Laminated

Engineering Failure Analysis 94 (2018) 412–437
Contents lists available at ScienceDirect
Engineering Failure Analysis

journal homepage: www.elsevier.com/locate/engfailanal
Review
Applications of finite element modelling in failure analysis of

T
laminated glass composites: A review
⁎
Meenu Teotia , R.K. Soni
Department of Chemistry, Chaudhary Charan Singh University, Meerut, UP 250004, India
A R T IC LE I N F O ABS TRA CT
Keywords: The subject of the paper is modelling of complex damage behaviour of laminated glass, which is a
Laminated glass sandwiched polymer composite. Last 20 years have witnessed extensive growth in computational
Modelling modelling of complex nonlinear behaviour of laminated glass panels with viscoelastic interlayers.
Simulation Present review article is an attempt to highlight applications of finite element approach in failure
CLPT
analysis of laminated glasses. Evolution of modelling theories such as equivalent single layer
FSDT
theories, layerwise theories and zig-zag models have been presented to insight the fundamental
Zigzag
FEM concepts used in modelling. Finite element techniques such as erosion, cohesive zone and ex-
Automobile windshields tended finite element methods available in FE packages are briefly reviewed alongwith material
Failure analysis models for glass and interlayer. The systematic growth in FE models in numerical simulation of
Blast loading human head impact on windshields and blast loading is specifically presented. Finite element
models have been proposed as useful guides to engineers for fabrication of laminated glasses with
more realistic data to be used in real world situations without going for expensive and time
consuming experimental set ups. Overall 223 research articles have been included exhibiting the
research performed in context of laminated glass.
1. Introduction
Damage prediction of polymer composites involves determination of impact forces & induced stresses and application of suitable
failure criterion for initial failures. Upto 1991, a large amount of experimental data was published and several important features of
impact damage were identified. Polymer composites have been reviewed by various authors detailing their damage behaviour and
modelling of damage characteristics. Abrate [1] organized all the experimental findings and presented a very comprehensive review
of impact of laminated composites. Several types of simple models such as spring mass models, energy balance models were ac-
counted for the behavior of structure, impactors and the local deformation of contact force. In further attempt, Abrate [2] organized
newer developments and focused on low velocity impact damage, the damage predictions and evaluation and prediction of residual
properties of damaged laminates. The author compared and analyzed various mathematical models and two & three dimensional
elasticity models. Impact damage, how it develops and what factors influence impact resistance were highlighted and a detailed
description of experimental techniques, failure modes, effect of material properties, effect of projectile characteristics, effect of layup
and preload and effect of environmental conditions were presented. Liu and Zheng [3] reviewed damage modelling and finite
element analysis for carbon fiber reinforced polymer matrix composite laminates. Finite element implementation of progressive
failure analysis in terms of the mechanical response for the variable stiffness composite laminates arising from the continuous failure
was presented. The nature and mechanics of damage in brittle layer structures with soft interlayers under concentrated loads was
⁎
Corresponding author.
E-mail address: teotiameenu@gmail.com (M. Teotia).
https://doi.org/10.1016/j.engfailanal.2018.08.016
Received 17 April 2018; Received in revised form 13 July 2018; Accepted 12 August 2018
Available online 16 August 2018
1350-6307/ © 2018 Elsevier Ltd. All rights reserved.
M. Teotia, R.K. Soni Engineering Failure Analysis 94 (2018) 412–437
reviewed extensively by Lawn in the year 2002 [4]. The results of contact studies on monolayer, bilayer, trilayer, and multilayer test
specimens that enable simple elucidation of fundamental damage mechanics and yet simulate essential function in a wide range of
practical structures were discussed in detail. Composite laminates involve a wide range of materials, underlayers and as well as fibres
used in their preparation and hence the prediction of their damage characteristics requires detailed knowledge of their mechanical
properties, orientation modes and corresponding materials characterization data. Most of the reported reviews detail damage
characteristics and impact behaviour of fibre reinforced composite laminates. The discussion of variety of composite laminates is
beyond the scope of our review. After extensive literature survey, we observed that no review is available on damage modelling of
laminated glass panels revealing the methods used for predictions of their damage characteristics under different types of impact
loadings.
Laminated glasses are sandwiched polymer composites, formed with brittle outerlayers and soft or compliant supporting inter-
layers which shield from potentially deleterious external forces. They have been used extensively in architectural and structural
applications [5–7], automotive windshields [8,9], safety and security endangered buildings [10], defence [11], railways, aerospace
applications [12], encapsulant materials for solar cell modules [13] to name a few. Because of their high strength and high modulus,
these glasses can be tailored for desired specific applications. The types of glass used in construction industry are annealed, heat
strengthened, toughened (fully tempered) and laminated. Annealed glass is cooled gradually from a high temperature to minimize
residual stresses during manufacturing process. However, heat strengthened and toughened glass undergo controlled heating and
cooling process to have 24-69 N/mm2 and over 69 N/mm2 residual compressive surface stresses, later involves high cooling rate.
Laminated glasses are commonly used as windshields and in building facades to resist high impact loads. The strength, breakage
performance and post failure behaviour of laminated glass depends upon glass type, thickness, dimensions and interlayer geometric
and material properties. The use of annealed glass and heat strengthened glass is suggested for fabrication of laminated panels [14].
Building industry is devoting full consideration to the performance of glazing materials in the event of explosive blasts and looks
forward to materials and glazing systems that can perform much better under such extreme loadings. Among all the type of safety
glasses, laminated glass is the preferred choice as it can resist high pressures and avoids the formation of large shards when glass
breaks. Laminated glass offers multifunctional benefits owing to its versatility and being used in a variety of innovative applications.
It possess huge architectural capabilities and physical barrier properties have contributed to make it an excellent safety glazing
providing solution to increasing demand in view of environmental threats and security. Remnikov et al., [15] experimentally in-
vestigated the load deformation response of these blast resistant façade glazing solutions adequately replicating the real world
situations. The authors concluded that laminated glass exhibits superior prefracture performance and higher peak resistance and also
favourable post fracture behaviour. Impact damage of laminated materials is a matter of great concern as it affects the mechanical
properties severely. Foraboschi [16] presented a criterion for fail safe design of glass plates loaded transversally. Three conditions
have been putforth; 1) any piece of glass that has a certain amount of redundancy, glass has two types of cracks, initial (during
manufacturing process) and contact induced cracks (due to live loads), the former is substantial however later is marginal and hence
can be neglected while assessing the load carrying capacity, 2) in the event of breaking no sharp shard should be fall to avoid injury
and 3) the glass must have postfracture load carrying capacity, means if one glass fails the structure must remain standing until total
failure. Laminated glass is the only product able to satisfy second and third fail safe condition. Specific glass types are recommended
to obtain a failsafe laminated glass system. A sacrificial ply (ply which is allowed to fracture in service due to live load) has to be
made of toughened glass as it can tolerate high local tension stress and also glass disintegrates into small pieces when it breaks.
Laminated glasses can be made by using different type of glasses and no manufacturing restrictions are involved. These are used as
primary (floor & roofs), secondary (staircases, overhead glazing supported by primary structures) or as non-bearing elements (facades
& partitions). Simulation of damage behaviour is not only important to structural glasses (load bearing) but is also important for non
structural (architectural/non-load bearing) as the later such as windows, glass façade panels, overhead glazings, skylights offer
structural defense to the building envelope in the event of naturally occurring phenomena such as earthquakes, windstorms,
snowstorms alongwith live loads.
Compression, bending, shear and residual strength of impact damaged specimens can be determined experimentally but the
assessment of all possible combinations of laminate construction experimentally would be expensive and time consuming. Despite of
profound advantages and applications of laminated glasses, still their use is limited because of the problems in strength calculations
at the stage of their design for a particular application. Hence, modelling of damage behaviour of laminated glasses has become an
essential pre requisite to stimulate an effective developmental process in context of their expanding use in aerospace and other
important industries. This provides effective working guidelines for predetermining optimum material combinations in applications
for desirable end use. Impact behaviour of laminated glasses can be modelled analytically or numerically. Analytical methods gives
exact solutions however numerical simulations give approximate solutions. The former may be useful for numerical solutions of the
formulated problem and later may be helpful in finding out solutions to differential equations that cannot be solved exactly. A lot of
research has been devoted to analytical modelling of glass using theoretical approaches such as classical plate theory or Von Karman's
large deflection theory. Foraboschi [17] put great efforts in developing analytical models for three layered plates and reported
analytical models to understand mechanical behaviour of three layered plates using a system of three exact and explicit equations
subjected to lateral static loading using Kirchoff's-Love plate assumptions. The models provides exact results and can be used to
validate numerical models and offers scientific understanding of deformation process and mechanical failure of TLPs and their
underlying physical mechanisms [18]. The approximate results obtained through FE models suffer inaccuracy due to high glass to
interlayer stiffness ratio. Hence validation of FE predictions either experimentally or analytically is a prominent requisite in order to
obtain accurate predictions. In continuation of this work, the author proposed an easy to use analytical closed form solution for three
layered plates which can be used to calibrate FE models in a straightforward manner [19]. Previous models were suitable to simulate
413
utterly compliant interlayers only as elastic modulus of interlayer was considered nil. The model can be applied to any interlayer as
both shear and elastic modulus were considered. It has been suggested as a benchmark for FE validations because zero elasticity
modulus cannot be assigned with actual shear modulus in FE simulations. A large amount of data is presented in [19] allowing the
engineers to choose optimal combinations of thickness of glass, type of glass, span sizes for particular loadings in terms of maximum
tensile stress in central and lower glass layer and also maximum deflection of the plates. The extremely low toughness, high elastic
modulus, utterly elastic constitutive law, low thermal conductivity and a high linear thermal expansion coefficient result in low
ability of glass to resist the propagation of flaw and hence more prone to thermal cracks and thermal shocks. Considering the intensity
of the problem, modelling/prediction of mechanical effects of temperature such as thermal stresses, thermal cracks and thermal
shocks becomes more demanding in order to present design guides to the architects and structural engineers.
In a recent work, Foraboschi [20] used analytical modelling to predict behaviour of laminated glass under thermal stresses
induced by non-uniform temperatures. The models are capable to predict most suitable glass types and glass systems without relying
upon experimental results which are far more difficult to perform for thermal loadings. The author emphasized the importance of FE
models and recommended them as a viable tool to represent the laminated sandwiches (glass plates, beams, panels) which do not
comply with Kirchoff's-Love plate assumptions and also to understand the behaviour of laminated glasses under dynamic loadings
and cracking mechanism. However the exact results obtained through analytical models are used to validate and calibrate the
parameters of FE models. Finite element method is a universally accepted method in many industrial sectors, which is helpful in
designing of structural laminates, saves time consumed in large experimental set ups and hence economic in prediction of damage
characteristics of investigated substrates [21,22]. This method has been widely employed for numerical simulation of various
polymer composite laminates such as fiber reinforced panels [23], carbon fiber reinforced polymer laminates [24], woven fabric
composites [25], metal beams [26], silicon sealant joints with glass [27], reinforced concrete structures [28]. With the advancement
in the capabilities of FE packages and advent of advanced techniques such as element deletion, XFEM etc., over the years, the method
can be efficiently applied to failure analysis of polymer composites.
FE modelling occupies a large section in the piechart of research on modelling of laminated glass and no state of art review is
available till date presenting an organized and uptodate information of applications of FE modelling in context of laminated glasses.
Detailed practice guides with huge information on using FE packages, exemplifying the key points are available with extending
possibilities to more accurate simulated predictions [29]. The purpose of the study is to present applicability of FE models in the
immediate field of failure analysis of laminated glasses. In this article, we limit our discussion to applications of Finite Element
Modelling (FEM) assisting the prediction of the damage characteristics of laminated glasses with polymeric supporting interlayers. A
first attempt has been made to summarize the extensive research conducted over the years in order to predict the damage behaviour
of laminated glass composites. First section prefaces the subject matter to the readers and insights the necessity of modelling of
damage behaviour of laminated glasses. A detailed account of theoretical approaches is discussed in section third in order to un-
derstand the suitability of a particular theoretical concept to the problem under investigation. Fundamental concepts of FE modelling,
methods being used in simulation of failure process and material models used for glass and interlayer with their merits and demerits
have been detailed in section four. This section also gives a systematic growth in finite element simulation of human head impact on
windshields and blast loading simulations.
2. Brief overview of modelling of laminated glass
Considerable attention has been paid to investigate the performance of laminated glasses in 1980s and 1990s using experimental
and theoretical approaches either alone or as combined approaches. Literature reports numerous experimental and theoretical in-
vestigations presenting analysis of damage behaviour of laminated glasses upon impact. The methodical growth in modelling of
laminated glass has been summarized in Tables 1, 2 and 3, highlighting the progressive research conducted over the years. PVB
laminated glass is the most investigated laminated glass for static and dynamic loadings under various boundary conditions and
support systems [30–32]. The pioneer work by Vallabhan in the year 1983 [33], reports iterative analysis of rectangular glass plates
using Von Karman's plate theory and then efforts of subsequent authors established the importance and utility of developed models in
prediction of damage behaviour of laminated glass. Laminated plates and beams have been inspected extensively for impact and
fracture strength, effect of thickness, effect of temperature, viscoelastic properties of interlayer, adhesion, human head impact, blast
performance, crack initiation, propagation and their mechanism using theoretical and numerical models and validated through
experimental findings. The use of FE modelling in damage predictions of laminated glass has gained momentum during the last
decade as evident from Table 3. Xu et al., [70–72,75,102,104] have investigated extensively laminated windshields for crack pro-
pagation, cracking mechanism and multicrack problem using high speed photography alongwith XFEM. Hidallana-Gamage and his
research group used 3D FEM modelling to describe the behavior of laminated glass panels under blast load accounting structural
supports also alongwith glass and interlayer [27,88,91,92,94,103]. 3D solid and layered shell finite elements have been used by the
researchers with different numerical techniques such as element erosion/deletion, smeared, cohesive zone (extrinsic & intrinsic)
entailing detailed precrack, impact fracture process and post breakage behavior of laminated glasses. It is a fact that numerical
findings have to be validated with experimental or exact results to avoid erroneous predictions, despite this limitation, FEM has been
a method of choice for the research groups providing valuable data with careful choice of numerical techniques, material models and
specifications for the specimen under consideration.
414
Table 1
Progressive research on complex nonlinear behaviour of laminated glasses (1983–2000).
Author name Specimen type Research Analysis Approach used
Vallabhan [33] Laminated plates Iterative analysis of rectangular glass plates; Theoretical; Von Karman's plate theory
Finite difference method was used to obtain
equilibrium equations
Behr et al. [34] PVB Laminated Laminated glass units under uniform lateral Experimental
plates pressure were compared with monolithic and
layered glass units under different temperatures
Behr & Linden [35] Laminated plates Load duration and interlayer thickness effects on Experimental
laminated glass.
Vallabhan et al. [36] Laminated plates Behavior of monolithic and layered systems was Theoretical; Theoretical stress analyses of thin glass plates
related by a “strength factor”
Das et al. [37] Monolithic glass Effect of fluctuating wind pressure Theoretical; Von Karman's non linear equations using
plate dynamic finite difference technique
Behr & Norville [38] Laminated plates Effect of temperature and load duration Experimental & Theoretical (Composite plate theory)
Vallabhan, et al. [39] Laminated plates Developed a mathematical model and verified Theoretical; Variational principle
experimentally
Edel [40] Laminated beams The effect of temperature on the bending of Experimental
laminated glass units.
Vallabhan, et al. [41] Glass plates Mathematical model is developed for four sided Theoretical
support system of structural glazings
Asık & Vallabhan [42] Glass plates Non linear behaviour of thin glass plates with Theoretical; von Karman equations using Airy stress
elastic supports function & lateral displacement and complete
displacement equations using lateral and in-plane
displacements
Flocker & Dharani Laminated Low velocity missile impact; effect of thickness & FEM; modelled fractures by using traditional wave
[43–46] architectural glass viscoelastic properties propagation codes through FEM
Dharani & Ji [47] Laminated glass Simulation of dynamic response of a human head Numerical method (a dynamic and non-linear FEM
plates to the windshield method)
Ji et al. [48] Laminated glass Missile impact was studied; damage factor was Cumulative damage theory; unknown constants in the
plates calculated through computed stress response damage model were determined experimentally
based on Weibull's probability distribution
Norville et al. [49] Laminated beams Development of a theoretical model accounting Theoretical; Engineering mechanics model
effect of temperature, thickness & composition of
interlayer, Fracture strength data for laminated
glass
Duser et al. [50] PVB laminated Investigated stress development and first FEM; 3D finite element model based on a visco-elastic
plates cracking in architectural laminated plates constitutive model for PVB deformation
Bennison et al. [51] PVB laminated Studied first cracking in commercial-scale 3D FEM
plates architectural laminated glass plates
Jagota et al. and Rahul Laminated plates Determined decohesion along glass-interlayer Interfaces were represented through Cohesive zone
kumar et al. [52, interfaces using compressive shear strength test models;experiments were also conducted
53]
Muralidhar et al. [54] Laminated plates Studied adhesion using through cracked tension Experimentally & theoretically
test, Developed a bridging model considering
interlayer as a hyperelastic material
3. Theoretical approaches for FE modelling
Numerical models are generally based on theoretical concepts [105,106] and the simplicity of modelling theories reduces the
model size and their demand on the computational resources. Owing to the complex behaviour of laminated glasses due to soft inner
plies, the three dimensional case of laminated glass can be modelled as two dimensional by using Equivalent single layer theories
with some assumptions. Numerical modelling was performed by researchers through coupling of different theories in order to arrive
at the accurate predictions depending upon the damage problem being investigated. Layerwise theories are also being used by the
research groups in order to make more accurate predictions. Zig Zag models are the emerging theoretical approaches in these
investigations. The present section reviews in brief the evolution of various modelling theories from very basic models to complex
models and their updates in context of laminated plates.
3.1. Classical plate theory (CLPT)
This theory is the oldest and was developed around 1850 [107], a simplest ESL modelling theory which neglects shear effects and
is applicable to any planar structural element having thickness to width ratio less than 1/10 that behaves as a plate [108]. Classical
plate theory, also known as Kirchhoff–Love theory of plates helps to represent the three-dimensional plate in two-dimensional form,
with the principal assumption that there is no change in thickness of the plate during a deformation i.e., no transverse shear or no
deflection and hence an overestimation of natural frequencies. The straight lines drawn normal to the mid surface remain straight and
normal to the midsurface after deformation, which implies that transverse normal strain and transverse shear strains are zero i.e.,
415
Table 2
Shesadri et al. [55] Cracked glass and Developed an analytical bridging model to predict the Theoretical, experimental and numerical
polymer plate compliant behavior of a cracked laminated plate for simulation (bridging/interface finite element)
adhesive strength, thickness and elastic properties of the
elastomer layer.
Dharani et al. [56] Laminated Calculation of damage probability of laminated glass plate a two-parameter Weibull distribution was used
architectural glass subjected to blast loading
panels
Asik [57] LG plate with PVB Complex stress fields and their propagation Variational approach, Central finite difference
interlayer method with iterative technique
Du Bois et al. [58] Laminated glass with Simulated the impact of sphere into a glass plate; Finite element simulation
PVB interlayer developed a laminated glass model consisting a shell
element for glass and membrane element for interlayer
Dharani & Wei [59] Laminated Effect of negative phase of blast loading on the dynamic Theoretical; CLPT, closed form solutions
architectural glass responses of laminated glass panel subjected to blast
panels loading
Dharani & Wei [60] Laminated glass plate A failure criterion for laminated glass panel subjected to Theoretical; Von Karman's large deflection with
blast loading was introduced, Development of a damage trigonometric functions,
factor for predicting failure of Laminated glass Griffith's energy balance approach
Asik & Tezcan [61] Laminated glass beams A mathematical model was developed to predict behaviour Theoretical; closed form solutions, minimum
of laminated glass beams under large deflections, Model potential energy approach & variational principle
was validated through experiments and Finite element
models
Wei et al., 2006 [62] Monolithic and Cumulative failure probability of monolithic and 3-D nonlinear dynamic finite element model is
laminated glass laminated architectural glazing subjected to blast loading used, cumulative damage theory based two
was investigated parameter weibull distribution was adopted
Ivanov [63] PVB laminated glass Development of mathematical model for triplex glass Linear Finite elements, Euler–Bernoulli beam
beam beam, mechanics of laminated glass beams, validation of theory
model through FEM
Wei & Dharani [64] Architectural Simulated blast loading, Investigated the significance of Theoretical; CLPT, Von Karman's theory and
laminated plates negative loading phase on dynamic response of laminate dynamic nonlinear three-dimensional finite
under blast loading element code LS-DYNA 3D
Wei et al. [65] Architectural Numerical simulation of blast loading from a bomb 3-D nonlinear dynamic finite element modelling
laminated plates explosion, Effect of viscoelastic parameters of PVB on
dynamic response of Laminated glass was investigated
Lee et al. [66] Glass/epoxy/Glass Dominance of crack reinitiation mode of failure was Experimental; Vicker's indentor was used for starter
laminate studied, critical strains for failure accounting geometrical crack, Specimens were subjected to Flexure loading
and material parameters were quantified
Timmel et al. [67] PVB laminated glass Coincident finite elements are used to model laminated Finite element modelling; Newton's equation was
plates glass; shell elements for glass and membrane elements for solved via a central difference method
PVB interlayer
Koustsawa & Daya Laminated beams An analytical solution is derived for optimization of Variational principle,
[68] laminated glass beam with viscoelastic supports, FEM Linear FEM
method was used to model the modal properties of the
glass
Belis et al. [69] Laminated beams Post breakage behaviour of laminates composed with SGP Experimental; destructive inplane four-points
interlayer, failure mechanism and residual capacity was bending tests at room temperature were performed
investigated experimentally
Xu et al. [70] Laminated windshield A model of vehicle impact speed versus deflection of Theoretical; physical, material and dynamic
windshield glazing is developed models were developed and combined
Xu et al. [71] Laminated windshield Analysis of low-speed head impact on a windshield plate Numerical simulation; Extended finite elements
with characterization of radial crack and circumferential
crack propagations
ϵzz = 0 and γxz = 0/γyz = 0. The Z- expansions for displacements are assumed as,
Uα(x,z,t) = u α(x,t) − Z∂α w(x, t) (1)
Uz(x,z,t) = w(x,t) , (2)

where uα and wα are inplane and transverse displacement components of a point located on reference plane (Z = 0) and subscript
α = 1,2. Zero transverse strains also neglects transverse stresses (σxz, σyz, σzz, σyy, σxy, σzy). This theoretical model can be represented
by the Eq. (3), where deflection is only a function of x:
∂ 2w ∂ 4w
ρ + k 4 = q(x, t)
∂t 2 ∂x (3)
CLPT has been widely used in the development of numerous theoretical, mathematical and numerical models for laminated
glasses (thin plates) [59,61,64,109]. For the laminates where larger magnitudes of deformation are observed, the middle surface of
416
Table 3
Xu et al. [72] Laminated windshield A constitutive model describing both quasi- Experimental; one dimensional quasi static
static and dynamic response based on Johnson- and dynamic compression experiment at
Cook model is established, dynamic stress different loading rates
intensity factor was calculated (β ≈ 1.5)
Kumar & Shukla [73] Monolithic glass, wired glass, Comparative analysis of five different panels Experimental; The high speed photography
tempered glass, Laminated glass, & under blast loading and 3D digital image correlation (DIC)
Laminated glass with film (XO technique was used
ARMOR®) on both the faces
Pyttel et al. [74] Laminated windshield & Laminated Rankine criterion was enhanced by a non local FEM; A failure criterion was developed &
glass approach, the influence of curvature and implemented in PAM crash and validated
boundary conditions was investigated experimentally
Xu et al. [75] Laminated windshield Dynamic crack patterns and crack propagation Experimental; High speed photography was
behaviour were investigated used
Foraboschi [17] Laminated glass plates Analytical model was developed, mechanical Analytical, based on Kirchoff's-Love plate
behaviour was investigated when subjected to assumptions; system of three exact & explicit
lateral loading equations were used
Larcher et al. [76] Laminated glass plates Comparative analysis of numerical models FEM; Smeared, layered shell and 3D solid
under the effect of an air blast models
Galuppi & Royer- Laminated glass beams Time dependent problem of a simply supported Analytical; viscoelastic response of interlayer
Carfagni [77] beam under constant loading is modelled using Maxwell-Wiechert model
Schulze et al. [78] Laminated glass beams Robust relationships between the maximum Theoretical; FSDT & a layerwise type beam
deflection, the transverse shear strain of the theory are applied, 3D FEM analysis is
core layer and the applied force in a three- performed to verify the applied beam theories,
point-bending test of laminated glass beam experimental; three-point-bending tests are
samples are derived performed.
Hooper et al. [79] PVB laminated windows A FEM model has been developed to predict FEM approach, High speed image correlation
the response of laminated glass under blast
loading in the precrack and postcrack phases
Ge et al. [80] Architectural glass panels A semi-theoretical method is developed to Analytical; based on energy conservation
predict the fragment projection velocity and principles
flight trajectory caused by blast loading
Zhang et al. [81] PVB laminated glass windows A 3D numerical model with dynamic material FEM; solid elements, erosion technique is used
properties was developed and pressure- to model glass cracking and interlayer failure
impulse diagrams were generated
Zhang et al., 2012 [82] Annealed float glass Dynamic material properties of annealed float Experimental; High speed photography&
glass, empirical formulas of dynamic modified SHPB tests
increment factor were derived
Zhang et al. [83] PVB laminated windows FEM model was developed to estimate glass Experimental, FEM with erosion technique
fragments, their mass and launching speed
with respect to debris impact mass and
velocity, crack initiation and propagation were
simulated
Galuppi & Royer- Laminated glass beams Time-dependent problem of a simply- Analytical; Maxwell-Wiechert model is used
Carfagni [84] supported three-layered sandwich-beam with
linear-viscoelastic interlayer under a loading/
unloading history
Serafinavicius et al. [85] Laminated glass plates Long term behaviour of structural glass plates Experimental; Four point bending tests
with different interlayer materials and at
different temperatures were made.
Peng et al. [86] Laminated windshield Comparative analysis of different patterns of FEM; layered shell and smeared models
FEM models to investigate mechanical
behaviour under pedestrian head impact
Zhang et al. [87] PVB laminated windows A detailed FEM model has been developed, FEM; dyanamic material models are used
effect of interlayer thickness, glass thickness,
boundary condition, window diamensions
have been analysed, empirical formulas are
derived using numerical results for
construction of P-I diagrams
Hidallana-Gamage et al. Laminated glass panels Comparative analysis of 2D and 3D FE FEM; 2D and 3D approaches with dynamic
[88] modelling approaches, validation with CLPT & material models
Von Karman's large deflection plate theoretical
models and experiments
Chen et al. [89] PVB laminated glass A simple cracking initiation law considering in- Experimental; high-speed photography system
plane cracking and the inter-plate cracking and drop-weight platform tests
process is elucidated
Gao et al. [90] Laminated glass beams A fracture model for 3D discrete element is Spherical discrete element method based on
developed to analyse fracture process cohesive model
Laminated glass panels
(continued on next page)
417
Table 3 (continued)
Hidallana-Gamage et al. FE model is developed to predict the failure of FEM; Glass, interlayer and sealant joints are
[27] different components in LG panels under blast modelled using 3D solid elements
load
Hidallana-Gamage et al. Laminated glass panels Post crack behaviour and effect of structural FEM; 3D constant stress solid elements
[91] sealants under blast loading is simulated
Hidallana-Gamage et al. Laminated glass panels Influence of structural sealant joints on FEM; 3D model including glass, interlayer and
[92] laminated glass performance under blast sealant joints
loading
Foraboschi [93] Monolithic glass An analytical model is developed accounting Analytical & Experimental ( Compression and
combined axial force and bending moment to three point bending tests)
predict load bearing capacity
Hidallana-Gamage et al. Laminated glass panels Effect of thickness & Young's modulus of FEM; Glass, interlayer and sealant joints are
[94] interlayer on the blast performance modelled using 3D solid elements
Chen et al. [95] Laminated glass plates A three dimensional computational model is FEM; Extrinsic cohesive zone modelling
developed, cohesive model is coupled with
frictional contact algorithm accounting
unilateral and frictional sliding effects
Jaskowiec [96] Laminated glass plates A three dimensional numerical model for Mathematical; XFEM is used to model PVB
delamination analysis is developed interlayer
Baraldi et al. [97] Laminated glass beams Prediction of load carrying capacity and Nonlinear Discrete Element modelling &
stiffness of broken laminated glass experimental (three point bending tests)
Pelfrene et al. [98] PVB laminated windows Post breakage behaviour, characterization of FEM; element deletion technique
dynamic behaviour of interlayer and
delamination under blast loading
Chen et al. [99] Laminated glass beams Mechanism of glass ply cracking under low FEM; cohesive zone modelling, brick elements
velocity impact is investigated, a laminated are used, intrinsic cohesive model for
glass model using brick elements is developed, adhesion
adhesion between glass and PVB is modelled
Biolzi et al. [100] Laminated glass beams (ionoplast Bending response under static and dynamic Experimental; static (three point bending test)
interlayer) inplane loadings was investigated & dynamic (piezoelectric accelerometer)
Gao et al. [101] Laminated glass plates Impact fracture behaviour under low velocity FEM; bilinear, intrinsic cohesive model
impact is studied, parametric investigations
were conducted
Xu et al. [102] Laminated glass plates Dynamic radial multicrack propagation Experimental; high speed photography, FEM;
behaviour was investigated to explore cracking XFEM is used to study multicrack problem
propagation mechanism
Hidallana-Gamage et al. Laminated glass panels Effect of negative phase of blast loading and FEM; Comprehensive FE model accounting for
[103] support flexibility on blast performance window frame, fasteners and flexible supports
Xu et al. [104] SGP laminated glass Effects of interface location and impact kinetic Experimental; optical acoustic method is used
energy upon the fracture morphology, crack
propagation velocity and the dynamic stress
intensity factor
the plate begins to stretch and geometrical nonlinearities becomes significant, CLPT fails to predict the accurate mechanical beha-
viour of the laminates [62]. When the rotations of the normal to the midsurface are in the range 10–15°, Von Karman's strains are
used to approximate the strain displacement relations. Von Karman's theory takes into account the effects of both bending and
stretching of the middle span of the plate based on the boundary conditions and utilizing the expansion theorem. The following
equations are nonlinear partial of differential equations and considered as large deflection theory of plates [110].
∂ 2ω ∂ 2∅ ∂ 2ω ∂ 2∅ ∂ 2ω ∂ 2∅ ∂ 2ω ⎞
ρ 2
+ D∇4 ω(x, t) = q(x, t) + h ⎜⎛ 2 2
+ 2 2
−2 ⎟
∂t ⎝ ∂y ∂x ∂x ∂y ∂x∂y ∂x∂y ⎠ (4)
2
∂ 4∅ ∂ 4∅ ∂ 4∅ ⎛ ∂ 2ω ⎞ ∂ 2ω ∂ 2ω ⎞
+2 2 2 + = ∇4 ∅ = E ⎜ ⎜⎛ −
∂x2 ∂y 2 ⎟
⎟
∂x 4 ∂x ∂y ∂y 4 ∂x∂y
⎝⎝ ⎠ ⎠ (5)
where, ∇ is biharmonic operator, ω is deflection of plate, ∅ is stress function, E is youngs modulus, h is plate thickness, D is flexural
4
rigidity, q is applied pressure and t is used for transverse direction. Vallabhan [33] used this geometrically nonlinear theory
(moderately large deflections) for thin plates to model the laminated plates under stress by obtaining a set of nonlinear partial
differential equations describing the large deflections of thin flat plates at discrete points using finite difference method (FDM), and
an iterative technique. In 2003, Asik et al. [57] used five coupled nonlinear partial differential equations obtained through the
variational approach for lateral and in-plane displacements of laminated glass plate and used iterative technique to predict the
displacements. Further, Wei and Dharani [60], expressed equations of motion for blast loading of laminated glass using Von Karman's
large deflection theory through Galerkin & Runge-Kutta method and developed a damage factor (value less than 1; plate is safe) to
predict the behaviour of laminates. The authors also predicted the size of glass shards from laminated glass panel upon impact
418
through the surface energy based failure model. Classical plate theory in combination with Von Karman's assumption has also been
used to understand the material behaviour of windshields by considering them as a composite plate [70,111,112]. In the last decade,
CLPT has been used in the development of layerwise theories for the modelling of outer faces of the laminates [113–115]. Extensive
work has been conducted by Foraboschi for the analysis of behaviour of laminated plates. The author used Kirchoff's Love plate
assumptions for analysis of outer glass layers however the shear layer has been modelled through other shear deformation theories
[17,18]. CLPT has its own limitations but suits best for predicting the behaviour of thin plated structures, having lesser number of
variables and equations, thus saving computational efforts. However for getting better predictions, other higher order ESL theories
are used (Fig. 1).
3.2. First order shear deformation theory (FSDT)
This theory also known as the Mindlin–Reissner theory of plates which considers shear deformations through the thickness of the
plates and can be applied to plate thicknesses of the order of 1/10th of the planar dimensions [105,116]. It is the extension of
Kirchhoff's love plate theory which incorporates first order shear effects implying a linear displacement variation through the la-
minated plate thickness, where the displacement field can be represented as;
Uα(x,z,t) = u α(x,t) + Z∅α(x,t) (6)
Uz(x,z,t) = w(x,t) (7)

where ∅α denotes rotation of the normal about the y axis. A shear correction factor/coefficient is used for computing transverse shear
force resultant which corrects the predicted transverse shear stress to the true transverse shear stress [117,118]. This theory assumes
that the transverse straight lines remains straight but are not normal to the midplane after deformation and the transverse stress is
constant through the thickness (Fig. 2). FSDT is widely used structural mechanics model for sandwich structures [119–125]. It is
generally applied to the laminated systems with soft cores in combination with a homogenisation approach. The governing differ-
ential equations can be solved in closed analytical forms for laminated plates with different shapes and boundary conditions. Re-
searchers have focussed on the finding solutions to the limitations of FSDT so that the complexity of the equations and hence
numerical modelling may be lessened [126–128]. Onsy et al. [129] presented a FSDT based model in which the shear strains across
the interfaces of adjacent layers were discontinuous. The model was further used for the development of two node and three node
finite strip elements.
The prediction of suitable shear correction coefficient is quite complicated and cumbersome [130]. In an attempt, Eisentrager
et al. [131], applied FSDT to laminated glasses and photovoltaic panels and evaluated its applicability range. The authors introduced
a shear rigidity parameter β, and concluded that for β > 10, the theory provides accurate results however for β < 10 (with very soft
interlayers) deflection was overestimated. Most of the finite element models are based on FSDT [132–134] but if there is large
difference between the shear stiffness then FSDT may not predict the deformation behaviour so accurately. Cugnoni et al. [135] have
modelled composite structures on the basis of FSDT alongwith higher order shear deformation theories. The theory is very ad-
vantageous in estimating the properties related to transverse shear deformation of the layered systems. Recent focus is on the use of
FSDT theory in layerwise theories for the modelling of individual layers, which assists in the formulation of less complex numerical
models. However for more accurate transverse shear calculations and analysis of complex thick laminates, with different boundary
conditions other higher order theories are in practice.
3.3. Higher order shear deformation theory (HSDT)
These theories take into account through the thickness distribution of transverse stress to correct the discrepancy in the shear
forces of FSDT theories [137–141]. Nonlinear distribution of transverse shear strains along the laminate thickness is approximated
properly which excludes the need of shear correction factors. High order ESL theories have been used extensively for the analysis of
static and dynamic responses of sandwich plates [142–143]. In HSDT models, the inplane displacement across the plate thickness is
expanded to polynomials or sinus expansion which only differs in the incorporation of shear deformation effects in the kinematic
formulations and equilibrium equations are obtained using the principle of virtual work giving more accurate transverse shear
stresses. High order shear deformation theories are able to predict the interlaminar stress distributions by quadratic variation of
transverse stress through each layer. These systems require C1-continuity of generalized displacements and derivation of second
derivative of transverse displacement becomes quite cumbersome when computational models are developed using the lower order
elements. Displacement fields for second order and third order shear deformation theories are given as;
Uα(x,z,t) = u α(x,t) + Z∅α(x,t) + Z2ψα (x,t) (Second order) (8)
Uz(x,z,t) = w(x,t) (9)
Uα(x,z,t) = u α(x,t) + Z∅α(x,t) + Z2ψα (x,t) + Z3λ α(x,t) (Third order) (10)
Uz(x,z,t) = w(x,t) (11)

1 0
To avoid complexities in numerical modelling due to C continuity, researchers have emphasized on the development of C type
higher order theories. A more refined form of HSDT was developed by Shankara and Iyengar [144] which required only C0-continuity
419
X
o Undef ormed
w (x,t)
CLPT
o
FSDT
o
HSDT
o
Fig. 1. Undeformed and deformed laminated plate in different equivalent single layer theoretical models [136].
of generalized displacements and hence simplified the computing of these systems. In another development Zhen and coworkers
[145–147] presented a C0 type global local higher order theory for sandwich plates which calculates the transverse shear stresses
accurately as obtained from C1 type higher order theories. The model is advantageous because the first derivatives of transverse
displacement have been taken out of the inplane displacements and only C0 interpolation function is required for FEM modelling. The
authors presented a finite element and analytical solution for the free vibrational analysis of sandwich plates. However, most of the
higher order models are very complicated algebrically and expensive numerically and require more effort in analysis and hence
should be used when necessary. In ESL theories, the interlaminar transverse stresses at the interfaces of layers of different stiffness are
obtained as discontinuous functions, hence can lead to errors in prediction of delamination and matrix cracking.
3.4. Layerwise shear deformation theory
Equivalent single layer models which we have discussed so far have been successfully and extensively applied to design multi-
layered structural components and predicts sufficiently accurate global response of thin to moderately thick laminates but fails when
laminates are thick, possess severe anisotropy, transverse shear modulus and inplane modulus ratio is low and longitudinal and
transverse Young's modulus ratio is high [137,148–150]. Layerwise theories are an alternative to ESL models which present full three
dimensional kinematics and constitutive relations [151]. These are a compromise between the three dimensional theories and
conventional plate theories which can predict the interlaminar stress very readily. A typical displacement assumption for layerwise
models can be represented by the following equation [152]:
Ukα (x,z,t) = u kα (x,t) + Z∅kα (x,t) + …Z nωkα (x,t) (12)
Ukα (x,z,t) = w(x,t) (13)
where the superscript ‘k’ means that the values are referred to the k layer. In layerwise theory, constitutive equations are presented
th
for individual layers and models based on constitutive assumptions for interaction forces and compatibility equations at the interface
to reduce the number of unknown variables are derived. This theory has been used extensively by numerous researchers in modelling
420
of laminated glass beams [63,68,78,153]. The close form solutions obtained through layerwise theories were found in good agree-
ment with the data obtained through three point bending experiments and as well as FEM analysis. Frostig et al. have worked
extensively for the development of layerwise theories using CLPT, FSDT and other higher order theories [154–156]. Detailed free
vibrational analysis of sandwich plates with soft inner cores has been presented in references [157–158].
Foraboschi [17,159] proposed layerwise theory for laminated plates and developed a system of three exact analytical equations
derived from Kirchoff's Love plate assumptions for the analysis of the behaviour of three layered plates. An exact mathematical model
was also developed for layered plates with discontinuous connection along the lines of classical plate theory [18]. Layerwise theories
provide easier equations with proper assumptions about stiffness/deformation of individual layers. Naumenko and Eremeyev [160]
have derived a layerwise plate theory for structural analysis of glass and photovoltaic modules. A system of four partial differential
equations has been proposed which is an extension to Kirchhoff plate and FSDT plate. Layerwise theories contribute to exact and
development of more refined FEM models. Eisentrager et al. in 2015 [161] developed a user defined finite element model based on an
improved layerwise theory which accounted shear deformation for all the layers in contrast to use of Kirchoff's plate theory for
modelling of skin layers. The main drawback of layerwise theories is that these theories require high computational time as the
number of variables depends upon the number of layers in the laminate.
3.5. Zig zag models
Zig zag theories has been emerging as more practical models to be used in engineering applications which overcomes the sig-
nificant flaws in previous theoretical foundations and their predictive capabilities. These theories are able to predict more accurate
behaviour of relatively thick laminated composite and sandwich structures. Zig zag theories with linear or high order local functions
were putforth by the researchers which proved to be more accurate and also reduced computational time and effort. Most of the
proposed Zig Zag models have been reviewed extensively by Carrerra [162–164]. The governing equations of 3D elasticity theory are
reduced to two diamensional equations using variational methods or asymptotic techniques. These theories represents the piecewise
continuous displacement in thickness direction of the sandwiched laminate with three displacement fields corresponding to each
layer [165]. Zig Zag effect, as explained in reference [166], is coined to the phenomenon when the displacement component show a
quick change in the thickness direction at the interfaces because of different tangential elastic compliance of the individual layers.
Hence the theories which justify the through the thickness piecewise behaviour of stress and displacement are called Zig Zag theories.
Zig Zag models use limited and fixed number of kinematic variables that reduces computational costs. The displacement assumptions
involve coarse kinematics as well as fine kinematics which represents the behaviour on total laminate thickness scale (ESL like) and
layer thickness scale (Zig-Zag function) respectively. Inclusion of Zig-Zag function in addition to ESL models results in constant
number of variables. The simplified Zig-Zag displacement fields can be represented as:
Ukα (x,z,t) = u α(x,t) + Z∅α(x,t) + …Z nωkα (x,t) + Πkα (z)Ψα(x,t) (14)
Ukα (x,z,t) = w(x,t) (15)

k
where the additional last term represents the zig-zag function (Πα (z)) and zig-zag amplitude (Ψα). The number of variables are
constant i.e., 2(n + 1)+3; 2(n + 1)+1from coarse kinematics & two zig-zag amplitudes and does not depend upon the number of
layers. Literature reveals that two types of zig-zag functions are used by the researchers for the development of zig-zag models, Di
Sciuva zig-zag function [167]; displacement based & Murakami zig-zag function [168]; Reissner mixed variational theorem based.
However, Di Sciuva's zig-zag function that involves the knowledge of mechanical properties of the layers was found more exact in
prediction of multilaminates behaviour [169–171].
The linear models having displacements as primary unknowns require C1 continuous function to approximate the deflection with
the finite element framework and predict erroneous results under clamped boundaries. These issues were resolved by researchers
subsequently by presenting more refined zig zag theories [172–178]. This research group have worked extensively on the devel-
opment of improved zig zag theories which are more efficient in the development of C0 continuous finite elements. They derived
novel zig zag functions without enforcing full transverse shear stress equilibrium along the interfaces of the laminate. The theory uses
FSDT with an improved zig zag kinematic field, independent of number of material layers and involves novel C0-continuous inplane
displacements.
The computational efforts have been reduced as compared to layerwise theories by choosing the appropriate degrees of freedom.
FEM models based on their work possess wide applicability range from moderately thick laminated composite and sandwich
structures exhibiting a high degree of transverse shear, flexibility, anisotropy, and heterogeneity. For soft core sandwich plates, third
order zig-zag theories have also been developed and used for the analysis of the nonlinear dynamic behaviour of laminated and
sandwich plates [179,180]. A cubic zig-zag theory accounting cubic distribution of inplane displacement and four order distribution
of transverse displacement was developed by Icardi [181–183]. The author further refined the assumed kinematics including higher
order terms to improve the prediction of local behaviour of laminates [184]. An improved cubic zig zag theory was proposed by
Pandit et al. [185] for the static analysis of soft inner core laminated sandwich plates. They assumed quadratic variation of transverse
displacement within the core and constant through the outer faces. Further, Kirchoff quadrilateral element was used to investigate the
free vibration sandwich plates based on third order zig zag theory [186]. In an another development a higher order theory was
putforth by adding zig zag functions to the existing theories and bending analysis of sandwich plates was performed [187]. Xiaohui
et al. [188,189] proposed a new zig zag theory and C0 plate bending element for composite and sandwich plates. The advantage of
this theory is that the first derivatives of transverse displacement have been taken out from the inplane displacement field that's why
421
only C0 interpolation is sufficient for its finite element implementation. More accurate transverse shear stresses were calculated by
integrating three diamensional equilibrium equations within one element. Six node triangular elements were used in FE modelling of
the proposed theory.
4. FE modelling of laminated glass
Damage predictions in a laminate involve determination of distribution of structural stresses which can be predicted with the help
of physical equations. But for complicated structures and distribution of stresses, these equations cannot be solved directly. Finite
element modelling offers solution to this problem by replacing the complicated structure to approximate simple closely related
shapes. A number of possible variations are involved in laminate construction, primarily the number and thickness of the constituent
layers can be varied with great effect on ballistic performance. Numerical simulations coupled with experiments have been reported
to provide a more effective way of studying the damage characteristics of laminated glasses [190–192]. Numerical modelling offers
easy and comprehensive study of all mechanical parameters and allow evaluation of large number of potential armour schemes
quickly and cheaply. It helps to understand how the laminate behaves during impact and how failure processes can be controlled to
our advantage. Numerical models allow simple design changes in the laminates and elimination of some casual and interferential
factors.
Flocker and Dharani [43,44,47] carried out pioneer work on numerical modelling of laminated glasses and developed dynamic,
asymmetric nonlinear finite element model to investigate the mechanical behaviour of laminated glasses used in automobiles.
Fractures upon low velocity missile impact were modelled using traditional wave propagation codes through FEM accounting the
effect of thickness & viscoelastic properties of interlayer [45,46]. Stress development and first cracking in Polyvinyl butyral (PVB)
laminated glass has been studied extensively using 3D finite element models [50,51].
In the same year, a numerical model was developed in the AUTODYN software to predict the response of sodalime glass/poly-
carbonate transparent laminates used as armour systems [193]. In order to support the development of a numerical modelling facility
for glass/polycarbonate laminates, a brief ballistic test program was undertaken with the aim of producing a small data set for
validation of numerical simulations. AUTODYN has a number of numerical processor techniques, the processors most suitable for
glass laminate impact modelling are the Lagrange and SPH processors [194,195]. Application of this numerical model has allowed
DERA (Defence Evaluation and Research Agency) to evaluate a large number of potential armour schemes and improve under-
standing of the interaction between laminate components under ballistic impact. DERA is an ongoing research program to investigate
improved light weight armour solutions.
In the year 2003, Du Bois et al. [58], simulated the crash behaviour of laminated safety glass using finite elements by considering
shell element and membrane element for glass and hyperelastic PVB-interlayer respectively neglecting the viscous nature of the
interlayer. The authors used “laminated glass model” in LS-DYNA, where piecewise linear plasticity was used for all layers and
investigated the impact of a rigid sphere onto the roof of windscreen. The behaviour of PVB was modelled by considering Blatz–Ko
and Mooney–Rivlin material laws (strains less than 100%), however a high order (six) Ogden material law was found suitable. The
model was further updated by Timmel et al. [67] in the year 2007 for impact simulations of laminated glasses. They putforth a
smeared model in order to obtain improved results for bending after fracture. Two coincident shell elements with same thickness
were considered and the membrane was modelled by bilinear plasticity law to smear the behaviour of interlayer and glass fragments.
The bending stiffness of the two elements with equivalent thickness and glass modulus was equivalent to bending stiffness of the PVB
laminated glazing. The authors also attempted modelling of glass failure by eroding one of the shell elements which helped to
compute the Young's modulus of the remaining shell element representing the interlayer and glass of compressive side of the
windscreen. This modification allows simulation of the acceleration of an impactor shot on a windscreen and the composite efficiency
of glass laminate. Koutsawa and Daya [68] used finite element method to determine modal properties of the glass in case of free
linear vibrations. The laminated beam was supported by a viscoelastic material at its ends and the complex eigen value problem (Eq.
(16)) was solved using MATLAB software to get the natural frequencies (ω) and associated loss factor (ƞ) through Eq. (17), where [K]
and [M] were element stiffness and mass matrices respectively. This study is helpful in designing of supports for laminated beams for
passive vibration control.
[K] − Ω2 [M]{q} = 0 (16)
Ω2 = ω2 (1 + jη) (17)
4.1. Fundamentals of FE modelling
FE models can be categorized into three types; 3D solid, layered shell and smeared model on the basis of type of finite elements
used in numerical simulation. These models are being used by the researchers to predict the complex damage behaviour of laminated
glasses alongwith different numerical techniques. The models have their own merits and shortcomings and are applicable to parti-
cular specimens with specified parameters. A comparative account of these models has been presented in Table 4 as a quick guide to
the readers. Layered shell is the most preferable model owing to less complexity, requiring less time and accounting failure of the
interlayer using erosion technique. Different approaches have been used by the researchers to model the adhesion between glass and
interlayer. Shared node algorithm, penalty based approach (tied connection) and intrinsic cohesive modelling are the widely used
approaches in interface modelling of laminated glasses. Shared node shows perfect bonding and there are no relative displacements
422
Table 4
Comparative account of 3D, layered shell and smeared models [76].
3D solid Layered shell Smeared
Elements 3D solid Shell Two superimposed shell elements

Implementation Easy Complex Complex
Material laws Linear elastic material laws Detailed material laws are needed Detailed material laws are needed
Strain distribution Nonlinear variation of inplane strain across Through the thickness, linear in plane Through the thickness, linear in plane
the laminate thickness can be obtained strain distribution can only be represented strain distribution can only be represented
Time required Requires large computational power and Requires less computational power and Requires less computational power and
hence more time hence less time hence less time
Failure of interlayer Failure of interlayer cannot be represented Failure of interlayer can be represented Failure of interlayer cannot be represented
Failure model Used to determine stresses in glass only, not Element erosion technique is used Behaviour of panel with at least one not
used in combination with a failure model broken glass ply can be represented
Applicability Require large computational power, hence Can be used efficiently for design studies Recommended for small displacements of
not recommended including failure of interlayer panels only
between glass and interlayer. Therefore, shared node algorithm is unable to describe flawed bonding between glass and interlayer in
practical situations and also crack propagation in top layer glass is hindered. Penalty based approach resist the relative displacements
between glass and interlayer by imposing penalty forces, hence different adhesion strength can be modelled by simply changing the
value of penalty stiffness. The approach can be used to model debonding in cohesive zone formulations. Chen et al., in 2016 studied
impact damage in laminated glasses and used intrinsic cohesive formulation for modelling of adhesion and compared with shared
node method [99]. The authors found it more appropriate than shared node method for modelling of adhesion. When crack initiates
in the glass pane, its evolution, propagation and the underlying mechanisms are the main features to capture the fracture process of
the laminated glass. A lot of research has been devoted to simulate the cracking process using numerical techniques such as element
erosion/deletion, cohesive zone and extended finite element to arrive at more accurate predictions of the fracture process.
4.1.1. Simulation of failure process

FE simulation of cracks initiation and propagation and their underlying mechanisms has been a challenging task, which en-
couraged the concerned researchers worldwide to find suitable methods accounting more accurate predictions. Various numerical
techniques, such as element erosion/deletion, cohesive zone modelling and extended finite element method (XFEM) have been
investigated for understanding the mode of cracking and the involved mechanisms.
4.1.1.1. Element erosion/deletion method. Element erosion method is a numerical manipulation which involves eroding of elements to
avoid element tangling. Elements are removed when reached the critical plastic strain set by the user within a defined material law.
When an element is removed, its mass is removed from the global mass matrix, hence violating the laws of mass and energy
conservations. Actually, element is not deleted, but the stress in the element is set to zero. Despite this fact, good results have been
observed by the researchers using erosion technique in FE models [83]. It assists in simulation of initiation and propagation of
cracking with careful selection of erosion criterion and material models in order to obtain reliable estimations. Various failure
criterion such as Rankine criterion [74], Von Mises yield criterion [52,61] strain rate dependent criterion [76] on strain energy have
been used to characterize the onset of the fracture. To the advantage, this is an efficient technique as it is directly applicable to initial
meshes and does not require any change or specified modelling techniques. Hence, the number of equations remains same
eliminating the need of additional computational efforts. Scaling of constitutive equation is very crucial to reflect element size so that
the released energy (fracture) is independent of size. The damage evolution law is given as equation [196]
∈0 ⎛ ∈1
D(∈∗) = 1 − − 1⎞
∈1 − ∈0 ⎝ ∈∗ ⎠ (18)
∗
where ∈0 and ∈1 are strain at peak stress and complete softening stage parameters respectively and ∈ is the effective strain. This
stress-strain law helps in equating the energy dissipation in an element to the surface energy of a developing crack passing through
the element parallel to the sides. Element erosion technique has been used to model glass cracking and interlayer failure of
rectangular laminated plates, laminated glass window response under blast loading [87] and impact of windborne debris [83]. As
further advancement, the mass of the deleted elements was retained in the simulation in order to maintain mass conservation. Larcher
et al., have compared layered elements with a special failure criterion with 3D and smeared model and found it more suitable for
design studies including failure of interlayer [76]. Further, Pelfrene et al., 2016, used element erosion technique to critically assess
the post breakage performance of blast loaded laminated glazings [98].
4.1.1.2. Cohesive zone method. In this method, fracture is represented as a global phenomenon which involves separation across a
cohesive zone or separation in the small region ahead of crack from the crack tip. This method involves addition of extra nodes and
the crack is modelled through separation along the edges of element. When the forces at the cohesive zone overcome cohesive
traction, crack initiates and propagates. This process is governed by phenomenological traction separation law, characterized by
cohesive strength. Maximum crack length is associated with zero bond strength and a maximum opening normal stress is associated
with an initial crack opening in a cohesive element. CZ models are usually used with explicit solvers as the implicit solvers are not
423
that much efficient in dealing with discontinuous functions. On the basis of FEM implementation, cohesive zone models can be
categorised as intrinsic (initially elastic) and extrinsic (initially rigid) models. Cohesive elements are inserted at all inter-element
boundaries before the beginning of simulation in intrinsic cohesive models. On the other hand, extrinsic cohesive models involves
adaptive insertion of cohesive elements during calculations. As intrinsic cohesive models suffer from artificial compliance, Chen
et al., [95] suggested extrinsic cohesive models for simulation of impact fracture process of laminated windshields. The classical
Camacho-Ortiz model [197]) is extrinsic and dissipates constant energy regardless of the opening path of a growing crack as it is
potential based, hence traction vector T can be obtained by differentiating free energy density function with respect to opening
displacement vector.
∂ψ
T=
∂∆ (19)
According to this model, elements are allowed to separate when the following criterion is met;
σ eff = σ 2n + β−2σ 2t ≥ σfr (20)

where σn, σt are normal and tangential stresses respectively and σfr is the material fracture strength. Snozzi et al., in the year 2013
modified Camacho-Ortiz model by introducing an independent parameter defining the ratio between the energy release rate of mode
I and II two dimensionally [198].
ψt
k=
ψn (21)
where ψt is tangential fracture energy and ψn is normal fracture energy. Further, Chen et al., [95] extend this model to three
dimensions to analyse the impact fracture behaviour of automotive laminated glass. With the help of this parameter, traction se-
paration law was modified and the effective opening displacement can be written as;
η2 2
Δeff = Δt + (Δn )2
k2 (22)
where Δt = + Δn is normal displacement component, Δt1 and Δt2 are tangential displacement components and η represents
Δ2t1 Δ2t2 ,
the mixed mode effect. The normal and tangent effective cohesive tractions are given as;
Δ Δn
Tn = σmax ⎛1 − eff ⎞
⎜ ⎟
⎝ δn ⎠ Δeff(max) (23)
η2 Δ Δt
Tt = σmax ⎛1 − eff ⎞
⎜ ⎟
k2 ⎝ δn ⎠ Δeff(max) (24)
where Δeff(max) is maximum displacement which is equal to Δeff under loading, representing the displacement at the turning point
from loading to unloading under unloading and reloading. The model has been implemented into an inhouse explicit software CDFP.
The cohesive model was also coupled with the frictional contact algorithm accounting the unilateral and frictional sliding effects of
cohesive elements under mixed mode bending. Simulation outcomes were in good agreement with experimental results in terms of
impact force and fracture patterns. Quite recently, Gao et al., [101] used intrinsic cohesive model to investigate the impact fracture
behaviour of laminated glass under low velocity impact. The authors divided the model in three phases, elastic deformation stage,
damage evolvement stage and complete fracture stage characterized by three parameters δm(effective separation), δm0 (effective
separation at damage initiation) and δmf (effective separation at complete failure). Macroscopic fracture was captured efficiently
although the model suffer from artificial compliance. The crack patterns of laminated glass including smashing cracks, radial cracks
and circular cracks were in good agreement with experimental observations.
4.1.1.3. Extended finite element method. It is a well-known tool which enhances tremendously the process of FE methods in simulation
of stationery as well as evolving cracks. Generally FE meshes have to be reconstructed for elements edges and faces which have to
coincide with crack surfaces and also the nodes must be on each side of the crack allowing separation of the material along the crack
surface which is quite cumbersome specially in 3D simulations. However, FE meshes are independent of the morphology of extended
finite element, eliminating the need of remeshing when a crack grows. XFEM is more efficient theoretically as the mesh generated is
independent of the crack, crack initiation and propagation path are arbitrary solution dependent and also improved convergence rate
for FE solution [199–200]. Extra enrichment functions are introduced to extend the piece-wise polynomial function space of
conventional FE methods. It was first applied by Beltschko et al., [201] to study dynamic crack propagation. Xu et al., [71] have used
XFEM to investigate the crack propagation mechanism in laminated windshields during quasi-static or low speed impact. A
discontinuous displacement field away from the crack tip across the crack is given as;
n 4 2
⎛ ⎞
Un = ∑ Ni (X ) ⎜Ui + bi H (X ) + ∑ ∑ Cil Fl j (X ) ⎟
i=1 ⎝ l=1 j=1 ⎠ (25)
where n is the number of nodes, Ni(X) is the shape function of the node, Ui are the classical degrees of freedom of node I, bi and Cil are
424
the degrees of freedom associated with Heaviside jump function H(X) with value 1 above and below the crack. The criterion of
maximum local hoop stress has been used for fracture propagation,
2 4 2 2
⎛ 3Kll + K1 + 8k1 Kll ⎞
θc = arc cos
⎜ 2 2
K1 + 9Kll ⎟
⎝ ⎠ (26)
The authors measured θc with respect to local polar coordinate system with its origin at the crack tip and aligned with the
direction of the crack tip. Mechanism of cracking was explored, initially radial crack grows at a particular angle whose length
depends upon plate displacement and boundary constraints. Afterwards circumferential cracks develop to release the extensive radial
stress. XFEM formulation was used to probe the effect of effective head mass and impact speed. Xu et al., [102] introduced XFEM to
investigate multifracture problem and validated the results experimentally. The mechanism of propagation of radial cracks was
studied, displacement field at crack tip and criterion of brittle solid was used to obtain brittle DSIFs of different failure modes which
could not be obtained through experimental observations.
Limitations of element deletion technique involves the use of finer meshes to get reliable results as the size of the crack is the same
as that of element, contact conditions cannot be well described since a lot of contact surfaces are eroded due to material failure and
also deletion of elements severely effects the simulation of evolution of cracks. Element deletion method is not able to simulate crack
paths which are not coincident with meshlines. These shortcomings can be handled with more finer meshes. XFEM involves inclusion
of discontinuities on the basis of failure criterion which favour the propagation of a single crack, hence it is quite cumbersome in
multifracture cases. Cohesive zone method also called as interelement method is more suitable in simulation of crack nucleation and
multifracture problems, but require much finer meshes to predict crack paths. Modelling of complex crack paths and achieving mesh
independence are the problems to be solved associated with current state of these numerical techniques in order to accurately capture
the fracture process. The use of multiscale approaches and modelling subscale phenomena may find solution to existing loopholes in
computational modelling of failure analysis.
4.1.2. Material models

Material models for glass and interlayer materials play an important role on prediction capabilities of the developed FE models. A
range of material models are available in FE packages, however the choice of suitable model for the problem under investigation is
the key factor to obtain more accurate numerical simulations.
4.1.2.1. Glass. Most of the previous research report the use of linear elastic material model for glass in FE modelling as the glass
under static loading is brittle and linear elastic till fracture. Glass can be modelled either static or dynamic, linear elastic model uses
static material properties and hence is only suitable for small deformations. The deviatoric and volumetric behaviours are governed
by equations [62,76];
E vεkk Eεij
Sij = ⎛
⎜ + p⎞ δij +
⎟
⎝ (1 + v)(1 − 2v) ⎠ (1 + v) (27)
E vεkk
p=−
3(1 − 2v) (28)
where E is Young's modulus, v is Poison's ratio and εij are strain components. Linear elastic model for glass gives reliable predictions
at low strain rates (quasi-static) and is not able to predict the failure processes involving high strain rates. Under dynamic loading,
glass behaves differently and hence dynamic material model is required to investigate the behaviour of laminated glasses subjected to
high strain rates. Johnson-Holmquist (JH2) model for brittle materials is the widely used material model which considers strain rate
effect, material damage and confinement effect. It is a macrolevel model which is implemented in LS-DYNA as material model 110
(MAT_HOLMQUIST_CERAMICS). Zhang et al., [83] used JH2 model to investigate the impact of flying glass debris on laminated
windows. In this model, the strength of glass is described as a function of damage level, intact strength, fractured strength and strain
rate as given by the equation;
σ ∗ = σ∗j − D(σ∗i − σ∗f ) (29)
where σ∗ is the normalized equivalent stress, σi∗ is the normalized intact equivalent stress, σf∗ is the normalized fracture stress and D
is material damage. All the stresses are normalized by dividing by equivalent stress at Hugoniot elastic limit (σHEL). Extensive
numerical investigations of laminated glasses have been performed by using JH2 material model (Ref). Hidallana-Gamage et al., [91]
analysed blast performance of laminated glass panels using JH2 material model for glass. The main limitation of this model is lack of
proper experimental data on dynamic material properties of glass leading to discrepancies between simulation and practical
situations. Initially it was developed for ballistic performance simulation, however various efforts are being made by the researchers
to improve the suitability of the model to construction glass panels [82,202,203]. Modification of material constants with more data
on dynamic properties of annealed glass has also been reported [204]. The accuracy of the model has also been verified for simulation
of behaviour of glass window subjected to shock and impact loads and found able to predict the response [205]. In a nutshell it can be
stated that, linear elastic model suits best to simulate the laminated glass behaviour under static or quasi static loadings and there is a
scope for improvement of JH2 model with more testing data on construction glass dynamic material properties for more accurate and
reliable predictions.
425
4.1.2.2. Interlayer. PVB is the most investigated interlayer material among the other variants such as polycarbonate, polyurethanes,
ethylvinyl acetate and ionoplast interlayers. Most of the numerical simulation investigations involve PVB laminated beams, panels
and plates. The mechanical behaviour of PVB at small strains has been extensively investigated and a linear viscoelastic model known
as Maxwell-Wiechert model is generally used to account the viscoelastic nature of the interlayer [50,51,79]. The shear modulus of
viscoelastic material decays with time which is interpreted by Prony's series of units arranged in Maxwell-Weichert model. Galuppi
and Royer Carfagani [77] compared the state of strain and stress calculated in the full viscoelastic and an equivalent elastic material
model for PVB for long duration loads. The authors recommended a full viscoelastic analysis for precise, non conservative design of
laminated glasses.
Wei et al., [62] modelled PVB as linear viscoelastic material and proposed a FE model for simulation of stress characteristics of
laminated architectural glazing under blast loading. The deviatoric and volumetric behaviours are governed by Eqs. (30) and (31)
respectively.
t
Sij (t) = 2 ∫0 G(t − τ)ė ijdτ (30)
p = −Kεkk (31)
where t is time, ė ij is deviatoric strain rate, K is the bulk modulus and G (t) is stress relaxation modulus given as;
G(t) = G∞ + (G0 − G∞) e−βt (32)
where G0 is short time shear modulus, G∞is long time shear modulus and β is the decay factor. The results obtained were found in
good agreement for static analysis. The viscous effects can be neglected under short duration loads as the change in shear modulus of
PVB is negligible (0–100 ms). However, dynamic material properties of PVB differ from quasi-static behaviour and hence use of PVB
static properties to analyse blast loading impact on laminated glasses leads to inaccurate simulations. PVB behaves differently at
different strain rates, it becomes more elastic at higher strain rates which is similar to its behaviour at low temperatures
[76,79,206–210]. Further below glass transition temperature, it becomes more elastic-plastic. Hence the choice of a material model
depends upon type of loading in terms of strain rates and temperature. PVB is a viscoelastic material, however its viscoelasticity can
be neglected under short duration loads, high strain rates, large strains (~200%) and below glass transition temperature. Du Bois
et al., [58] modelled PVB interlayer as hyperelastic membrane neglecting viscous effects using hyperelastic material models such as
Blatz-Ko, Mooney-Rivlin and Odgen material law in which stress can be derived from an energy function related to current state of
deformation. The first two material models suits for strains less than 100% however for higher strains Ogden material law with higher
order terms is suggested.
Blast investigations usually involves large strains and large strain rates, hence strain rate dependent elastic-plastic material model
is suggested for blast loadings however for lower strain rates (quasi-static) viscous effect can not be neglected [76,98]. Zhang et al.,
[83] used strain rate dependent elastic plastic material model (MAT_STRAIN_RATE_DEPENDENT_PLASTICITY) and formulated a
numerical simulation for laminated glass windows subjected to debris impact as under high speed tension PVB behaves as a brittle
material. The use of material model 24 (MAT_PIECEWISE_LINEAR_PLASTICITY) accounting elastic plastic behaviour of PVB which
can be implemented in LS-DYNA is the current practice for analysing laminated glass panels under blast loading [91]. This model
defines arbitrary stress vs plastic strain for different strain rates. It can be stated that linear viscoelastic models with static material
properties of PVB have been replaced by strain rate dependent elastic plastic material models especially for blast loadings and
loadings involving high strain rates for more accurate predictions.
4.2. Finite element simulation of head impact on laminated windshields
Numerical simulations of windshield are of great importance to safety engineers, vehicle manufactures and accident investigators
as the derived information is very useful in pedestrian & passenger protection and accident reconstruction [99,211]. Extensive studies
have been conducted in past by several researchers on finite element modelling of mechanical behaviour of laminated windshields
upon head impact. Dharani and Ji in 1998 [47] used DYNA2D, a dynamic, nonlinear finite element code to model the dynamic
responses of the occupants head and laminated glass. Simulation of human head is complex as its structure, geometry and mechanical
parameters are complicated. It was modelled as a solid aluminium sphere covered with a viscoelastic skin. The glass plies and
aluminium sphere were modelled as linear elastic materials. Maximum resultant head acceleration at different impact speeds and
laminate configurations before the failure of the laminated glass plate was investigated. Zhao et al. in the year 2005, [212] extended
this work and used finite element code wherein glass plies, aluminium sphere and PVB interlayer were modelled as linear elastic
materials. Strain energy release rate criterion (ℊ) was developed to calculate the crack initiation time and location.
dWs
= = 2γs
dA (33)
where, γsis the fracture surface energy & Ws is the work required to create a new surface. For a linear elastic material, f = , where f
integral is independent of path and its value was taken as 7.8 J/m2 and when its value reaches critical value (fic ), crack initiates. The
authors adopted global to submodel technique, the head impact problem was simulated as a 2D axisymmetric finite element model
with four node bilinear axisymmetric element CAX4R. The global model was run with the dynamic analysis using ABAQUS/Explicit
and the results were interpolated on the boundary of the submodel. The authors also studied the effect of PVB thickness, glass ply
thickness and also compared laminated and monolithic glass and found that under the simulated head impact the crack always
426
Fig. 2. Model used for FEM simulation of human head impact on windshield [216] (a) Quarter model of human head impact on windshield [52] (b)
and typical blast pressure curve [41].
initiates in the impact side ply first. The PVB thickness and non impact side ply thickness do not affect the cracking behaviour. In
another work, the authors [213] developed a three dimensional featureless headform model to simulate the occupants head impact
on the automotive laminated glazing. Von karman's large deflection theory was used to develop an analytical solution and the results
were compared with a 3D non linear finite element model. The headform and the laminated glass were modelled with eight node
linear brick solid elements, C3D8R with hourglass control and reduced integration using the ABAQUS EXPLICIT package. The closed
form analytical solutions were found in good agreement with FE model and could be used to predict failure in automotive wind-
shields upon head impact. Constitutive model based on continuum damage mechanics has also been developed to predict the failure
resistance of windshield upon head impact by implementing it to an axisymmetric FEM model [214]. The authors modelled lami-
nated glass failure by using elastic damage tensor (anisotropic) with a linear damage evolution law and predicted damage patterns
and zone sizes. In an another approach, Xu and Li [215] developed new constitutive relations on the basis of classical plate theory in
combination with Green strain and Von Karman assumptions able to describe the large deformation of laminated windshield glazings.
The authors proposed a 3D FE model with hybrid mesh of triangular and quadrilateral elements to describe the behaviour of
windshield upon head impact. Damage mechanism with crack evolution stage and crack patterns were characterized through non
linear dynamic finite element model considering windshield as a composite plate. Three types of crack patterns, circumferential,
spider web and perforation crack were identified in PVB laminated windshield through FEM modelling. The maximum deflection of
the order of 27.734 mm obtained through FEA model which was found very close to 24.611 mm calculated through constitutive
relations.
As further improvement in prediction of crack evolution and the types of stresses involved in initiation and propagation of these
cracks, the authors proposed a damage-modified nonlinear viscoelastic constitutive relations model of laminated windshield [70].
Lagrangian method based constitutive model was used to develop FEA model with proper material properties. FEM model was used
to investigate pedestrian head impact with vehicle windshield in traffic accident and internal stresses. This research presented the
role of shear, compressive and tensile stresses in evolution of different types of cracks after head impact on automotive windshield.
Good similarities in the results were obtained between the computational FEA model and the classical Hertzian pressure model. With
new advances in numerical modelling Xu et al. [71], investigated the mechanism of crack propagation in windshield upon head
impact. Extended finite element modelling was used which eliminated remeshing during crack propagation and only the additional
degree of freedoms associated with the nodes intersected by propagating crack were included. A featureless spherical headform was
used to model human head. Eight node linear brick elements were used to model PVB laminated glazing and head with a friction
coefficient of 0.1 between head and PVB laminate. The authors decoupled the crack simulation in two steps; 1) radial cracks were
developed by introducing initial flaws at the centre of plate and 2) new circumferential cracks were introduced on the radially
cracked specimen and simulated by XFEM. This research concluded that initially radial crack grows and to release the radial stress,
circumferential cracks are developed at the contact point and the perimeter of radial crack. The impact speed can also be deduced
through reverse analysis from the crack patterns which showed good agreement with the real world accidents.
Munsch et al. in the year 2009 [217], introduced a FE model for laminated side glazing based on experimental validation to
investigate the mechanical behaviour during head impact. The headform was hybrid III dummy head composed of an aluminium
structure with outer covering of vinyl synthetic skin having a total mass of 4.53 kg. The proposed FE model consists of 16613 shell
elements for laminated glazing and 8443 brick elements for the viscoelastic support. Shell was a composite of three individual layers
of laminated glass where the glass followed elastic brittle law and PVB interlayer followed elastoplastic law. Head was modelled with
408 shells for skull and 1224 bricks for outer vinyl skin following linear elastic law. Errors of 20%, 37% and 20% was observed for
maximum linear acceleration, head injury criteria and window permanent strain respectively when compared with experimental data
at a velocity of 5 m/s. Pyttel et al. [74] suggested a “Non failure criterion” i.e. a critical threshold energy is required before failure of
laminated glass under impact loading. The developed criterion was implemented in PAM CRASH, a finite element solver. The head
427
Glass
Interlayer (PVB)
(a)
(b) (c)
(d) (e)
(f) (g)
Fig. 3. Different types of models used in FEM simulations of laminated glasses: (a)–(e) Layered elements; (a) Monolayer element with shared nodes,
(b) Bilayered element with shared nodes, (c) Bilayered element with tied connections, (d) Trilayered element with shared nodes & (e) Trilayered
element with tied connections. (f) 3D solid element and (g) Smeared model [67,86].
impactor was aluminium skull covered with vinyl skin which was modelled as solid elements to investigate the failure of plane
laminated glass and real windshields. The experiments involved the impact against windshield both from outside and inside of the
car. Crack initiation was based on local Rankine criterion enhanced by a non local approach which offered good agreement between
experiments and simulations. Authors preferred simplest modelling approach with two layers of shells for external glass and PVB
interlayer. The non linear viscoelastic model following G'Sell equation was used to model PVB interlayer. The proposed model is able
to calculate deformation and impactor acceleration. Critical stress, critical energy and critical radius for failure alongwith Young's
modulus and Poisson's ratio were the only input parameters required for glass.
In the previous reference [74], three possibilities were discussed to model the windshields, however Peng et al. [86] in the year
2013 reported five windshield finite element models (Fig. 3) categorized into single, double and triple layer models set up in LS DYNA
code. A single shell element was used for both glass and PVB interlayer in single layered glass model, where the shell follows glass
parameters upto fracture strain of glass and then onwards PVB characteristics are followed. Two double layered models were pro-
posed consisting one layer of glass and another of PVB, wherein shell elements are joined either through shared node or tied element
connections. The glass and PVB were modelled using piecewise linear plasticity and modified piecewise linear plasticity respectively.
In three layered models, two shell elements for outer glasses and one for PVB layer were considered with either shared node or tied
element connections. In these models glass was modelled using linear elastic behaviour and PVB interlayer was modelled using
Mooney–Rivlin material laws. Experiments were conducted using standard European Experimental Vehicles Committee adult
428
Table 5
Parameters involved in smeared modelling technique used for windshield FE models [67,86].
1. Equivalent thickness of two coincident shells of Glass and PVB 3 EPVB 3
tE = 3 tG + 3tG (tG + tPVB )2 + t
2EG PVB
2. Modified density 1
⎛ρG tG + ρPVB tPVB ⎞
⎝ 2 ⎠
ρE =
tE
3. Young' modulus of the modified PVB after fracture EII =
1 3 2 3
+ 3tPVB tG2 )]
tG3 [EG (tG + 3tG tPVB ) + EPVB (tPVB
4. Young' modulus of the modified glass EG, modified=2EG−EPVB, modified
EG and EPVB are Young's modulus of glass and PVB; tG and tPVB are thickness of glass and PVB; ρGand ρPVBare density of glass and PVB.
headform impactor with a weight of 4.8 Kg at a speed of 11.1 m/s. Clamped windshield experiments were also performed at 6 m/s.
Numerical models were validated in terms of linear acceleration at headform centre of gravity and crack propagation. The authors
concluded that double layered model with tied connection and mesh size of 5 mm is the best model in terms of headform linear
acceleration and fracture of windshield to simulate the mechanical behaviour of windshields. In this model, Timmel's concept of
smearing modelling technique was applied to account the composite efficiency of windshield. Before fracture, the behaviour of
windshield is controlled by glass shell and after fracture by PVB shell and the glass shell is deleted alongwith the tied elements as only
the compression stress remains in the fractured glass. The modified density and modified Young's modulus of PVB and glass are given
in Table 5.
4.3. Finite element simulation of blast loading
Due to prevalent threats from bomb explosions, gaseous blasts, windborne debris, earthquakes etc., structural analysis and design
for laminated glazings becomes all the more important to resist such unwanted and sudden extreme loading conditions. Ngo et al.,
[218] reviewed blast loading and blast effects and summarized various computer codes such as ABAQUS AUTODYN, DYNA3D and
LSDYNA, used to simulate blast effects on structures and able to implement FE models accounting structural motions while calcu-
lations proceed. FEM analysis has been used extensively by the researchers for prediction of failure of laminated architectural
glazings when subjected to blast loading. Dharani et al. [56] used a 3D non linear finite element model to predict the failure of
laminated glass and describe cumulative damage probability of inner glass pane of architectural glazing upon blast through bomb
explosion on ground. The pressure time curve for the shock wave consists of positive and negative pressure phases, however the
authors accounted the structural damage due to positive phase through an empirical quasi exponential form of the positive phase.
−∝ t
( )
p(t) = p0 1 − t t d e t d
(34)
where, p(t) is instantaneous free air overpressure at time t, p0 is the peak overpressure when t is zero, td is the duration of positive
pressure and α is the decay factor (Fig. 2). The tabulated values of these parameters by Kinney and Graham [219] were fitted for
efficient application of the equation in numerical simulation. Uniform pressure simulating blast loading was applied and the dynamic
problem was solved through a dynamic nonlinear 3D FEM, LS DYNA with basic governing equations given by Hallquist [220].
Stresses were calculated through finite element method using 8 node solid elements for the laminated glass using the following
equation;
σij = Sij − pδij (35)
where σij represents stress components, Sij; deviatoric stress components, p; pressure and δij; Kronecker delta. Glass was modelled as
linear elastic material and PVB interlayer as linear viscoelastic material with simply supported boundary condition. The study finds
that new laminated glasses were able to carry higher blast impulse load than the aged samples and the thickness of inner ply of glass
can be increased, keeping the total laminate thickness constant for efficient designing of laminated glass.
The model was further used to characterize the stress distributions and effects of blast loading parameters and PVB interlayer's
viscoelastic parameters on dynamic response of laminated architectural glazing [62]. Ten, 3D 8 node solid elements were used to
model the glass, four for each glass pane and two for PVB with 40 × 40 meshes in x-y plane. Central deflection and maximum
principal stress calculated through static solutions were found in good agreement with earlier works. Table 6 shows the material
properties of glass and PVB generally used for FEM calculations. Dynamic response of the glass was effectively predicted with the
Table 6
Common parameters of glass and PVB interlayer used for FEM modelling [63,215].
Parameter Glass PVB Human head
Modulus of elasticity(E) (GPa) 70 20 6.5

Poisson's ratio (ν) 0.23 0.45 0.22
Density (ρ) (kg/m3) 2,500 1100 1412
Shear modulus (GPa) – 6.9 –
429
proposed model and no effects of bulk modulus of PVB in the range 0.3–200 GPa was observed. However the authors concluded that
short time shear modulus significantly affects the dynamic response. This study promoted the applicability of computational efforts in
numerical modelling of complex behaviour of laminated glass.
The negative phase of the blast load can also cause breakage of laminated glass if the positive phase is not enough violent for
failure [60]. A failure criterion was developed using Von Karman's large deflection equation to predict the failure of glass. The results
were validated through 3D FE model treating PVB as viscoelastic as well as elastic material. For a typical blast load wave, results were
very close for both the cases. In another study, FEM model was used to validate the results of analytical models developed using
classical plate theory (CLPT) and Von Karman's large deflection theory [64]. The significance of negative phase of blast loading has
been established by the authors as midspan deflection and maximum tensile stress were found double for negative phase loading than
positive phase. They also suggested that a rectangular laminated panel is more blast resistant than a square panel. In another work,
authors plotted two principal stresses at the centre of the bottom surface of panel (inner ply) against time showing that the panel
exhibits forced response (t ≤ td) during positive blast loading phase and free vibration (t ≥ td) in negative phase, where td is the blast
loading duration [46]. The equivalent 60 s constant stress was calculated for the duration tf in which both principal stresses were
tensile, which represents the dynamic state of loading. The authors determined λ, the ratio of minimum to maximum principal stress
for three locations and it was assumed to be constant by averaging its values from FEA stress outputs. The tensile stress controlled
mainly the failure of glass panel and σx & σy were calculated where the stresses were tensile. These outputs were used to study the
failure probability of laminated panels. The effect of thickness of glass panel and PVB interlayer on the blast resistance of the panel
was also investigated and the authors suggested increased thickness of inner glass ply and adequate thickness range of 0.76 to
1.52 mm for PVB for economical and efficient architectural laminated panels.
Zhao et al. in the year 2008, [221] developed a constitutive model based on continuum damage mechanics and studied the blast
resistance of laminated panel by implementing it into a finite element code. Damage patterns and zone sizes were predicted using
anisotropic elastic damage tensor with a linear evolution law and studied the effects of various geometric parameters on failure
behaviour of laminated glass. Larcher et al. [76] reviewed and assessed 3D model with solid elements, layered model with shell and
solid elements and smeared model with two coincident shell elements used for investigating blast resistance of architectural glazings
by implementing in the finite element code EUROPLEXUS. Experiments were also conducted and compared with simulation results in
order to enlist the advantages and disadvantages associated with these widely employed numerical models. A triangle shell element
with five integration points was used in the layered model where the stresses were set to zero if the strains are greater than zero after
glass failure, further the element is eroded when the interlayer reaches failure criterion of PVB. Smeared model, with coincident
elements detailed above and five solid elements with full integration were used in 3D model. Glass was considered as linear elastic
material and PVB followed elastoplastic material law. The elastic support of the laminated panel was modelled by solid elements
following an elastic material law with Young's modulus of 3.5 MPa. The authors concluded that 3D solid model is very time con-
suming and requires detailed material laws with large computational efforts, however it can produce detailed results. Layered model
can be used efficiently and is able to reproduce experimental results as it also accounts the failure of interlayer. Smeared model can
only be used for small displacements of the panel as it does not represent failure of PVB. Hidallana-Gamage et al., [88] compared 2D
shell element and 3D solid element models for analysis of laminated glass panels under blast loads and found that both models are
able to predict the experimental results with considerable accuracy. 3D model is more accurate and is suggested to investigate the
influence of controlling parameters such as material, geometric restraints and support conditions and then 2D model can be used to
analyse blast response under different loadings.
A new finite element model in ABAQUS finite element software package for predicting blast resistance of laminated glasses is
reported by Hooper et al. [79], which was further validated using laboratory tests. In order to simplify the modelling, authors split the
response of laminated glass into precrack model and postcrack model. The simplified shell element model involving 3D quadrilateral
finite membrane strain shell element and more complex 3D solid element model involving C3D81 solid continuum element were used
to describe the precrack behaviour. Effect of viscoelasticity of the interlayer on the blast response was studied by solving these two
models with and without viscoelasticity in the PVB interlayer. Post crack model involved shell elements with modified material
properties and Johnson Cook plasticity model was used to measure the stress-strain response of fractured laminated glass. The blast
load was simulated using Air 3D software package using variant of Air 3D code with adaptive mesh refinement and detonation
modelling. Detailed pressure impulse analysis was conducted to predict the damage by running the model with peak reflected
pressures in the range 1 KPa–1 MPa and positive impulse in the range of 10 KPa ms to 10 MPa ms. 20% maximum strain was sug-
gested which can be safely handled by the glass laminate. The isodamage model adopted various strain criteria at different strain
levels and used to predict the isodamage curves. The predictions obtained from simulated results showed fair agreement with ex-
perimental data initially, however the JC plasticity model was not found suitable to simulate the exact behaviour of cracked lami-
nated glass. The authors find the FE model suitable to capture the deformation behaviour upto maximum deflection adequately for
full scale blast assessment of laminated glass panels when subjected to different charge and standoff conditions and the model can
also be applied to laminated glass under other impact conditions. In continuation of this work, Zhang et al. [87] performed numerical
simulations and constructed pressure-impulse diagrams to correlate dynamic response of laminated glass and blast loadings. FE
model was developed in LS-DYNA with solid continuum elements having mesh size of 5 mm × 5 mm × 3 mm for glass and
5 mm × 5 mm × 1.52 mm for interlayer. 8 nodes solid 164 element with full integration were used and erosion technique was
employed to simulate glass cracking and interlayer damage. The widely adapted Johnson Holmquist ceramic constitutive model and
strain rate dependent elastic plastic model were used for glass and PVB interlayer respectively. The potential debonding between
glass and interlayer was modelled using automatic surface to surface tiebreak in LS-DYNA. Under strong impulse maximum deflection
in numerical simulation was 13.884 mm, very close to experimental value i.e., 14.583 mm, however slight variations were observed
430
for weak impulse results. Failure criterion based on breakage of interlayer was used to construct unified P-I diagrams of laminated
glass involving bending failure and shear failure modes. P-I diagrams of laminated glasses are given by the equation:
P I β
(P − Po )(I − Io) = A ⎛ o + o ⎞
⎝2 2⎠ (36)
where, Po and Io are the pressure asymptote and impulse asymptote respectively. A and β were insensitive to change in PVB thickness
and hence to reduce the number of parameters these were taken as constants (2.4 and 1.2 respectively). The effect of interlayer
thickness, glass thickness, window dimensions, boundary conditions and glass strength on the P-I diagrams has been investigated and
the numerical results are used to derive an empirical formula for P-I diagrams as given by the Eqs. (37) and (38):
Po = 2.433 − 0.007L + 13.123dPVB + 1.844dg + 0.177ft,glass (37)
Io = 113.487 − 0.008L + 118.380dPVB + 44.645dg + 0.579ft,glass (38)
where, dPVBand dgare thickness of PVB and glass, L is window length and ft, glass is glass strength. These equations alongwith Eq. (21),
can be used to construct P-I diagrams for laminated glasses. The authors also compared these P-I curves with previous works [79] and
found the proposed model accurate as it involves detailed dynamic material properties of glass and PVB. The numerical model can
give reliable predictions of P-I diagrams of laminated glasses of different parameters when subjected to blast loading.
Hidallana-Gamage et al., have conducted extensive research on numerical simulation of blast response of laminated glasses
[27,91,92]. The researchers considered sealant joints also alongwith glass and interlayer to present more useful information to the
engineers. They developed comprehensive numerical models using FE techniques to account post crack load carrying capacity of
laminated glass as well as the effect of sealant joints. 3D constant stress elements were used with material models 110 for glass and 24
for PVB and sealant joints. Effect of sealant joints was investigated by varying their thickness, width and Young's modulus under two
different blast loads. The authors find out that increasing the number of elements in the plane of panel leads to more accurate results
than increasing the elements through the thickness of the panel. JH2 model (110) is able to predict the failure of glass as its strength is
reduced when it is damaged. As the dynamic breaking strength of annealed glass is increased under high strain rates when subjected
to blast loads, therefore FE models are sensitive to tensile strength values and care must be taken while choosing T values in the
analysis. The authors concluded that interlayer properties play an important role under severe blast loads. Increased flexibility of the
supports reduces stress and damage to the glass panes, however it has little impact under severe blast loading conditions. In con-
tinuation of this work, influence of interlayer properties on blast performance of laminated glass panels is also been studied by
varying thickness and Young's modulus of the interlayer [95]. The FE models were used to understand their effect on total energy
absorption, centre deflection, glass panes fracture and stress-strain behaviour of interlayer. The authors find out that there is a critical
interlayer thickness for a particular interlayer material and recommended to choose the thickness as close as possible to this value
which will enhance energy absorption and reduce support reactions. Hidallana-Gamage et al., [222] reviewed latest design standards
in blast resistant glazing designs and compared with comprehensive FE modelling. The reported methods are expensive due to large
experimental set ups and can be used with limited dimensions of glass however numerical analysis offers a comprehensive set of
information useful to design laminated panels to blast loads reducing cost, avoiding safety issues and environmental pollution as-
sociated with blast testing. In a recent work, Hidallana-Gamage et al., [103] established the importance of structural sealant joints by
investigating laminated glass panels fixed to rigid and flexible supports. A comprehensive FE model accounting window frame,
fasteners, and flexible cable net has been presented and effect of negative phase on flexible supports was investigated. Negative phase
was found to have little impact on laminated glass panels with rigid supports, on the contrary it has significant effect on flexible
supports as the former are considerably damaged during positive phase. The authors suggested that smaller dimensions of sealant
joints could be used in flexible supports because a large amount of blast energy is absorbed by flexible supports.
Pelfrene et al., have investigated [98], fracture and dynamic post breakage performance of PVB laminated glass panel under blast
loading using FE method. Shell elements with full integration, following linear elastic law with Rankine criterion characterized
failure and solid elements with full integration following strain rate dependent hyperelasticity law were used to model glass and PVB
respectively. Critical value of, σcr = 81 MPa was used under dynamic loading for fracture criterion. The interface between glass and
PVB was modelled through tiebreak contact to study delamination at the interface. For boundary conditions, solid elements with full
integration were involved with linear elastic model for steel and Neo-Hookean hyperelastic law for rubber. The authors tried to
reproduce the experimental results with the FE model using blast experiments by EN 13541. Different PVB interlayer materials from
softer to stiff were examined for blast response and they find that the interlayer with low adhesion level relieves local PVB strains and
thus delay the tearing of interlayer. The shortcomings of FE method have been putforth in order to develop improved models capable
of more realistic predictions of blast response of laminated glass fracture. Glass fracture was represented by deleting a finite element
upon reaching the failure criterion which developed unstable stress wave propagation in the glass part. The model is able to re-
produce crack formation and also the global deformation behaviour at load levels, ER 1 and ER 1.6, however, the deletion technique
posed serious limitations to predict the failure of interlayer. Mesh size greatly affects the simulation results because the crack appears
in the same size with element deletion method. Finer meshes allows higher strains of PVB but failed to reach the tearing limit and
hence more narrow simulations of glass cracks have been suggested for more real picture of delamination accounting the tearing of
interlayer. Spiller et al., [223] compared various software tools based either on explicit finite element analysis or single degree of
freedom analysis for their applications in designing or analysis of glass windows under blast loading. The authors reported that
nonlinear material behaviour, complex support conditions and higher modes of vibrations under dynamic analysis can be easily
accounted in FEM modelling hence it is able to predict more accurate results than SDOF method.
431
Quite recently, Gao et al., [101] developed a bilinear intrinsic cohesive model to simulate the impact fracture behaviour of
laminated glasses. The authors divided the model in three phases; the elastic deformation stage, damage evolvement stage and
complete failure stage. Three parameters i.e. δm, δm0 and δmf representing effective separation, effective separation at damage
evolution and effective separation at complete failure between cohesive elements surfaces were used to capture the impact fracture
process. At complete failure stage, surfaces of the CE separate and a new crack surface is formed. ABAQUS/Explicit is used to simulate
the developed model, an error of 10.47% was observed when validated experimentally, however the force peak appeared almost at
the same time. To the advantage, finer meshes with 156420 & 380280 and coarser meshes with 79344 & 192442 solid elements and
cohesive elements respectively showed similar results. The model suits best to laminated glass plates and can also be applied to
curved laminated glasses such as windshields with further experimental testings.
5. Conclusion
The above discussion reveals that the emphasis of the research groups is on the development of such models which can predict the
transverse shear stress more accurately and efficiently. CLPT and FSDT are not suitable for accurate computations of transverse shear
of moderately thick and thick laminated plates. However higher order models are able to predict transverse shear effects more
accurately but require more computational effects. In all these ESL theories, the laminated plate is analyzed as two dimensional
equivalent single layer with continuous transverse strains. Because of that, the interlaminar transverse stresses are obtained as
discontinuous functions at the interfaces of adjacent layers having different stiffness properties. Also ESL theories are not able to
predict local behaviour of sandwich plates and are suitable for thin laminated plates only. Layerwise theories offer the researchers to
develop different models using any of the ESL theory suitable for individual layers of the laminate. This approach can help to compute
these theories with minimum number of variables by selecting appropriate theories i.e., CLPT, FSDT, HSDT etc suiting best to the type
of laminate under investigation or to be analyzed. The main drawback of the layerwise theories is that for laminates with multiple
layers, the number of variables increases and hence the numerical modelling becomes very complex and time consuming. Consistent
efforts are seen in the development of more refined and robust zig zag theories that can predict the behaviour of sandwich plates more
accurately where the transverse shear stresses are obtained by integration of equilibrium equations. Finite element implementation of
these models have been extensively assessed and found authentic when compared with experimental findings. Recent trends are on
evolution of mixed plate theories with main focus on added accuracy to transverse shear predictions, development of simplified
theoretical approaches that can be easily adopted for numerical modelling. The researchers have combined different theories to arrive
at unified approaches suiting best for the considered problem with superior predictive capabilities and abilities to model wide range
of laminates under different boundary conditions.
Finite element modelling has been used by the researchers to assess the damage behaviour of laminated glass effectively. The
numerical results are helpful in designing of laminated glass of desired damage characteristic. Continuous research has been con-
ducted on numerical simulation of human head modelled through 3D headform models and its impact on windshields using CLPT,
Von Karman's assumptions, layerwise theories and smearing modelling techniques. FE modelling techniques have been advanced
over the years from 2D to 3D models comprising triangular and quadrilateral elements and experimentally validated accurate models
have been proposed. The type of cracks, their evolution and propagation have been investigated through extended nonlinear dynamic
FE models. Extensive research has been conducted on simulation of blast loading for predicting failure of laminated architectural
glazings. Both phases of blast loading i.e., positive and negative have been investigated effectively through numerical simulations.
Adequate thickness of the interlayer and glass plies has also been proposed by the authors for economical and efficient architectural
laminated glasses. Among different models, layered model was found most efficient and accurate when validated experimentally. The
researchers constructed pressure-impulse diagrams and correlated the dynamic behaviour of laminated glass upon blast loadings.
Numerical simulations has been used to investigate the effect of interlayer thickness, glass thickness, window dimensions, boundary
conditions and glass strength on the P-I diagrams. More reliable predictions of P-I diagrams of laminated glasses of different para-
meters can be made through numerical models when subjected to blast loading.
Skilful attempts are required to evaluate the outputs of particular computer codes in terms of accuracy and appropriateness to
modelled situation. Modelling errors and poor interpretation may lead to erroneous results questioning the applicability of the
proposed models to be used as practical design guides. Hence, a good understanding of computational mechanics, structural dy-
namics and material properties is essentially required to obtain accurate results from developed FE models. Another problem is mesh
size convergence, high frequency energies will be filtered into the structure if mesh size is large therefore small mesh sizes are
required for proper simulation of structural responses. Finer meshes require huge computational efforts and enormous computer
resources, a reasonable compromise between accuracy and simulation is recommended through mesh convergence check. Use of
appropriate dynamic models is another challenge to FE modelling of laminated glasses. A large amount of experimental data is
required accounting dynamic behaviour of glass and interlayer. Further chemical composition and thermal treatment of glass leads to
significant variations in the material properties of glass.
Despite of these limitations, FE modelling is an appropriate choice owing to its flexible nature as 2D models can be used for
comparative analysis and 3D models where greater accuracy is required. Substrates can be easily changed like type of glass and
interlayer with an essential requisite of proper material model. FE models are able to capture the fracture process i.e., initiation and
propagation of cracks with the use of advanced techniques such as element deletion, cohesive modelling and XFEM. Effect of geo-
metric variables, structural frames and sealants can also be accounted through FE modelling techniques offering more realistic data to
be used in real world situations. FE modelling has been suggested to be most effective when compared with various software tools.
432
Acknowledgement
The financial support given by Department of Science and Technology, Govt. of India, through the project DST/TSG/Ceramic/
2011/77-G is gratefully appreciated.
References
[1] S. Abrate, Impact on laminated composite materials, Appl. Mech. Rev. 44 (4) (1991) 155–190, https://doi.org/10.1115/1.3119500.
[2] S. Abrate, Impact on laminated composites: recent advances, Appl. Mech. Rev. 47 (11) (1994) 517–544, https://doi.org/10.1115/1.3111065.
[3] P.F. Liu, J.Y. Zheng, Recent developments on damage modeling and finite element analysis for composite laminates: A review, Mater. Des. 31 (2010) 3825–3834,
https://doi.org/10.1016/j.matdes.2010.03.031.
[4] B.R. Lawn, Y. Deng, P. Miranda, A. Pajares, H. Chai, D.K. Kim, Overview: Damage in brittle layer structures from concentrated loads, J. Mater. Res. 17 (2002)
3019–3036, https://doi.org/10.1557/JMR.2002.0440.
[5] M.S. Shetty, J. Wei, L.R. Dharani, D.S. Stutts, Analysis of damage in laminated architectural glazing subjected to wind loading and windborne debris impact,
Buildings 3 (2) (2013) 422–441, https://doi.org/10.3390/buildings3020422.
[6] L. Biolzi, S. Cattaneo, G. Rosati, Progressive damage and fracture of laminated glass beams, Constr. Build. Mater. 24 (4) (2010) 577–584, https://doi.org/10.
1016/j.conbuildmat.2009.09.007.
[7] D. Delince, D. Callewaert, J. Belis, R. Van Impe, Post-breakage behaviour of laminated glass in structural applications, in: F. Bos, C. Louter, F. Veer (Eds.),
Challenging Glass, Conference on Architectural and Structural Applications of Glass, IOS Press, Netherlands, 2008, pp. 459–467.
[8] H. Kunert, Automotive windshield, U.S. Patent 4,378,484 (1983).
[9] Y. Furukawa, T. Igarashi, Laminated glass windshield and method of making same, U.S. Patent 4,075,381, 1978.
[10] G. Savineau, F. Serruys, Durability and postbreakage behaviour of laminated safety glass, Glass Processing Days 2001 Conference Proceedings Book, 2001, pp.
328–330.
[11] W.P. Stephinson, Bullet resistant glass, U.S. Patent 5,368,904, 1994.
[12] N.P. Bolton, S. Novis, Laminated safety glass, U.S. Patent 5,002,820, 1991.
[13] J.I. Hanoka, Encapsulant material for solar cell module and laminated glass applications, U.S. Patent 6,114,046, 2000.
[14] H.S. Norville, E.J. Conrath, Considerations for blast-resistant glazing design, J. Arch. Eng. 7 (3) (2001) 80–86.
[15] A. Remennikov, L. Chipperfield, N. McInerney, Experimental investigation of load-deformation response for blast-resistant facade glazing solutions,
Australasian Structural Engineering Conference 2008: Engaging with Structural Engineering. Meeting Planners, 2008, pp. 854–864.
[16] P. Foraboschi, Optimal design of glass plates loaded transversally, Mater. Des. 62 (2014) 443–458, https://doi.org/10.1016/j.matdes.2014.05.030.
[17] P. Foraboschi, Analytical model for laminated-glass plate, Compos. Part B Eng. 43 (2012) 2094–2106, https://doi.org/10.1016/j.compositesb.2012.03.010.
[18] P. Foraboschi, Three-layered sandwich plate: Exact mathematical model, Compos. Part B Eng. 45 (2013) 1601–1612, https://doi.org/10.1016/j.compositesb.
2012.08.003.
[19] P. Foraboschi, Three-layered plate: elasticity solution, Compos. Part B: Eng. 60 (2014) 764–776, https://doi.org/10.1016/j.compositesb.2013.06.037.
[20] P. Foraboschi, Analytical modeling to predict thermal shock failure and maximum temperature gradients of a glass panel, Mater. Des. 134 (2017) 301–319,
https://doi.org/10.1016/j.matdes.2017.08.021.
[21] M.I. Friswell, J.E. Mottershead, Finite element modelling, Finite Element Model Updating in Structural Dynamics, Solid Mechanics and its Applications, 38
Springer, Dordrecht, 1995, , https://doi.org/10.1007/978-94-015-8508-8_2.
[22] A. Benjeddou, Advances in piezoelectric finite element modelling of adaptive structural elements: a survey, Comput. Struct. 76 (1–3) (2000) 347–363.
[23] T.T. Nguyen, Analysis and Design of Innovative Composite Storm Shelter Against Wind Hazards, PhD diss University of Alabama at Birmingham, Graduate
School, 2009.
[24] X.C. Sun, M.R. Wisnom, S.R. Hallett, Interaction of inter-and intralaminar damage in scaled quasi-static indentation tests, Part 2–Numerical simulation, Compos.
Struct. 136 (2016) 727–742, https://doi.org/10.1016/j.compstruct.2015.09.062.
[25] S.P. Ng, P.C. Tse, K.J. Lau, Numerical and experimental determination of in-plane elastic properties of 2/2 twill weave fabric composites, Compos. Part B: Eng.
29 (6) (1998) 735–744, https://doi.org/10.1016/S1359-8368(98)00025-0.
[26] S. Pashah, M. Massenzio, E. Jacquelin, Prediction of structural response for low velocity impact, Int. J. Impact Eng. 35 (2) (2008) 119–132, https://doi.org/10.
1016/j.ijimpeng.2006.12.006.
[27] H.D. Hidallana-Gamage, D.P. Thambiratnam, N.P. Perera, Failure analysis of laminated glass panels subjected to blast loads, Eng. Fail. Anal. 36 (2014) 14–29,
https://doi.org/10.1016/j.engfailanal.2013.09.018.
[28] D.I. Kachlakev, T.H. Miller, T. Potisuk, S.C. Yim, K. Chansawat, Finite Element Modeling of Reinforced Concrete Structures Strengthened with FRP Laminates,
No. FHWA-OR-RD-01-XX, Oregon. Dept. of Transportation. Research Group, 2001.
[29] L. Crocker, Good practice Guide No. 146.
[30] Y. Sha, C.Y. Hui, E.J. Kramer, P.D. Garrett, J.W. Knapczyk, Analysis of adhesion and interface debonding in laminated safety glass, J. Adhes. Sci. Technol. 11
(1997) 49–63, https://doi.org/10.1163/156856197X01010.
[31] R. Iwasaki, C. Sato, J. Lataillade, P. Viot, Experimental study on the interface fracture toughness of PVB (polyvinyl butyral)/glass at high strain rates, Int. J.
Crashworthiness. 12 (3) (2007) 293–298, https://doi.org/10.1080/13588260701442249.
[32] P. Del Linz, P.A. Hooper, H. Arora, Y. Wang, D. Smith, B.R.K. Blackman, J.P. Dear, Delamination properties of laminated glass windows subject to blast loading,
Int. J. Impact Eng. 105 (2017) 39–53, https://doi.org/10.1016/j.ijimpeng.2016.05.015.
[33] C.V.G. Vallabhan, Iterative analysis of rectangular glass plates, J. Struct. Eng. ASCE. 109 (2) (1983) 489–502.
[34] R.A. Behr, J.E. Minor, M.P. Linden, C.V. Vallabhan, Laminated glass units under uniform lateral pressure, J. Struct. Eng. 111 (5) (1985) 1037–1050.
[35] R.A. Behr, J.E. Minor, M.P. Linden, Load duration and interlayer thickness effects on laminated glass, J. Struct. Eng. 112 (6) (1986) 1441–1453.
[36] C.G. Vallabhan, J.E. Minor, S.R. Nagalla, Stresses in layered glass units and monolithic glass plates, J. Struct. Eng. 113 (1) (1987) 36–43.
[37] Y.C. Das, C.V.G. Vallabhan, Y.M. Zhang, Nonlinear dynamic response of window glass plates subjected to fluctuating wind pressures, J. Wind Eng. Ind. Aerodyn.
44 (1992) 2437–2447, https://doi.org/10.1016/0167-6105(92)90035-9.
[38] R.A. Behr, J.E. Minor, H.S. Norville, Structural behavior of architectural laminated glass, J. Struct. Eng. 119 (1) (1993) 202–222.
[39] C.G. Vallabhan, Y.C. Das, M. Magdi, M. Asik, J.R. Bailey, Analysis of laminated glass units, J. Struct. Eng. ASCE. 119 (5) (1993) 1572–1585.
[40] M.T. Edel, The Effect of Temperature on the Bending of Laminated Glass Beam, Dissertation Texas A&M University, 1997.
[41] C.V. Vallabhan, M.Z. Aşik, K. Kandil, Analysis of structural glazing systems, Comput. Struct. 65 (2) (1997) 231–239.
[42] M.Z. Asik, C.V. Vallabhan, On the convergence of nonlinear plate solutions, Comput. Struct. 65 (2) (1997) 225–229.
[43] F.W. Flocker, L.R. Dharani, Stresses in laminated glass subject to low velocity impact, Eng. Struct. 19 (10) (1997) 851–856.
[44] F.W. Flocker, L.R. Dharani, Modeling fracture in laminated architectural glass subject low velocity impact, J. Mater. Sci. 32 (1997) 2587–2594, https://doi.org/
10.1023/A:101869890.
[45] F.W. Flocker, L.R. Dharani, Modelinginterplydebonding in laminated architectural glass subject to low velocity impact, Struct. Eng. Mech. 6 (5) (1998)
485–496, https://doi.org/10.12989/sem.1998.6.5.485.
[46] F.W. Flocker, L.R. Dharani, Low velocity impact resistance of laminated architectural glass, J. Archit. Eng. 4 (1) (1998) 12–17.
[47] L. Dharani, F. Ji, Dynamic Analysis of Normal Impact of Occupant head on Laminated Glass, SAE Technical Paper 980862, (1998), https://doi.org/10.4271/
980862.
[48] F.S. Ji, L.R. Dharani, R.A. Behr, Damage probability in laminated glass subjected to low velocity small missile impacts, J. Mater. Sci. 33 (19) (1998) 4775–4782,
https://doi.org/10.1023/A:100445762.
[49] H.S. Norville, K.W. King, J.L. Swofford, Behavior and strength of laminated glass, J. Eng. Mech. 124 (1) (1998) 46–53.
[50] A.V. Duser, A. Jagota, S.J. Bennison, Analysis of glass/polyvinyl butyral laminates subjected to uniform pressure, J. Eng. Mech. 125 (1999) 435–442.
433
[51] S.J. Bennison, A. Jagota, C.A. Smith, Fracture of glass/ polyvinyl butyral (Butacite) laminates in biaxial flexure, J. Am. Ceram. Soc. 82 (7) (1999) 1761–1770.
[52] A. Jagota, S.J. Bennison, C.A. Smith, Analysis of a compressive shear test for adhesion between elastomeric polymers and rigid substrates, Int. J. Fract. 104 (2)
(2000) 105–130, https://doi.org/10.1023/A:100761710.
[53] P. Rahul-Kumar, A. Jagota, S.J. Bennison, S. Saigal, Interfacial failures in a compressive shear strength test of glass/polymer laminates, Int. J. of Solids Struct.
37 (48) (2000) 7281–7305.
[54] S. Muralidhar, A. Jagota, S.J. Bennison, S. Saigal, Mechanical behaviour in tension of cracked glass bridged by an elastomeric ligament, Acta Mater. 48 (18)
(2000) 4577–4588.
[55] M. Seshadri, S.J. Bennison, A. Jagota, S. Saigal, Mechanical response of cracked laminated plates, Acta Mater. 50 (18) (2002) 4477–4490.
[56] L.R. Dharani, J. Wei, F.S. Ji, Failure analysis of laminated architectural glass panels subjected to blast loading, WIT Trans. Built Environ. 63 (2002) 10, https://
doi.org/10.2495/SU020041.
[57] M.A. Asık, Laminated glass plates: revealing of nonlinear behaviour, Comput. Struct. 81 (2003) 2659–2671.
[58] P.A. Du Bois, S. Kolling, W. Fassnacht, Modelling of safety glass for crash simulation, Comput. Mater. Sci. 28 (3–4) (2003) 675–683, https://doi.org/10.1016/j.
commatsci.2003.08.023.
[59] L.R. Dharani, J. Wei, Dynamic response of laminated glass under blast loading: effect of negative phase, in: N. Jones, C.A. Brebbia (Eds.), Structures Under
Shock and Impact VIII, Vol 73 WIT Press, Southampton, UK, 2004, pp. 181–190, , https://doi.org/10.2495/SU040181.
[60] J. Wei, L.R. Dharani, Fracture mechanics of laminated glass subjected to blast loading, Theor. Appl. Fract. Mec. 44 (2) (2005) 157–167, https://doi.org/10.
1016/j.tafmec.2005.06.004.
[61] M.Z. Asık, S. Tezcan, A mathematical model for the behavior of laminated glass beams, Comput. Struct. 83 (2005) 1742–1753, https://doi.org/10.1016/j.
compstruc.2005.02.020.
[62] J. Wei, M.S. Shetty, L.R. Dharani, Stress characteristics of a laminated architectural glazingsubjected to blast loading, Comput. Struct. 84 (2006) 699–707,
https://doi.org/10.1016/j.compstruc.2005.11.007.
[63] I.V. Ivanov, Analysis, modelling, and optimization of laminated glasses as plane beam, Int. J. Solids Struct. 43 (2006) 6887–6907, https://doi.org/10.1016/j.
ijsolstr.2006.02.014.
[64] J. Wei, L.R. Dharani, Response of laminated architectural glazing subjected to blast loading, Int. J. Impact Eng. 32 (12) (2006) 2032–2047, https://doi.org/10.
1016/j.ijimpeng.2005.05.012.
[65] J. Wei, M.S. Shetty, L.R. Dharani, Failure analysis of architectural glazing subjected to blast loading, Eng. Fail. Anal. 13 (7) (2006) 1029–1043, https://doi.org/
10.1016/j.engfailanal.2005.07.010.
[66] J.J.W. Lee, H. Chai, I.K. Lloyd, B.R. Lawn, Crack propagation across an adhesive interlayer in flexural loading, Scr. Mater. 57 (2007) 1077–1080, https://doi.
org/10.1016/j.scriptamat.2007.08.031.
[67] M. Timmel, S. Kolling, P. Osterrieder, P.A. Du Bois, A finite element model for impact simulation with laminated glass, Int. J. Impact Eng. 34 (2007)
1465–1478, https://doi.org/10.1016/j.ijimpeng.2006.07.008.
[68] Y. Koutsawa, E.M. Daya, Static and free vibration analysis of laminated glass beam on viscoelastic supports, Int. J. Solids Struct. 44 (2007) 8735–8750, https://
doi.org/10.1016/j.ijsolstr.2007.07.009.
[69] J. Belis, J. Depauw, D. Callewaert, D. Delince, R.V. Impe, Failure mechanisms and residual capacity of annealed glass/SGP laminated beams at room tem-
perature, Eng. Fail. Anal. 16 (2009) 1866–1875, https://doi.org/10.1016/j.engfailanal.2008.09.023.
[70] J. Xu, Y. Li, Crack analysis in PVB laminated windshield impacted by pedestrian head in traffic accident, Int. J. Crashworthiness. 14 (1) (2009) 63–71, https://
doi.org/10.1080/13588260802462427.
[71] J. Xu, Y. Li, X. Chen, Y. Yan, D. Ge, M. Zhu, B. Liu, Characteristics of windshield cracking upon low-speed impact: numerical simulation based on the extended
finite element method, Comput. Mater. Sci. 48 (3) (2010) 582–588, https://doi.org/10.1016/j.commatsci.2010.02.026.
[72] J. Xu, Y. Li, B. Liu, M. Zhu, D. Ge, Experimental study on mechanical behavior of PVB laminated glass under quasi-static and dynamic loadings, Compos. Part B:
Eng. 42 (2) (2011) 302–308, https://doi.org/10.1016/j.compositesb.2010.10.009.
[73] P. Kumar, A. Shukla, Dynamic response of glass panels subjected to shock loading, J. Non-Cryst. Solids. 357 (24) (2011) 3917–3923, https://doi.org/10.1016/j.
jnoncrysol.2011.08.009.
[74] T. Pyttel, H. Liebertz, J. Cai, Failure criterion for laminated glass under impact loading and its application in finite element simulation, Int. J. Impact Eng. 38 (4)
(2011) 252–263, https://doi.org/10.1016/j.ijimpeng.2010.10.035.
[75] J. Xu, Y. Sun, B. Liu, M. Zhu, X. Yao, Y. Yan, Y. Li, X. Chen, Experimental and macroscopic investigation of dynamic crack patterns in PVB laminated glass sheets
subject to light-weight impact, Eng. Failure Anal. 18 (6) (2011) 1605–1612, https://doi.org/10.1016/j.engfailanal.2011.05.004.
[76] M. Larcher, G. Solomos, F. Casadei, N. Gebbeken, Experimental and numerical investigations of laminated glass subjected to blast loading, Int. J. Impact Eng. 39
(1) (2012) 42–50, https://doi.org/10.1016/j.ijimpeng.2011.09.006.
[77] L. Galuppi, G. Royer-Carfagni, Laminated beams with viscoelastic interlayer, Int. J. Solids Struct. 49 (18) (2012) 2637–2645, https://doi.org/10.1016/j.
ijimpeng.2011.09.006.
[78] S.H. Schulze, M. Pander, K. Naumenko, H. Altenbach, Analysis of laminated glass beams for photovoltaic applications, Int. J. of Solids Struct. 49 (15–16) (2012)
2027–2036, https://doi.org/10.1016/j.ijsolstr.2012.03.028.
[79] P.A. Hooper, R.A.M. Sukhram, B.R.K. Blackman, J.P. Dear, On the blast resistance of laminated glass, Int. J. Solids Struct. 49 (6) (2012) 899–918, https://doi.
org/10.1016/j.ijsolstr.2011.12.008.
[80] J. Ge, G.Q. Li, S.W. Chen, Theoretical and experimental investigation on fragment behavior of architectural glass panel under blast loading, Eng. Failure Anal
26 (2012) 293–303, https://doi.org/10.1016/j.engfailanal.2012.07.022.
[81] X. Zhang, H. Hao, G. Ma, Development of PI curve for laminate glass windows, Australasian Structural Engineering Conference 2012: The Past, Present and
Future of Structural Engineering, 2012, pp. 534–541.
[82] X. Zhang, Y. Zou, H. Hao, X. Li, G. Ma, K. Liu, Laboratory test on dynamic material properties of annealed float glass, Int. J. Protect. Strut. 3 (4) (2012) 407–430,
https://doi.org/10.1260/2041-4196.3.4.407.
[83] X. Zhang, H. Hao, G. Ma, Laboratory test and numerical simulation of laminated glass window vulnerability to debris impact, Int. J. Impact Eng. 55 (2013)
49–62, https://doi.org/10.1016/j.ijimpeng.2013.01.002.
[84] L. Galuppi, G. Royer-Carfagni, The design of laminated glass under time-dependent loading, Int. J Mech. Sci. (2013) 67–75, https://doi.org/10.1016/j.ijmecsci.
2012.12.019.
[85] T. Serafinavicius, J.P. Lebet, C. Louter, T. Lenkimas, A. Kuranovas, Long-term laminated glass four point bending test with PVB, EVA and SG interlayers at
different temperatures, Procedia Eng. 57 (2013) 996–1004, https://doi.org/10.1016/j.proeng.2013.04.126.
[86] Y. Peng, J. Yang, C. Deck, R. Willinger, Finite element modeling of crash test behavior for windshield laminated glass, Int. J. Impact Eng. 57 (2013) 27–35,
https://doi.org/10.1016/j.ijimpeng.2013.01.010.
[87] X. Zhang, H. Hao, G. Ma, Parametric study of laminated glass window response to blast loads, Eng. Struct. 56 (2013) 1707–1717, https://doi.org/10.1016/j.
engstruct.2013.08.007.
[88] H.D. Hidallana-Gamage, D.P. Thambiratnam, N.J. Perera, Computational analysis of laminated glass panels under blast loads: a comparison of two dimensional
and three dimensional modelling approaches, Int. J. Eng. Sci. 2 (8) (2013) 69–79.
[89] J. Chen, J. Xu, X. Yao, X. Xu, B. Liu, Y. Li, Different driving mechanisms of in-plane cracking on two brittle layers of laminated glass, Int. J. of Impact Eng. 69
(2014) 80–85, https://doi.org/10.1016/j.ijimpeng.2014.02.014.
[90] W. Gao, M. Zang, The simulation of laminated glass beam impact problem by developing fracture model of spherical DEM, Eng. Anal. Bound. Elem. 42 (2014)
2–7, https://doi.org/10.1016/j.enganabound.2013.11.011.
[91] H.D. Hidallana-Gamage, D.P. Thambiratnam, N.J. Perera, Numerical modelling and analysis of the blast performance of laminated glass panels and the influence of
material parameters, Eng. Fail. Anal. 45 (2014) 65–84.
[92] H.D. Hidallana-Gamage, D.P. Thambiratnam, N.J. Perera, Influence of structural sealant joints on the blast performance of laminated glass panels, J. Perform.
Constr. Fac. 29 (6) (2014).
[93] P. Foraboschi, Experimental characterization of non-linear behavior of monolithic glass, Int. J. Non-Linear Mech. 67 (2014) 352–370.
[94] H.D. Hidallana-Gamage, D.P. Thambiratnam, N.J. Perera, Influence of interlayer properties on the blast performance of laminated glass panels, Constr. Build.
434
Mater. 98 (2015) 502–518, https://doi.org/10.1016/j.conbuildmat.2015.08.129.

[95] S. Chen, M. Zang, W. Xu, A three-dimensional computational framework for impact fracture analysis of automotive laminated glass, Comput. Methods Appl.
Mech. Eng. 294 (2015) 72–99, https://doi.org/10.1016/j.cma.2015.06.005.
[96] J. Jaskowiec, Numerical modeling mechanical delamination in laminated glass by XFEM, Procedia Eng. 108 (2015) 293–300, https://doi.org/10.1016/j.
proeng.2015.06.150.
[97] D. Baraldi, A. Cecchi, P. Foraboschi, Broken tempered laminated glass: Non-linear discrete element modeling, Compos. Struct. 140 (2016) 278–295, https://doi.
org/10.1016/j.compstruct.2015.12.050.
[98] J. Pelfrene, J. Kuntsche, S. Van Dam, W. Van Paepegem, J. Schneider, Critical assessment of the post-breakage performance of blast loaded laminated glazing:
experiments and simulations, Int. J. Impact Eng. 88 (2016) 61–71, https://doi.org/10.1016/j.ijimpeng.2015.09.008.
[99] S. Chen, M. Zang, D. Wang, Z. Zheng, C. Zhao, Finite element modelling of impact damage in polyvinyl butyral laminated glass, Compo. Struct. 138 (2016)
1–11, https://doi.org/10.1016/j.compstruct.2015.11.042.
[100] L. Biolzi, M. Orlando, L.R. Piscitelli, P. Spinelli, Static and dynamic response of progressively damaged ionoplast laminated glass beams, Compo. Struct. 157
(2016) 337–347, https://doi.org/10.1016/j.compstruct.2016.09.004.
[101] W. Gao, J. Xiang, S. Chen, S. Yin, M. Zang, X. Zheng, Intrinsic cohesive modelling of impact fracture behavior of laminated glass, Mater. Des. 127 (2017)
321–335, https://doi.org/10.1016/j.matdes.2017.04.059.
[102] X. Xu, J. Xu, J. Chen, P. Li, B. Liu, Y. Li, Investigation of dynamic multi-cracking behavior in PVB laminated glass plates, Int. J. Impact Eng. 100 (2017) 62–74,
[103] H.D. Hidallana-Gamage, D.P. Thambiratnam, N.J. Perera, Influence of the negative phase and support flexibility on the blast response of laminated glass panels,
Constr. Build. Mater. 154 (2017) 462–481, https://doi.org/10.1016/j.conbuildmat.2017.07.231.
[104] C. Xu, Y. Yuan, C. Zhao, P.J. Tan, X. Xu, Y. Li, Dynamic crack-interface interactions in SGP laminated glass: An experimental investigation, Mech. of Mater. 122
(2018) 76–84.
[105] J.N. Reddy, Mechanics of Laminated Composite Plates and Shells - Theory and Analysis, second ed., CRC Press, Boca Raton, 2004.
[106] E. Carrera, Theories and finite elements for laminated, anisotropic, composite plates and shells, Arch. Comput. Methods Eng. 9 (2) (2002) 87–140, https://doi.
org/10.1007/BF02736649.
[107] G. Kirchhoff, On the balance and the movement of an elastic cries, J Pure Appl. Math. 40 (1850) 51–58.
[108] A.E.H. Love, The Mathematical Theory of Elasticity, Cambridge University Press, Cambridge, 1934.
[109] I.V. Ivanov, D. Velchev, T. Sadowski, M. Knec, Computational models of laminated glass plate under transverse static loading, in: H. Altenbach, V.A. Eremeyey
(Eds.), s Shell-like Structures, Springer, Berlin, Heidelberg, 2011, pp. 469–490, , https://doi.org/10.1007/978-3-642-21855-2_30.
[110] V.V. Nishawala, A Study of Large Deflection of Beams and Plates, Dissertation The State University of New Jersey, 2011, https://doi.org/10.7282/T3WW7H91.
[111] J. Xu, Y. Li, Study of damage in windshield glazing subject to impact by a pedestrian's head, Proc. Inst. Mech. Eng. 223 (1) (2009) 77–84.
[112] J. Xu, Y. Li, G. Lu, W. Zhou, Reconstruction model of vehicle impact speed in pedestrian-vehicle accident, Int. J. Impact Eng. 36 (6) (2009) 783–788.
[113] Y. Frostig, Buckling of sandwich panels with a transversely flexible core, high-order theory, Int. J. Solids Struct. 35 (1998) 183–204, https://doi.org/10.1016/
S0020-7683(97)00078-4.
[114] J. Hohe, L. Librescu, Advances in the structural modeling of elastic sandwich panels, Mech. Adv. Mater. Struc. 11 (2004) 395–424, https://doi.org/10.1080/
15376490490451561.
[115] J. Hohe, L. Librescu, S.Y. Oh, Dynamic buckling of flat and curved sandwich panels with transversely compressible core, Compos. Struct. 74 (2006) 10–24,
https://doi.org/10.1016/j.compstruct.2005.03.003.
[116] E. Reissner, A consistent treatment of transverse shear deformation in laminated anisotropic plates, AIAA J 10 (5) (1972) 716–718, https://doi.org/10.2514/3.
50194.
[117] D.D. Fox, Transverse shear and normal stresses in nonlinear shell theory, Comput. Struct. 75 (2000) 313–319, https://doi.org/10.1016/S0045-7949(99)
00139-X.
[118] M. Aydogdu, A new shear deformation theory for laminated composite plates, Compos. Struct. 89 (2009) 94–101, https://doi.org/10.1016/j.compstruct.2008.
07.008.
[119] E. Reissner, The effect of transverse shear deformation on the bending of elastic plates, J. Appl. Mech. T ASME 12 (1945) 69–76.
[120] R.D. Mindlin, Influence of rotatory inertia and shear in flexural motions of isotropic elastic plates, J Appl. Mech. T ASME 18 (1951) 1031–1036.
[121] J.N. Reddy, Energy and Variational Methods in Applied Mechanics, Wiley, London, 1984.
[122] J.N. Reddy, A.A. Khdeir, L. Librescu, Levy type solutions for symmetrically laminated rectangular plates using first order shear deformation theories, J. Appl.
Mech. T ASME 54 (5) (1987) 740–742, https://doi.org/10.1115/1.3173104.
[123] K.W. Liew, Solving the vibration of thick symmetric laminates by Reissner/Mindlin plate theory and the p-Ritz method, J. Sound Vib. 198 (1996) 343–360,
https://doi.org/10.1006/jsvi.1996.0574.
[124] H. Altenbach, J. Altenbach, K. Naumenko, Level Surface Supports: The Basics of Modeling and Calculation of Disks and Plates, Springer, 1998.
[125] R. Szilard, Theories and Applications of Plate Analysis, John Wiley & Sons, Hoboken, New Jersey, 2004.
[126] M. Levinson, An accurate, simple theory of the statics and dynamics of elastic plates, Mech. Res. Commun. 7 (6) (1980) 343–350, https://doi.org/10.1016/
0093-6413(80)90049-X.
[127] M.V.V. Murthy, An improved transverse shear deformation theory for laminated anisotropic plates, NASA Technical paper No 1903, 1981.
[128] J.N. Reddy, Exact solutions of moderately thick laminated shells, J. Eng. Mech. ASCE 110 (5) (1984) 794–809.
[129] R.L. Onsy, T. Tran-Cong, Finite strip elements for laminated composite plates with transverse shear strain discontinuities, Compos. Struct. 56 (3) (2002) 249–258,
https://doi.org/10.1016/S0263-8223(02)00010-7.
[130] F.A. Moleiro, C.M. MotaSoares, C.A. MotaSoares, J.N. Reddy, Mixed least-squares finite element models for static and free vibration analysis of laminated
composite plates, Comput. Methods Appl. Mech. Eng. 198 (2009) 1848–1856, https://doi.org/10.1016/j.cma.2008.12.023.
[131] J. Eisentrager, K. Naumenko, H. Altenbach, H. Koppe, Application of the first-order shear deformation theory to the analysis of laminated glasses and pho-
tovoltaic panels, Int. J. Mech. Sci. 96–97 (2015) 163–171, https://doi.org/10.1016/j.ijmecsci.2015.03.012.
[132] P. Krawczyk, F. Frey, A.P. Zielinski, Large deflections of laminated beams with interlayer slips: Part 1: model development, Eng. Computations 24 (1) (2007)
17–32, https://doi.org/10.1108/02644400710718556.
[133] H. Nguyen-Van, N. Mai-Duy, T. Tran-Cong, Free vibration analysis of laminated plate/shell structures based on FSDT with a stabilized nodal-integrated
quadrilateral element, J. Sound Vib. 313 (2008) 205–223, https://doi.org/10.1016/j.jsv.2007.11.043.
[134] A.J.M. Ferreira, L.M.S. Castro, S.A. Bertoluzza, High order collocation method for the static and vibration analysis of composite plates using a first-order theory,
Compos. Struct. 89 (3) (2009) 424–432, https://doi.org/10.1016/j.compstruct.2008.09.006.
[135] J. Cugnoni, T. Gmur, A. Schorderet, Identification by modal analysis of composite structures modelled with FSDT and HSDT laminated shell finite elements,
Compos. Part A: Appl. S35 (7–8) (2004) 977–987, https://doi.org/10.1016/j.compositesa.2004.01.014.
[136] C.M. Wang, J.N. Reddy, K.H. Lee, Shear Deformable Beams and Plates: Relationships with Classical Solutions, Elsevier, 2000.
[137] J.N. Reddy, A simple higher-order theory for laminated composite plates, J. Appl. Mech. 51 (1984) 745–752, https://doi.org/10.1115/1.3167719.
[138] M. Touratier, An efficient standard plate theory, Int. J. Eng. Sci. 29 (1991) 901–916, https://doi.org/10.1016/0020-7225(91)90165-Y.
[139] H. Matsunaga, Vibration and stability of cross-ply laminated composite plates according to a global higher-order plate theory, Compos. Struct. 48 (2000)
231–244, https://doi.org/10.1016/S0263-8223(99)00110-5.
[140] T. Kant, K. Swaminathan, Analytical solutions for free vibration of laminated composite and sandwich plates based on a higher order refined theory, Compos.
Struct. 53 (2001) 73–85, https://doi.org/10.1016/S0263-8223(00)00180-X.
[141] L. Liu, L.P. Chua, D.N. Ghista, Mesh-free radial basis function method for static, free vibration and buckling analysis of shear deformable composite laminates,
Compos. Struct. 78 (2007) 58–69, https://doi.org/10.1016/j.compstruct.2005.08.010.
[142] H. Matsunaga, Assessment of a global higher-order deformation theory for laminated composite and sandwich plates, Compos. Struct. 56 (3) (2002) 279–291,
https://doi.org/10.1016/S0263-8223(02)00013-2.
[143] K. Swaminathan, D. Ragounadin, Analytical solutions using a higher-order refined theory for the static analysis of antisymmetric angle-ply composite and
sandwich plates, Compos. Struct. 64 (2004) 405–417.
435
[144] C.A. Shankara, N.G.R. Iyengar, A C0 element for the free vibration analysis of laminated composite plates, J. Sound Vib. 191 (1996) 721–738, https://doi.org/
10.1006/jsvi.1996.0152.
[145] W. Zhen, C. Wanji, A global-local higher order theory including interlaminar stress continuity and C0 plate bending element for cross-ply laminated composite
plates, Comput. Mech. 45 (5) (2010) 387–400, https://doi.org/10.1007/s00466-009-0460-x.
[146] W. Zhen, C. Wanji, A C0-type higher-order theory for bending analysis of laminated composite and sandwich plates, Compos. Struct. 92 (2010) 653–661,
[147] W. Zhen, C. Wanji, R. Xiaohui, An accurate higher-order theory and C0 finite element for free vibration analysis of laminated composite and sandwich plates,
Compos. Struct. 92 (2010) 1299–1307, https://doi.org/10.1016/j.compstruct.2009.11.011.
[148] A.K. Noor, W.S. Burton, Assessment of shear deformation theories for multilayered composite plates, Appl. Mech. Rev. 41 (1) (1989) 113.
[149] A.K. Noor, W.S. Burton, Assessment of computational models for multilayered anisotropic plates, Compos. Struct. 14 (1990) 233–265, https://doi.org/10.
1016/0263-8223(90)90050-O.
[150] A.K. Noor, W.S. Burton, C.W. Bert, Computational models for sandwich panels and shells, Appl. Mech. Rev. 49 (3) (1996) 155–199, https://doi.org/10.1115/1.
3101923.
[151] H. Hu, S. Belouettar, M. Potier-Ferry, M. Daya, Review and assessment of various theories for modeling sandwich composites, Compos. Struct. 84 (2008)
282–292, https://doi.org/10.1016/j.compstruct.2007.08.007.
[152] L. Iurlaro, Development of Refined Models for Multilayered Composite and Sandwich Structures: Analytical Formulation, FEM Implementation and
Experimental Assessment, Dissertation Politecnico di Torino, 2015.
[153] M. Weps, K. Naumenko, H. Altenbach, Unsymmetric three-layer laminate with soft core for photovoltaic modules, Compos. Struct. 105 (2013) 332–339,
[154] Y. Frostig, Buckling of sandwich panels with a transversely flexible core, high-order theory, Int. J. Solids Struct. 35 (1998) 183–204, https://doi.org/10.1016/
S0020-7683(97)00078-4.
[155] Y. Frostig, O.T. Thomsen, High-order free vibration of sandwich panels with a flexible core, Int. J. Solids Struct. 41 (2004) 1697–1724, https://doi.org/10.
1016/j.ijsolstr.2003.09.051.
[156] Y. Frostig, O.T. Thomsen, On the free vibration of sandwich panels with a transversely flexible and temperature-dependent core material – Part I: Mathematical
formulation, Compos. Sci. Tech. 69 (2009) 856–862, https://doi.org/10.1016/j.compscitech.2008.03.003.
[157] K. Malekzadeh, M.R. Khalili, R.K. Mittal, Damped vibrations of sandwich plates with a viscoelastic soft flexible core, an improved high-order approach,
Proceeding of 12th International Conference of Mechanical Engineers, Tehran, Iran, 2004.
[158] K. Malekzadeh, M.R. Khalili, R.K. Mittal, Prediction of low velocity impact response of composite sandwich panels using new three degrees-of-freedom model,
Proceeding of 12th International Conference of Mechanical Engineers, Esfahan, Iran, 2005.
[159] P. Foraboschi, Layered plate with discontinuous connection: Exact mathematical model, Compos. Part B Eng. 47 (2013) 365–378, https://doi.org/10.1016/j.
compositesb.2012.11.004.
[160] K. Naumenko, V.A. Eremeyev, A layer-wise theory for laminated glass and photovoltaic panels, Compos. Struct. 112 (2014) 283–291, https://doi.org/10.1016/
j.compstruct.2014.02.009.
[161] J. Eisentrager, K. Naumenko, H. Altenbach, J. Meenen, A user-defined finite element for laminated glass panels and photovoltaic modules based on a layer-wise
theory, Compos. Struct. 133 (2015) 265–277, https://doi.org/10.1016/j.compstruct.2015.07.049.
[162] E. Carrera, Historical review of zig-zag theories for multilayered plates and shells, Appl. Mech. Rev. 56 (2) (2003) 287–308, https://doi.org/10.1115/1.
1557614.
[163] E. Carrera, A. Ciuffreda, A unified formulation to assess theories of multilayered plates for various bending problems, Compos. Struct. 69 (3) (2005) 271–293,
[164] E. Carrera, S. Brischetto, P. Nali, Plates and Shells for Smart Structures: Classical and Advanced Theories for Modelling and Analysis, Wiley, Chichester, 2011.
[165] E.J. Barbero, J.N. Reddy, Modelling of delamination in composite laminates using a layerwise plate theory, Int. J. Solid Struct. 28 (3) (1991) 373–388, https://
doi.org/10.1016/0020-7683(91)90200-Y.
[166] R. Khandan, S. Noroozi, P. Sewell, J. Vinney, The development of laminated composite plate theories: a review, J. Mater. Sci. 47 (2012) 5901–5910, https://
doi.org/10.1007/s10853-012-6329-y.
[167] M. Di Sciuva, A refinement of the transverse shear deformation theory for multilayered orthotropic plates, Proc. AIDAA National Congress, 1983, pp. 84–92.
[168] H. Murakami, Laminated composite plate theory with improved in-plane responses, J. Appl. Mech. 53 (3) (1986) 661–666, https://doi.org/10.1115/1.
3171828.
[169] M. Gherlone, On the use of zigzag functions in equivalent single layer theories for laminated composite and sandwich beams: a comparative study and some
observations on external weak layers, J. Appl. Mech. 80 (6) (2013) 061004, https://doi.org/10.1115/1.4023690.
[170] L. Iurlaro, M. Gherlone, M. Di Sciuva, A. Tessler, Assessment of the Refined Zigzag Theory for bending, vibrations, and buckling of sandwich plates: a
comparative study of different theories, Compos. Struct. 106 (2013) 777–792, https://doi.org/10.1016/j.compstruct.2013.07.019.
[171] L. Iurlaro, M. Gherlone, M. Di Sciuva, Bending and free vibration analysis of functionally graded sandwich plates using the Refined Zigzag Theory, J. Sandw
Struct. Mater. 16 (2014) 669–699.
[172] A. Tessler, M. Di Sciuva, M. Gherlone, Refined zigzag theory for laminated composite and sandwich plates, Technical report TP-2009-215561, NASA Langley
Research Center, Hampton, VA, 2009.
[173] A. Tessler, M. Di Sciuva, M. Gherlone, A homogeneous limit methodology and refinements of computationally efficient zigzag theory for homogeneous, laminated
composite, and sandwich plates, Numer. Meth. Part DE 27 (1) (2011) 208–229, https://doi.org/10.1002/num.20646.
[174] A. Tessler, A multi-scale theory for laminated composite and sandwich beams based on refined zigzag kinematics and mixed-field approximations, NASA/TP
(2014).
[175] M. Gherlone, A. Tessler, M. Di Sciuva, C0 beam elements based on the refined zigzag theory for multilayered composite and sandwich laminates, Compos.
Struct. 93 (11) (2011) 2882–2894, https://doi.org/10.1016/j.compstruct.2011.05.015.
[176] D. Versino, Refined Theories and Discontinuous Galerkin Methods for the Analysis of Multilayered Composite Structures, Dissertation Politecnico di Torino,
Italy, 2012.
[177] D. Versino, M. Gherlone, M. Mattone, M. Di Sciuva, A. Tessler, An efficient, C0 triangular elements based on the Refined Zigzag Theory for multilayered
composite and sandwich plates, Compos. Part B (1) (2013) 218–230, https://doi.org/10.1016/j.compositesb.2012.05.026.
[178] D. Versino, M. Gherlone, M. Di Sciuva, Four-node shell element for doubly curved multilayered composites based on the Refined Zigzag Theory, Compos. Struc.t
118 (2014) 392–402, https://doi.org/10.1016/j.compstruct.2014.08.018.
[179] M. Ganapathi, B.P. Patel, D.P. Makhecha, Nonlinear dynamic analysis of thick composite/sandwich laminates using an accurate higher-order theory, Compos.
Part B Eng 35 (2004) 345–355, https://doi.org/10.1016/S1359-8368(02)00075-6.
[180] S. Kapuria, G.G.S. Achary, An efficient higher-order zigzag theory for laminated plates subjected to thermal loading, Int. J. Solids Struct. 41 (2004) 4661–4684,
https://doi.org/10.1016/j.ijsolstr.2004.02.020.
[181] U. Icardi, Higher-order zig-zag model for analysis of thick composite beams with inclusion of transverse normal stress and sublaminates approximation,
Compos. Part B Eng 32 (2001) 343–354, https://doi.org/10.1016/S1359-8368(01)00016-6.
[182] U. Icardi, A three-dimensional zig-zag theory for analysis of thick laminated beams, Compos. Struct. 52 (2001) 123–135, https://doi.org/10.1016/S0263-
8223(00)00189-6.
[183] U. Icardi, C0 plate element for global/local analysis of multilayered composites based on a 3D zig-zag model and strain energy updation, Int. J. Mech. Sci. 47
(2001) (2001) 1561–1594, https://doi.org/10.1016/j.ijmecsci.2005.06.002.
[184] U. Icardi, Higher-order zig-zag model for analysis of thick composite beams with including transverse normal stress and sublaminates approximation, Compos.
Part B Eng. 32 (2001) 343–354, https://doi.org/10.1016/S1359-8368(01)00016-6.
[185] M.K. Pandit, A.H. Sheikh, B.N. Singh, An improved higher order zigzag theory for the static analysis of laminated sandwich plate with soft core, Finite Elem.
Anal. Des. 44 (2008) 602–610, https://doi.org/10.1016/j.finel.2008.02.001.
[186] S.D. Kulkarni, S. Kapuria, Free vibration analysis of composite and sandwich plates using an improved discrete Kirchhoff quadrilateral element based on third-
order zigzag theory, Comput. Mech. 42 (2008) 803–824, https://doi.org/10.1007/s00466-008-0285-z.
436
[187] S. Brischetto, E. Carrera, L. Demasi, Improved bending analysis of sandwich plates using a zig-zag function, Compos. Struct. 89 (3) (2009) 408–415, https://doi.
org/10.1016/j.compstruct.2008.09.001.
[188] R. Xiaohui, C. Wanji, W. Zhen, A new zig-zag theory and C0 plate bending element for composite and sandwich plates, Arch. Appl. Mech. 81 (2011) 185–197,
https://doi.org/10.1007/s00419-009-0404-0.
[189] R. Xiaohui, C. Wanji, W. Zhen, A C0-type zig-zag theory and finite element for laminated composite and sandwich plates with general configuration, Arch. Appl.
Mech. 82 (2012) 391–406, https://doi.org/10.1007/s00419-011-0563-7.
[190] W.J. Cantwell, J. Morton, The impact resistance of composite materials — a review, Composite 22 (5) (1991) 347–362, https://doi.org/10.1016/0010-
4361(91)90549-V.
[191] M. Richards, R. Clegg, S. Howlett, Ballistic performance assessment of glass laminates through experimental and numerical investigation, 18th International
Symposium and Exhibition on Ballistics San Antonio, Texas USA, 1999, pp. 1123–1130.
[192] D.J. Elder, R.S. Thomson, M.Q. Nguyen, M.L. Scott, Review of delamination predictive methods for low speed impact of composite laminates, Compos. Struct.
66 (1–4) (2004) 677–683, https://doi.org/10.1016/j.compstruct.2004.06.004.
[193] L.J. Deka, S.D. Bartus, U.K. Vaidya, Numerical Modeling of Simultaneous and Sequential Multi-Site Impact Response of S-2 Glass/Epoxy Composite Laminates,
COMPOSITES & POLYCON 2007 American Composites Manufacturers Association, Tampa, FL USA, 2007.
[194] C.J. Hayhurst, R.A. Clegg, The Application of SPH Techniques in AUTODYN-2D to ballistic impact problems, Proc. 16th in ‘Symposium on Ballistics, San
Francisco, 1996, pp. 409–417.
[195] R.A. Clegg, C.J. Hayhurst, Numerical modeling of the compressive and tensile response of brittle materials under high pressure dynamic loading, AIP
Conference Proceedings, Vol. 505 (1) 2000, pp. 321–324, , https://doi.org/10.1063/1.1303483.
[196] J.H. Song, H. Wang, T. Belytschko, A comparative study on finite element methods for dynamic fracture, Computat. Mech. 42 (2) (2008) 239–250, https://doi.
org/10.1007/s00466-007-0210-x.
[197] G. Camacho, M. Ortiz, Computational modelling of impact damage in brittle materials, Int. J. Solids Struct. 33 (1996) 2899–2938, https://doi.org/10.1016/
0020-7683(95)00255-3.
[198] L. Snozzi, J.F. Molinari, A cohesive element model for mixed mode loading with frictional contact capability, Internat. J. Numer. Methods Engrg. 93 (2013)
510–526, https://doi.org/10.1002/nme.4398.
[199] T. Belytschko, T. Black, Elastic crack growth in finite elements with minimal remeshing, Int. J. Numer. Methods Eng. 45 (1999) 601–620.
[200] N. Moes, J. Dolbow, T. Belytschko, A finite element method for crack growth without remeshing, Int. J. Numer. Meth. Eng. 46 (1999) 131–150.
[201] T. Belytschko, H. Chen, J. Xu, G. Zi, Dynamic crack propagation based on loss of hyperbolicity with a new discontinuous enrichment, Int. J. Numer. Meth. Eng.
58 (2003) 1873–1905, https://doi.org/10.1002/nme.941.
[202] X. Zhang, H. Hao, Z. Wang, Experimental study of laminated glass window responses under impulsive and blast loading, Int. J. Impact Eng. 78 (2015) 1–19,
[203] X. Zhang, H. Hao, Z. Wang, Experimental investigation on monolithic tempered glass windows responses to blast loads, Int. J. Protective Struct. 6 (2015)
287–309.
[204] H. Hao, X. Zhang, Analysis and design of glass windows against blast and impact loads, Proceedings of the 11th International Conference on Shock & Impact
Loads on Structures, 2015, pp. 1–32.
[205] X. Zhang, H. Hao, G. Ma, Dynamic material model of annealed soda-lime glass, Int. J. Impact Eng. 77 (2015) 108–119, https://doi.org/10.1016/j.ijimpeng.
2014.11.016.
[206] S. Bennison, J. Sloan, D. Kistunas, P. Buehler, T. Amos, C. Smith, Laminated glass for blast mitigation: role of interlayer properties, Glass Processing Days, 2005.
[207] C. Morison, M. Zobec, A. Frenceschet, The measurement of PVB properties at high strain rates, and their application in the design of laminated glass under
bomb blast, ISIEMS 2007, International Symposium on Interaction of the Effects of Munitions with Structures, 17–21 September, Orlando, US, 2007.
[208] R. Iwasaki, C. Sato, J.L. Lataillade, P. Viot, Experimental study on the interface fracture toughness of PVB (polyvinyl butyral)/glass at high strain rates, Int. J.
Crashworthiness 12 (3) (2007) 293–298, https://doi.org/10.1080/13588260701442249.
[209] S.J. Bennison, C.A. Smith, C.C. Anderson, J.G. Sloan, Expanding bomb blast performance of architectural glass, Glass Processing Days, 2003.
[210] C. Morison, The Resistance of Laminated Glass to Blast Pressure Loading and the Coefficients for Single Degree of Freedom Analysis of Laminated Glass, PhD
thesis Cranfield University, 2010.
[211] S. Chen, M. Zang, D. Wang, S. Yoshimura, T. Yamada, Numerical analysis of impact failure of automotive laminated glass: A review, Compos. Part B: Eng. 122
(2017) 47–60, https://doi.org/10.1016/j.compositesb.2017.04.007.
[212] S. Zhao, L.R. Dharani, X. Liang, L. Chai, S.D. Barbat, Crack initiation in laminated automotive glazing subjected to simulated head impact, Int. J.
Crashworthiness. 10 (3) (2005) 229–236, https://doi.org/10.1533/ijcr.2005.0340.
[213] S. Zhao, L.R. Dharani, L. Chai, S.D. Barbat, Dynamic response of laminated automotive glazing impacted by spherical featureless headform, Int. J.
Crashworthiness 11 (2) (2006) 105–114, https://doi.org/10.1533/ijcr.2005.0386.
[214] S. Zhao, L.R. Dharani, L. Chai, S.D. Barbat, Analysis of damage in laminated automotive glazing subjected to simulated head impact, Eng. Fail. Anal. 13 (4)
(2006) 582–597, https://doi.org/10.1016/j.engfailanal.2004.12.038.
[215] J. Xu, Y. Li, X. Chen, Y. Yan, D. Ge, M. Zhu, B. Liu, Numerical study of PVB laminated windshield cracking upon human head impact, CMC-Comput. Mater. Con.
18 (2) (2010) 183–211.
[216] J.U. Xu, Y.B. Li, Study of damage in windshield glazing subject to impact by a pedestrian's head, in ‘Proceedings of the Institution of Mechanical Engineers, Part
D, J. Automobile Eng. 223 (1) (2009) 77–84.
[217] M. Munsch, N. Bourdet, C. Deck, R. Willinger, Lateral glazing characterization under head impact: experimental and numerical investigation, Proceedings of
The 21st (ESV) International Technical Conference on the Enhanced Safety of Vehicles, Stuttgart, Germany, 2009.
[218] T. Ngo, P. Mendis, A. Gupta, J. Ramsay, Blast loading and blast effects on structures–an overview, Elect. J. Struct. Eng. 7 (2007) 76–91.
[219] G.F. Kinney, K.J. Graham, Explosive Shocks, Air, Macmillan, New York, 1962.
[220] J.O. Haliquist, Theoretical Manual for DYNA3D, Lawrence Livermore National Laboratory, UCRL-521 12 (1983).
[221] S. Zhao, L. Dharani, X. Liang, Analysis of damage in laminated architectural glazing subjected to blast loading, Adv. Struct. Eng. 11 (1) (2008) 129–134,
https://doi.org/10.1260/136943308784069432.
[222] H.D. Hidallana-Gamage, D.P. Thambiratnam, N.J. Perera, Design guidance for blast-resistant glazing, J. Architect. Eng. 21 (3) (2015) 04015003.
[223] K. Spiller, J.A. Packer, M.V. Seica, D.Z. Yankelevsky, Prediction of annealed glass window response to blast loading, Int. J. Impact Eng. 88 (2016) 189–200.
437

Laminated

Diunggah oleh

Informasi Dokumen

Judul Asli

Hak Cipta

Format Tersedia

Bagikan dokumen Ini

Bagikan atau Tanam Dokumen

Opsi Berbagi

Apakah menurut Anda dokumen ini bermanfaat?

Apakah konten ini tidak pantas?

Hak Cipta:

Format Tersedia

Laminated

Diunggah oleh

Hak Cipta:

Format Tersedia

Engineering Failure Analysis 94 (2018) 412–437

Contents lists available at ScienceDirect

Engineering Failure Analysis

Applications of ﬁnite element modelling in failure analysis of

2. Brief overview of modelling of laminated glass

3. Theoretical approaches for FE modelling

3.1. Classical plate theory (CLPT)

Uz(x,z,t) = w(x,t) , (2)

Author name Specimen type Research Analysis Approach used

3.2. First order shear deformation theory (FSDT)

Uz(x,z,t) = w(x,t) (7)

3.3. Higher order shear deformation theory (HSDT)

Uz(x,z,t) = w(x,t) (9)

Uz(x,z,t) = w(x,t) (11)

3.4. Layerwise shear deformation theory

Ukα (x,z,t) = u kα (x,t) + Z∅kα (x,t) + …Z nωkα (x,t) (12)

Ukα (x,z,t) = w(x,t) (13)

3.5. Zig zag models

Ukα (x,z,t) = u α(x,t) + Z∅α(x,t) + …Z nωkα (x,t) + Πkα (z)Ψα(x,t) (14)

Ukα (x,z,t) = w(x,t) (15)

4. FE modelling of laminated glass

4.1. Fundamentals of FE modelling

Elements 3D solid Shell Two superimposed shell elements

4.1.1. Simulation of failure process

σ eff = σ 2n + β−2σ 2t ≥ σfr (20)

4.1.2. Material models

⎝ (1 + v)(1 − 2v) ⎠ (1 + v) (27)

4.2. Finite element simulation of head impact on laminated windshields

4. Young' modulus of the modified glass EG, modified=2EG−EPVB, modified

4.3. Finite element simulation of blast loading

Modulus of elasticity(E) (GPa) 70 20 6.5

Io = 113.487 − 0.008L + 118.380dPVB + 44.645dg + 0.579ft,glass (38)

Mater. 98 (2015) 502–518, https://doi.org/10.1016/j.conbuildmat.2015.08.129.

Anda mungkin juga menyukai