Effective Laminate Thickness for the Design of

Laminated Glass
Dr. Ignatius Calderone,* Mr. Phillip S. Davies,** Dr. Stephen J. Bennison,** Professor Huang Xiaokun*** and Mr. Liu Gang***
* Calderone and Associates Pty Ltd, Australia
** E.I. DuPont de Nemours & Co Inc., Wilmington, DE USA
*** China Academy of Building Research (CABR), Beijing China

Keywords
calculating the thickness of a monolithic beam with eq
1=Laminated Glass 2=Ionoplast 3=PVB 4=Effective Thickness 5=AS 1288 6=Design
laminated beam. This thickness then can be used in place
equations for deformation of beams and simplified finite
calculating the thickness of a monolithic beam with eq
Abstract temperature on the polymer calculating
equations
constitutive the glass
thickness
describe the shearof laminate
properties, a monolithic
coupling between
geometry beam
and twowith eq
glass
laminated beam. This thickness then can be used in place
In the design of laminated glass to laminated
properties. The upshot of shear
this body beam.
coupling the This
lengththickness
depends scale then
in the
primarily oncanthebe
problem. Theused in place
interlayer shea
equations
of work is the capability to equations
model for deformation of beamsΓand
shear transfer coefficient, simplified
, which is a finite
laminate for deformation
geometry and of beams and simplified finiteT
meet various structural loads, engineers equations
accurately the structural behaviour of describe
measuretheofthe
shearlength scale
coupling
the transfer
in the
between
of shear
problem.
stressestwo glass
equations
which describe ofthe
iscoupling
a measure theshear coupling
transfer between two glass
isofgiven
shear thestresses across
generally make simplifying assumptions
laminated glass using modern shearfinite depends
across the primarily
interlayer, on by: interlayer she
or make reference to procedures
element methods. However, shear
the coupling depends primarily on the interlayer she
given in various standards. The current laminate geometry and the length scale in the problem.
glass design industry often laminate geometry and the1length scale in the problem.
Australian standard, AS1288, which
approach of using simplified
takes the
which
which is a measure of the
calculation Γ=
is a measure of the transfer
1+ 9.transfer
of shear stresses across
6 2s v2 of shear stresses
EI h (1) across
was published in 2006, allows the total
methods for engineering laminated Ghs a
glass thickness to be used for short-
term load duration but for medium
glass due to the slow adoption of finite
With:
element technology. These simplified With:
Γ=
Γ = 1
1EI h
12+ 9.6 EI s2s h2vv2
or long-term load durations it gives

Laminated Glass / Process / Design

design approaches are often inaccurate, I s = h1h1s+;29+.6hGh Gh 2h s2 as2;1
relationships for load sharing factors sa
for the glass components. However,
although usually conservatively
With:so. Such I s = h1hs2;2 + h2hs2;1 (2)
With:result
overly-conservative approaches 2 2
these methods may not give correct I = hhhh2 + h h 2
results and the Australian standard
in much over-design of laminated glass hI s = hh1hshs ;12 + h22hss;;11
along with an associated, unnecessary hss;1;s1 = h1 s+s1;h2 (3)
alternatively allows a full non-linear
cost penalty. Accordingly, there is an h11 + h22
analysis, modelling the glass-interlayer hh
industry need to develop calculation hs;1 = hssh11
sheets behaviour, to be undertaken. h = hh h+
h22h2
Since such an analysis is not readily
methods that capture accurately the hhss;;s22;1== hh11ss+ h
available to all designers, a Working
mechanical response of laminated glass hh11 ++ hh222 (4)
while being relatively straightforward
Group was formed in Australia in order hh
to determine if a procedure using the
to implement in standards and existing
h hs= ;2 0 =.5(hhssh22+ h ) + h
calculation methodologies. In this hhs
ss ; 2= =
0 .h (+
5 h1 h+ h2 ) + hvv (5)
effective thickness of the laminated
contribution we examine the so-called h11 +1h22 2
glass could be adopted for the design Where:
“effective-thickness” method for Where:
of laminated glass. This paper presents
the design of laminated glass. This hhhhν s = 0=.5Interlayer ( h1 + h2 )=thickness + hInterlayer thickness
ν s = 0.5( h1 + h2 ) + hv
v
the results of the investigation carried h = Glass ply 1 minimum thic
method has been proposedWhere:for several h1 1 = Glass ply 1 minimum thickness
out by the authors for the purpose Where: h = Glass ply 2 minimum thic
developing standards in Europe, USA, hh2 ν
2 = Glass ply 2=minimum thickness
of assisting the Working Group in Interlayer thickness
China and Australia. We will present Ehν
E = Glass == Glass
Young’s InterlayerYoung’s
modulus thickness
(70modulus
adopting revised design procedures h = Glass ply 1 minimum thi
for use in the Australian Standard.
evidence that the method has the ah11 GPa) == Length
Glass ply scale 1 minimum
(shortestthi
be
capability of being used in structural h 2 = Glass ply 2 minimum thi
The investigation includes a review of G
a h2 = Length scale == Interlayer
Glass
(shortest plybending
2 minimum
shear thi
modulus
calculations for the performance of E = Glass Young’s modulus
data from test programs in which the E direction) = Glass Young’s modulus
laminated glass and captures many of a = Length scale (shortest b
laminate effective thickness has been
the important variables thatThe G
shear transfer
influence a coefficient, = Interlayer Γ ,shear=varies
Length from
modulus scale to(shortest
0shear 1.0. b
measured directly and compared to G = Interlayer modulu
performance. G = Interlayer shear modulus
simplified calculation methods based
For calculations The shear transfer
of laminate deflection, coefficient, theΓlaminate
, varies effective
on the effective thickness approach. The shear transfer
from coefficient,
0 to 1.0. Γ , varies from 0 to 1.0.
Additionally, finite element methods
Effective Thickness Concept The shear transferFor coefficient,
calculations of laminate
Γ , varies from 0 to 1.0.
are compared to an effective thickness The concept of the “effective”
For thickness hof 3 h 3 laminate
h23 + 12 ΓI
calculationsdeflection, w = 1 +deflection,
of laminate the effective
For gained
calculationsthickness,
ef ;laminate sthe laminate effectiv
approach to calculating laminated glass of laminated glass has recently hef; w,deflection,
is given by: the laminate effectiv
structural behaviour. traction in the design community and
For calculations hofef ;wthe
is based on analysis of composite h133 + h233 + 12
= 33maximum glass
ΓI bending(6)stress, th
Introduction sandwich structures originally developed h
(one for each glass ply) are1 given
ef ; w = h + h2 +by: 12 ΓI ss
by Wölfel [5]. The analysis proposes For calculations of the maximum glass
The structural behaviour of laminated For calculations of thestress, maximum glass bending stress, t
analytic equations that provide
glass is a complex topic and many For acalculationsbending
of the maximum 3
the laminate
hefgiven glasseffective
bending stress, t
(one for each glass
method of calculating the thickness thicknesses
h ply)
= are(one ;wfor each by: glass ply) are
factors influence the response (one
of a monolithic beam with equivalentfor each glass 1;ef ;σ ply) are given by:
of a laminated plate or beam to given by: h1 + 2 Γhs ;2
bending properties to a laminated
an imposed load. Despite this
beam. This thickness then can be h3
complexity, much progress has been h1;ef ;σ = hefef3 ;;ww
used in place of the actual thickness h1;ef ;σ = h1 + h 32 Γhs ;2 (7)
made in understanding laminated
in analytic equations for deformation h2;ef ;σ = h1 +ef2;wΓhs ;2
glass in the last decade [1-3]. This
of beams and simplified finite element h2 + 2 Γhs ;1
progress is primarily attributable to
analysis. The analytic equations hef33 ;w
advances in mechanics and associated
describe the shear coupling between h = h ef ;w
The h22;;efef ;;σσ = hneeds
calculation normally only to be performed (8) for
computational tools [4] and the 2 + 2 Γhs ;1
two glass plies through the interlayer. h + 2
different types of glass in the laminate that have different
2 Γh s ;1
development of appropriate interlayer
The shear coupling depends primarily
property information that accurately interlayer
G, property that influences the laminate deformatio
The calculation
on the interlayer shear stiffness, normally needs only to be performed for
capture the effects of load duration and The calculation
shear modulus isnormally
a measure needs of only to be performed
the plastic interlayer’sforsh
different types of glass in the laminate that have different
different
shear types of glass
resistances, in theeffectively
the more laminate that haveglass
the two different
plie
GLASS PERFORMANCE DAYS 2009 | www.gpd.fi interlayer property that influences the laminate deformatio
interlayer
under property
loading. The that influences
effective laminatethe thickness
laminate deformatio
approache
1
shear modulus is a measure of the plastic interlayer’s s
shear modulus(Γis →a 1)
measure
and of the plastic layered
interlayer’s s(
stiff
shearinterlayers
resistances, the moreapproaches
effectively the
the two glass limitpli
shear resistances,
interlayers (Γ → 0). the
Note more
that effectively
the polymer the
doestwonotglass
need pli
to
under loading. The effective laminate thickness approach
 The calculation normally needs only to
be performed for the thickest ply, unless
there are different types of glass in the
laminate that have different allowable
stresses. The primary interlayer
property that influences the laminate
deformation is the shear modulus, G.
The shear modulus is a measure of the
plastic interlayer’s shear resistance. The
greater the shear resistances, the more
effectively the two glass plies couple and
resist deformation under loading. The
effective laminate thickness approaches
the total laminate thickness for stiff
interlayers (Γ → 1) and approaches the
layered limit (IGU approximation) for
compliant interlayers (Γ → 0). Note that
the polymer does not need to achieve
the glass modulus value to impart
efficient structural coupling.
Strictly, the analysis only applies
to pure bending of beams and it is
important to identify the correct length
scale to use in a design calculation
(usually the shortest dimension
along a bending direction). Despite
these restrictions, the approach is a
great simplification to the analysis of
laminated glass and when properly
used, can provide efficient design
Laminated Glass / Process / Design
solutions with minimal computation.
Key to the use of this approach is the
availability of comprehensive interlayer
modulus data and knowledge of the
temperature/load duration conditions
for the loading actions.
Experimental Measurements of
Laminate Effective Thickness
Here we propose a method to
experimentally determine the effective
thickness of a laminated glass ply
under well-defined loading and
support conditions. These tests have
been carried using a four-point bend
methodology based on the proposed
ISO 1288-3 standard. The method
comprises of three steps: 1) measure
the maximum glass stress (or deflection)
for a range of applied loads (at a
specified loading rate and temperature);
2) Fit a straight line to the stress-load
characteristic using a linear least squares
method and determine the slope, Φ, of
the fitted line; 3) Extract the effective
thickness, heff,σ using:
3( Ls − Lb )
heff ,s = (9)
2 BΦ
Where: Ls = support dimension
(1,000 mm) Lb = loading dimension
(200 mm), and B = specimen width (360
mm).
Figure 1 plots the maximum principal
glass stress (measured with strain
gages) as a function of applied load for
nominal 5 mm annealed glass / 0.76
mm Ionoplast interlayer (DuPont™
SentryGlas® Plus) / 5 mm annealed
glass. Note tests have been run for
temperature spanning 25 ºC to 80 ºC.
The laminate effective thickness as
a function of temperature extracted
using the procedure described above
2 GLASS PERFORMANCE DAYS 2009 | www.gpd.fi
Figure 1 Figure 2
Maximum principal glass stress in a nominal 5 Effective thickness for a nominal 5 mm glass /
mm glass / 0.76 mm DuPont™ SentryGlas® / 5 0.76 mm DuPont™ SentryGlas® / 5 mm glass
mm glass laminate under four point bending for laminate under four point bending for a range
a range of temperatures of temperatures.
is graphed in Figure 2. Note that SentryGlas® Shear
the effective thickness is essentially Temperature, T (oC) Modulus, G – 1 minute
load duration (MPa)
10 mm for this ionoplast laminate at
temperatures of 50 ºC and below. 10 225
Above 50 ºC, some reduction in 20 195
effective thickness is seen as the 24 173
interlayer begins to soften. 30 110
The proposed methodology has good 40 30.7
capability for measuring the effective 50 11.3
thickness of laminated glass under a 60 3.64
specified deformation state and allows 70 1.88
study of the role of load duration and 80 0.83
temperature in laminate behaviour.
More extensive results from these Table 1
studies will be presented elsewhere. SentryGlas® shear moduli for a range of tem-
perature (1 minute load duration).
Comparisons of Effective Thickness
Predictions with Experimental
results
We now use the effective thickness
methodology described above to
calculate the effective thickness of
the laminates used in the bending
experiments. Firstly, accurate
measurements of the glass and
interlayer thickness are needed. The
measured laminate construction used in
the test was: 4.63 mm annealed glass /
0.76 mm DuPont™ SentryGlas® Plus /
4.63 mm annealed glass (total laminate Figure 3
thickness = 10.02 mm). The length Cantilevered Laminated Glass Balustrade
scale chosen for calculation purposes
was set at 1 m, the support rolls
span in the four point bend test. The Analysis with the Effective
remaining piece of information needed Thickness: Cantilevered Laminated
for calculations is the interlayer shear Glass Balustrade
modulus for the various temperatures We now examine the use of the
at the appropriate time scale (1 minute effective thickness approach for
in this case). Table 1 shows the shear calculating glass stress and laminate
modulus properties of the Ionoplast deflection of a cantilevered laminated
interlayer (DuPont™ SentryGlas®). glass balustrade.
Figure 3 Cantilevered
The effective laminate thickness FigureLaminated
3 shows theGlass Balustrade
example of
for stress behaviour calculated using interest. The standard analytical
equations 7 and 8 and is plotted
Figurein3 showsformulae
the example
can be usedof interest. The standard analytic
to determine
Figure 2. As can be seen, the predicted
determine the maximum glass stress, σσmax
the maximum glass stress, < and
max < and laminate deflectio
effective thickness behaviour is in laminate deflection, δmax:
close agreement with the measured
behaviour. We have also carried out 6 Pa
similar experiments using PVB laminates σ max = 2 (10)
h1,ef ;σ
and have found the calculation method
to be accurate in describing the effective 4 Pa 3
thickness of laminates in the four point δ max = 3 (11)
Ehef ;w
bend test.
For the example of:
• P = 0.75 kN/m
• a=1m
• hglass = 6 mm (use 5.8 mm ISO minimum)
For the example of: Maximum Glass Stress, σmax (MPa) Maximum Laminate Deflection, δmax (mm)
• P = 0.75 kN/m Finite Elements Effective Thickness Finite Elements Effective Thickness
• a = 1 m 27.8 26.6 18.5 19.8
• hglass = 6 mm (use 5.8 mm ISO
minimum) Table 2
• hinterlayer = 1.52 mm Computed glass stress and laminate deflection for a cantilevered laminated glass balustrade. Effec-
• Ginterlayer = 59.9 MPa (SentryGlas®: 60 tive thickness predictions and finite element predictions are compared.
mins at 30 °C)
• Eglass = 70 GPa
•
The effective thickness values computed
are:
• Laminate Effective thickness
(deflection), heff; w = 12.94 mm
• Laminate Effective thickness (stress),
h1;ef;σ = 13.02 mm

Inserting these effective thickness

values into equations 10 and 11, gives
the stress and deflection values. These
values are presented in Table 2.
As can be seen from Table 2, the
values predicted using the effective
thickness approach with known analytic
formulae compare well with those
derived from the finite element method.

Conclusions
We have demonstrated that the

Laminated Glass / Process / Design

effective thickness method is a useful
approach to calculating glass stress
and deflection of laminated glass.
The approach provides a simplified
methodology for design work and
adequately incorporates the major
variables that influence the behaviour
of laminated glass. The methodology is
currently being reviewed by standards
bodies globally.

References
[1] Bennison, S.J., Jagota, A. & Smith, C.A., (1999)
“Fracture of Glass/Polyvinyl Butyral (Butacite®)
Laminates in Biaxial Failure,” J. Am. Ceram.
Soc., 82[7] 1761-70
[2] Bennison, S.J., Smith, C.A., Van Duser, A. and
Jagota, A., (2001) “Structural Performance
of Laminated Safety Glass Made with “Stiff”
Interlayers”, proceedings of the seventh
international conference on architectural and
automotive glass Glass Processing Days 2001,
Tampere (Finland), published by Tamglass Ltd.
Oy, Vehmaistenkatu 5, Tampere.
[3] Van Duser, A., Jagota, A., Bennison, S.J. (1999)
“Analysis of Glass/Polyvinyl Butyral (Butacite)
Laminates Subjected to Uniform Pressure”
Journal of Engineering Mechanics, ASCE,
125[4] 435-42
[4] ABAQUS version 6.8, (2009) Simulia,
Pawtucket, R.I. 02860 USA.
[5] Wölfel, E., (1987) Nachgiebiger Verbund
Eine Näherungslösung und deren
Anwendungsmöglichkeiten. In: Stahlbau
6/1987 pp. 173-180

GLASS PERFORMANCE DAYS 2009 | www.gpd.fi 3