove that the (weighted) average displacement fields, defined as
h2 h2
U (x, y) = u 1 (x, y) + u 2 (x, y), V (x, y) = v1 (x, y) + v2 (x, y), (2-17)
h1 h1
are identically zero.
To illustrate, notice first that equations (2-4), (2-5), (2-6) and (2-7) may be rearranged in the form
12D1  1−ν 1+ν  12D2  1−ν 1+ν 
u ,
1 xx + u ,
1 yy + v ,
1 xy = − u ,
2 xx + u ,
2 yy + v2 xy ,
, (2-18)
h 21 2 2 h 22 2 2
12D1  1−ν 1+ν  12D2  1−ν 1+ν 
v ,
1 yy + v ,
1 xx + u ,
1 xy = − v ,
2 yy + v ,
2 xx + u ,
2 xy . (2-19)
h 21 2 2 h 22 2 2

In terms of U (x, y) and V (x, y), these can be rewritten as

U,x x + 21 (1 − ν)U, yy + 21 (1 + ν)V,x y = 0,

(2-20)
V, yy + 21 (1 − ν)V,x x + 12 (1 + ν)U,x y = 0,

which is a system of partial differential equations defined on a connected domain . If the glass plies
are constrained so that u i = vi = 0, i = 1..2, on the boundary ∂, then from (2-17) also U = V = 0 on
∂. In general, however, the in-plane displacements of laminated glass are not constrained; in this case,
 THE EFFECTIVE THICKNESS OF LAMINATED GLASS PLATES 383

from (2-9) and (2-10) one finds the boundary conditions

(U,x +νV, y )n x + 12 (1 − ν)(U, y +V,x )n y = 0 on ∂,

1
(2-21)
2 (1 − ν)(U, y +V,x )n x + (V, y +νV,x )n y = 0 on ∂.

We can prove that the system of partial differential equations (2-20) with boundary condition U = V =
0 on ∂ or, equivalently, with boundary condition given by (2-21), implies U = V ≡ 0 in . Indeed, this
is a standard argument in PDEs, because the system (2-20) can be proved to be strongly elliptic. Here,
however, we use a more practical argument, by showing that this problem is equivalent to a classical
boundary value problem in linear elasticity, for which well-known results hold. To this aim, we imagine
that U (x, y) and V (x, y) represents the displacement field of a fictitious two-dimensional linear-elastic
body , with Young’s modulus E and Poisson’s ration ν, for which the (fictitious) stress components
read
E E E
σ̂x = (U,x +νV, y ), σ̂ y = (V, y +νU,x ), σ̂x y = (U, y +V,x ). (2-22)
1 − ν2 1 − ν2 2(1 + ν)

It can then be shown that equations (2-20) with boundary conditions (2-21) can be rewritten as

σ̂x ,x +σ̂x y , y = 0
in , (2-23)
σ̂x y ,x +σ̂ y , y = 0

σ̂x n x + σ̂x y n y = 0
on ∂. (2-24)
σ̂x y n x + σ̂ y n y = 0

In the language of linear elasticity, these represent the equilibrium of a body in generalized plane stress
with null boundary traction. Kirchhoff’s theorem [Knops and Payne 1971] states that there is at most
one solution to the Dirichlet boundary value problems in plane elasticity provided −∞ < ν < 12 , ν 6 = −1,
E 6 = 0; in the traction boundary value problem there is uniqueness to within a rigid body displacement.
In the considered case of null body forces and null boundary traction, since E > 0 and −1 < ν < 12 , the
solution is unique and it consist in a null stress field, leading to a displacement field of the form
     
U (x, y) −ωy c
= + 1 , (2-25)
V (x, y) ωx c2

with constants c1 , c2 and ω, that represents an infinitesimal rigid body displacement. It is easy to show
that such constants are null for the case at hand if the laminated glass package is properly constraint in
order to prevent its rigid displacements. In conclusion, one finds U (x, y) ≡ 0 and V (x, y) ≡ 0, for which
the expected identities

h1 h1
u 2 (x, y) = − u 1 (x, y), v2 (x, y) = − v1 (x, y). (2-26)
h2 h2

This is because the in-plane forces in the two glass plies, which are due to the mutual shear forces
transmitted by the interlayer, must balance one another.
384 LAURA GALUPPI AND GIANNI ROYER-CARFAGNI

2.4. Layered and monolithic limits. When the shear modulus of the interlayer vanishes, i.e., G → 0,
conditions (2-3), (2-4), (2-5), (2-6) and (2-7) take the form (i = 1, 2)

 (D1 + D2 )△△w − p(x, y) = 0,


 12Di 
 1−ν 1+ν

u i ,x x + u, + v, = 0, (2-27)
 h i2 2 i yy 2 i xy

  
 12Di vi , yy + 1−ν vi ,x x + 1+ν u i ,x y = 0.

h i2 2 2
The first equation clearly corresponds to the flexural equilibrium of two frictionless sliding glass plates,
with flexural bending D1 + D2 , subject to a distributed load p(x, y), while the others are the equilibrium
conditions for the in-plane forces in x and y direction of the upper and lower glass plies, respectively.
This is the layered limit, i.e., the laminated glass plate behaves as two free-sliding glass plies.
When G → +∞, then the shear strain in the interlayer tends to zero, i.e., γ̃zx = 0, γ̃zy = 0 in (2-1).
Using the relationships (2-26), such conditions give:
D2 h 21 h2
γ̃zx = [u 1 − u 2 + w,x H ]/t = 0 ⇒ u1 = − H w,x = − H w,x ,
D1 h 22 + D2 h 21 h1 + h2
(2-28)
D2 h 21 h2
γ̃zy = [v1 − v2 + w, y H ]/t = 0 ⇒ v1 = − H w, y = − H w, y .
D1 h 22 + D2 h 21 h1 + h2
Substituting in the Euler–Lagrange equations, one finds


 Dtot △△w − p(x, y) = 0,

    

 12D1 1−ν 1+ν 12D2 1−ν 1+ν

 u ,
1 xx + u , + v , = − u , + u , + v , ,


 h 21 2 1 yy 2 1 xy h 22
2 xx
2 2 yy 2 2 xy

 12D1 
 1−ν 1+ν

12D2

1−ν 1+ν

2
v ,
1 yy + v ,
1 xx + u ,
1 xy = − 2
v ,
2 yy + v ,
2 xx + u 2 xy ,
, (2-29)
 h 1
2 2 h 2
2 2

  

 12D1 1−ν 1+ν
 G γ̃ = u ,
1 xx + u , + v , ,



zx
h 21 2 1 yy 2 1 xy

  

G γ̃zy = 12D
 1
v ,
1 yy +
1−ν
v , +
1+ν
u , ,
h 21 2 1 xx 2 1 xy
where the quantity Dtot , defined as
12D1 D2 E h1h2
Dtot = D1 + D2 + 2 2
H 2 = D1 + D2 + 2
H 2, (2-30)
D1 h 2 + D2 h 1 1 − ν h1 + h2
represents the flexural stiffness of a monolithic plate, whose flexural inertia is that of the two glass plies
properly spaced of the gap given by the thickness of the interlayer. This is indeed the monolithic limit of
laminated glass [Bennison et al. 1999].

3. The enhanced effective thickness approach

It is not possible to solve the system of differential equations (2-3), (2-4), (2-5), (2-6) and (2-7) in closed
form, but an approximation can be found by choosing an appropriate class of shape functions for the
unknown fields w(x, y), u 1 (x, y), u 2 (x, y), v1 (x, y) and v2 (x, y) defined up to parameters that will be
 THE EFFECTIVE THICKNESS OF LAMINATED GLASS PLATES 385

determined from energy minimization. The shape functions must be compatible with the qualitative
properties of the exact solution and, in particular, must comprehend the monolithic-limit solution, when
G → ∞, and the layered-limit solution, when G → 0. In terms of the field w(x, y), such borderline cases
correspond, respectively, to the fields w M (x, y) and w L (x, y) that, being the solutions of the differential
equations
Dtot △△w M (x, y) − p(x, y) = 0, (D1 + D2 )△△w L (x, y) − p(x, y) = 0, (3-1)
are of the form
g(x, y) g(x, y)
w M (x, y) ≡ , w L (x, y) ≡ , (3-2)
Dtot D1 + D2
where g(x, y) is a function that depends upon the boundary conditions of the problem at hand. Hence-
forth, we may define the equivalent (reduced) flexural rigidity of the laminate plate
1 η 1−η
= + , (3-3)
DR Dtot D1 + D2
being the parameter η a nondimensional quantity, tuning the plate response from the layered limit (η = 0)
to the monolithic limit (η = 1). An approximating class of solutions can thus be sought in the form
g(x, y)
w(x, y) = , (3-4)
DR
where g(x, y) is the shape function for the vertical displacement, uniquely determined by the shape of
the laminated glass plate in x − y plane, by the external load p(x, y) and by the geometric boundary
conditions.
The shape functions for the in-plane displacements should also guarantee that γ̃zx = 0, γ̃zy = 0 in the
borderline monolithic case. Recalling (2-18) and (2-19), we select the form
1 h2 1 h1
u 1 (x, y) = −β H g,x (x, y), u 2 (x, y) = β H g,x (x, y),
Dtot h 1 +h 2 Dtot h 1 +h 2
(3-5)
1 h2 1 h1
v1 (x, y) = −β H g, y (x, y), v2 (x, y) = β H g, y (x, y),
Dtot h 1 +h 2 Dtot h 1 +h 2
where β is another nondimensional parameter, again tuning the response from the layered limit (β = 0,
implying null in-plane force in the glass layers) to the monolithic limit (β = 1, leading to γ̃zx = γ̃zy = 0).
The corresponding total strain energy (2-2) can thus be rewritten as a function of the parameters η and
β to give

E[w, u 1 , u 2 , v1 , v2 ] = e
E[η, β] =
Z  2 1 
1 D1 + D2 2 12D1 D2 H

2
+β 2 2
(g,x x +g, yy )2 − 2(1 − ν)(g,x x g, yy −g,2x y )
2  DR D1 h 2 +D2 h 1 Dtot

GH2  1 β

2 2 p(x, y)
+ − [g,x +g, y ] + 2 g d x dy. (3-6)
t D R Dtot DR
This expression can be simplified by observing from (3-1) and (3-2) that w(x, y) = g(x, y)/D R of
(3-4) is the exact solution of the elastic bending of a plate with constant flexural rigidity D R under the
load p(x, y), with the same domain  and the geometric boundary condition of the problem at hand.
386 LAURA GALUPPI AND GIANNI ROYER-CARFAGNI

Consider the virtual work equality for this elastic body, in which the aforementioned w(x, y) is selected
as the strain/displacement field, whereas the stress/force field in equilibrium with p(x, y) is given by

 Mx = −D R (w,x x +νw, yy ) = −(g,x x +νg, yy ),
M = −D R (w, yy +νw,x x ) = −(g, yy +νg,x x ), (3-7)
 y
Mx y = D R (1 − ν)w,x y = (1 − ν)g,x y .
The external and internal virtual works, L ve and L vi , can be written as
Z
1
L ve = − p(x, y) g d x dy,
DR 
Z
L vi = [Mx w,x +M y w, y +Mx y γx y ] d x dy (3-8)
 Z
1  
=− (g,x x +g, yy )2 − 2(1 − ν)(g,x x g, yy −g,2x y ) d x dy.
DR 
Equality of external and internal virtual work then gives
Z Z
 
p(x, y)g d x dy = (g,x x +g, yy )2 − 2(1 − ν)(g,x x g, yy −g,2x y ) d x dy. (3-9)
 
This condition allows a drastic simplification of the energy (3-6). In fact, substituting (3-9) into (3-6),
the following expression for the strain energy can be found:

E[w, u 1 , u 2 , v1 , v2 ] = e
E[η, β] =
Z  2 1 
1 D1 + D2 2 12D1 D2 H
+β p(x, y) g
2  D 2R D1 h 22 +D2 h 21 Dtot

GH2  1 β

2 2 p(x, y) g
+ − [g,x +g, y ] + d x dy. (3-10)
t D R Dtot DR
Since g(x, y) is supposed to have been determined from (3-1), the integral in (3-10) depends upon the
free parameters η and β only, whose optimal value, say η∗ and β ∗ , is obtained by direct minimization.
The final result is that
1
η∗ = β ∗ = , (3-11)
D1 + D2 12D1 D2
1+ 9
(G/t)Dtot D1 h 12 + D2 h 21
where the coefficient R
p(x, y) g d x dy
9= R 2 2
(3-12)
 [g,x +g, y ] d x dy

depends upon the geometry of the plate and on its boundary and loading condition.
Deflection-effective thickness. The coefficient η, appearing in the definition of D R (3-3), is in some
sense analogous to the parameter Ŵ of (1-5) in the Bennison–Wölfel model [Wölfel 1987; Bennison
2009; Calderone et al. 2009], because the layered limit corresponds to Ŵ = η = 0 and the monolithic
limit to Ŵ = η = 1. From (3-3), the deflection-effective thickness ĥ ef;w , associated with the value η∗ , can
be written in the form
 THE EFFECTIVE THICKNESS OF LAMINATED GLASS PLATES 387

!−1/3
η∗ 1 − η∗
ĥ w = + 3 3
. (3-13)
h1h2
h 31 + h 32 + 12 H 2 h1 + h2
h 1 +h 2

Stress-effective thickness. The stress-effective thickness may be defined as the (constant) thickness ĥ i;ef;σ
of a monolithic plate for which the maximum bending stress is equal to the maximum stresses in the i-th
glass layer of the laminated plate. The stresses in a monolithic plate are associated with the moments
per-unit-length Mx , M y , Mx y , defined by (3-7). On the other hand, the i-th glass ply of the laminated
plate is subjected to moments and in-plane forces per unit length given by (2-13) and (2-11), respectively.
Then, ĥ i;ef;σ can be found by equating the two contributions, i.e.,

Ni x (x, y) 6Mi x (x, y) max | Mx (x, y)|

|σx x |max = max ± 2
= 1 2
,
hi hi ĥ
6 i;ef;σ

Ni y (x, y) 6Mi y (x, y) max | M y (x, y)|

|σ yy |max = max ± 2
= 1 2
, (3-14)
hi hi 6 ĥ i;ef;σ

Ni x y (x, y) 6Mi x y (x, y) max | Mx y (x, y)|

|σx y |max = max ± 2
= 1 2
.
hi hi ĥ
6 i;ef;σ

Recalling (3-4) and (3-5), the moments and in-plane forces per unit length can be rewritten as a function
of the shape function g(x, y) in the form
Di 12 D1 D2 H
Mi x = − (g,x x +νg, yy ), Ni x = −(−1)i η∗ 2 2 D
(g,x x +νg, yy ),
DR D2 h 1 + D1 h 2 tot
Di 12 D1 D2 H
Mi y = − (g, yy +νg,x x ), Ni y = −(−1)i η∗ 2 2 D
(g, yy +νg,x x ), (3-15)
DR D2 h 1 + D1 h 2 tot
Di 12 D1 D2 H
Mi x y = (1 − ν) g,x y , Ni x y = −(−1)i η∗ (1 − ν) 2 2 D
g,x y .
DR D2 h 1 + D1 h 2 tot
After defining, as in [Bennison 2009],
h1 H h2 H
h s;1 = , h s;2 = , (3-16)
h1 + h2 h1 + h2
one finds from (3-14) expressions analogous to that defined in (1-7) in the form
1 2η∗ h s;2 h1 1 2η∗ h s;1 h2
= + , = + . (3-17)
ĥ 21;σ 3 3
h 1 + h 2 + 12
h1 h2
H ĥ 3
w ĥ 22;σ 3 3
h 1 + h 2 + 12
h1 h2
H ĥ 3
w
h 1 +h 2 h 1 +h 2
Notice that the expressions for the effective thickness (3-13) and (3-17) are of the same type obtained
in [Galuppi and Royer-Carfagni 2012] for the one dimensional case.
In the following, the method based upon formulas (3-13) and (3-17) will be referred to as the enhanced
effective thickness (EET) approach.
388 LAURA GALUPPI AND GIANNI ROYER-CARFAGNI

4. Examples

The four paradigmatic cases of rectangular laminated plates {0 ≤ x ≤ a; 0 ≤ y ≤ b} with various boundary
conditions, as shown in Figure 5, are now examined. We set b = 2000 mm, while for a the three values
a = 2000 mm, a = 3000 mm and a = 6000 mm are considered. Parameters for glass plates are E = 70 GPa,
ν = 0.22, h 1 = h 2 = 10 mm, the thickness of the interlayer is t = 0.76 mm while its shear modulus G is
continuously varied from 0.01 MPa to 10 MPa, in order to evaluate its influence on the shear-coupling of
the glass plies. All the laminates are subjected to a uniformly distributed pressure p = 0.75 kN/m2 , but
since all materials are linear elastic, stress and strain depend linearly upon p. The results obtained with
the approximate methods will in each case be compared with an accurate numerical analysis performed
with ABAQUS, using a 3-D mesh with 110000 solid 20-node quadratic bricks with reduced integration,
available in the ABAQUS program library.
The shape function g(x, y) of (3-4) is assumed according to the form (uniform distribution) of the
external load p(x, y) and the geometric boundary conditions. The coefficient η∗ , which allows to calcu-
late the stress and deflection-effective thickness, as per (3-13) and (3-17), is calculated by using (3-11),
evaluating the parameter 9 through (3-12). But since for the case at hand h 1 = h 2 = h, as customary in
the design practice, the expression for η∗ can be simplified:
1
η∗ = . (4-1)
t E h3
1+ 9
G 1 − ν 2 2(h 2 + 3H 2 )
To facilitate the analysis of rectangular plates of any size, the values of 9 are recorded in tables as a
function of the length a and of the aspect ratio λ = b/a.

4.1. Simply supported rectangular plates. Consider a rectangular laminated glass under a uniformly
distributed load p with four simply supported edges (Figure 5a). The classical Navier solution [Timo-
shenko and Woinowsky-Krieger 1971] gives the elastic deflection of a monolithic plate with flexural

y y
z

z
a) b)

y x

x
z c) y
d)
z

Figure 5. Representative examples of laminated glass plates under various boundary

and load conditions.
 THE EFFECTIVE THICKNESS OF LAMINATED GLASS PLATES 389

l
a[mm] . . . . . . . . .
39.8732 10.2644 4.7813 2.8622 1.9739 1.4914 1.2005 1.0116 0.8822 0.7896
9.9683 2.5661 1.1953 0.7155 0.4935 0.3729 0.3001 0.2529 0.2205 0.1974
4.4304 1.1405 0.5313 0.3180 0.2193 0.1657 0.1334 0.1124 0.0980 0.0877
2.4921 0.6415 0.2988 0.1789 0.1234 0.0932 0.0750 0.0632 0.0551 0.0493
1.5949 0.4106 0.1913 0.1145 0.0790 0.0597 0.0480 0.0405 0.0353 0.0316
1.1076 0.2851 0.1328 0.0795 0.0548 0.0414 0.0333 0.0281 0.0245 0.0219

Table 1. Coefficient 9 (in mm−2 × 104 ) for rectangular plates with four edges simply supported.

rigidity D R in the form

mπ x nπ y
16 p X X sin a sin b
∞ ∞
w(x, y) = 6  2 2 2
. (4-2)
π DR m n
m=1 n=1 nm + 2
a2 b
Partial sums of a finite number of terms of the series can be used as approximations of the entire
function. By taking just the first term in the series (first-order approximation), the shape function g(x, y)
of (3-4) is
16 p 1 πx πy
g(x, y) = 6  2 sin sin , (4-3)
π 1 1 a b
+
a 2 b2
and the corresponding graph is drawn in Figure 6.
From this, the coefficient η∗ , evaluated through (3-11) or (4-1), reads
1
η∗ = . (4-4)
D1 + D2 12D1 D2 π 2 (a 2 + b2 )
1+
(G/t)Dtot D1 h 12 + D2 h 21 a 2 b2
It has been directly verified that higher-order approximations, obtained by considering more terms
of the series (4-2), do not substantially increase the level of accuracy. The coefficient 9[mm−2 ] that
appears in (3-11) and (4-1) is tabulated in Table 1 as a function of the length a [mm] and of the aspect
ratio λ = b/a.
For the case a = 3000 mm, b = 2000 mm and for a shear modulus of the polymeric interlayer G varying
from 0, 01 MPa to 10 MPa, the graphs in Figure 7 compare the deflection- and stress-effective thickness

7
x 10

-5
0
500
-10
1000
3000
2500
2000 1500
1500
1000
500 2000
0

Figure 6. Shape function for simply supported rectangular laminated plates. Case a =
3000 mm, b = 2000 mm.
390 LAURA GALUPPI AND GIANNI ROYER-CARFAGNI

Deflection-effective thickness Stress-effective thickness

22 22
MONOLITHIC LIMIT MONOLITHIC LIMIT
21 21

20 20

19 19

18 18
[mm]

[mm]
17 17

16 16

15 15

14 14
EET LAYERED LIMIT EET
13 LAYERED LIMIT B-W 13 B-W
Numerical Numerical
12 12
-2 -1 0 1 -2 -1 0 1
10 10 10 10 10 10 10 10
G[MPa] G[MPa]

Figure 7. Simply supported rectangular plate, case a = 3000 mm, b = 2000 mm. Com-
parison of the effective thicknesses obtained with the Bennison–Wölfel approach (B-W),
the enhanced effective thickness approach (ETT), and numerical simulations.

calculated according to the proposed approach of Section 3, here referred to as EET (enhanced effective
thickness), with the effective thicknesses calculated through (1-6) and (1-7) for the Bennison–Wölfel
model, recalled in the Introduction.
It is evident here that the proposed enhanced effective-thickness (ETT) approach and the Bennison–
Wölfel (B-W) formulation give qualitatively different results. However, B-W is on the side of safeness,
because it predicts deflection and stress values higher than those predicted by the EET approach. The
numerical simulations show that the EET approach provides a better approximation than B-W, but the
difference is not substantial, at least for the case at hand. The analytical approach recently proposed in
[Foraboschi 2012] for the particular case at hand gives results in good agreement with the EET results.
The case of Figure 8 corresponds to a = 6000 mm and b = 2000 mm, that is, the plate is a long
rectangle whose deformation tends to be cylindrical in a neighborhood of the center. In such a case, the
behavior predicted by the EET approach is close to Bennison–Wölfel’s. This is not surprising because the
aspect ratio is such that plate response is similar to the response of a beam (λ = b/a ≫ 1), and Bennison–
Wölfel’s model is calibrated on the case of simply supported beams under uniformly distributed load
[Galuppi and Royer-Carfagni 2012]. Numerical simulations confirm the accuracy.
On the contrary, the greatest differences between the EET and B-W approaches are obtained when
the plate is square, i.e., when the deflections of beam and plate differ the most. This case is illustrated
in Figure 9 for a plate with a = 2000 mm and b = 2000 mm. It is evident, here, that the results achieved
through the ETT approach are closer to the numerical data.

4.2. Rectangular plates with two opposite simply supported sides. For the case of rectangular plates
with two opposite simply supported sides (Figure 5b), following [Timoshenko and Woinowsky-Krieger
1971] and reasoning as in the previous case, the shape function g(x, y) may be chosen in the form
X∞  
4 4 mπ y mπ y mπ y mπ x
g(x, y) = pa 5 5
+ Am cosh + Bm sinh sin , (4-5)
π m a a a a
m=1,3,5,...
 THE EFFECTIVE THICKNESS OF LAMINATED GLASS PLATES 391

Deflection-effective thickness Stress-effective thickness

22 22

21 MONOLITHIC LIMIT 21 MONOLITHIC LIMIT

20 20

19 19

18 18
[mm]

[mm]
17 17

16 16

15 15

14 14
EET LAYERED LIMIT EET
13 LAYERED LIMIT B-W 13 B-W
Numerical Numerical
12 12
-2 -1 0 1 -2 -1 0 1
10 10 10 10 10 10 10 10
G[MPa] G[MPa]

Figure 8. Simply supported rectangular plate, case a = 6000 mm, b = 2000 mm (see
Figure 7 for legend).
Deflection-effective thickness Stress-effective thickness
22 22

21 MONOLITHIC LIMIT 21 MONOLITHIC LIMIT

20 20

19 19

18 18
[mm]

[mm]

17 17

16 16

15 15

14 14
EET LAYERED LIMIT EET
13 LAYERED LIMIT B-W 13 B-W
Numerical Numerical
12 12
-2 -1 0 1 -2 -1 0 1
10 10 10 10 10 10 10 10
G[MPa] G[MPa]

Figure 9. Simply supported square plate, case a = 2000 mm, b = 2000 mm (see Figure 7
for legend).

where
mπb mπb mπb
ν(1+ν) sinh −ν(1−ν) cosh
4 2a 2a 2a
Am = ,
π 5m5 mπb mπ b mπ b
(3+ν)(1−ν) sinh cosh −(1−ν)2
2a 2a 2a
(4-6)
mπb
ν(1−ν) sinh
4 2a
Bm = .
π 5m5 mπb mπ b 2 mπ b
(3+ν)(1−ν) sinh cosh −(1−ν)
2a 2a 2a
We take a first-order approximation just keeping the first term of the series, whose graph is plotted in
Figure 10.
392 LAURA GALUPPI AND GIANNI ROYER-CARFAGNI

8
x 10

-0.5

-1

-1.5
2000
1500
-2 1000
1500 1000 500
500 0 -500 -1000 0
-1500

Figure 10. Shape function for rectangular plate with two opposite edges simply sup-
ported. Case a = 3000 mm, b = 2000 mm.

The coefficient η∗ is given by (4-1), where 9, when evaluated by taking the first term only in the
series, is of the form

9=
(4-7)
8π 2 4b + B1 π 5 bC + 2aSπ 4 (A1 − B1 )

,
a B12 π 11 b2 SC + 32ab + abB1 (−B1 + 4A1 S 2 )π 10 + 2a 2 SC(2A21 + B12 )π 9 + 16B1 π 5 abC + 32a 2 Sπ 4 (A1 − B1 )

mπb mπb
with C = cosh and S = sinh . The value of 9 is reported in Table 2 as a function of a and
2a 2a
λ = b/a.
Figure 11 shows the comparison between the deflection- and stress-effective thicknesses, calculated
according to the enhanced effective thickness approach and to Bennison–Wölfel model as a function
of G.
In this case the EET and B-W approaches give results that in practice coincide. A plate under these
particular boundary conditions presents in fact a cylindrical deformed shape very similar to that of a
beam with equivalent cross-sectional inertia. The good approximation that can be achieved is evidenced
by the comparison with the numerical results.

4.3. Rectangular plates with three simply supported sides. For the case of rectangular plates with three
simply supported sides under uniform pressure (Figure 5c), Timoshenko and Woinowsky-Krieger [1971]
furnishes the general form for the elastic deflection of a monolith. Then, the shape function can be

l
a[mm] 0.2 0.4 0.6 0.8 1 1.25 1.667 2.5 5
500 0.4233 0.3908 0.3816 0.3770 0.3742 0.3718 0.3690 0.3653 0.3579
1000 0.1058 0.0977 0.0954 0.0943 0.0935 0.0929 0.0922 0.0913 0.0895
1500 0.0470 0.0434 0.0424 0.0419 0.0416 0.0413 0.0410 0.0406 0.0398
2000 0.0265 0.0244 0.0238 0.0236 0.0234 0.0232 0.0231 0.0228 0.0224
2500 0.0169 0.0156 0.0153 0.0151 0.0150 0.0149 0.0148 0.0146 0.0143
3000 0.0118 0.0109 0.0106 0.0105 0.0104 0.0103 0.0102 0.0101 0.0099

Table 2. Coefficient 9 (in mm−2 × 104 ) for rectangular plates with two opposite edges
simply supported.
 THE EFFECTIVE THICKNESS OF LAMINATED GLASS PLATES 393

Deflection-effective thickness Stress-effective thickness

22 22
MONOLITHIC LIMIT MONOLITHIC LIMIT
21 21

20 20

19 19

18 18
[mm]

[mm]
17 17

16 16

15 15

14 14
EET LAYERED LIMIT EET
13 LAYERED LIMIT B-W 13 B-W
Numerical Numerical
12 12
-2 -1 0 1 -2 -1 0 1
10 10 10 10 10 10 10 10
G[MPa] G[MPa]

Figure 11. Rectangular plates with two opposite edges simply supported, case a =
3000 mm, b = 2000 mm. Comparison of the effective thicknesses obtained with:
Bennison–Wölfel approach (B-W); the enhanced effective thickness approach (EET);
the numerical simulations.

written as
g(x, y) = p a 4 (4-8)
X∞  
4 mπ y mπ y mπ y mπ y mπ y mπ y mπ x
× 5 5
+Am cosh +Bm sinh +Cm sinh +Dm cosh sin
π m a a a a a a a
m=1,3,5,...

By imposing the relevant boundary conditions, one finds with some effort the values of the constants
appearing in (4-8) in the form
4 2 (2((3 − ν)S 2 + 2νC(C − 1)))a
Am = − , Bm = , Dm = ,
π m5
5 π m5
5 ((3 + ν)C Sa + (1 − ν)π bm)π 5 m 5
(4-9)
2(π 2 m 2 (1 − ν)2 b2 + 2ν(1 − ν)mπ Sab + (2νC(C − 1)(1 + ν) − 2S 2 (3 − ν))a 2 )
Cm = ,
(1 − ν)((3 + ν)C Sa + (1 − ν)π bm)π 5 m 5 a
where C = cosh(mπb/a) and S = sinh(mπ b/a). Figure 12 shows the graph of the first-order approxi-
mation of the shape function. The coefficient 9 to calculate η∗ from (3-12) is tabulated in Table 3 again
as a function of a and λ = b/a.
It is evident, from Figure 13, that the enhanced effective thickness approach and the Bennison–Wölfel
model give, in the case at hands, slightly different results; the data obtained numerically are in favor of
the approach proposed here.

4.4. Rectangular plates resting on corner points. The case of rectangular plates point-wise supported at
the corners under uniform pressure (Figure 5d) is of particular interest for frameless glazing, but presents
some difficulty because even the elastic solution for the monolith is not simple. The first attempts of
analytical solutions were given by Galerkin [1915] and Nádai [1922]. Then Wang et al. [2002] proved
394 LAURA GALUPPI AND GIANNI ROYER-CARFAGNI

8
x 10

-0.5

-1

-1.5

0
-2
0 500
500 1000
1000
1500 1500
2000
2500
3000 2000

Figure 12. Shape function for rectangular plates with three simply supported sides.
Case a = 3000 mm, b = 2000 mm.

l
a[mm] 0.2 0.4 0.6 0.8 1 1.25 1.667 2.5 5
500 0.5982 0.5578 0.5168 0.4855 0.4640 0.4465 0.4300 0.4156 0.4166
1000 0.1495 0.1394 0.1292 0.1214 0.1160 0.1116 0.1075 0.1039 0.1041
1500 0.0665 0.0620 0.0574 0.0539 0.0516 0.0496 0.0478 0.0462 0.0441
2000 0.0374 0.0349 0.0323 0.0303 0.0290 0.0279 0.0269 0.0260 0.0260
2500 0.0239 0.0223 0.0207 0.0194 0.0186 0.0179 0.0172 0.0166 0.0171
3000 0.0166 0.0155 0.0144 0.0135 0.0129 0.0124 0.0119 0.0115 0.0110

Table 3. Coefficient 9 (in mm−2 × 104 ) for rectangular plates with three edges simply supported.

Deflection-effective thickness Stress-effective thickness

22 22

21 MONOLITHIC LIMIT 21 MONOLITHIC LIMIT

20 20

19 19

18 18
[mm]

[mm]

17 17

16 16

15 15
LAYERED LIMIT
14 14
EET EET
13 LAYERED LIMIT B-W 13 B-W
Numerical Numerical
12 12
-2 -1 0 1 -2 -1 0 1
10 10 10 10 10 10 10 10
G[MPa] G[MPa]

Figure 13. Rectangular plates with three simply supported edges, case a = 3000 mm,
b = 2000 mm. Comparison of the effective thicknesses obtained with: Bennison–Wölfel
approach (B-W); the enhanced effective thickness approach (EET); the numerical simu-
lations.

that such works, focused on determining the deflection, produce rather inaccurate results in terms of
stress. Batista [2010] presented a solution in form of trigonometric series , where the coefficients of the
series form a regular infinite system of linear equations, providing accurate results for deflection, moment
 THE EFFECTIVE THICKNESS OF LAMINATED GLASS PLATES 395

and shear forces. Therefore, following Batista, the deflection of a monolith of flexural stiffness D is here
expressed in the form

p  4 
w(x, y) = 5(a + b4 ) − 6νa 2 b2 − 6(a 2 − νb2 )x 2 − 6(b2 − νa 2 )y 2 + x 4 − 6νx 2 y 2 + y 4
48(1 − ν)D
p X
∞ h sinh(α b) 2  i cos(α x)
n n
+ (−1)n Bn + b cosh(αn y) − y sinh(αn y)
D cosh(αn b) (1 − ν)αn b a cosh(αn b)
n=0
∞
p X h sinh(β a) 2  i cos(β y)
n n
+ (−1)n Dn + a cosh(βn y) − x sinh(βn y) , (4-10)
D cosh(βn a) (1−ν)βn a b cosh(βn a)
n=0

where αn = (2n + 1)π/(2a) and βn = (2n + 1)π/(2b). The coefficients Bn and Dn can be found by
expanding into a trigonometric series the boundary condition of null vertical forces on the free edges.
By taking, as in the previous cases, just the first-order approximation, the shape function g(x, y) reads

p  4 
g(x, y) = 5(a + b4 ) − 6νa 2 b2 − 6(a 2 − νb2 )x 2 − 6(b2 − νa 2 )y 2 + x 4 − 6νx 2 y 2 + y 4
48(1 − ν)
h sinh(α b) 2  i cos(α x)
0 0
+ p B0 + b cosh(α0 y) − y sinh(α0 y)
cosh(α0 b) (1 − ν)α0 b a cosh(α0 b)
h sinh(β a) 2  i cos(β y)
0 0
+ p D0 + a cosh(β0 y) − x sinh(β0 y) , (4-11)
cosh(β0 a) (1−ν)β0 a b cosh(β0 a)
where the coefficients B0 and D0 are given by
h 16ab4 i.
B0 = 16a 4 b Cba
2
(1 − ν) −a(1 − ν)π 2 + 2b Sab Cab (3 + ν)π + C 2
(1 − ν) Q,
(b2 + a 2 )2 ab
h i. (4-12)
16a 4 b
D0 = 16a 4 b Cba
2
(1 − ν) −b(1 − ν)π 2 + 2a Sba Cba (3 + ν)π + 2 C 2
(1 − ν) Q,
(b + a 2 )2 ba
with

Q = π 2 (ab(1 − ν)2 π 4 − 2(3 + ν)(1 − ν)(Cba Sba a 2 + Sab Cab b2 )π 3

256a 5 b5
+ 4ab Sab Sba Cba Cab (3 + ν)2 π 2 − C 2 C 2 (1 − ν)2 ),
(b2 + a 2 )4 ab ba
Cab = cosh2 (α0 b), Sab = sinh2 (α0 b), Cba = cosh(β0 a), Sba = sinh(β0 a). (4-13)

The shape function thus obtained is plotted in Figure 14.

The coefficient η∗ may be determined through (3-11) or (4-1) as a function of the material properties
and of the coefficient 9 of (3-12), tabulated in Table 4.
Figure 15 shows the comparison between the deflection- and stress-effective thickness calculated
according to EET and B-W approaches, for the case a = 3000 mm, b = 2000 mm. From this, it is
evident that the EET and B-W give substantially different results. For what concerns the stress-effective
thickness, numerical experiments are in favor of our present proposal.
396 LAURA GALUPPI AND GIANNI ROYER-CARFAGNI

8
x 10

-2

-4

-6

1000
-8
500
0
-10
1500 -500
1000 500 0 -500 -1000
-1000 -1500

Figure 14. Shape function for rectangular plates resting on corner points. Case a =
3000 mm, b = 2000 mm.

l
a[mm] 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
500 0.6112 0.5982 0.5793 0.5578 0.5363 0.5168 0.4998 0.4855 0.4737 0.4640
1000 0.1528 0.1495 0.1448 0.1394 0.1341 0.1292 0.1249 0.1214 0.1184 0.1160
1500 0.0679 0.0665 0.0644 0.0620 0.0596 0.0574 0.0555 0.0539 0.0526 0.0516
2000 0.0382 0.0374 0.0362 0.0349 0.0335 0.0323 0.0312 0.0303 0.0296 0.0290
2500 0.0244 0.0239 0.0232 0.0223 0.0215 0.0207 0.0200 0.0194 0.0189 0.0186
3000 0.0170 0.0166 0.0161 0.0155 0.0149 0.0144 0.0139 0.0135 0.0132 0.0129

Table 4. Coefficient 9 (in mm−2 × 104 ) for rectangular plates resting on corner points.
Deflection-effective thickness Stress-effective thickness
22 22

21
MONOLITHIC LIMIT 21 MONOLITHIC LIMIT

20 20

19 19

18 18
[mm]

[mm]

17 17

16 16

15 15

14 14
EET LAYERED LIMIT EET
13 LAYERED LIMIT B-W 13 B-W
Numerical Numerical
12 12
-2 -1 0 1 -2 -1 0 1
10 10 10 10 10 10 10 10
G[MPa] G[MPa]

Figure 15. Rectangular plates supported at corner points, case a = 3000 mm, b =
2000 mm. Comparison of the effective thicknesses obtained with: Bennison–Wölfel
approach (B-W); the enhanced effective thickness approach (EET); the numerical simu-
lations.

Figure 16 shows the deflection- and stress-effective thickness for a 2000 mm ×2000 mm square plate
supported on corner points. Once again, in this case experimental results are better fitted through the
EET approach.
Whenever b ≫ a, the plate deformed shape tends to be cylindrical and, consequently, the behavior
close to that of a beam. This is why for the case a = 6000 mm and b = 2000 mm, recorded in Figure 17,
B-W and EET give results that in practice coincide, in agreement with the numerical simulations.
 THE EFFECTIVE THICKNESS OF LAMINATED GLASS PLATES 397

Deflection-effective thickness Stress-effective thickness

22 22
MONOLITHIC LIMIT MONOLITHIC LIMIT
21 21

20 20

19 19

18 18
[mm]

[mm]
17 17

16 16

15 15

14 14
EET LAYERED LIMIT EET
13 LAYERED LIMIT B-W 13 B-W
Numerical Numerical
12 12
-2 -1 0 1 -2 -1 0 1
10 10 10 10 10 10 10 10
G[MPa] G[MPa]

Figure 16. Square plates supported at corner points, case a = 2000 mm, b = 2000 mm.
Comparison of the effective thicknesses obtained with: Bennison–Wölfel approach (B-
W); the enhanced effective thickness approach (EET); the numerical simulations.

Deflection-effective thickness Stress-effective thickness

22 22

21 MONOLITHIC LIMIT 21 MONOLITHIC LIMIT

20 20

19 19

18 18
[mm]

[mm]

17 17

16 16

15 15

14 14
EET LAYERED LIMIT EET
13 LAYERED LIMIT B-W 13 B-W
Numerical Numerical
12 12
-2 -1 0 1 -2 -1 0 1
10 10 10 10 10 10 10 10
G[MPa] G[MPa]

Figure 17. Rectangular plates supported at corner points, case a = 6000 mm, b =
2000 mm. Comparison of the effective thicknesses obtained with: Bennison–Wölfel
approach (B-W); the enhanced effective thickness approach (EET); the numerical simu-
lations.

5. Discussion and conclusions

Although it is possible to calculate numerically, and with excellent precision, the state of strain and
stress in laminated glass of any composition, size and shape, under the most various boundary and load
conditions, nevertheless simplified methods based upon the notion of effective thickness still remain
a very powerful tool, especially in the first preliminary phases of the design procedure. The designers
need simple expressions that allow to readily determine the structural response of laminated glass, leaving
398 LAURA GALUPPI AND GIANNI ROYER-CARFAGNI

more sophisticated computer methods to the final verifications. The most important result of this study
has been the definition of simple formulas for the effective thickness that can be applied, in principle, to
the two-dimensional problems of plates of any shape and size, under any boundary and load conditions.
This method is not recommended to evaluate local effects, such as stress concentrations around holes
and/or at contact points, but can be conveniently used to calculate maximum deflection and stress with
good accuracy, as confirmed by numerical and analytical approaches [Foraboschi 2012].
The key point is the assumption of a shape-function g(x, y) for the flexural-deformation surface of the
laminate, which, in accord with (3-4), is taken equal to the deflection of a monolithic plate of arbitrary
thickness with equal boundary and load conditions. Such a shape function may be estimated analytically,
when an analytical though approximate solution is available, or even numerically, when this is not the
case. From this, one can define the deflection- and stress-effective thicknesses of the laminate, by using
the simple formulas (3-13) and (3-17). These are defined by a parameter η∗ , indicating the shear coupling
offered by the polymeric interlayer, whose form can be calculated from expression (3-11), which in turn
is defined by the parameter 9, given by (3-12) as a function of the pressure distribution on the plate,
p(x, y), and of the shape function g(x, y). Indeed, the only “difficulty” of the proposed method, here
referred to as the enhanced effective thickness approach (EET), consists in calculating 9 from (3-12),
which involves integration of p(x, y) and g(x, y) over the referential domain of the plate, in accord with
(3-12). Here, we have exemplified this procedure for the case, very important in the design practice, of
a rectangular laminated glass plates under uniformly distributed pressure.
The shape function g(x, y) has been approximated by the first term of the series expansion for the
deflection surface of a monolithic plate, arriving at simple expressions for 9, whose values have been
recorded in the tables of figures 1, 2, 3 and 4 for various boundary conditions at the borders. Comparisons
with the results obtained with the classical formulas for the effective thickness à la Bennison–Wölfel (B-
W), in accordance with (1-5), (1-6) and (1-7), and with the results from accurate numerical models,
highlight the better accuracy that is obtained with the proposed EET approach. Indeed, the B-W ap-
proach assumes that laminated glass is a simply supported beam under uniformly distributed load; it
then turns out to be reliable only when the flexural deformation of the plate is cylindrical, i.e., in the
case of rectangular plates simply supported at two opposite sides. We have also shown that the use of
B-W formulation is not always on the side of safeness because they are cases, like that of a laminated
plate point-wise supported at the corners, where B-W gives effective thicknesses that underestimate both
deflection and stress.
A more general and comprehensive treatment of other relevant problems for laminated glass design
has been recorded in [Galuppi et al. 2012]. Here, the EET method is applied to the most common cases of
the design practice including plates under pseudoconcentrated loads, providing synthetic tables for ease
of reference and immediate applicability. The extension of the enhanced effective thickness approach to
other cases, like that of curved plates and shells, presents in principle no further conceptual difficulty,
and it will be the subject of future work.

Acknowledgements

Gianni Royer-Carfagni acknowledges the Italian MURST for its partial support under the PRIN2008
program.
 THE EFFECTIVE THICKNESS OF LAMINATED GLASS PLATES 399

References
[Aşik 2003] M. Z. Aşik, “Laminated glass plates: revealing of nonlinear behavior”, Comput. Struct. 81:28–29 (2003), 2659–
2671.
[Aşik and Tezcan 2005] M. Z. Aşik and S. Tezcan, “A mathematical model for the behavior of laminated glass beams”, Comput.
Struct. 83:21–22 (2005), 1742–1753.
[Batista 2010] M. Batista, “New analytical solution for bending problem of uniformly loaded rectangular plate supported on
corner point”, IES J. A Civ. Struct. Eng. 3:2 (2010), 75–84.
[Behr et al. 1993] R. A. Behr, J. E. Minor, and H. S. Norville, “Structural behavior of architectural laminated glass”, J. Struct.
Eng. (ASCE) 119:1 (1993), 202–222.
[Bennison 2009] S. J. Bennison, “Structural properties of laminated glass”, short course presented at Glass performance days
(Tampere), 2009.
[Bennison et al. 1999] S. J. Bennison, A. Jagota, and C. A. Smith, “Fracture of glass/poly(vinyl butyral) (Butacite®) laminates
in biaxial flexure”, J. Am. Ceram. Soc. 82:7 (1999), 1761–1770.
[Bennison et al. 2001] S. J. Bennison, C. A. Smith, A. Van Duser, and A. Jagota, “Structural performance of laminated glass
made with a ‘stiff’ interlayer”, pp. 283–287 in Glass processing days (Tampere, 2001), edited by J. Vitkala, Tamglass, Tampere,
2001.
[Bennison et al. 2005] S. J. Bennison, J. G. Sloan, D. F. Kristunas, P. J. Buehler, T. Amos, and C. A. Smith, “Laminated glass
for blast mitigation: role of interlayer properties”, in Glass processing days (Tampere, 2005), Glaston Finland/GPD, Tampere,
2005.
[Bennison et al. 2008] S. J. Bennison, M. H. X. Qin, and P. S. Davies, “High-performance laminated glass for structurally
efficient glazing”, paper presented at conference on Innovative light-weight structures and sustainable facades (Hong Kong),
2008, available at http://tinyurl.com/Bennison-et-al-2008.
[Calderone et al. 2009] I. Calderone, P. S. Davies, S. J. Bennison, X. Huang, and L. Gang, “Effective laminate thickness for
the design of laminated glass”, in Glass performance days (Tampere, 2009), Glaston Finland/GPD, Tampere, 2009.
[Cas et al. 2007] B. Cas, M. Saje, and I. Planinc, “Buckling of layered wood columns”, Adv. Eng. Softw. 38:8–9 (2007), 586–
597.
[Foraboschi 2007] P. Foraboschi, “Behavior and failure strength of laminated glass beams”, J. Eng. Mech. (ASCE) 133:12
(2007), 1290–1301.
[Foraboschi 2012] P. Foraboschi, “Analytical model for laminated-glass plate”, Compos. B Eng. 43:5 (2012), 2094–2106.
[Galerkin 1915] B. G. Galerkin, “Sterжin i plastinki: rdy v nekotoryh volrosah uprugogo ravnovesi
sterжneĭ i plastinok”, Vestn. Inzh. Tekh. (USSR) 1:19 (1915), 897–908. In Russian; translated as Rods and plates:
series in some questions of elastic equilibrium of rods and plates, National Technical Information Service, Springfield, VA,
1968.
[Galuppi and Royer-Carfagni 2012] L. Galuppi and G. F. Royer-Carfagni, “Effective thickness of laminated glass beams: new
expressions via a variational approach”, Eng. Struct. 38 (2012), 53–67.
[Galuppi et al. 2012] L. Galuppi, G. Manara, and G. F. Royer-Carfagni, “Practical expressions for the design of laminated
glass”, 2012, available at http://hdl.handle.net/1889/1720.
[Hooper 1973] J. A. Hooper, “On the bending of architectural laminated glass”, Int. J. Mech. Sci. 15:4 (1973), 309–323.
[Ivanov 2006] I. V. Ivanov, “Analysis, modelling, and optimization of laminated glasses as plane beam”, Int. J. Solids Struct.
43:22-23 (2006), 6887–6907.
[Knops and Payne 1971] R. J. Knops and L. E. Payne, Uniqueness theorems in linear elasticity, Springer, New York, 1971.
[Le Grognec et al. 2012] P. Le Grognec, Q.-H. Nguyen, and M. Hjiaj, “Exact buckling solution for two-layer Timoshenko
beams with interlayer slip”, Int. J. Solids Struct. 49:1 (2012), 143–150.
[Louter et al. 2010] C. Louter, J. Belis, F. Bos, D. Callewaert, and F. Veer, “Experimental investigation of the temperature effect
on the structural response of SG-laminated reinforced glass beams”, Eng. Struct. 32:6 (2010), 1590–1599.
[Murakami 1984] H. Murakami, “A laminated beam theory with interlayer slip”, J. Appl. Mech. (ASME) 51:3 (1984), 551–559.
400 LAURA GALUPPI AND GIANNI ROYER-CARFAGNI

[Nádai 1922] A. Nádai, “Über die Biegung durchlaufender Platten und der rechteckigen Platte mit freien Rändern”, Z. Angew.
Math. Mech. 2:1 (1922), 1–26.
[Newmark et al. 1951] N. M. Newmark, C. P. Siess, and I. M. Viest, “Test and analysis of composite beams with incomplete
interaction”, Proc. Soc. Exp. Stress Anal. 9:1 (1951), 75–92.
[Norville et al. 1998] H. S. Norville, K. W. King, and J. L. Swofford, “Behavior and strength of laminated glass”, J. Eng. Mech.
(ASCE) 124:1 (1998), 46–53.
[Sagan 1969] H. Sagan, Introduction to the calculus of variations, McGraw-Hill, New York, 1969. Reprinted Dover, New York,
1992.
[Schnabl and Planinc 2011] S. Schnabl and I. Planinc, “The effect of transverse shear deformation on the buckling of two-layer
composite columns with interlayer slip”, Int. J. Non-Linear Mech. 46:3 (2011), 543–553.
[Timoshenko and Woinowsky-Krieger 1971] S. P. Timoshenko and S. Woinowsky-Krieger, Theory of plates and shells, Mc-
Graw-Hill, New York, 1971.
[Wang et al. 2002] C. M. Wang, Y. C. Wang, and J. N. Reddy, “Problems and remedy for the Ritz method in determining stress
resultants of corner supported rectangular plates”, Comput. Struct. 80:2 (2002), 145–154.
[Wölfel 1987] E. Wölfel, “Nachgiebiger Verbund: eine Näherungslösung und deren Anwendungsmöglichkeiten”, Stahlbau
56:6 (1987), 173–180.
[Xu and Wu 2007] R. Xu and Y. Wu, “Static, dynamic, and buckling analysis of partial interaction composite members using
Timoshenko’s beam theory”, Int. J. Mech. Sci. 49:10 (2007), 1139–1155.

Received 30 Nov 2011. Revised 1 Mar 2012. Accepted 16 Mar 2012.

L AURA G ALUPPI : [email protected]
Civil-Environmental Engineering and Architecture, University of Parma, Parco Area delle Scienze 181/A, I, 43124 Parma,
Italy
G IANNI ROYER -C ARFAGNI : [email protected]
Civil-Environmental Engineering and Architecture, University of Parma, Parco Area delle Scienze 181/A, I, 43124 Parma,
Italy

mathematical sciences publishers msp

JOURNAL OF MECHANICS OF MATERIALS AND STRUCTURES
jomms.net

Founded by Charles R. Steele and Marie-Louise Steele

EDITORS
C HARLES R. S TEELE Stanford University, USA
DAVIDE B IGONI University of Trento, Italy
I WONA JASIUK University of Illinois at Urbana-Champaign, USA
YASUHIDE S HINDO Tohoku University, Japan

EDITORIAL BOARD
H. D. B UI École Polytechnique, France
J. P. C ARTER University of Sydney, Australia
R. M. C HRISTENSEN Stanford University, USA
G. M. L. G LADWELL University of Waterloo, Canada
D. H. H ODGES Georgia Institute of Technology, USA
J. H UTCHINSON Harvard University, USA
C. H WU National Cheng Kung University, Taiwan
B. L. K ARIHALOO University of Wales, UK
Y. Y. K IM Seoul National University, Republic of Korea
Z. M ROZ Academy of Science, Poland
D. PAMPLONA Universidade Católica do Rio de Janeiro, Brazil
M. B. RUBIN Technion, Haifa, Israel
A. N. S HUPIKOV Ukrainian Academy of Sciences, Ukraine
T. TARNAI University Budapest, Hungary
F. Y. M. WAN University of California, Irvine, USA
P. W RIGGERS Universität Hannover, Germany
W. YANG Tsinghua University, China
F. Z IEGLER Technische Universität Wien, Austria

PRODUCTION [email protected]
S ILVIO L EVY Scientific Editor

Cover design: Alex Scorpan Cover photo: Mando Gomez, www.mandolux.com

See http://jomms.net for submission guidelines.
JoMMS (ISSN 1559-3959) is published in 10 issues a year. The subscription price for 2012 is US $555/year for the electronic
version, and $735/year (+$60 shipping outside the US) for print and electronic. Subscriptions, requests for back issues, and changes
of address should be sent to Mathematical Sciences Publishers, Department of Mathematics, University of California, Berkeley,
CA 94720–3840.
JoMMS peer-review and production is managed by EditF LOW® from Mathematical Sciences Publishers.
PUBLISHED BY
mathematical sciences publishers
http://msp.org/
A NON-PROFIT CORPORATION
Typeset in LATEX
Copyright ©2012 by Mathematical Sciences Publishers
Journal of Mechanics of Materials and Structures
Volume 7, No. 4 April 2012

Analytical study of plastic deformation of clamped circular plates subjected to

impulsive loading H ASHEM BABAEI and A BOLFAZL DARVIZEH 309
Thereotical solutions of adhesive stresses in bonded composite butt joints
G ANG L I 323
Micromechanical study of dispersion and damping characteristics of granular
materials N IELS P. K RUYT 347
Buckling instabilities of elastically connected Timoshenko beams on an elastic layer
subjected to axial forces
V LADIMIR S TOJANOVI Ć , P REDRAG KOZI Ć and G ORAN JANEVSKI 363
The effective thickness of laminated glass plates
L AURA G ALUPPI and G IANNI ROYER -C ARFAGNI 375
Elastic solution in a functionally graded coating subjected to a concentrated force
ROBERTA S BURLATI 401

1559-3959(2012)7:4;1-A