International Journal of Mechanical Sciences 280 (2024) 109620

Contents lists available at ScienceDirect

International Journal of Mechanical Sciences

journal homepage: www.elsevier.com/locate/ijmecsci

3D electro-elastic static analysis of advanced plates and shells

S. Brischetto ∗, D. Cesare
Department of Mechanical and Aerospace Engineering, Politecnico di Torino, Torino, Italy

ARTICLE INFO ABSTRACT

Keywords: The proposed three-dimensional (3D) coupled electro-elastic shell model allows the static analysis of smart
Smart structures structures embedding classical piezoelectric and functionally graded piezoelectric layers. Plates, cylindrical
Actuator configuration and spherical panels are investigated in both sensor and actuator configurations. The primary variables of
Sensor configuration
the coupled model are displacement components and the electric potential. Therefore, displacements, stresses,
Classical and functionally graded plates
strains, electric potential and electric displacements are calculated for smart structures used as sensors (applied
Piezoelectric materials
Plates and shells
mechanical load) and smart structures used as actuators (applied electric potential). In the case of spherical
Electro-elastic coupling shells, 3D equilibrium equations are coupled with the 3D divergence electric displacement equation. The
3D model obtained coupled system is also valid for cylindrical shells and plates thanks to the use of a mixed curvilinear
Shell model orthogonal reference system. The partial differential governing equations are solved using the Navier harmonic
Exponential matrix method form and the exponential matrix method for simply supported structures. Preliminary assessments are proposed
to validate the model and further benchmarks are analyzed to discuss the effects connected with the thickness
ratio, the lamination scheme, the load conditions, the geometry of the structures and the employed materials.
The main innovation point of the proposed model is the possibility to analyze several geometries including
different materials and applied loads by means of a unique and general formulation where partial differential
equations are solved in closed form. The use of functionally graded piezoelectric materials allows to eliminate
all stress component discontinuities at each layer interface.

1. Introduction is general because governing equations are able to describe several ge-
ometries and several lamination schemes including different materials
Future aircraft and spacecraft will use conventional and unconven- in a unified manner. Works proposed in literature are not so general
tional multilayered structures. An example of unconventional struc- because they are developed for particular geometries, materials and
tures are those employing piezoelectric materials, which are commonly lamination schemes. In fact, we found 3D or 2D models developed
used in so-called smart structures as sensor or actuators depending on in a specified manner for classical piezoelectric plates, classical piezo-
whether they use the direct piezoelectric effect or the converse piezo- electric shells, functionally graded piezoelectric plates or functionally
electric effect. These piezoelectric materials could be also functionally graded piezoelectric shells.
graded in order to have a continuous variation of physical properties A large amount of papers are devoted to the analysis of plates or
in a particular direction to combine the desirable properties of the con- shells including classical piezoelectric layers. Alessandroni et al. [5]
stituent phases to obtain a superior performance, avoiding the problem proposed a dynamic passive controller for thin plate vibrations, the
of interfacial stresses typical of classical multilayered configurations. governing equations were derived via the Lagrangian functional and
Appropriate three-dimensional (3D) and two-dimensional (2D) models the Kirchhoff–Love theory. Ballhause et al. [6] developed refined two-
must be developed for the analysis of smart structures and in particular dimensional plate models in layer-wise and equivalent single layer
they must consider an adequate electro-elastic coupling as described forms for the electro-mechanical analysis of multilayered plates em-
in [1,2]. All these models are applicable if elastic and electric material bedding piezo-layers. A three-dimensional mixed variational principle
properties are appropriately defined, some examples of these studies was used in [7] to derive a Kth-order two-dimensional linear theory for
were proposed in [3,4]. the analysis of anisotropic homogeneous piezoelectric plates, Legendre
The present paper proposes a closed-form three-dimensional shell polynomials were employed as basis functions. The two-dimensional
model for the coupled electro-elastic analysis of plates, cylinders, cylin- closed-form solution for free-vibration analyses of simply-supported
drical panels and spherical panels embedding piezoelectric, function- piezoelectric sandwich plates proposed in [8] was a full layer-wise
ally graded piezoelectric and composite materials. The proposed model

∗ Correspondence to: Department of Mechanical and Aerospace Engineering, Politecnico di Torino, corso Duca degli Abruzzi 24, 10129 Torino, Italy.
E-mail address: [email protected] (S. Brischetto).

https://doi.org/10.1016/j.ijmecsci.2024.109620
Received 18 June 2024; Received in revised form 30 July 2024; Accepted 31 July 2024
Available online 5 August 2024
0020-7403/© 2025 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
S. Brischetto and D. Cesare International Journal of Mechanical Sciences 280 (2024) 109620

first-order shear deformation theory employing a through-the-thickness and a general 3D finite element formulation. Saravanos et al. [30]
quadratic electric potential. The converse piezoelectric effect was used presented analyses for laminated composite plates with piezoelectric
in [9] to suppress the vibrations of an orthotropic plate stiffened actuators and sensors, finite-element formulations were developed for
with piezoelectric actuators. Exact three-dimensional analyses were the quasi-static and dynamic analysis via layerwise representations
proposed in [10] for natural frequencies and mode shapes of simply of displacements and electric potential. A finite element model for
supported rectangular piezoelectric plates in both closed and open the analysis of fiber-reinforced laminated composite plates embedding
circuit configurations. Deu and Benjeddou [11] presented a three- piezoelectric layers and subjected to mechanical loadings and/or elec-
dimensional exact mixed state space solution and a parametric analysis tric potentials was presented in [31]. A single layer theory based on
for the free-vibration of simply-supported laminates with embedded Mindlin’s hypotheses was used for deformations; a layer-wise theory
transverse shear mode piezoceramic layers. The free vibration analysis was applied for the electric potential. An accurate finite element model
of piezoelectric coupled annular plates using the Kirchhoff and Mindlin for laminated piezoelectric composite plates in cylindrical bending
plate models was proposed in [12] where the distribution of electric was developed in [32] for free vibration analyses. Zhang et al. [33]
potential along the thickness direction was a sinusoidal function that extended the Hellinger–Reissner variational principle for elasticity to
satisfied the Maxwell static electricity equation. Multilayer plates with the systems of piezoelectric multi-materials; the numerical solutions
piezoelectric layers were modeled in [13] considering the deformation of the governing equations were proposed for two-dimensional elec-
of the adhesive layers. Special finite elements, that used linear Lagrange tromechanical coupling problems of plates embedding piezoelectric
and conforming Hermite type interpolation functions, were developed. patches. Heyliger [34] presented three-dimensional exact solutions for
A 2D approach for piezoelectric laminated plates was proposed in [14] a hybrid laminated plate containing both piezoelectric and elastic
where displacement was modeled via a higher order shear deformation layers under applied surface tractions and surface potential. The same
theory and the electric potential was a function of the thickness coor- author also proposed in [35,36] the same methodology used to solve 3D
dinate. The same authors [15] also proposed an efficient and accurate equations of linear piezoelectricity for the static response of finite lam-
approach to piezoelectric bimorphs, based on a refined expansion of inated piezoelectric cylinders and plates with simply supported sides.
the elastic displacement and electric potential. The field approximation Even if classical piezoelectric layers are considered, works devoted
accounted for a shear correction and a layerwise modeling was con- to shell analyses are less numerous; some examples are given in the
sidered for the electric potential. A higher order displacement theory following. An exact elasticity solution for an orthotropic cylindrical
was used in [16] to define the mechanical displacement field in the
shell with piezoelectric layers was obtained in [37]. The stress and
case of coupled and uncoupled thermo-electro-mechanical analysis of
displacement distributions were presented by considering the influence
composite plates embedding piezoelectric actuators. He et al. [17]
of the piezoelectric layers on the mechanical behavior of structures
presented a uniformizing method for the free vibration analysis of
in the case of both direct and converse piezoelectric effects. Dumir
metal–piezoceramic composite thin plates via a classical thin plate
et al. [38] presented an exact piezoelastic solution of an infinitely
theory. Governing equations for the linear response of piezo-thermo-
long, simply-supported, orthotropic, piezoelectric, radially polarized,
elastic plates were outlined in [18], the Hamilton principle and the
circular cylindrical shell panel in cylindrical bending under pressure
finite element method were employed using linear shape functions
and electrostatic excitations. A semi-analytical axisymmetric shell finite
and the first-order shear deformation theory. A third-order zig-zag
element model with embedded and/or surface bonded piezoelectric
approximation for the in-plane displacements and a sublayer piecewise
ring actuators and/or sensors for active damping vibration control of
linear approximation for the electric potential were developed in [19]
the structure was presented in [39]. The mixed finite element approach
for the static analysis of hybrid cross-ply plates with some piezoelectric
combined the equivalent single-layer higher order shear deformation
layers. The same method was extended to the dynamic analysis of
theory for the mechanical behavior with a layerwise discretization
piezoelectric composite and sandwich beams with damping in [20].
for the electric potential. A semi-analytical finite element formula-
A finite element formulation based on Reissner–Mindlin assumptions
tion was presented in [40] for the analysis of active controlled shells
was developed in [21] for the accurate analysis of thin and moderately
thick piezoelectric plates. Shear locking was eliminated via the mixed of revolution taking into account a steady state temperature field
interpolation of tensorial components approach. The active control of over the piezoelectric and elastic structural continuum. First order
laminated composite plates with piezoelectric sensor/actuator patches, shear deformation theory was used to model the elastic shells of rev-
using an efficient mesh-free method, was proposed in [22] employing olution. Comparisons between classical lamination shell theory and
the first-order shear deformation plate theory and the principle of first-order shear deformation theory were presented in [41] for sim-
virtual displacements. Liu et al. [23] developed an analytical model ply supported finite circular cylindrical hybrid shells with cross-ply
for free vibration analyses of piezoelectric coupled moderately thick composite laminate as elastic substrate under electromechanical static
circular plates. The Mindlin plate theory was employed and differential loads, Navier-type solutions were used. Oh and Lee [42] proposed the
equations of motion were analytically solved. A three-node triangular aerothermoelastic analysis of cylindrical piezolaminated shells via ge-
finite element based on cubic zig-zag plate theory was developed ometrically nonlinear finite elements based on the multifield layerwise
in [24] to refine the predictions of fully coupled and uncoupled be- theory. The dynamic solution of a multilayered orthotropic piezoelec-
haviors of thick smart composite plates under mechanical, thermal and tric infinite hollow cylinder was obtained in [43] via the superposition
electrical loadings. Analytical solutions were derived in [25] for free method where the solution was divided in a quasi-static and in a
vibrations of three-dimensional, linear anisotropic, magneto-electro- dynamic part. Wu and Heyliger [44] analyzed the free asymmetric and
elastic and multilayered rectangular simply supported sandwich plates. axisymmetric vibrations of layered piezoelectric spherical shells using a
A modified mixed variational principle for piezoelectric plates was two-dimensional first order shear deformable shell theory where piece-
established in [26], the exact solution was given for the static analysis. wise Hermite interpolation polynomials and Fourier approximations
Ramirez et al. [27] proposed an approximate solution for the free were employed. A set of Mathieu–Hill equations was derived in [45]
vibration problem of two-dimensional magneto-electro-elastic plates for the dynamic stability analysis of piezoelectric circular cylindrical
by combining a discrete layer approach with the Ritz method. Static shells subjected to combined periodic axial compression and radial
analysis of laminated smart composite plates was carried out in [28] electric field loadings. Carrera and Brischetto [46] proposed 2D exact
using a generalized-energy-based finite element model employing a solutions in both layer wise and equivalent single layer forms for the
first-order shear deformation theory and an eight node isoparamet- static analysis of smart piezoelectric plates and shells. Displacement
ric serendipity element. Samal et al. [29] modeled nonlinearities in and mixed models were employed. The same authors proposed an
piezoceramic plates via a nonlinear electric enthalpy density function extension of the Reissner’s mixed variational theorem in [47] in order

2
S. Brischetto and D. Cesare International Journal of Mechanical Sciences 280 (2024) 109620

to have displacements, transverse stresses, electric potential and trans- functionally graded plates integrated with a piezoelectric fiber re-
verse normal electric displacement as primary variables of the problem. inforced composite layer. An exact three-dimensional analysis was
Vel and Batra [48,49] presented an analytical solutions for static 3D presented in [70] for a functionally gradient piezoelectric simply sup-
deformations of multilayered piezoelectric rectangular plates using the ported rectangular plate when grounded along its four edges. The state
Eshelby–Stroh formalism. The equations of static and linear piezoelec- equations were developed using the state space approach. Brischetto
tricity were exactly satisfied at each point in the body considering a and Carrera [71] proposed 2D exact solutions in layer wise and equiv-
solution in terms of infinite series. The same authors [50] extended alent single layer forms for static analyses of functionally graded smart
piezoelectric plates. Governing equations were derived from an appro-
previous works developing a generalized plane strain thick laminated
priate extension of the principle of virtual displacements. Considering
plates model for thermopiezoelectric deformations using the Eshelby–
shells employing functionally graded piezoelectric layers, less works
Stroh formalism. Kulikov and Plotnikova [51] proposed an extension
can be discussed. Wu et al. [72] proposed a higher order theory
of the sampling surfaces method to shells to solve the static problem of for the electromechanical behavior of piezoelectric shells with graded
3D electroelasticity for cylindrical and spherical piezoelectric laminated material properties in the thickness direction. The same authors [73]
shells. also developed an analytical model for piezothermoelastic analysis of
Researchers have worked on the analysis of plates or shells com- functionally graded piezoelectric cylindrical shells subjected to axisym-
prising functionally piezoelectric layers for the past twenty years pro- metric thermal or mechanical loadings. Nonlinear dynamic analysis
ducing a substantial number of papers. A hierarchical system of two- and vibrations of imperfect functionally graded thick double curved
dimensional approximate equations was systematically derived in [52] shallow shells with piezoelectric actuators were investigated in [74].
for vibrations of functionally graded piezoelectromagnetic plates. Yang Liew et al. [75] developed a first-order shear deformation theory in
et al. [53] presented a large amplitude vibration analysis of pre-stressed the framework of a generic finite element formulation to consider
functionally graded laminated plates. Nonlinear governing equations of the coupling between mechanical and electrical fields in functionally
motion were derived within the context of Reddy higher-order shear graded shells with piezoelectric sensor and actuator layers. A eight-
deformation plate theory and Galerkin technique. A modified classical node shell element based on the first order shear deformation theory
lamination theory to account electro-elastic coupling for piezoelectric for the bending analysis of structures embedding piezoelectric layers
was developed in [76]. Classical thin shell theory, Von Karman–Donnell
functionally graded microstructure bi-morphs was developed in [54]
geometrical nonlinearity assumptions and the Galerkin method were
by using the finite element method. Static analyses and free vibra-
employed in [77] to derive governing equations to investigate the
tion studies were conducted in [55,56], respectively, for functionally
nonlinear dynamical behaviors of shells embedding piezolectric and
graded, anisotropic and linear magneto-electro-elastic plates by means functionally graded layers. The coupled vibroacoustic equations were
of a semi-analytical finite element method. Chen and Ding [57] derived obtained in [78] through the implementation of the Hamilton principle
two independent state equations by using three-dimensional theory and the Kirchhoff–Love thin shell theory. Fourier series expansions
equations of transversely isotropic piezoelasticity. The free vibration were used to solve the governing partial differential equations. An
problem of a piezoelectric rectangular plate with a functionally graded electro-mechanical coupling model was proposed in [79] to examine
property was then investigated. The element free Galerkin method to the stochastic vibration characteristics of smart piezoelectric shells
analyze buckling of piezoelectric functionally graded rectangular plates based on the Spectral–Tchebychev method. A three-dimensional elas-
subjected to non-uniformly distributed loads, heat and voltage was ticity solution for static analysis of a simply supported FGM cylindrical
employed in [58], Mindlin’s plate assumptions were used. A meshfree panel bonded with piezoelectric layers was developed in [80]. Kulikov
model (based on Galerkin method) was presented in [59] for the active and Plotnikova [81] showed the implementation of the sampling sur-
shape control of functionally graded material (FGM) plates containing faces method for the 3D exact solutions for FG piezoelectric laminated
distributed piezoelectric sensors and actuators; the first-order shear shells.
deformation theory was employed. A finite element formulation based The main novelty of this paper is a closed-form three-dimensional
shell model, here proposed for the static analysis of plates and shells
on the classical laminated plate theory was presented in [60] for the
embedding both piezoelectric and functionally graded piezoelectric
shape and vibration control of FGM plates with integrated piezoelectric
sensors and actuators, which couples the 3D equilibrium equations
sensors and actuators. He et al. [61] developed the mixed-variable
with the 3D divergence electric displacement equation. A layer wise
state space formulation for functionally graded piezoelectric material approach is applied and the partial differential equations in the thick-
(FGPM) strips and laminates considering the material inhomogeneity. ness direction are solved by means of the exponential matrix method.
The problem of a functionally graded, transversely isotropic, magneto– Navier-type closed form solution is proposed in in-plane directions.
electro-elastic circular plate subjected to a uniform load was considered The model was used for the static analysis of isotropic, orthotropic,
in [62] where the displacements and electric potential were represented composite, sandwich and functionally graded plates and shells in [82,
by appropriate polynomials in the radial coordinate. A new optimiza- 83]. Functionally graded piezoelectric materials have elastic, piezoelec-
tion algorithm based on computational intelligence combined with a tric and permittivity properties that change with continuity along the
modified finite element formulation was employed in [63] for the shape thickness direction in order to avoid stress discontinuities at each layer
control of FGM plates that contained piezoelectric sensor and actuator interface. The 3D equilibrium equations have been coupled with the
patches. Lu et al. [64] presented an alternative method for the deriva- 3D Fourier heat conduction equation to analyze the thermo-mechanical
tion of exact solutions of simply supported rectangular functionally coupling in multilayered composite and functionally graded plates
graded piezoelectric plates or laminates where related equations and and shells in [84,85]. The three-dimensional solution via layer wise
formulas were developed in terms of an elegant Stroh formalism. The approach and the exponential matrix method was applied in [86] for
vibration analyses of smart piezoelectric composite and functionally
same methodology was also applied in the case of cylindrical bend-
graded plates and shells. In [86], the 3D divergence electric displace-
ing [65]. Ootao and Tanigawa [66] developed the theoretical analysis
ment equation replaced the 3D Fourier heat conduction equation or
of a three-dimensional transient piezothermoelasticity problem for a
the 3D Fick diffusion law. The present work extends the model in [86]
functionally graded rectangular plate. Pan and Han [67] presented an to the static analysis. The Section 2 proposes the 3D coupled electro-
exact solution for multilayered rectangular plates embedding function- elastic governing equations for spherical shells. Geometrical and consti-
ally graded, anisotropic, and linear magneto-electro-elastic materials. tutive equations, governing equations for static analysis and solutions
The same authors [68] derived Green’s functions for the analysis of for sensor and actuator configurations are deeply discussed. Section 3 is
transversely isotropic, piezoelectric functionally graded and multilay- devoted to results and it is divided in two parts: the first part proposes
ered plates. A finite element model based on the first order shear validation results and the second part shows new benchmarks. Main
deformation theory was developed in [69] for the static analysis of conclusions are given in Section 4.

3
S. Brischetto and D. Cesare International Journal of Mechanical Sciences 280 (2024) 109620

Fig. 1. Geometries and the orthogonal mixed curvilinear reference system (𝛼, 𝛽, 𝑧) for model developing, assessments and benchmarks. The 3D model is able to consider the
following geometries: (a) Spherical Shell, (b) Cylindrical Shell, (c) Cylinder, (d) Plate.

2. 3D coupled electro-elastic governing equations for spherical two involved physical fields. Geometrical equations link strains with
shells the three displacement components and the electric field vector with
the electric potential (see Eqs. (1)–(2)). Constitutive equations con-
In the present section, a complete description of the 3D coupled nect stresses with strains and electric field components, and electric
electro-elastic shell model is proposed. This model permits to analyze displacements with strains and electric field components (see Eqs. (3)–
behavior of several geometries such as plates, cylinders, cylindrical (4)). The geometrical and constitutive Eqs. (1)–(4) in matrix form for
shells and spherical shells thanks to the use of the mixed curvilinear the electro-elastic problem can be written as follows:
reference system (𝛼, 𝛽, 𝑧) and via proper considerations on the radii
𝜺𝑘 = 𝜟𝑘 (𝑧)𝒖𝑘 (1)
of curvature. In-plane directions 𝛼 and 𝛽 follow the curvature of the
geometry. The 𝑧 direction is always perpendicular to the reference
surface 𝛺0 . In Fig. 1 all the analyzable geometries are shown. The  𝑘 = −𝜟𝑘𝐸 (𝑧)𝜙𝑘 (2)
present model is coupled because the 3D equilibrium equations for
shells and the 3D divergence equation for the electric displacement are 𝝈 𝑘 = 𝑪 𝑘 (𝑧)𝜺𝑘 − 𝒆𝑘𝑇 (𝑧) 𝑘 (3)
solved together. The coupling between the two physical fields allows
that an electric input gives a mechanical output and vice-versa. For this  𝑘 = 𝒆𝑘 (𝑧)𝜺𝑘 + 𝝐 𝑘 (𝑧) 𝑘 (4)
reason, the primary variables of the model are the three displacement 𝑘 𝜖 𝑘 𝜖 𝑘 𝛾 𝑘 𝛾 𝑘 𝛾 𝑘 }𝑇 , 𝑘
components 𝑢, 𝑣 and 𝑤 and the electric potential 𝜙. where 𝜺𝑘 is the 6 × 1 strain vector {𝜖𝛼𝛼 𝛽𝛽 𝑧𝑧 𝛽𝑧 𝛼𝑧 𝛼𝛽
𝜟 (𝑧) is a
6 × 3 matrix for spherical shells, 𝒖𝑘 is the 3 × 1 displacement vector
2.1. Geometrical and constitutive equations {𝑢𝑘 𝑣𝑘 𝑤𝑘 }𝑇 ,  𝑘 is the 3 × 1 electric field vector {𝛼𝑘 𝛽𝑘 𝑧𝑘 }𝑇 , 𝜟𝑘𝐸 (𝑧) is a
3 × 1 matrix for spherical shells, 𝜙𝑘 is the scalar electric potential, 𝝈 𝑘 is
the 6 × 1 stress vector {𝜎𝛼𝛼 𝑘 𝜎 𝑘 𝜎 𝑘 𝜎 𝑘 𝜎 𝑘 𝜎 𝑘 }𝑇 , 𝑪 𝑘 (𝑧) is the 6 × 6 elas-
For the present 3D coupled electro-elastic model, the geometri- 𝛽𝛽 𝑧𝑧 𝛽𝑧 𝛼𝑧 𝛼𝛽
cal and constitutive relations are fundamental in order to couple the 𝑘
tic coefficient matrix, 𝒆 (𝑧) is the 3 × 6 piezoelectric coefficient matrix,

4
S. Brischetto and D. Cesare International Journal of Mechanical Sciences 280 (2024) 109620

 𝑘 is the 3 × 1 electric displacement vector {𝑘𝛼 𝑘𝛽 𝑘𝑧 }𝑇 and 𝝐 𝑘 (𝑧) is

the 3 × 3 permittivity coefficient matrix. All matrices and vectors are 𝑘 𝑘
𝜕𝜎𝛽𝑧 𝑘 𝐻𝛽 (𝑧)
𝜕𝜎𝛼𝑧 𝜕𝜎𝑧𝑧 𝑘 𝐻𝛼 (𝑧) 𝑘
written in the structural reference system. 𝑘 and 𝑇 superscripts specify a 𝐻𝛽 (𝑧) + 𝐻𝛼 (𝑧) + 𝐻𝛼 (𝑧)𝐻𝛽 (𝑧) − 𝜎𝛼𝛼 − 𝜎
𝜕𝛼 𝜕𝛽 𝜕𝑧 𝑅𝛼 𝑅𝛽 𝛽𝛽
generic physical layer of the multilayered structure and the transpose
( 𝐻𝛽 (𝑧) 𝐻𝛼 (𝑧) ) 𝑘
of a matrix/vector, respectively. The explicit form of all the matrices + + 𝜎𝑧𝑧 = 0,
involved in Eqs. (1)–(4) are: 𝑅𝛼 𝑅𝛽

⎡ 1 𝜕
0 1
⎤ (13)
𝐻𝛼 (𝑧) 𝜕𝛼 𝐻𝛼 (𝑧)𝑅𝛼
⎢ 1 𝜕 1 ⎥
⎢ 𝛽,𝛽
𝐻𝛽 (𝑧)𝑅𝛽 ⎥
0 𝛼,𝛼
𝐻𝛽 (𝑧) 𝜕𝛽 + + 𝑧,𝑧 = 0. (14)
⎢ 𝜕 ⎥ 𝐻𝛼 (𝑧) 𝐻𝛽 (𝑧)
⎢ 0 0 𝜕𝑧 ⎥
𝜟𝑘 (𝑧) = ⎢ 𝜕 1 1 𝜕 ⎥, (5)
0 − This set of four differential equations is solved in closed-form to analyze
⎢ 𝜕𝑧 𝐻𝛽 (𝑧)𝑅𝛽 𝐻𝛽 (𝑧) 𝜕𝛽 ⎥
⎢𝜕 − 1
0 1 𝜕 ⎥ the electro-elastic behavior of multilayered classical and functionally
⎢ 𝜕𝑧 𝐻𝛼 (𝑧)𝑅𝛼 𝐻𝛼 (𝑧) 𝜕𝛼 ⎥
graded plates and shells. Eqs. (11)–(14) are written for spherical shells,
⎢ 1 𝜕 1 𝜕
0 ⎥
⎣ 𝐻𝛽 (𝑧) 𝜕𝛽 𝐻𝛼 (𝑧) 𝜕𝛼 ⎦ they are a particular case of general double curved shells. Constant
1
⎡ 𝐻 (𝑧) 𝜕𝛼 ⎤
𝜕 coefficients in 𝛼 and 𝛽 directions are shown in these equations because
𝑘 ⎢ 𝛼1 𝜕 ⎥ 𝑅𝛼 = 𝑅𝛽 .
𝜟𝐸 (𝑧) = ⎢ 𝐻 (𝑧) 𝜕𝛽 ⎥ , (6)
⎢ 𝛽𝜕 ⎥
⎣ 𝜕𝑧 ⎦ 2.3. Solution for sensor and actuator configurations
⎡𝐶 𝑘 (𝑧) 𝑘 (𝑧)
𝐶12 𝑘 (𝑧)
𝐶13 0 0 𝑘 (𝑧)⎤
𝐶16
⎢𝐶11
𝑘 𝑘 𝑘 𝑘 (𝑧)⎥ Geometrical Eqs. (1)–(2) and constitutive Eqs. (3)–(4) are intro-
⎢ 12 (𝑧) 𝐶22 (𝑧) 𝐶23 (𝑧) 0 0 𝐶26 ⎥ duced in the 3D governing Eqs. (11)–(14) of the electro-elastic problem.
⎢𝐶13
𝑘 (𝑧) 𝑘 (𝑧)
𝐶23 𝑘 (𝑧)
𝐶33 0 0 𝑘 (𝑧)⎥
𝐶36
𝑪 𝑘 (𝑧) = ⎢ (7) In this way, Eqs. (11)–(14) are expressed in terms of the primary
0 ⎥⎥
𝑘 (𝑧) 𝑘 (𝑧) ,
⎢ 0 0 0 𝐶44 𝐶45
variables (𝑢, 𝑣, 𝑤 and 𝜙) and related derivatives in 𝛼, 𝛽 and 𝑧. Harmonic
⎢ 0 0 0 𝑘 (𝑧)
𝐶45 𝑘 (𝑧)
𝐶55 0 ⎥
⎢𝐶 𝑘 (𝑧) 𝑘 (𝑧) 𝑘 (𝑧) 𝑘 (𝑧)⎥
forms for primary variables are:
⎣ 16 𝐶26 𝐶36 0 0 𝐶66 ⎦
𝑢𝑘 (𝛼, 𝛽, 𝑧) = 𝑈 𝑘 (𝑧) cos(𝛼𝛼) ̄ ,
̄ sin(𝛽𝛽) (15)
⎡ 0 0 0 𝑒𝑘14 (𝑧) 𝑒𝑘15 (𝑧) 0 ⎤
𝒆𝑘 (𝑧) = ⎢ 0 0 0 𝑒𝑘24 (𝑧) 𝑒𝑘25 (𝑧) 0 ⎥, (8)
⎢ 𝑘 ⎥ 𝑣𝑘 (𝛼, 𝛽, 𝑧) = 𝑉 𝑘 (𝑧) sin(𝛼𝛼) ̄ ,
̄ cos(𝛽𝛽) (16)
⎣𝑒31 (𝑧) 𝑒𝑘32 (𝑧) 𝑒𝑘33 (𝑧) 0 0 𝑒36 (𝑧)⎦
𝑘

𝑘 𝑘 (𝑧)
⎡𝜖11 (𝑧) 𝜖12 0 ⎤
𝑤𝑘 (𝛼, 𝛽, 𝑧) = 𝑊 𝑘 (𝑧) sin(𝛼𝛼) ̄ ,
̄ sin(𝛽𝛽) (17)
𝝐 𝑘 (𝑧) = ⎢𝜖12
𝑘 (𝑧) 𝑘 (𝑧)
𝜖22 0 ⎥. (9)
⎢ 𝑘 (𝑧)⎥
⎣ 0 0 𝜖33 ⎦
𝜙𝑘 (𝛼, 𝛽, 𝑧) = 𝛷𝑘 (𝑧) sin(𝛼𝛼) ̄
̄ sin(𝛽𝛽), (18)
The 𝑧 dependence in matrices presented in Eqs. (5)–(9) regards material
properties and curvature coefficients. Curvature terms are 𝐻𝛼 , 𝐻𝛽 and where 𝑈 𝑘 (𝑧), 𝑉 𝑘 (𝑧),
𝑊 and 𝑘 (𝑧) 𝛷𝑘 (𝑧)
are the maximum amplitudes of
𝐻𝑧 that can be written as follows: each primary variable and 𝛼̄ and 𝛽̄ are algebraic coefficients:
( ) ( ) 𝑚𝜋 𝑛𝜋
𝑧 𝑧 𝛼̄ = , 𝛽̄ = , (19)
𝐻𝛼 (𝑧) = 1 + , 𝐻𝛽 (𝑧) = 1 + , 𝐻𝑧 (𝑧) = 1. (10) 𝑎 𝑏
𝑅𝛼 𝑅𝛽
where 𝑚 and 𝑛 are half-wave numbers along the 𝛼 and 𝛽 directions; 𝑎
where 𝑅𝛼 and 𝑅𝛽 are the radii of curvature of the reference surface. and 𝑏 are in-plane dimensions. From this point forward, the amplitudes
In order to obtain a plate geometry, 𝑅𝛼 and 𝑅𝛽 are infinite. The total of the primary variables are the unknowns of this 3D electro-elastic
range of the thickness 𝑧 coordinate is −ℎ∕2 < 𝑧 < ℎ∕2. It is also possible problem. These harmonic forms satisfy the boundary conditions in the
to introduce the 𝑧̃ coordinate, which range is 0 < 𝑧̃ < ℎ. simply-supported case. Thanks to the harmonic form impositions in
In order to perform a closed-form solution, only 0◦ or 90◦ or- Eqs. (15)–(18), the derivatives along in-plane directions 𝛼 and 𝛽 can
thotropic angles are possible. For this reason, terms 𝐶16 𝑘 (𝑧), 𝐶 𝑘 (𝑧),
26 be exactly computed because derivatives of harmonic functions are
𝑘 𝑘 𝑘 𝑘 𝑘 𝑘
𝐶36 (𝑧), 𝐶45 (𝑧) in the 𝑪 (𝑧) matrix, 𝑒14 (𝑧), 𝑒25 (𝑧) and 𝑒36 (𝑧) in the 𝒆𝑘 (𝑧) known. In this way, the explicit form of equations is:
matrix, and 𝜖12 𝑘 (𝑧) in the 𝜺𝑘 (𝑧) matrix are set to zero.
𝐴𝑘1 𝑈 𝑘 +𝐴𝑘2 𝑉 𝑘 +𝐴𝑘3 𝑊 𝑘 +𝐴𝑘4 𝛷𝑘 +𝐴𝑘5 𝑈,𝑧𝑘 +𝐴𝑘6 𝑊,𝑧𝑘 +𝐴𝑘7 𝛷,𝑧
𝑘
+𝐴𝑘8 𝑈,𝑧𝑧
𝑘
= 0 , (20)
2.2. Governing equations for static analysis
𝐴𝑘9 𝑈 𝑘 +𝐴𝑘10 𝑉 𝑘 +𝐴𝑘11 𝑊 𝑘 +𝐴𝑘12 𝛷𝑘 +𝐴𝑘13 𝑉,𝑧𝑘 +𝐴𝑘14 𝑊,𝑧𝑘 +𝐴𝑘15 𝛷,𝑧
𝑘
+𝐴𝑘16 𝑉,𝑧𝑧
𝑘
= 0,
In the present subsection, the complete set of equations for the 3D
(21)
coupled electro-elastic problem is presented. 3D equilibrium equations
and the 3D divergence electric displacement equation are written for
𝐴𝑘17 𝑈 𝑘 + 𝐴𝑘18 𝑉 𝑘 + 𝐴𝑘19 𝑊 𝑘 + 𝐴𝑘20 𝛷𝑘 + 𝐴𝑘21 𝑈,𝑧𝑘 + 𝐴𝑘22 𝑉,𝑧𝑘 + 𝐴𝑘23 𝑊,𝑧𝑘
the spherical shell case. The 3D divergence electric displacement equa-
tion for spherical shells in the mixed curvilinear reference system has + 𝐴𝑘24 𝛷,𝑧
𝑘
+ 𝐴𝑘25 𝑊,𝑧𝑧
𝑘
+ 𝐴𝑘26 𝛷,𝑧𝑧
𝑘
= 0, (22)
been derived as already presented in [86]. The set of equations for the
electro-elastic problem is: 𝐴𝑘27 𝑈 𝑘 + 𝐴𝑘28 𝑉 𝑘 + 𝐴𝑘29 𝑊 𝑘 + 𝐴𝑘30 𝛷𝑘 + 𝐴𝑘31 𝑈,𝑧𝑘 + 𝐴𝑘32 𝑉,𝑧𝑘
𝑘
𝜕𝜎𝛼𝛼
𝑘
𝜕𝜎𝛼𝛽 𝑘
𝜕𝜎𝛼𝑧 ( 2𝐻𝛽 (𝑧) 𝐻𝛼 (𝑧) ) 𝑘 + 𝐴𝑘33 𝑊,𝑧𝑘 + 𝐴𝑘34 𝑊,𝑧𝑧
𝑘
+ 𝐴𝑘35 𝛷,𝑧𝑧
𝑘
= 0. (23)
𝐻𝛽 (𝑧) + 𝐻𝛼 (𝑧) + 𝐻𝛼 (𝑧)𝐻𝛽 (𝑧) + + 𝜎𝛼𝑧 = 0,
𝜕𝛼 𝜕𝛽 𝜕𝑧 𝑅𝛼 𝑅𝛽 Eqs. (20)–(23) is a set of second-order differential equations in 𝑧 where
(11) each 𝐴𝑘𝑖 term depends on the variable 𝑧. The 𝑧 dependence comes
from both the curvature terms (𝐻𝛼 and 𝐻𝛽 ) and the possibility of
𝑘
𝜕𝜎𝛼𝛽 𝑘
𝜕𝜎𝛽𝛽 𝑘
𝜕𝜎𝛽𝑧 ( 2𝐻 (𝑧) 𝐻𝛽 (𝑧) ) including FGPM (their 𝑪 𝑘 , 𝒆𝑘 and 𝝐 𝑘 matrices are 𝑧-dependent). In
𝛼 𝑘
𝐻𝛽 (𝑧) + 𝐻𝛼 (𝑧) + 𝐻𝛼 (𝑧)𝐻𝛽 (𝑧) + + 𝜎𝛽𝑧 = 0, order to solve Eqs. (20)–(23), constant 𝐴𝑘𝑖 terms are mandatory. For
𝜕𝛼 𝜕𝛽 𝜕𝑧 𝑅𝛽 𝑅𝛼
this reason, each 𝑘 physical layer is divided into 𝑗 fictitious layers; this
(12) method allows to calculate 𝐻𝛼 , 𝐻𝛽 and the FGPM properties within

5
S. Brischetto and D. Cesare International Journal of Mechanical Sciences 280 (2024) 109620

the fictitious layers and to obtain constant coefficients. In the case of The matrix 𝑻 𝒋−𝟏,𝒋 is the transfer matrix and it links the unknown
classical materials, each 𝑘 physical layer already has constant material variables at interfaces. Thanks to the definition of this matrix, the
properties but the introduction of the fictitious layers is important to layer-wise approach is correctly implemented in the present model.
eliminate the 𝑧 dependence coming from the curvature terms 𝐻𝛼 and The solution along the 𝑧 thickness direction is obtained thanks a
𝐻𝛽 . From this point forward, all the equations for the 3D exact coupled recursive introduction of Eq. (31) in Eq. (26):
electro-elastic problem will be written for the fictitious layers, adopting
𝑿𝑴
𝒕 =𝑨
∗∗𝑴 𝑴,𝑴−𝟏
𝑻 … 𝑻 𝟐,𝟏 𝑨∗∗𝟏 𝑿 𝒃𝟏 = 𝑯 𝒎 𝑿 𝒃𝟏 . (32)
the 𝑗 superscript in place of 𝑘. A total number of 𝑀 fictitious layers is
used to describe the structure. 𝑯 𝒎 matrix includes all the material and geometrical features of the
Introducing fictitious layers, 𝐴𝑗𝑖 terms are now constant. It is possi- functionally graded multilayered structure. The main peculiarity of the
ble to solve the set of second-order differential equations considering 𝑯 𝒎 matrix is the 8 × 8 dimension, independently by the number 𝑀 of
the standard techniques for differential equations. Eqs. (20)–(23) can fictitious layers and the order 𝑁 of the exponential matrix.
be rewritten as a set of eight first-order differential equations sim- The simply-supported constraints along the 𝛼 and 𝛽 directions can
ply doubling the number of primary variables: 𝑈,𝑧 , 𝑉,𝑧 , 𝑊,𝑧 and 𝛷,𝑧 be written as:
are now directly obtained from the model. This peculiarity allows to 𝜙 = 0, 𝑤 = 𝑣 = 0, 𝜎𝛼𝛼 = 0 for 𝛼 = 0, 𝑎, (33)
calculate very accurate stress and electric displacement trends in the
post-processing.
𝜙 = 0, 𝑤 = 𝑢 = 0, 𝜎𝛽𝛽 = 0 for 𝛽 = 0, 𝑏. (34)
The resulting set of the eight first-order differential equations can
be written in matrix form as: that fulfill the harmonic forms imposed for the primary variables at the
𝒋
edges.
𝑫 𝒋 𝑿 ,𝒛 = 𝑨𝒋 𝑿 𝒋 , (24) The transverse normal mechanical load 𝑝𝑧 is also written in har-
monic form:
that becomes:
𝒋 𝑝𝑧 (𝛼, 𝛽, 𝑧) = 𝑃𝑧 (𝑧) sin(𝛼𝛼) ̄ ,
̄ sin(𝛽𝛽) (35)
𝑿 ,𝒛 = (𝑫 𝒋 )−1 𝑨𝒋 𝑿 𝒋 = 𝑨∗𝒋 𝑿 𝒋 . (25)
where 𝑃𝑧 is the maximum amplitude which can be imposed at the top
The solution for Eq. (25) can be found by adopting the exponential
(t) or at the bottom (b) of the structure.
matrix method. So, the explicit solution form is:
The load boundary conditions have to be imposed on the external
( )
𝑿 𝒋 (ℎ𝑗 ) = 𝑒𝑥𝑝 𝑨∗𝒋 ℎ𝑗 𝑿 𝒋 (0) 0 ≤ 𝑧̃ 𝑗 ≤ ℎ𝑗 (26) surfaces. In the present 3D coupled exact electro-elastic formulation,
two load boundary conditions are possible: sensor configuration and
where 𝑧̃ 𝑗 is the thickness coordinate for each 𝑗 fictitious layer. 𝑧̃ 𝑗 actuator configuration. In the case of sensor configuration:
coordinate can vary from 0 (the bottom of the 𝑗th layer) to ℎ𝑗 (the
𝜎𝑧𝑧 = 𝑝𝑧𝑡 , 𝜎𝛼𝑧 = 0, 𝜎𝛽𝑧 = 0, 𝜙 = 0 for 𝑧 = +ℎ∕2, (36)
top of the 𝑗th layer).
The exponential matrix computed along the 𝑧̃ direction for the
solution is: 𝜎𝑧𝑧 = 0, 𝜎𝛼𝑧 = 0, 𝜎𝛽𝑧 = 0, 𝜙 = 0 for 𝑧 = −ℎ∕2, (37)
( ) ∑𝑁 𝑖
(𝑨∗𝒋 ) 𝑖 for the actuator configuration:
𝑒𝑥𝑝 𝑨∗𝒋 ℎ𝑗 = 𝑨∗∗𝒋 = 𝑰 + ℎ𝑗 (27)
𝑖=1
𝑖! 𝜎𝑧𝑧 = 0, 𝜎𝛼𝑧 = 0, 𝜎𝛽𝑧 = 0, 𝜙 = 𝜙𝑡 for 𝑧 = +ℎ∕2, (38)
where 𝑰 indicates the 8×8 identity matrix and 𝑁 is the maximum order
of expansion used to approximate the exponential matrix. 𝜎𝑧𝑧 = 0, 𝜎𝛼𝑧 = 0, 𝜎𝛽𝑧 = 0, 𝜙 = 𝜙𝑏 for 𝑧 = −ℎ∕2. (39)
As the present formulation adopts a layer-wise approach, proper The explicit forms of Eqs. (36) and (37) for sensor configuration are:
impositions of the interlaminar continuity conditions are mandatory.
The following imposed relations link variables 𝑢, 𝑣, 𝑤, 𝜙, 𝜎𝛼𝑧 , 𝜎𝛽𝑧 , 𝜎𝑧𝑧 ⎧ 𝑈𝑡𝑀 ⎫
⎪ 𝑀⎪
𝑉
and 𝑧 at the top (t) of the 𝑗 − 1 fictitious layer with the same variables ⎤⎪ 𝑡 ⎪
𝑀 𝑀 𝑀 𝑀
⎡ − 𝐶13 𝛼̄ − 𝐶23 𝛽̄ 𝐶13
+
𝐶23
0 0 0 𝑀
𝐶33 𝑒𝑀
⎢ 𝐻𝛼𝑡 𝐻𝛽𝑡 𝐻𝛼𝑡 𝑅𝛼 𝐻𝛽𝑡 𝑅𝛽 33 ⎥ ⎪𝑊 𝑀 ⎪
at the bottom (b) of the 𝑗 fictitious layer. The interlaminar continuity ⎪ 𝑡 ⎪
⎢ 0 𝐶𝑀
− 𝐻44 𝑅𝛽
𝑀
𝐶44
𝛽̄
𝑒𝑀
24
𝛽̄ 0 𝑀
𝐶44 0 0 ⎥ ⎪ 𝛷𝑡𝑀 ⎪
conditions written in explicit form are: ⎢ 𝛽𝑡 𝐻𝛽𝑡 𝐻𝛽𝑡 ⎥⎨ 𝑀 ⎬ =
⎢ 𝐶55𝑀 𝑀
𝐶55 𝑒𝑀 ⎥ 𝑈
⎢− 𝐻𝛼𝑡 𝑅𝛼 0 𝛼̄ 15
𝛼̄ 𝑀
𝐶55 0 0 0 ⎥ ⎪ ,𝑧𝑡 ⎪
𝑢𝑗−1
𝑡 = 𝑢𝑗𝑏 , 𝑣𝑗−1
𝑡 = 𝑣𝑗𝑏 , 𝑤𝑗−1
𝑡 = 𝑤𝑗𝑏 , 𝜙𝑗−1
𝑡 = 𝜙𝑗𝑏 (28) ⎢
𝐻𝛼𝑡 𝐻𝛼𝑡
⎥ ⎪ 𝑉,𝑧𝑡 ⎪
𝑀
⎣ 0 0 0 1 0 0 0 0 ⎦⎪ 𝑀 ⎪
(40)
⎪𝑊,𝑧𝑡 ⎪
𝑗−1 𝑗 𝑗−1 𝑗 𝑗−1 𝑗 ⎪ 𝛷𝑀 ⎪
𝜎𝛼𝑧 = 𝜎𝛼𝑧 , 𝜎𝛽𝑧 = 𝜎𝛽𝑧 , 𝜎𝑧𝑧 = 𝜎𝑧𝑧 , 𝑗−1
𝑧 = 𝑗𝑧 (29) ⎩ ,𝑧 ⎭
𝑡
𝑡 𝑏 𝑡 𝑏 𝑡 𝑏 𝑡 𝑏
⎧𝑃 ⎫
Developing Eqs. (28)–(29) and compacting them in a matrix form, ⎪ 𝑧𝑡 ⎪
it is possible to write the following equation for the interlaminar ⎪0⎪
=⎨ ⎬,
continuity conditions: ⎪0⎪
⎪0⎪
⎩ ⎭
𝑗 𝑗−1,𝑗 𝑗−1
⎧𝑈 ⎫ ⎡ 1 0 0 0 0 0 0 0 ⎤ ⎧𝑈 ⎫ ⎧ 𝑈𝑏1 ⎫
⎪𝑉 ⎪ ⎢ 0 ⎪ 1⎪
1 0 0 0 0 0 0 ⎥ ⎪𝑉 ⎪ 1 1 1 1
⎪ 𝑉𝑏 ⎪
⎪ ⎪ ⎢ ⎥ ⎪ ⎪ ⎡ − 𝐶13 𝛼̄ − 𝐶23 𝛽̄ 𝐶13
+
𝐶23
0 0 0 1
𝐶33 𝑒133 ⎤ ⎪ 𝑊 1 ⎪
⎪ 𝑊 ⎪ ⎢ 0 0 1 0 0 0 0 0 ⎥ ⎪𝑊 ⎪ ⎢ 𝐻𝛼𝑏 𝐻𝛽𝑏 𝐻𝛼𝑏 𝑅𝛼 𝐻𝛽𝑏 𝑅𝛽 ⎥⎪ 𝑏 ⎪
⎪𝛷⎪ ⎢ 0 0 0 1 0 0 0 0 ⎥ ⎪𝛷⎪ ⎢ 0 𝐶1
− 𝐻44 𝑅𝛽
1
𝐶44
𝛽̄
𝑒124
𝛽̄ 0 1
𝐶44 0 0 ⎥ ⎪ 𝛷𝑏1 ⎪
⎨𝑈 ⎬ = ⎢ 𝑇 ⎢ ⎥⎨ 1 ⎬ =
0 ⎥⎥ ⎨𝑈 ⎬ 𝛽𝑏 𝐻𝛽𝑏 𝐻𝛽𝑏
0 𝑇2 𝑇3 𝑇4 0 0 ⎢ 𝐶551 1
𝑒115 ⎥ 𝑈
⎪ ,𝑧 ⎪ ⎢ 1 ⎪ ,𝑧 ⎪ ⎢− 𝐻𝛼 𝑅𝛼 0
𝐶55
𝛼̄ 1
𝛼̄ 𝐶55 0 0 0 ⎥ ⎪ ,𝑧𝑏 ⎪
⎪ 𝑉,𝑧 ⎪ ⎢ 0 0 ⎥ ⎪ 𝑉,𝑧 ⎪ ⎥ ⎪ 𝑉,𝑧 ⎪
𝐻𝛼𝑏 𝐻𝛼𝑏 1
𝑇5 𝑇6 𝑇7 0 𝑇8 0 ⎢ 𝑏

⎪𝑊 ⎪ ⎢𝑇 𝑇10 𝑇11 0 0 0 𝑇12 𝑇13 ⎥ ⎪𝑊 ⎪ ⎣ 0 0 0 1 0 0 0 0 ⎦ ⎪ 1𝑏 ⎪ (41)

⎪ ,𝑧 ⎪ ⎢ 9 ⎥ ⎪ ,𝑧 ⎪ ⎪𝑊,𝑧𝑏 ⎪
⎩ 𝛷,𝑧 ⎭𝑏 ⎣𝑇14 𝑇15 𝑇16 0 0 0 𝑇17 𝑇18 ⎦ ⎩ 𝛷,𝑧 ⎭𝑡 ⎪ 𝛷1 ⎪
⎩ ,𝑧 ⎭
𝑏

(30) ⎧0⎫
⎪ ⎪
⎪0⎪
that can be further compacted as: =⎨ ⎬.
⎪0⎪
𝑿 𝑗𝒃 = 𝑻 𝑗−1,𝑗 𝑿 𝑗−1 (31) ⎪0⎪
𝒕 . ⎩ ⎭

6
S. Brischetto and D. Cesare International Journal of Mechanical Sciences 280 (2024) 109620

Table 1
Elastic, piezoelectric and permittivity coefficients for the materials involved in assessments and benchmarks.
PZT-4 CFRP-A Piezo CFRP-B PVDF FGPM Al-2024 Foam
𝐸1 (GPa) 81.3 132.38 2 172 2 81.3 73 0.180
𝐸2 (GPa) 81.3 10.756 2 6.9 2 81.3 73 0.180
𝐸3 (GPa) 64.5 10.756 2 6.9 2 64.5 73 0.180
𝜈12 (–) 0.329 0.24 0.29 0.25 1/3 0.329 0.3 0.37
𝜈13 (–) 0.432 0.24 0.29 0.25 1/3 0.432 0.3 0.37
𝜈23 (–) 0.432 0.49 0.29 0.25 1/3 0.432 0.3 0.37
𝐺12 (GPa) 30.6 5.6537 0.7752 3.4 0.75 30.6 28.077 65.693
𝐺13 (GPa) 25.6 5.6537 0.7752 3.4 0.75 25.6 28.077 65.693
𝐺23 (GPa) 25.6 3.6060 0.7752 1.4 0.75 25.6 28.077 65.693
𝑒15 (C∕m2 ) 12.72 0 0 0 0 12.7 0 0
𝑒24 (C∕m2 ) 12.72 0 0 0 0 12.7 0 0
𝑒31 (C∕m2 ) −5.20 0 0.046 0 −0.0015 −5.20 0 0
𝑒32 (C∕m2 ) −5.20 0 0 0 0.0285 −5.20 0 0
𝑒33 (C∕m2 ) 15.08 0 0 0 −0.051 15.1 0 0
𝜖1 (nF∕m) 13.06 0.030989 0.106 13.06 0.1062 6.46 0.008854 0.008854
𝜖2 (nF∕m) 13.06 0.026563 0.106 13.06 0.1062 6.46 0.008854 0.008854
𝜖3 (nF∕m) 11.51 0.026563 0.106 13.06 0.1062 5.62 0.008854 0.008854

The explicit forms of Eqs. (38) and (39) for actuator configuration are: Table 2
Geometrical data and imposed half-wave numbers 𝑚 and 𝑛 for the four assessments
(A1-A4).
⎧ 𝑈𝑀 ⎫
A1 A2 A3 A4
⎪ 𝑡𝑀 ⎪
𝑉
⎤⎪ 𝑡 𝑀⎪
𝑀 𝑀 𝑀 𝑀
⎡ − 𝐶13 𝛼̄ − 𝐶23 𝛽̄ 𝐶13
+
𝐶23
0 0 0 𝑀
𝐶33 𝑒𝑀 a (m) 0.4 – – 1
33 ⎪𝑊
⎢ 𝐻𝛼𝑡 𝐻𝛽𝑡 𝐻𝛼𝑡 𝑅𝛼 𝐻𝛽𝑡 𝑅𝛽 ⎥⎪ 𝑡 ⎪ ⎪ b (m) 0.4 𝜋 𝜋
1
⎢ 0 −
𝑀
𝐶44
𝑅
𝑀
𝐶44
𝛽̄
𝑒𝑀
24
𝛽̄ 0 𝑀
𝐶44 0 0 ⎥⎥ ⎪ 𝛷𝑡 ⎪
𝑀
ℎ (m) 0.1
3
Variable
3
Variable Variable
⎢ 𝐻𝛽𝑡 𝛽 𝐻𝛽𝑡 𝐻𝛽𝑡
⎨ 𝑀 ⎬=
⎢ 𝐶55
𝑀 𝑀 𝑒𝑀 ⎥ 𝑈,𝑧 R𝛼 (m) ∞ ∞ ∞ ∞
0 ⎥ ⎪ 𝑀𝑡 ⎪
𝐶55 𝑀
⎢− 𝐻 𝑅𝛼
15
0 𝛼̄ 𝛼̄ 𝐶55 0 0 R𝛽 (m) ∞ 1 1 ∞
⎥ ⎪ 𝑉,𝑧𝑡 ⎪
𝐻𝛼𝑡 𝐻𝛼𝑡
⎢ 𝛼𝑡
0 ⎦ ⎪𝑊 𝑀 ⎪
m 1 0 0 1
⎣ 0 0 0 1 0 0 0 (42)
⎪ ,𝑧𝑡 ⎪ n 1 1 1 1
⎪ 𝛷𝑀 ⎪
⎩ ,𝑧 ⎭
𝑡

⎧ ⎫
⎪0⎪
⎪0⎪ layer to the top of the last fictitious layer. In this way, all displacement,
=⎨ ⎬,
⎪0⎪ electric potential, strain, stress and electric displacement trends can be
⎪𝜙𝑡 ⎪ evaluated along the 𝑧 thickness direction of the multilayered structure.
⎩ ⎭
The present mathematical formulation has been developed in Mat-
⎧ 𝑈1 ⎫
⎪ 𝑏1 ⎪ lab environment to obtain an in-house academic code called 3D expo-
1
𝐶13 1
𝐶23 1
𝐶13 1
𝐶23 𝑉
1 ⎤⎪ 𝑏 ⎪
⎡− 𝛼̄ − 𝐻 𝛽̄ + 0 0 0 1
𝐶33 𝑒33 ⎪ 1 ⎪ nential multifield solutions (3DEMS).
⎢ 𝐻𝛼𝑏 𝛽𝑏 𝐻𝛼𝑏 𝑅𝛼 𝐻𝛽𝑏 𝑅𝛽 ⎥ ⎪ 𝑊𝑏 ⎪
⎢ 0 −
1
𝐶44
𝑅
1
𝐶44
𝛽̄
𝑒124
𝛽̄ 0 1
𝐶44 0 0 ⎥⎥ ⎪ 𝛷𝑏 ⎪
1
⎢ 𝐻𝛽𝑏 𝛽 3. Results
⎥⎨ 𝑈,𝑧1 ⎬
𝐻𝛽𝑏 𝐻𝛽𝑏 =
⎢ 𝐶55
1 1 𝑒115
0 ⎥ ⎪ 1𝑏 ⎪
𝐶55 1
⎢− 𝐻 𝑅𝛼 0 𝛼̄ 𝛼̄ 𝐶55 0 0
⎥ ⎪ 𝑉,𝑧𝑏 ⎪
𝐻𝛼𝑏 𝐻𝛼𝑏
⎢ 𝛼𝑏 In this section some results regarding the present 3D closed form
⎣ 0 0 0 1 0 0 0 0 ⎦ ⎪𝑊 1 ⎪ (43)
⎪ ,𝑧𝑏 ⎪ coupled electro-elastic model (called as 3D-u-𝜙) are shown. The name
⎪ 𝛷1 ⎪
⎩ ,𝑧𝑏 ⎭ of the model intuitively represents all its main characteristics: a three-
⎧ ⎫ dimensional formulation where primary variables are displacements
⎪0⎪ and the electric potential. This section is divided in two subsections:
⎪0⎪
=⎨ ⎬. the first one is related to the validation of the 3D-u-𝜙 model us-
⎪0⎪ ing results from other 3D exact electro-elastic theories or quasi-3D
⎪𝜙𝑏 ⎪
⎩ ⎭ electro-elastic formulations. The validation is performed for several
Superscript 𝑀 indicates the last layer and superscript 1 indicates the laminations (FGPM and classical multilayered materials) and geome-
first layer. Subscripts 𝑡 and 𝑏 indicate top and bottom, respectively. A tries. The second subsection is devoted to new benchmarks where
compact form of Eqs. (40)–(41) or Eqs. (42)–(43) is: results for FGPM and multilayered plates, cylinders, cylindrical shells
and spherical shells are proposed in both tabular and graphical forms.
𝑩𝑴 𝑴
𝒕 𝑿𝒕 = 𝒫 𝒕 , (44) An in-house academic Matlab code (called as 3D exponential multifield
solutions (3DEMS)) has been used to obtain tabular and graphical results
𝑩 𝟏𝒃 𝑿 𝟏𝒃 = 𝒫 𝒃 , (45) discussed in the present section.
introducing Eq. (32) in Eq. (44), we obtain:
3.1. Validation results
𝟏
𝑩𝑴
𝒕 𝑯 𝒎𝑿𝒃 = 𝒫 𝒕. (46)
Compacting in a unique matrix form Eqs. (45) and (46), the final linear In the present subsection four assessment cases are taken into
system to be solved is: account regarding multilayered/FGPM plates and multilayered/one-
[ 𝑴 ] layered cylindrical panels. Sensor and actuator configurations are in-
𝑩𝒕 𝑯 𝒎
𝟏 𝑿 𝟏𝒃 = 𝒫 ⇒ 𝑬𝑿 𝟏𝒃 = 𝒫 . (47) vestigated. A number 𝑀 = 300 of mathematical layers are used for the
𝑩𝒃 four assessments together with the exponential matrix truncated at the
Eq. (47) gives the solution at the bottom of the structure. Then, the third order (𝑁 = 3). In past authors’ works, a convergence analysis
solution can be reconstructed along the 𝑧 direction recursively by sub- was performed for the pure elastic case to opportunely choose these
stituting Eq. (31) in Eq. (26) to go from the bottom of the first fictitious two parameters. The proposed 𝑀 and 𝑁 values are a conservative

7
S. Brischetto and D. Cesare International Journal of Mechanical Sciences 280 (2024) 109620

Table 3
Assessment 1, simply-supported four-layered 𝑃 𝑍𝑇 − 4∕0◦ ∕90◦ ∕𝑃 𝑍𝑇 − 4 thick square plate in sensor configuration (zero electric potential at
external surfaces (𝛷𝑏 = 𝛷𝑡 = 0 V) and transverse normal load applied at the top of the structure (𝑃𝑧𝑡 = 1 Pa and 𝑃𝑧𝑏 = 0 Pa) with imposed
half-wave numbers 𝑚 = 𝑛 = 1).
𝑧∕ℎ
̃ 3D [34] Quasi-3D [47] 3D-u-𝜙 3D [34] Quasi-3D [47] 3D-u-𝜙 3D [34] Quasi-3D [47] 3D-u-𝜙
𝜎𝛼𝛼 Pa 𝜎𝑧𝑧 Pa 𝐷𝑧 (10−13 C∕m2 )
1.0 6.5643 6.5642 6.5642 1.0000 1.0000 1.0000 160.58 160.58 160.58
0.9 3.6408 3.6408 3.6408 0.9515 0.9515 0.9515 −0.3382 −0.3348 −0.3383
0.9 2.8855 2.8858 2.8802 0.9515 0.9518 0.9515 −0.3382 −0.3384 −0.3383
0.8 1.4499 1.4551 1.4501 0.8520 0.8517 0.8520 −0.1276 −0.1277 −0.1277
0.7 0.2879 0.2880 0.2880 0.7375 0.7376 0.7375 0.0813 0.0813 0.0813
0.6 −0.7817 −0.7829 −0.7818 0.6169 0.6168 0.6169 0.2913 0.2914 0.2914
0.5 −1.9266 −1.9266 −1.9267 0.4983 0.4986 0.4983 0.5052 0.5053 0.5061
0.5 0.0991 0.0991 0.0988 0.4983 0.4982 0.4983 0.5052 0.5053 0.5061
0.4 −0.0149 −0.0150 −0.0150 0.3805 0.3805 0.3805 0.7259 0.7262 0.7261
0.3 −0.1280 −0.1281 −0.1280 0.2614 0.2613 0.2614 0.9563 0.9565 0.9565
0.2 −0.2426 −0.2427 −0.2427 0.1482 0.1485 0.1482 1.1995 1.2000 1.1997
0.1 −0.3616 −0.3617 −0.3617 0.0487 0.0485 0.0487 1.4587 1.4590 1.4590
0.1 −4.2348 −4.2349 −4.2434 0.0487 0.0487 0.0487 1.4587 1.4559 1.4590
0.0 −6.8658 −6.8658 −6.8658 0.0000 0.0000 0.0000 −142.46 −142.46 −142.46

Table 4 Table 5
Assessment 1, simply-supported four-layered 𝑃 𝑍𝑇 −4∕0◦ ∕90◦ ∕𝑃 𝑍𝑇 −4 thick square plate Assessment 2, simply-supported five-layered 𝑃 𝑖𝑒𝑧𝑜∕0◦ ∕90◦ ∕0◦ ∕𝑃 𝑖𝑒𝑧𝑜 cylindrical panel in
in actuator configuration (electric potential applied at the top of the structure (𝛷𝑡 = 1 V sensor configuration (zero electric potential at external surfaces (𝛷𝑏 = 𝛷𝑡 = 0 V) and
and 𝛷𝑏 = 0 V) and zero transverse normal load at external surfaces (𝑃𝑧𝑡 = 𝑃𝑧𝑏 = 0 Pa) transverse normal load applied at the top of the structure (𝑃𝑧𝑡 = 1 Pa and 𝑃𝑧𝑏 = 0 Pa)
with imposed half-wave numbers 𝑚 = 𝑛 = 1.) with imposed half-wave numbers 𝑚 = 0, 𝑛 = 1). Non-dimensional displacements
10 𝐸 𝑤 ℎ3 100 𝐸 𝑣 ℎ3
𝑧∕ℎ
̃ 3D [34] Quasi-3D [47] 3D-u-𝜙 3D [34] Quasi-3D [47] 3D-u-𝜙 𝑤̄ = 𝑃 3 𝑅4 and 𝑣̄ = 𝑃 3𝑅4 (𝐸3 is related to composite CFRP-B in Table 1).
𝑧𝑡 𝛽 𝑧𝑡 𝛽

𝜙 (V) 𝑢 (10−12 m) 𝑅𝛽 ∕ℎ 3D [37] Quasi-3D [47] 3D-u-𝜙 3D [37] Quasi-3D [47] 3D-u-𝜙
1.0 1.0000 1.0000 1.0000 −32.764 −32.765 −32.765 𝑤̄ at 𝑧̃ = 0.5ℎ 𝑣̄ at 𝑧̃ = 0.5ℎ
0.9 0.9929 0.9929 0.9929 4.7356 4.7352 4.7352
2 1.440 1.443 1.441 5.294 5.308 5.250
0.9 0.9929 0.9929 0.9929 4.7356 4.7352 4.7352
4 0.459 0.458 0.458 1.549 1.547 1.540
0.8 0.8415 0.8418 0.8416 2.9808 2.9902 2.9806
10 0.144 0.144 0.144 0.480 0.479 0.480
0.7 0.7014 0.7015 0.7015 1.7346 1.7346 1.7346
50 0.0808 0.0810 0.0810 0.269 0.270 0.270
0.6 0.5707 0.5709 0.5707 0.8008 0.8045 0.8010
100 0.0785 0.0787 0.0787 0.262 0.262 0.263
0.5 0.4476 0.4477 0.4477 0.0295 0.0297 0.0297
0.5 0.4476 0.4477 0.4477 0.0295 0.0297 0.0297
0.4 0.3305 0.3307 0.3306 −0.4404 −0.4395 −0.4401
0.3 0.2179 0.2179 0.2179 −0.8815 −0.8811 −0.8811
0.2 0.1081 0.1082 0.1082 −1.3206 −1.3207 −1.3202 three composite layers have thickness ℎ2 = ℎ3 = ℎ4 = ℎ∕3 where ℎ =
0.1 −0.0010 −0.0010 −0.0001 −1.7839 −1.7834 −1.7834 ℎ2 + ℎ3 + ℎ4 ; the additional piezoelectric layers are very thin and their
0.1 −0.0010 −0.0010 −0.0001 −1.7839 −1.7834 −1.7834 additional thickness is ℎ1 = ℎ5 = ℎ∕100. Geometrical data are reported
0.0 0.0000 0.0000 0.0000 −2.8625 −2.8618 −2.8618
in Table 2 (A2). The in-plane dimension 𝑎 is not given because the half-
wave number in this direction is zero (𝑚 = 0 and 𝑛 = 1). A sensor
case has been evaluated considering a top load pressure 𝑃𝑧𝑡 = 1 Pa and
choice because the proposed method is not time consuming. Validation a short-circuit configuration (that means 𝛷𝑡 = 𝛷𝑏 = 0 V). The 3D-u-
of the 3D-u-𝜙 model is performed comparing results with different 3D 𝜙 model has been compared with the 3D reference solution by Chen
electro-elastic models and quasi-3D theories. et al. [37] and a quasi-3D layer wise analytical model [47]. The results
are compared considering different 𝑅𝛽 ∕ℎ thickness ratios, from thick
For the first assessment (A1), a multilayered 𝑃 𝑍𝑇 −4∕0◦ ∕90◦ ∕𝑃 𝑍𝑇 −
(𝑅𝛽 ∕ℎ = 2) to thin (𝑅𝛽 ∕ℎ = 100) cases. Non-dimensional displacements
4 thick square plate is evaluated. The two external layers in piezo-
𝑤̄ and 𝑣,
̄ presented in Table 5, show very good accordance with the two
ceramic lead zirconate titanate material (PZT-4) and the two 0◦ ∕90◦
reference solutions for all the thickness ratios analyzed. Therefore, all
internal composite layers in carbon fiber reinforced plastic - version
the effects regarding the curvature of the cylindrical panel are properly
A (CFRP-A) have elastic, piezoelectric and permittivity properties for
depicted.
sensor and actuator cases as shown in Table 1. Geometrical data are
given in Table 2 (A1). ℎ is the total thickness, the two external layers In the third assessment (A3), a single-layered polyvinylidene fluo-
have thickness ℎ1 = ℎ4 = 0.1ℎ and the two internal layers have ride (PVDF) cylindrical panel is investigated. Material properties for
thickness ℎ2 = ℎ3 = 0.4ℎ. In the sensor case, a top pressure 𝑃𝑧𝑡 = 1 Pa is the PVDF are collected in Table 1 and geometrical data can be seen in
applied with 𝛷𝑡 = 𝛷𝑏 = 0 V. The actuator case is performed imposing Table 2 (A3). A sensor configuration has been studied, considering the
electric potential at the top 𝛷𝑡 = 1 V and at the bottom 𝛷𝑏 = 0 V same load conditions already described for A2. The reference solutions
without any pressure load at the external surfaces. Half-wave numbers are the 3D model presented by Dumir et al. [38] and a quasi-3D layer
are 𝑚 = 𝑛 = 1. The present model has been validated with the 3D wise analytical model [47]. Results, presented in Table 6, are proposed
exact electro-elastic model by Heyliger [34] and a quasi-3D analytical for several thickness ratios 𝑅𝛽 ∕ℎ. Non-dimensional variables in Table 6
model in layer wise form [47]. Results show a perfect match between show satisfactory results for each 𝑅𝛽 ∕ℎ considered. Therefore, electro-
the three models along the thickness direction for both sensor (Table 3) elastic coupling is perfectly described for moderately thick and thin
and actuator (Table 4) cases. Material layer effect, thickness layer effect cylindrical panels.
and electro-elastic coupling have been correctly depicted in terms of The fourth assessment (A4) is a one-layered functionally graded
𝜎𝛼𝛼 , 𝜎𝑧𝑧 , 𝐷𝑧 , 𝑢 and 𝜙. piezoelectric square plate. Both sensor and actuator cases have been
The second assessment (A2) is focused on a three-layered 0◦ ∕90◦ ∕0◦ considered. The values proposed in the sixth column of Table 1 are
cylindrical panel with total thickness ℎ (see carbon fiber reinforced valid for the bottom (b) of the FGPM plate (called as 𝐸𝑖𝑏 , 𝐺𝑖𝑗 𝑏 , 𝑒𝑖𝑗 𝑏
plastic - version B (CFRP-B) in Table 1). Two piezoelectric patches and 𝜖𝑖𝑗 𝑏 in Eqs. (48)–(51)). They have been obtained from the PZT-4
(see Piezo in Table 1) have been applied at the external surfaces. The coefficients presented by Lu et al. in [64]. The exponential laws adopted

8
S. Brischetto and D. Cesare International Journal of Mechanical Sciences 280 (2024) 109620

Table 6
Assessment 3, simply-supported single-layered PVDF cylindrical panel in sensor configuration (zero electric potential at external surfaces
(𝛷𝑏 = 𝛷𝑡 = 0 V) and transverse normal load applied at the top of the structure (𝑃𝑧𝑡 = 1 Pa and 𝑃𝑧𝑏 = 0 Pa) with imposed half-wave numbers
ℎ4 𝜎 𝜎 ℎ −9 𝐸 𝜙 ℎ 𝐷 ℎ
̄ 𝑣̄ >= 100 𝐸 <𝑤,𝑣>
𝑚 = 0, 𝑛 = 1). Non-dimensional unknowns < 𝑤, , 𝜎̄ 𝑧𝑧 = 𝑃𝑧𝑧 , 𝜎̄ 𝛽𝑧 = 𝑅 𝛽𝑧𝑃 , 𝜙̄ = 30×10 and 𝐷̄ 𝛽 = 30×10𝛽−11 𝑅 (𝐸 is related to PVDF
ℎ 𝑃 𝑅4 𝑧𝑡 𝛽 𝑧𝑡 𝛽 𝑧𝑡 𝑃 𝑅2
𝑧𝑡 𝛽 𝛽

in Table 1).
𝑅𝛽 /h 2 4 6 10 20 100 500
𝑤̄ at 𝑧̃ = 0
3D [38] −28.65 −20.55 −18.72 −17.60 −16.96 −16.55 −16.48
Quasi-3D [47] −28.63 −20.55 −18.72 −17.60 −16.95 −16.55 −16.48
3D-u-𝜙 −28.65 −20.55 −18.72 −17.60 −16.96 −16.55 −16.48
𝑤̄ at 𝑧̃ = 0.5ℎ
3D [38] −31.45 −21.10 −18.96 −17.68 −16.98 −16.55 −16.48
Quasi-3D [47] −31.47 −21.10 −18.96 −17.68 −16.98 −16.55 −16.48
3D-u-𝜙 −31.46 −21.10 −18.96 −17.68 −16.98 −16.55 −16.48
𝑤̄ at 𝑧̃ = ℎ
3D [38] −31.31 −20.65 −18.73 −17.60 −16.95 −16.55 −16.48
Quasi-3D [47] −31.30 −20.65 −18.73 −17.60 −16.95 −16.55 −16.48
3D-u-𝜙 −31.31 −20.65 −18.73 −17.60 −16.96 −16.55 −16.48
𝑣̄ at 𝑧̃ = 0
3D [38] −23.61 −13.09 −10.23 −8.177 −6.778 −5.738 −5.539
Quasi-3D [47] −23.59 −13.09 −10.23 −8.177 −6.778 −5.738 −5.539
3D-u-𝜙 −23.60 −13.09 −10.23 −8.177 −6.778 −5.738 −5.539
𝑣̄ at 𝑧̃ = ℎ
3D [38] 2.046 −0.8806 −2.331 −3.572 −4.528 −5.297 −5.451
Quasi-3D [47] 2.045 −0.8806 −2.331 −3.572 −4.528 −5.297 −5.451
3D-u-𝜙 2.044 −0.8807 −2.331 −3.572 −4.528 −5.297 −5.451
𝜎̄ 𝑧𝑧 at 𝑧̃ = 0.5ℎ
3D [38] −0.2906 0.2170 0.6356 1.420 3.319 18.34 93.34
Quasi-3D [47] −0.2359 0.2453 0.6541 1.430 3.325 18.34 93.34
3D-u-𝜙 −0.2906 0.2171 0.6356 1.420 3.319 18.34 93.34
𝜎̄ 𝛽𝑧 at 𝑧̃ = 0.5ℎ
3D [38] −0.6653 −0.6238 −0.6055 −0.5893 −0.5762 −0.5653 −0.5631
Quasi-3D [47] −0.6742 −0.6261 −0.6065 −0.5896 −0.5763 −0.5653 −0.5631
3D-u-𝜙 −0.6653 −0.6238 −0.6055 −0.5893 −0.5762 −0.5653 −0.5631
𝜙̄ at 𝑧̃ = 0.5ℎ
3D [38] 1.734 2.443 2.541 2.560 2.540 2.504 2.494
Quasi-3D [47] 1.729 2.442 2.540 2.560 2.540 2.504 2.494
3D-u-𝜙 1.810 2.450 2.542 2.560 2.540 2.504 2.494
𝐷̄ 𝛽 at 𝑧̃ = 0.5ℎ
3D [38] −0.3070 −0.4324 −0.4497 −0.4531 −0.4496 −0.4431 −0.4415
Quasi-3D [47] −0.3060 −0.4323 −0.4497 −0.4531 −0.4496 −0.4431 −0.4414
3D-u-𝜙 −0.3204 −0.4337 −0.4491 −0.4531 −0.4496 −0.4431 −0.4415

for elastic, piezoelectric and permittivity coefficients are: respectively. The validation results for the sensor case in Tables 7 and
[ 𝑧̃ ] 8 show a very good accordance for all the variables presented. Some
𝑝
̃ = 𝐸𝑖𝑏 𝑒 ℎ𝐹 𝐺𝑀 ,
𝐸𝑖 (𝑧) (48) slight differences occur for exponential terms 𝑝 = ±1 in thick plates. The
[ ] electro-elastic coupling is very well depicted for all 𝑝 values. Moreover,
𝑝 ℎ 𝑧̃ it is possible to notice the perfect accordance of the 3D-u-𝜙 model
𝐺𝑖𝑗 (𝑧)
̃ = 𝐺𝑖𝑗 𝑏 𝑒 𝐹 𝐺𝑀 , (49)
with the 3D electroelastic reference solution by Lu et al. [64] also for
[ ]
𝑝 ℎ 𝑧̃ the transverse normal stress 𝜎𝑧𝑧 . In Tables 9 and 10, no differences
𝑒𝑖𝑗 (𝑧)
̃ = 𝑒𝑖𝑗 𝑏 𝑒 𝐹 𝐺𝑀 , (50) occur for the primary variables (𝑢, 𝑤 and 𝜙) for thick plates but some
[ ] slight differences are shown for in-plane normal/shear stresses. Electro-
𝑝 ℎ 𝑧̃
𝜖𝑖𝑖 (𝑧)
̃ = 𝜖𝑖𝑗 𝑏 𝑒 𝐹 𝐺𝑀 , (51) elastic coupling is perfectly evaluated in this case. It is clear the perfect
depiction of the electro-elastic coupling and the correct evaluation of
where 𝑝 = −1, −0.5, 0, 0.5, 1 and 0 ≤ 𝑧̃ ≤ ℎ𝐹 𝐺𝑀 . 𝑝 = 0 means classical
the transverse normal stress 𝜎𝑧𝑧 for each considered thickness value.
material. Geometrical data are collected in Table 2 (A4). The load
This last assessment case has been useful to validate the capability of
boundary conditions for the two cases are: 𝑃𝑧𝑡 = −1 Pa and 𝛷𝑡 =
the 3D-u-𝜙 model to evaluate the behavior of the FGPMs.
𝛷𝑏 = 0 V for sensor configuration; 𝛷𝑡 = 1 V and 𝛷𝑏 = 0 V and zero
mechanical loads for actuator configuration. In both cases, half-wave Considering all the assessment cases here presented and the ob-
numbers are 𝑚 = 𝑛 = 1. Adopted reference solutions are the 3D electro- tained results, it is possible to confirm the validation of the 3D-u-𝜙
elastic model by Lu et al. [64] and the quasi-3D layer wise model model for both multilayered and functionally graded structures. In ad-
in [71]. For each ℎ thickness value (0.1 m, 0.01 m, 0.15 m and 0.25 m), dition, geometries with curvature are also validated. The 3D-u-𝜙 model
different exponential terms 𝑝 are investigated for both load conditions. is validated considering 300 mathematical layers and an expansion
Tables 7 and 8 show results for sensor configurations for thin and for the exponential matrix 𝑁 = 3. These two parameters permit the
moderately thick structures, respectively. Tables 9 and 10 show results depiction of all the possible peculiarities in terms of electro-elastic
for actuator configurations for thin and moderately thick structures, coupling and material/thickness layer effects.

9
S. Brischetto and D. Cesare International Journal of Mechanical Sciences 280 (2024) 109620

Table 7
Assessment 4, simply-supported one-layered FGPM thin plate in sensor configuration (zero electric potential at external surfaces (𝛷𝑏 = 𝛷𝑡 = 0 V)
and transverse normal load applied at the top of the structure (𝑃𝑧𝑡 = −1 Pa and 𝑃𝑧𝑏 = 0 Pa) with imposed half-wave numbers 𝑚 = 𝑛 = 1). All
the values are computed in 𝑥 = 0.25𝑎 and 𝑦 = 0.25𝑏.
𝑝 −1 −0.5 0 0.5 1
ℎ = 0.01 m
𝑢 (10−11 m) at 𝑧̃ = 0.5ℎ
3D [64] −0.500 −0.500 −0.500 −0.500 −0.500
Quasi-3D [71] −0.511 −0.507 −0.500 −0.490 −0.480
3D-u-𝜙 −0.500 −0.500 −0.500 −0.500 −0.500
ℎ = 0.1 m
𝑢 (10−11 m) at 𝑧̃ = 0.5ℎ
3D [64] 0.650 0.247 −0.031 −0.197 −0.275
Quasi-3D [71] 0.647 0.243 −0.029 −0.192 −0.273
3D-u-𝜙 0.672 0.246 −0.029 −0.195 −0.284
𝑤 (10−9 m) at 𝑧̃ = 0.5ℎ
3D [64] −0.251 −0.196 −0.151 −0.119 −0.094
Quasi-3D [71] −0.248 −0.194 −0.151 −0.118 −0.092
3D-u-𝜙 −0.253 −0.195 −0.151 −0.119 −0.094
𝜙 (10−2 V) at 𝑧̃ = 0.5ℎ
3D [64] −0.596 −0.492 −0.393 −0.299 −0.220
Quasi-3D [71] −0.591 −0.493 −0.393 −0.299 −0.217
3D-u-𝜙 −0.601 −0.496 −0.394 −0.301 −0.222
𝜎𝛼𝛼 (Pa) at 𝑧̃ = 0
3D [64] 15.23 13.05 10.92 9.138 7.529
Quasi-3D [71] 15.21 12.97 10.96 9.156 7.504
3D-u-𝜙 15.38 13.00 10.96 9.190 7.643
𝜎𝛼𝛽 (Pa) at 𝑧̃ = ℎ
3D [64] 3.041 3.616 4.247 4.959 5.830
Quasi-3D [71] 3.039 3.612 4.241 4.952 5.773
3D-u-𝜙 3.108 3.634 4.245 4.968 5.836
𝜎𝑧𝑧 (Pa) at 𝑧̃ = ℎ
3D [64] −0.500 −0.500 −0.500 −0.500 −0.500
Quasi-3D [71] −0.511 −0.507 −0.500 −0.491 −0.481
3D-u-𝜙 −0.500 −0.500 −0.500 −0.500 −0.500

Table 8 ered and functionally graded piezoelectric. The following four bench-
Assessment 4, simply-supported one-layered FGPM thick plate in sensor configuration
marks are presented considering different thickness ratios (from thick
(zero electric potential at external surfaces (𝛷𝑏 = 𝛷𝑡 = 0 V) and transverse normal load
applied at the top of the structure (𝑃𝑧𝑡 = −1 Pa and 𝑃𝑧𝑏 = 0 Pa) with imposed half-wave
to thin configurations). For each benchmark, the behavior of the struc-
numbers 𝑚 = 𝑛 = 1). All the values are computed in 𝑥 = 0.25𝑎 and 𝑦 = 0.25𝑏. ture is investigated in both sensor and actuator cases. Six different
𝑝 −1 −0.5 0 0.5 1 electro-elastic variables (𝑢, 𝑤, 𝜎𝛼𝛼 , 𝜎𝑧𝑧 , 𝜙 and 𝐷𝑧 ) are given at different
ℎ = 0.15 m
thickness locations. For all benchmarks, 300 mathematical layers have
been used together with the exponential matrix considered at the third
𝑤 (10−10 m) at 𝑧̃ = 0.5ℎ
order (𝑁 = 3). These values are a very conservative choice suggested
3D [64] −0.780 −0.615 −0.480 −0.375 −0.294
by past authors’ works about the study of convergence in the case of
Quasi-3D [71] −0.777 −0.611 −0.479 −0.375 −0.293
3D-u-𝜙 −0.792 −0.613 −0.479 −0.377 −0.299 pure elastic problems.
𝜎𝑧𝑧 (Pa) at 𝑧̃ = ℎ The first benchmark (B1) is a sandwich (PZT-4/Al2024/Foam/
3D [64] −0.500 −0.500 −0.500 −0.500 −0.500 Al2024/PZT-4) square plate. The thickness of each physical layer is:
Quasi-3D [71] −0.511 −0.507 −0.500 −0.490 −0.480 ℎ𝑃 𝑍𝑇 −4 = ℎ𝐴𝑙2024 = 0.1ℎ and ℎ𝐹 𝑜𝑎𝑚 = 0.6ℎ with ℎ as total thickness.
3D-u-𝜙 −0.500 −0.500 −0.500 −0.500 −0.500
Material properties and geometrical data can be seen in Tables 1 and
ℎ = 0.25 m 11, respectively. For the sensor configuration, a top pressure load is
𝑤 (10−10 m) at 𝑧̃ = 0.5ℎ imposed as 𝑃𝑧𝑡 = 10000 Pa with a short-circuit configuration (𝛷𝑡 = 𝛷𝑏 =
3D [64] −0.200 −0.157 −0.124 −0.098 −0.078 0 V). The actuator case is performed considering an imposed electric
Quasi-3D [71] −0.196 −0.156 −0.124 −0.098 −0.076 potential at the top 𝛷𝑡 = 100 V and at the bottom 𝛷𝑏 = 0 V, without
3D-u-𝜙 −0.200 −0.157 −0.124 −0.098 −0.078 any pressure load applied. Imposed half-wave numbers are 𝑚 = 𝑛 = 1.
𝜎𝑧𝑧 (Pa) at 𝑧̃ = ℎ Different thickness ratios 𝑎∕ℎ are presented for both load conditions. In
3D [64] −0.500 −0.500 −0.500 −0.500 −0.500 Tables 12 and 13 six variables (𝑢, 𝑤, 𝜎𝛼𝛼 , 𝜎𝑧𝑧 , 𝜙 and 𝐷𝑧 ) are presented
Quasi-3D [71] −0.511 −0.507 −0.500 −0.490 −0.480 at different thickness locations. It is possible to notice, for the sensor
3D-u-𝜙 −0.500 −0.500 −0.500 −0.500 −0.500
configuration, the increasing trend of the transverse in-plane/normal
stresses, electric potential and electric displacement going from thick
to thin thickness ratios. This effect happens because the stiffness of the
3.2. Benchmarks structure decreases, so the stresses and electric field increase. In the
actuator configuration, the trend is different because pressure loads are
not applied. In Figs. 2–5, variables 𝑢, 𝑤, 𝜙, 𝐷𝑧 , 𝜎𝛼𝛼 , 𝜎𝑧𝑧 , 𝜎𝛼𝑧 and 𝜎𝛽𝑧
In the present subsection four new benchmarks are proposed for are plotted along the thickness direction for 𝑎∕ℎ = 20 case for both
each possible geometry. The material configurations can be multilay- sensor and actuator cases. Changing in slope between two different

10
S. Brischetto and D. Cesare International Journal of Mechanical Sciences 280 (2024) 109620

Table 9 Table 11
Assessment 4, simply-supported one-layered FGPM thin plate in actuator configuration Geometrical data and imposed half-wave numbers 𝑚 and 𝑛 for the four benchmarks
(electric potential applied at the top of the structure (𝛷𝑡 = 1 V and 𝛷𝑏 = 0 V) and zero (B1-B4).
transverse normal load at external surfaces (𝑃𝑧𝑡 = 𝑃𝑧𝑏 = 0 Pa) with imposed half-wave B1 B2 B3 B4
numbers 𝑚 = 𝑛 = 1). All the values are computed in 𝑥 = 0.25𝑎 and 𝑦 = 0.25𝑏. 𝜋 𝜋
a (m) 1 3
10 2𝜋10 3
10
𝑝 −1 −0.5 0 0.5 1 𝜋
b (m) 1 1 100 3
10
ℎ = 0.01 m h (m) Variable Variable Variable Variable
𝑢 (10−11 m) at 𝑧̃ = 0.5ℎ R𝛼 (m) ∞ 10 10 10
R𝛽 (m) ∞ ∞ ∞ 10
3D [64] 0.000 0.000 0.000 0.000 0.000 m 1 1 2 1
Quasi-3D [71] 0.000 0.000 0.000 0.000 −0.002 n 1 0 1 1
3D-u-𝜙 0.000 0.000 0.000 0.000 0.000
ℎ = 0.1 m
Table 12
𝑝 −1 −0.5 0 0.5 1
Benchmark 1, simply-supported five-layered PZT-4/Al2024/Foam/Al2024/PZT-4 sand-
−9
𝑢 (10 m) at 𝑧̃ = 0.5ℎ wich square plate in sensor configuration (zero electric potential at external surfaces
3D [64] −0.129 −0.129 −0.129 −0.129 −0.129 (𝛷𝑏 = 𝛷𝑡 = 0 V) and transverse normal load applied at the top of the structure
Quasi-3D [71] −0.125 −0.128 −0.128 −0.128 −0.125 (𝑃𝑧𝑡 = 10 000 Pa and 𝑃𝑧𝑏 = 0 Pa) with imposed half-wave numbers 𝑚 = 𝑛 = 1).
3D-u-𝜙 −0.129 −0.129 −0.129 −0.129 −0.129 a/h 4 10 20 50 100
𝑤 (10−9 m) at 𝑧̃ = 0.5ℎ 𝑢 (10−7 m) at 𝑧̃ = 0.75ℎ
3D [64] 0.327 0.140 −0.074 −0.280 −0.480 8.6313 12.619 6.2546 −59.111 −296.59
Quasi-3D [71] 0.325 0.139 −0.072 −0.283 −0.466 𝑤 (10−6 m) at 𝑧̃ = 0.5ℎ
3D-u-𝜙 0.338 0.141 −0.072 −0.287 −0.487 8.9184 54.445 156.24 853.45 4703.4
𝜙 (V) at 𝑧̃ = 0.5ℎ 𝜎𝛼𝛼 (103 Pa) at 𝑧̃ = 0.75ℎ
3D [64] 0.181 0.210 0.242 0.272 0.300 2.5461 3.1429 4.0125 9.9881 31.321
Quasi-3D [71] 0.183 0.212 0.243 0.273 0.303 𝜎𝑧𝑧 (103 Pa) at 𝑧̃ = 0.95ℎ
3D-u-𝜙 0.183 0.213 0.243 0.273 0.303 9.3376 9.7171 9.8427 9.8884 9.8954
𝜎𝛼𝛼 (Pa) at 𝑧̃ = 0 𝜙 (V) at 𝑧̃ = 0.15ℎ
3D [64] −18.06 −19.71 −23.79 −31.75 −44.27 51.426 77.653 102.79 208.25 401.46
Quasi-3D [71] −18.04 −19.71 −23.68 −31.75 −45.61 𝐷𝑧 (10−9 C∕m2 ) at 𝑧̃ = 0.85ℎ
3D-u-𝜙 −17.82 −19.66 −23.81 −31.38 −43.45 −1.1991 −0.2870 −0.0552 0.0235 0.0354
𝜎𝛼𝛽 (Pa) at 𝑧̃ = ℎ
3D [64] −14.22 −20.31 −27.19 −34.84 −41.72
Table 13
Quasi-3D [71] −13.80 −19.99 −27.16 −34.56 −41.00
Benchmark 1, simply-supported five-layered PZT-4/Al2024/Foam/Al2024/PZT-4 sand-
3D-u-𝜙 −14.36 −20.26 −27.27 −34.84 −41.83
wich square plate in actuator configuration (electric potential applied at the top of
𝜎𝑧𝑧 (Pa) at 𝑧̃ = ℎ the structure (𝛷𝑡 = 100 V and 𝛷𝑏 = 0 V) and zero transverse normal load at external
3D [64] 0.000 0.000 0.000 0.000 0.000 surfaces (𝑃𝑧𝑡 = 𝑃𝑧𝑏 = 0 Pa) with imposed half-wave numbers 𝑚 = 𝑛 = 1).
Quasi-3D [71] 0.002 0.003 0.005 0.008 0.001 a/h 4 10 20 50 100
3D-u-𝜙 0.000 0.000 0.000 0.000 0.000
𝑢 (10−11 m) at 𝑧̃ = 0.75ℎ
6.8261 16.335 8.4438 −0.70485 −8.0823

Table 10 𝑤 (10−9 m) at 𝑧̃ = 0.5ℎ

Assessment 4, simply-supported one-layered FGPM thick plate in actuator configuration −5.4491 −2.7851 −1.6696 −1.2583 −1.1948
(electric potential applied at the top of the structure (𝛷𝑡 = 1 V and 𝛷𝑏 = 0 V) and zero 𝜎𝛼𝛼 (Pa) at 𝑧̃ = 0.75ℎ
transverse normal load at external surfaces (𝑃𝑧𝑡 = 𝑃𝑧𝑏 = 0 Pa) with imposed half-wave −1.9234 −0.29305 −0.09300 0.00623 0.07299
numbers 𝑚 = 𝑛 = 1). All the values are computed in 𝑥 = 0.25𝑎 and 𝑦 = 0.25𝑏.
𝜎𝑧𝑧 (10−5 Pa) at 𝑧̃ = 0.95ℎ
𝑝 −1 −0.5 0 0.5 1
−92196 −7080.2 −887.69 −36.241 5.1601
ℎ = 0.15 m
𝜙 (V) at 𝑧̃ = 0.15ℎ
𝜙 (V) at 𝑧̃ = 0.5ℎ 5.4381 6.1172 6.2215 6.2509 6.2552
3D [64] 0.176 0.203 0.232 0.263 0.292 𝐷𝑧 (10−9 C∕m2 ) at 𝑧̃ = 0.85ℎ
Quasi-3D [71] 0.177 0.205 0.234 0.263 0.292 −5.2884 −11.431 −22.316 −55.403 −110.70
3D-u-𝜙 0.177 0.205 0.234 0.263 0.292
𝜎𝑧𝑧 (Pa) at 𝑧̃ = ℎ
3D [64] 0.000 0.000 0.000 0.000 0.000
Quasi-3D [71] 0.005 0.009 0.016 0.028 0.032
in the 3D-u-𝜙 model is clearly shown; displacements, electric potential,
3D-u-𝜙 0.000 0.000 0.000 0.000 0.000 transverse normal electric displacement, transverse normal stress and
ℎ = 0.25 m transverse shear stresses are continuous along the thickness direction
and at layer interfaces because interlaminar continuity conditions have
𝜙 (V) at 𝑧̃ = 0.5ℎ
been correctly imposed in the model. In-plane normal stress shows
3D [64] 0.159 0.184 0.209 0.235 0.262
Quasi-3D [71] 0.159 0.184 0.209 0.236 0.262
typical discontinuity at interfaces meanwhile transverse normal stress
3D-u-𝜙 0.159 0.184 0.210 0.236 0.262 and transverse shear stresses are continuous. Sensor configuration (see
𝜎𝑧𝑧 (Pa) at 𝑧̃ = ℎ
Figs. 2 and 3) has 𝜎𝑧𝑧 = 10000 Pa at the top and 𝜎𝑧𝑧 = 0 Pa at the
bottom and 𝛷 = 0 V at external surfaces. No mechanical loads are
3D [64] 0.000 0.000 0.000 0.000 0.000
Quasi-3D [71] 0.021 0.036 0.073 0.130 0.201 present in 𝛼 and 𝛽 directions, for this reason 𝜎𝛼𝑧 = 0 Pa and 𝜎𝛽𝑧 = 0 Pa
3D-u-𝜙 0.000 0.000 0.000 0.000 0.000 at the outer surfaces. Actuator configuration (see Figs. 4 and 5) has
𝜎𝑧𝑧 = 𝜎𝛼𝑧 = 𝜎𝛽𝑧 = 0 Pa at external surfaces and 𝛷 = 100 V at the top
and 𝛷 = 0 V at the bottom. Therefore, load boundary conditions for
actuator and sensor cases have been correctly imposed by the 3D-u-𝜙
physical layers is a clear effect of the multilayered configuration and model.
transverse anisotropy (zigzag effect). The layer-wise approach adopted

11
S. Brischetto and D. Cesare International Journal of Mechanical Sciences 280 (2024) 109620

Fig. 2. Benchmark 1, simply-supported five-layered PZT-4/Al2024/Foam/Al2024/PZT-4 sandwich square plate in sensor configuration (zero electric potential at external surfaces
(𝛷𝑏 = 𝛷𝑡 = 0 V) and transverse normal load applied at the top of the structure (𝑃𝑧𝑡 = 10000 Pa and 𝑃𝑧𝑏 = 0 Pa) with imposed half-wave numbers 𝑚 = 𝑛 = 1). Thickness ratio 𝑎∕ℎ = 20.
Evaluation of: (a) in-plane displacement 𝑢, (b) transverse normal displacement 𝑤, (c) electric potential 𝜙 and (d) transverse normal electric displacement 𝑧 through the thickness
direction 𝑧.

The second benchmark (B2) presents a single-layered FGPM cylin- conditions. Figs. 8 and 9 for actuator case shows 𝜎𝑧𝑧 = 𝜎𝛼𝑧 = 𝜎𝛽𝑧 = 0 Pa
drical shell. Material properties are presented in Table 1, and geomet- at the external surfaces as no mechanical loads have been applied and
rical data are reported in Table 11. The same load boundary conditions 𝛷 = 100 V at the top and 𝛷 = 0 V at the bottom, confirming the load
proposed in the B1 are valid for present sensor and actuator cases but boundary conditions.
with half-wave number impositions 𝑚 = 1 and 𝑛 = 0. The exponential In the third benchmark (B3), a multilayered FGPM cylinder is pro-
laws used for the material properties variation along the thickness posed. The thickness of each physical layer (from bottom to top) is: ℎ1 =
direction are those in Eqs. (48)–(51) considering 𝑝 = 0.5. Different ℎ3 = 0.1ℎ and ℎ2 = 0.8ℎ with ℎ as total thickness. Material properties
thickness ratios 𝑅𝛼 ∕ℎ are considered. In Tables 14 and 15 the same at the bottom layer with thickness ℎ1 are the same shown in Table 1 in
six variables (𝑢, 𝑤, 𝜎𝛼𝛼 , 𝜎𝑧𝑧 , 𝜙 and 𝐷𝑧 ) of the B1 case are proposed the FGPM column: this is a classical layer. In the central layer, material
at different thickness locations. The same considerations done for B1 properties vary with the exponential laws defined in Eqs. (48)–(51)
sensor and actuator cases are also valid for the present benchmark. In and considering 𝑝 = −0.5 and ℎ𝐹 𝐺𝑀 = 0.8ℎ. The third layer at the
Figs. 6–9, the trends along the thickness direction of the 𝑢, 𝑤, 𝜙, 𝐷𝑧 , top maintains the constant material properties reached by the FGPM
𝜎𝛼𝛼 , 𝜎𝑧𝑧 , 𝜎𝛼𝑧 and 𝜎𝛽𝑧 variables are proposed for 𝑅𝛼 ∕ℎ = 20 for both layer at the top of the central layer. Geometrical data are collected
load conditions. In this benchmark, the main peculiarity of the FGPMs in Table 11. Sensor and actuator cases are investigated for different
is highlighted: the use of a through-the-thickness continuous material thickness ratios 𝑅𝛼 ∕ℎ considering the same load conditions described
in place of multilayered configurations reduces the zigzag effect. In in the B1, 𝑚 = 2 and 𝑛 = 1 are imposed in this case. The values of
Figs. 7 and 9, it must be noted that 𝜎𝛽𝑧 trend is null due to an imposed the six variables 𝑢, 𝑤, 𝜎𝛼𝛼 , 𝜎𝑧𝑧 , 𝜙 and 𝐷𝑧 at specific thickness locations
null half-wave number 𝑛 in 𝛽 direction. Figs. 6 and 7 for sensor case are collected in Tables 16 and 17. For both sensor and actuator cases,
confirms the load boundary conditions because 𝜎𝑧𝑧 = 10000 Pa at the displacements increase when the thickness decreases. Figs. 10–13 show
top and 𝜎𝑧𝑧 = 0 Pa at the bottom and no mechanical loads are applied the thickness trends for 𝑅𝛼 ∕ℎ = 20 of the same investigated variables
in the in-plane directions. 𝛷𝑡 = 𝛷𝑏 = 0 V confirms the short circuit in Tables 16 and 17 and also for transverse shear stresses 𝜎𝛼𝑧 and

12
S. Brischetto and D. Cesare International Journal of Mechanical Sciences 280 (2024) 109620

Fig. 3. Benchmark 1, simply-supported five-layered PZT-4/Al2024/Foam/Al2024/PZT-4 sandwich square plate in sensor configuration (zero electric potential at external surfaces
(𝛷𝑏 = 𝛷𝑡 = 0 V) and transverse normal load applied at the top of the structure (𝑃𝑧𝑡 = 10000 Pa and 𝑃𝑧𝑏 = 0 Pa) with imposed half-wave numbers 𝑚 = 𝑛 = 1). Thickness ratio 𝑎∕ℎ = 20.
Evaluation of: (a) in-plane normal stress 𝜎𝛼𝛼 , (b) transverse normal stress 𝜎𝑧𝑧 , (c) transverse shear stress 𝜎𝛼𝑧 and (d) transverse shear stress 𝜎𝛽𝑧 through the thickness direction 𝑧.

Table 14 Table 15
Benchmark 2, simply-supported single layered FGPM (𝑝 = 0.5) cylindrical shell in sensor Benchmark 2, simply-supported single layered FGPM (𝑝 = 0.5) cylindrical shell in
configuration (zero electric potential at external surfaces (𝛷𝑏 = 𝛷𝑡 = 0 V) and transverse actuator configuration (electric potential applied at the top of the structure (𝛷𝑡 = 100 V
normal load applied at the top of the structure (𝑃𝑧𝑡 = 10 000 Pa and 𝑃𝑧𝑏 = 0 Pa) with and 𝛷𝑏 = 0 V) and zero transverse normal load at external surfaces (𝑃𝑧𝑡 = 𝑃𝑧𝑏 = 0 Pa)
imposed half-wave numbers 𝑚 = 1, 𝑛 = 0). with imposed half-wave numbers 𝑚 = 1, 𝑛 = 0).
𝑅𝛼 /h 4 10 20 50 100 𝑅𝛼 /h 4 10 20 50 100
𝑢 (10−5 m) at 𝑧̃ = ℎ 𝑢 (10−8 m) at 𝑧̃ = ℎ
0.09547 3.2468 31.414 543.32 4487.84 −3.1151 −5.1429 −7.1937 −0.86708 51.985
𝑤 (10−4 m) at 𝑧̃ = 0 𝑤 (10−7 m) at 𝑧̃ = 0
0.11336 1.4939 11.470 175.63 1396.7 0.06845 0.40726 1.4810 8.5166 32.959
𝜎𝛼𝛼 (104 Pa) at 𝑧̃ = 0.5ℎ 𝜎𝛼𝛼 (102 Pa) at 𝑧̃ = 0.5ℎ
−1.2993 −6.9228 −26.513 −161.59 −641.02 −1.2447 −1.4485 −1.7893 −2.8503 −4.6368
𝜎𝑧𝑧 (103 Pa) at 𝑧̃ = 0.75ℎ 𝜎𝑧𝑧 (Pa) at 𝑧̃ = 0.75ℎ
2.8012 −6.9489 −22.610 −69.202 −146.72 −6.7694 −11.025 −12.675 −13.713 −14.067
𝜙 (103 V) at 𝑧̃ = 0.25ℎ 𝜙 (V) at 𝑧̃ = 0.25ℎ
0.56208 1.2259 2.3477 5.7217 11.348 27.774 30.368 30.800 30.942 30.969
𝐷𝑧 (10−5 C∕m2 ) at 𝑧̃ = 0.5ℎ 𝐷𝑧 (10−6 C∕m2 ) at 𝑧̃ = 0.5ℎ
0.28824 1.1123 4.1592 25.687 102.82 −0.78131 −1.9958 −4.0138 −10.062 −20.139

𝜎𝛽𝑧 . The correct imposition of the interlaminar continuity conditions use of a FGPM core, even in-plane stress 𝜎𝛼𝛼 is continuous at interfaces
is clearly shown for both load conditions. In this case, thanks to the but zig-zag effect is present because the changing material properties

13
S. Brischetto and D. Cesare International Journal of Mechanical Sciences 280 (2024) 109620

Fig. 4. Benchmark 1, simply-supported five-layered PZT-4/Al2024/Foam/Al2024/PZT-4 sandwich square plate in actuator configuration (electric potential applied at the top of the
structure (𝛷𝑡 = 100 V and 𝛷𝑏 = 0 V) and zero transverse normal load at external surfaces (𝑃𝑧𝑡 = 𝑃𝑧𝑏 = 0 Pa) with imposed half-wave numbers 𝑚 = 𝑛 = 1). Thickness ratio 𝑎∕ℎ = 20.
Evaluation of: (a) in-plane displacement 𝑢, (b) transverse normal displacement 𝑤, (c) electric potential 𝜙 and (d) transverse normal electric displacement 𝑧 through the thickness
direction 𝑧.

between layers. Transverse normal stress 𝜎𝑧𝑧 and transverse shear Table 16
Benchmark 3, simply-supported multilayered sandwich cylinder (embedding an FGPM
stresses 𝜎𝛼𝑧 , 𝜎𝛽𝑧 are continuous thanks to the correct imposition of the
core with 𝑝 = −0.5) in sensor configuration (zero electric potential at external surfaces
interlaminar continuity conditions. The load boundary conditions for (𝛷𝑏 = 𝛷𝑡 = 0 V) and transverse normal load applied at the top of the structure
sensor and actuator configurations are confirmed in Figs. 11 and 13 in (𝑃𝑧𝑡 = 10 000 Pa and 𝑃𝑧𝑏 = 0 Pa) with imposed half-wave numbers 𝑚 = 2, 𝑛 = 1).
terms of 𝜎𝑧𝑧 , 𝜎𝛼𝑧 , 𝜎𝛽𝑧 and 𝜙 values obtained at external surfaces. 𝑅𝛼 /h 4 10 20 50 100
The fourth benchmark (B4) is devoted to a multilayered PZT- 𝑢 (10−3 m) at 𝑧̃ = ℎ
4/0◦ /90◦ /0◦ /PZT-4 spherical shell. The lamination scheme is composed 0.88593 2.0677 4.0294 9.9104 19.711
as follows: ℎ𝑃 𝑍𝑇4 = 0.05ℎ and ℎ𝐶𝐹 𝑅𝑃 −𝐴 = 0.3ℎ. Involved material 𝑤 (10−3 m) at 𝑧̃ = 0
properties can be seen in Table 1 (PZT-4 and CFRP-A columns) and 0.87016 2.0302 3.9555 9.7269 19.345
geometrical data are collected in Table 11. Impositions for the sensor 𝜎𝛼𝛼 (105 Pa) at 𝑧̃ = 0.5ℎ
configuration are top pressure load 𝑃𝑧𝑡 = 10000 Pa and short-circuited 0.44234 1.0247 1.9882 4.8747 9.6846
electric potentials at outer surfaces (𝛷𝑡 = 𝛷𝑏 = 0 V). Impositions for 𝜎𝑧𝑧 (103 Pa) at 𝑧̃ = 0.75ℎ
actuator configuration are electric potential at the top 𝛷𝑡 = 100 V, 9.0392 8.9095 8.8562 8.8214 8.8093
electric potential at the bottom 𝛷𝑏 = 0 V and no pressure loads 𝜙 (102 V) at 𝑧̃ = 0.25ℎ
applied at outer surfaces. Imposed half-wave numbers are 𝑚 = 𝑛 = 1. −4.7139 −6.0345 −6.3984 −6.5971 −6.6599
Several 𝑅𝛼 ∕ℎ ratios are investigated, as presented in Tables 18 and 𝐷𝑧 (10−4 C∕m2 ) at 𝑧̃ = 0.5ℎ
19, for different variables. In them, the thickness ratio effects are −0.58632 −1.3882 −2.7183 −6.7048 −13.348
clearly shown. Figs. 14–17 display the 𝑢, 𝑤, 𝜎𝛼𝛼 , 𝜎𝑧𝑧 , 𝜎𝛼𝑧 , 𝜎𝛽𝑧 , 𝜙 and
𝐷𝑧 plot along the thickness direction for a moderately thick spherical
shell (𝑅𝛼 ∕ℎ = 20). The multilayered configuration of this structure is
clear considered by showing the discontinuities for the 𝜎𝛼𝛼 and the

14
S. Brischetto and D. Cesare International Journal of Mechanical Sciences 280 (2024) 109620

Fig. 5. Benchmark 1, simply-supported five-layered PZT-4/Al2024/Foam/Al2024/PZT-4 sandwich square plate in actuator configuration (electric potential applied at the top of the
structure (𝛷𝑡 = 100 V and 𝛷𝑏 = 0 V) and zero transverse normal load at external surfaces (𝑃𝑧𝑡 = 𝑃𝑧𝑏 = 0 Pa) with imposed half-wave numbers 𝑚 = 𝑛 = 1). Thickness ratio 𝑎∕ℎ = 20.
Evaluation of: (a) in-plane normal stress 𝜎𝛼𝛼 , (b) transverse normal stress 𝜎𝑧𝑧 , (c) transverse shear stress 𝜎𝛼𝑧 and (d) transverse shear stress 𝜎𝛽𝑧 through the thickness direction 𝑧.

Table 17
change in slope at the interfaces for 𝑢, 𝑣, 𝜎𝑧𝑧 , 𝜎𝛼𝑧 , 𝜎𝛽𝑧 , 𝜙 and 𝐷𝑧 .
Benchmark 3, simply-supported multilayered sandwich cylinder (embedding an FGPM They are typical zigzag effects of transverse anisotropic multilayered
core with 𝑝 = −0.5) in actuator configuration (electric potential applied at the top of structures. Interlaminar continuity and load boundary conditions have
the structure (𝛷𝑡 = 100 V and 𝛷𝑏 = 0 V) and zero transverse normal load at external been correctly evaluated by the 3D-u-𝜙 model in Figs. 14 and 15
surfaces (𝑃𝑧𝑡 = 𝑃𝑧𝑏 = 0 Pa) with imposed half-wave numbers 𝑚 = 2, 𝑛 = 1).
and Figs. 16 and 17 for both sensor and actuator configurations,
𝑅𝛼 /h 4 10 20 50 100 respectively. Load boundary conditions are clearly indicated by the 𝜎𝑧𝑧 ,
𝑢 (10−6 m) at 𝑧̃ = ℎ 𝜎𝛼𝑧 , 𝜎𝛽𝑧 and 𝜙 values calculated at external surfaces (for sensor case,
1.2524 2.0011 3.2166 6.8477 12.895 𝜎𝑧𝑧 = 10000 Pa imposed at the top surface and 𝜎𝑧𝑧 = 0 Pa imposed at
𝑤 (10−6 m) at 𝑧̃ = 0 the bottom surface with no applied loads in the in-plane directions;
1.3018 2.1198 3.45378 7.4424 14.086
for actuator case, 𝛷 = 100 V imposed at the top surface and 𝛷 = 0 V
𝜎𝛼𝛼 (Pa) at 𝑧̃ = 0.5ℎ imposed at the bottom surface with 𝜎𝑧𝑧 = 𝜎𝛼𝑧 = 𝜎𝛽𝑧 = 0 Pa conditions).
−5.3033 −18.883 −41.409 −108.68 −220.66
𝜎𝑧𝑧 (Pa) at 𝑧̃ = 0.75ℎ
10.461 9.0701 8.4034 7.9490 7.7881 4. Conclusions
𝜙 (V) at 𝑧̃ = 0.25ℎ
17.610 18.031 18.194 18.297 18.332 This work has proposed a three-dimensional shell solution for the
𝐷𝑧 (10−6 C∕m2 ) at 𝑧̃ = 0.5ℎ static analysis of plates, cylinders, cylindrical panels and spherical
−0.56756 −1.3567 −2.6685 −6.6022 −13.158 panels embedding isotropic, orthotropic, composite, piezoelectric and
functionally graded layers. Both sensor and actuator configurations
have been investigated.
3D equilibrium equations for spherical shells have been coupled
with the 3D divergence equation of electric displacement in order to

15
S. Brischetto and D. Cesare International Journal of Mechanical Sciences 280 (2024) 109620

Fig. 6. Benchmark 2, simply-supported single layered FGPM (𝑝 = 0.5) cylindrical shell in sensor configuration (zero electric potential at external surfaces (𝛷𝑏 = 𝛷𝑡 = 0 V) and
transverse normal load applied at the top of the structure (𝑃𝑧𝑡 = 10000 Pa and 𝑃𝑧𝑏 = 0 Pa) with imposed half-wave numbers 𝑚 = 1, 𝑛 = 0). Thickness ratio 𝑅𝛼 ∕ℎ = 20. Evaluation of:
(a) in-plane displacement 𝑢, (b) transverse normal displacement 𝑤, (c) electric potential 𝜙 and (d) transverse normal electric displacement 𝑧 through the thickness direction 𝑧.

Table 18 Table 19
Benchmark 4, simply-supported multilayered PZT-4/0◦ /90◦ /0◦ /PZT-4 spherical shell in Benchmark 4, simply-supported multilayered PZT-4/0◦ /90◦ /0◦ /PZT-4 spherical shell in
sensor configuration (zero electric potential at external surfaces (𝛷𝑏 = 𝛷𝑡 = 0 V) actuator configuration (electric potential applied at the top of the structure (𝛷𝑡 = 100 V
and transverse normal load applied at the top of the structure (𝑃𝑧𝑡 = 10 000 Pa and and 𝛷𝑏 = 0 V) and zero transverse normal load at external surfaces (𝑃𝑧𝑡 = 𝑃𝑧𝑏 = 0 Pa)
𝑃𝑧𝑏 = 0 Pa) with imposed half-wave numbers 𝑚 = 𝑛 = 1). Composite layers are in with imposed half-wave numbers 𝑚 = 𝑛 = 1). Composite layers are in CFRP-A.
CFRP-A. 𝑅𝛼 /h 4 10 20 50 100
𝑅𝛼 /h 4 10 20 50 100
𝑢 (10−10 m) at 𝑧̃ = 0.75ℎ
𝑢 (10−6 m) at 𝑧̃ = 0.75ℎ 4.5586 6.4563 9.1110 11.592 12.591
2.2747 7.2142 18.227 51.566 106.04
𝑤 (10−10 m) at 𝑧̃ = 0.5ℎ
𝑤 (10−6 m) at 𝑧̃ = 0.5ℎ 6.1164 17.331 28.274 37.035 41.351
8.1441 29.243 69.761 185.52 372.91
𝜎𝛼𝛼 (Pa) at 𝑧̃ = 0.75ℎ
𝜎𝛼𝛼 (104 Pa) at 𝑧̃ = 0.75ℎ −9.3659 −2.7655 1.2040 3.0518 4.8817
2.1226 10.213 20.405 42.092 75.084
𝜎𝑧𝑧 (Pa) at 𝑧̃ = 0.95ℎ
𝜎𝑧𝑧 (103 Pa) at 𝑧̃ = 0.95ℎ 0.16537 −0.02447 −0.00976 0.01703 0.02348
8.4501 8.0275 8.2123 8.6185 8.8043
𝜙 (V) at 𝑧̃ = 0.15ℎ
𝜙 (V) at 𝑧̃ = 0.15ℎ 9.4888 10.841 11.050 11.108 11.117
73.405 61.114 18.425 −24.442 −40.096
𝐷𝑧 (10−9 C∕m2 ) at 𝑧̃ = 0.85ℎ
𝐷𝑧 (10−9 C∕m2 ) at 𝑧̃ = 0.85ℎ −1.4074 −3.0520 −5.9535 −14.775 −29.519
−0.90406 −2.9563 −7.4954 −20.662 −41.931

consider the electro-mechanical coupling. Governing equations have

directions by means of Navier hypotheses and in the thickness direction
been developed in layer-wise form and then solved in the in-plane
by means of the exponential matrix method.

16
S. Brischetto and D. Cesare International Journal of Mechanical Sciences 280 (2024) 109620

Fig. 7. Benchmark 2, simply-supported single layered FGPM (𝑝 = 0.5) cylindrical shell in sensor configuration (zero electric potential at external surfaces (𝛷𝑏 = 𝛷𝑡 = 0 V) and
transverse normal load applied at the top of the structure (𝑃𝑧𝑡 = 10000 Pa and 𝑃𝑧𝑏 = 0 Pa) with imposed half-wave numbers 𝑚 = 1, 𝑛 = 0). Thickness ratio 𝑅𝛼 ∕ℎ = 20. Evaluation of:
(a) in-plane normal stress 𝜎𝛼𝛼 , (b) transverse normal stress 𝜎𝑧𝑧 , (c) transverse shear stress 𝜎𝛼𝑧 and (d) transverse shear stress 𝜎𝛽𝑧 through the thickness direction 𝑧.

Several preliminary assessments have been proposed to validate the the literature (in particular, the exact ones) are not so general because
model and new benchmarks have been discussed. The proposed model they are limited to particular geometries, load conditions, lamination
correctly describe in 3D sense and in layer wise form all the pos- schemes and/or materials.
sible electro-elastic variables: displacements, strains, stresses, electric Benchmarks here proposed are useful comparison results for those
potential and electric displacements. scientists involved in the development of numerical 3D models, ex-
Zigzag effects through the thickness of the transverse anisotropic act/numerical 2D models and experimental investigations related to
and multilayered structures and interlaminar continuity in terms of the behavior of multilayered piezoelectric structures. Therefore, more
equilibrium and congruence conditions have been correctly described. efficient smart structures could be produced.
Moreover, the use of some functionally graded layers (where elastic,
CRediT authorship contribution statement
piezoelectric and permittivity properties continuously change through
the thickness direction) allows the interlaminar continuity of all stress
S. Brischetto: Writing – review & editing, Writing – original draft,
components at the interfaces. The correct results in terms of electric
Methodology, Conceptualization. D. Cesare: Writing – original draft,
potential and electric displacements will allow an appropriate modeling
Validation, Software, Investigation, Formal analysis, Data curation.
of both sensor and actuator configurations. The model is also able to
correctly impose the load boundary conditions at external surfaces of Declaration of competing interest
the structures for both sensor and actuator configurations.
This new work has proposed a novel and unique model that is The authors declare no conflict of interest for the paper.
capable, by means of general equations for spherical shells, to in-
vestigate several geometries (plates, cylinders, cylindrical panels and Data availability
spherical panels) and lamination schemes (classical and functionally
graded layers) in a unified manner. All the other 3D models proposed in Data will be made available on request.

17
S. Brischetto and D. Cesare International Journal of Mechanical Sciences 280 (2024) 109620

Fig. 8. Benchmark 2, simply-supported single layered FGPM (𝑝 = 0.5) cylindrical shell in actuator configuration (electric potential applied at the top of the structure (𝛷𝑡 = 100 V
and 𝛷𝑏 = 0 V) and zero transverse normal load at external surfaces (𝑃𝑧𝑡 = 𝑃𝑧𝑏 = 0 Pa) with imposed half-wave numbers 𝑚 = 1, 𝑛 = 0). Thickness ratio 𝑅𝛼 ∕ℎ = 20. Evaluation of: (a)
in-plane displacement 𝑢, (b) transverse normal displacement 𝑤, (c) electric potential 𝜙 and (d) transverse normal electric displacement 𝑧 through the thickness direction 𝑧.

18
S. Brischetto and D. Cesare International Journal of Mechanical Sciences 280 (2024) 109620

Fig. 9. Benchmark 2, simply-supported single layered FGPM (𝑝 = 0.5) cylindrical shell in actuator configuration (electric potential applied at the top of the structure (𝛷𝑡 = 100 V
and 𝛷𝑏 = 0 V) and zero transverse normal load at external surfaces (𝑃𝑧𝑡 = 𝑃𝑧𝑏 = 0 Pa) with imposed half-wave numbers 𝑚 = 1, 𝑛 = 0). Thickness ratio 𝑅𝛼 ∕ℎ = 20. Evaluation of: (a)
in-plane normal stress 𝜎𝛼𝛼 , (b) transverse normal stress 𝜎𝑧𝑧 , (c) transverse shear stress 𝜎𝛼𝑧 and (d) transverse shear stress 𝜎𝛽𝑧 , through the thickness direction 𝑧.

19
S. Brischetto and D. Cesare International Journal of Mechanical Sciences 280 (2024) 109620

Fig. 10. Benchmark 3, simply-supported multilayered sandwich cylinder embedding an FGPM core (with 𝑝 = −0.5) in sensor configuration (zero electric potential at external
surfaces (𝛷𝑏 = 𝛷𝑡 = 0 V) and transverse normal load applied at the top of the structure (𝑃𝑧𝑡 = 10000 Pa and 𝑃𝑧𝑏 = 0 Pa) with imposed half-wave numbers 𝑚 = 2, 𝑛 = 1). Thickness
ratio 𝑅𝛼 ∕ℎ = 20. Evaluation of: (a) in-plane displacement 𝑢, (b) transverse normal displacement 𝑤, (c) electric potential 𝜙 and (d) transverse normal electric displacement 𝑧
through the thickness direction 𝑧.

20
S. Brischetto and D. Cesare International Journal of Mechanical Sciences 280 (2024) 109620

Fig. 11. Benchmark 3, simply-supported multilayered sandwich cylinder embedding an FGPM core (with 𝑝 = −0.5) in sensor configuration (zero electric potential at external
surfaces (𝛷𝑏 = 𝛷𝑡 = 0 V) and transverse normal load applied at the top of the structure (𝑃𝑧𝑡 = 10000 Pa and 𝑃𝑧𝑏 = 0 Pa) with imposed half-wave numbers 𝑚 = 2, 𝑛 = 1). Thickness
ratio 𝑅𝛼 ∕ℎ = 20. Evaluation of: (a) in-plane normal stress 𝜎𝛼𝛼 , (b) transverse normal stress 𝜎𝑧𝑧 , (c) transverse shear stress 𝜎𝛼𝑧 and (d) transverse shear stress 𝜎𝛽𝑧 through the
thickness direction 𝑧.

21
S. Brischetto and D. Cesare International Journal of Mechanical Sciences 280 (2024) 109620

Fig. 12. Benchmark 3, simply-supported multilayered sandwich cylinder embedding an FGPM core (with 𝑝 = −0.5) in actuator configuration (electric potential applied at the top
of the structure (𝛷𝑡 = 100 V and 𝛷𝑏 = 0 V) and zero transverse normal load at external surfaces (𝑃𝑧𝑡 = 𝑃𝑧𝑏 = 0 Pa) with imposed half-wave numbers 𝑚 = 2, 𝑛 = 1). Thickness ratio
𝑅𝛼 ∕ℎ = 20. Evaluation of: (a) in-plane displacement 𝑢, (b) transverse normal displacement 𝑤, (c) electric potential 𝜙 and (d) transverse normal electric displacement 𝑧 through
the thickness direction 𝑧.

22
S. Brischetto and D. Cesare International Journal of Mechanical Sciences 280 (2024) 109620

Fig. 13. Benchmark 3, simply-supported multilayered sandwich cylinder embedding an FGPM core (with 𝑝 = −0.5) in actuator configuration (electric potential applied at the top
of the structure (𝛷𝑡 = 100 V and 𝛷𝑏 = 0 V) and zero transverse normal load at external surfaces (𝑃𝑧𝑡 = 𝑃𝑧𝑏 = 0 Pa) with imposed half-wave numbers 𝑚 = 2, 𝑛 = 1). Thickness ratio
𝑅𝛼 ∕ℎ = 20. Evaluation of: (a) in-plane normal stress 𝜎𝛼𝛼 , (b) transverse normal stress 𝜎𝑧𝑧 , (c) transverse shear stress 𝜎𝛼𝑧 and (d) transverse shear stress 𝜎𝛽𝑧 through the thickness
direction 𝑧.

23
S. Brischetto and D. Cesare International Journal of Mechanical Sciences 280 (2024) 109620

Fig. 14. Benchmark 4, simply-supported multilayered PZT-4/0◦ /90◦ /0◦ /PZT-4 spherical shell in sensor configuration (zero electric potential at external surfaces (𝛷𝑏 = 𝛷𝑡 = 0 V) and
transverse normal load applied at the top of the structure (𝑃𝑧𝑡 = 10000 Pa and 𝑃𝑧𝑏 = 0 Pa) with imposed half-wave numbers 𝑚 = 𝑛 = 1). Composite layers are in CFRP-A. Thickness
ratio 𝑅𝛼 ∕ℎ = 20. Evaluation of: (a) in-plane displacement 𝑢, (b) transverse normal displacement 𝑤, (c) electric potential 𝜙 and (d) transverse normal electric displacement 𝑧
through the thickness direction 𝑧.

24
S. Brischetto and D. Cesare International Journal of Mechanical Sciences 280 (2024) 109620

Fig. 15. Benchmark 4, simply-supported multilayered PZT-4/0◦ /90◦ /0◦ /PZT-4 spherical shell in sensor configuration (zero electric potential at external surfaces (𝛷𝑏 = 𝛷𝑡 = 0 V)
and transverse normal load applied at the top of the structure (𝑃𝑧𝑡 = 10000 Pa and 𝑃𝑧𝑏 = 0 Pa) with imposed half-wave numbers 𝑚 = 𝑛 = 1). Composite layers are in CFRP-A.
Thickness ratio 𝑅𝛼 ∕ℎ = 20. Evaluation of: (a) in-plane normal stress 𝜎𝛼𝛼 , (b) transverse normal stress 𝜎𝑧𝑧 , (c) transverse shear stress 𝜎𝛼𝑧 and (d) transverse shear stress 𝜎𝛽𝑧 , through
the thickness direction 𝑧.

25
S. Brischetto and D. Cesare International Journal of Mechanical Sciences 280 (2024) 109620

Fig. 16. Benchmark 4, simply-supported multilayered PZT-4/0◦ /90◦ /0◦ /PZT-4 spherical shell in actuator configuration (electric potential applied at the top of the structure
(𝛷𝑡 = 100 V and 𝛷𝑏 = 0 V) and zero transverse normal load at external surfaces (𝑃𝑧𝑡 = 𝑃𝑧𝑏 = 0 Pa) with imposed half-wave numbers 𝑚 = 𝑛 = 1). Composite layers are in CFRP-A.
Thickness ratio 𝑅𝛼 ∕ℎ = 20. Evaluation of: (a) in-plane displacement 𝑢, (b) transverse normal displacement 𝑤, (c) electric potential 𝜙 and (d) transverse normal electric displacement
𝑧 through the thickness direction 𝑧.

26
S. Brischetto and D. Cesare International Journal of Mechanical Sciences 280 (2024) 109620

Fig. 17. Benchmark 4, simply-supported multilayered PZT-4/0◦ /90◦ /0◦ /PZT-4 spherical shell in actuator configuration (electric potential applied at the top of the structure
(𝛷𝑡 = 100 V and 𝛷𝑏 = 0 V) and zero transverse normal load at external surfaces (𝑃𝑧𝑡 = 𝑃𝑧𝑏 = 0 Pa) with imposed half-wave numbers 𝑚 = 𝑛 = 1). Composite layers are in CFRP-A.
Thickness ratio 𝑅𝛼 ∕ℎ = 20. Evaluation of: (a) in-plane normal stress 𝜎𝛼𝛼 , (b) transverse normal stress 𝜎𝑧𝑧 , (c) transverse shear stress 𝜎𝛼𝑧 and (d) transverse shear stress 𝜎𝛽𝑧 through
the thickness direction 𝑧.

References [8] Benjeddou A, Deu J-F. A two-dimensional closed-form solution for the
free-vibrations analysis of piezoelectric sandwich plates. Int J Solids Struct
2002;39:1463–86.
[1] Altay G, Dokmeci MC. The consistent Mindlin’s thermopiezoelectric equations
[9] Birman V, Adali S. Vibration damping using piezoelectric stiffener-actuators with
and the principle of virtual work. Mech Res Commun 2005;32:115–9.
application to orthotorpic plates. Compos Struct 1996;35:251–61.
[2] Benjeddou A, Andrianarison O. A thermopiezoelectric mixed variational theorem
[10] Cupial P. Three-dimensional natural vibration analysis and energy considerations
for smart multilayered composites. Comput Struct 2005;83:1266–76.
for a piezoelectric rectangular plate. J Sound Vib 2005;283:1093–113.
[3] Araujo AL, Mota Soares CM, Herskovits J, Pedersen P. Development of a finite [11] Deu J-F, Benjeddou A. Free-vibration analysis of laminated plates with embedded
element model for the identification of mechanical and piezoelectric properties shear-mode piezoceramic layers. Int J Solids Struct 2005;42:2059–88.
through gradient optimisation and experimental vibration data. Compos Struct
[12] Duan WH, Quek ST, Wang Q. Free vibration analysis of piezoelectric coupled
2002;58:307–18.
thin and thick annular plate. J Sound Vib 2005;281:119–39.
[4] Takagi K, Lia J-F, Yokoyama Si, Watanabe R. Fabrication and evaluation of [13] Erturk CL, Tekinalp O. A layerwise approach to piezo-electric plates accounting
PZT/Pt piezoelectric composites and functionally graded actuators. J Eur Ceram for adhesive flexibility and delaminated regions. Comput Struct 2005;83:279–96.
Soc 2003;23:1577–83. [14] Fernandes A, Pouget J. Two-dimensional modelling of laminated piezoelectric
[5] Alessandroni S, Andreaus U, dell’Isola F, Porfiri M. A passive electric controller composites: Analysis and numerical results. Thin-Walled Struct 2001;39:3–22.
for multimodal vibrations of thin plates. Comput Struct 2005;83:1236–50. [15] Fernandes A, Pouget J. Analytical and numerical approaches to piezoelectric
[6] Ballhause D, D’Ottavio M, Kroplin B, Carrera E. A unified formulation to assess bimorph. Int J Solids Struct 2003;40:4331–52.
multilayered theories for piezoelectric plates. Comput Struct 2005;83:1217–35. [16] Gu H, Chattopadhyay A, Li J, Zhou X. A higher order temperature theory for
[7] Batra RC, Vidoli S. Higher-order piezoelectric plate theory derived from a coupled thermo-piezoelectric-mechanical modeling of smart composites. Int J
three-dimensional variational principle. AIAA J 2002;40:91–104. Solids Struct 2000;37:6479–97.

27
S. Brischetto and D. Cesare International Journal of Mechanical Sciences 280 (2024) 109620

[17] He SY, Chen WS, Chen ZL. A uniformizing method for the free vibration analysis [52] Altay G, Dokmeci MC. Variational principles and vibrations of a functionally
of metal–piezoceramic composite thin plates. J Sound Vib 1998;217:261–81. graded plate. Comput Struct 2005;83:1340–54.
[18] Heidary F, Eslami MR. Piezo-control of forced vibrations of a thermoelastic [53] Yang J, Kitipornchai S, Liew KM. Large amplitude vibration of thermo-electro-
composite plate. Compos Struct 2006;74:99–105. mechanically stressed FGM laminated plates. Comput Methods Appl Mech Engrg
[19] Kapuria S. A coupled zig-zag third-order theory for piezoelectric hybrid cross-ply 2003;192:3861–85.
plates. J Appl Mech 2004;71:604–14. [54] Almajid A, Taya M, Hudnut S. Analysis of out-of-plane displacement and stress
[20] Kapuria S, Ahmed A, Dumir PC. An efficient coupled zigzag theory for dynamic field in a piezocomposite plate with functionally graded microstructure. Int J
analysis of piezoelectric composite and sandwich beams with damping. J Sound Solids Struct 2001;38:3377–91.
Vib 2005;279:345–71. [55] Bhangale RK, Ganesan N. Static analysis of simply supported functionally graded
[21] Kogl M, Bucalem ML. A family of piezoelectric MITC plate elements. Comput and layered magneto-electro-elastic plates. Int J Solids Struct 2006;43:3230–53.
Struct 2005;83:1277–97. [56] Bhangale RK, Ganesan N. Free vibration of simply supported functionally graded
[22] Liew KM, He XQ, Tan MJ, Lima HK. Dynamic analysis of laminated composite and layered magneto-electro-elastic plates by finite element method. J Sound Vib
plates with piezoelectric sensor/actuator patches using the FSDT mesh-free 2006;294:1016–38.
method. Int J Mech Sci 2004;46:411–31. [57] Chen WQ, Ding HJ. On free vibration of a functionally graded piezoelectric
[23] Liu X, Wang Q, Quek ST. Analytical solution for free vibration of piezoelectric rectangular plate. Acta Mech 2002;153:207–16.
coupled moderately thick circular plates. Int J Solids Struct 2002;39:2129–51. [58] Chen XL, Zhao ZY, Liew KM. Stability of piezoelectric FGM rectangular plates
[24] Oh J, Cho M. A finite element based on cubic zig-zag plate theory for subjected to non-uniformly distributed load, heat and voltage. Adv Eng Softw
the prediction of thermo-electric-mechanical behaviors. Int J Solids Struct 2008;39:121–31.
2004;41:1357–75. [59] Dai KY, Liu GR, Han X, Lim KM. Thermomechanical analysis of functionally
[25] Pan E, Heyliger PR. Free vibrations of simply supported and multilayered graded material (FGM) plates using element-free Galerkin method. Comput Struct
magneto-electro-elastic plates. J Sound Vib 2002;252:429–42. 2005;83:1487–502.
[26] Qing G, Qiu J, Liu Y. Semi-analytical solution for static and dynamic analysis of [60] He XQ, Ng TY, Sivashanker S, Liew KM. Active control of FGM plates
plates with piezoelectric patches. Int J Solids Struct 2006;43:1388–403. with integrated piezoelectric sensors and actuators. Int J Solids Struct
[27] Ramirez F, Heyliger PR, Pan E. Free vibration response of two-dimensional 2001;38:1641–55.
magneto-electro-elastic laminated plates. J Sound Vib 2006;292:626–44. [61] He X, Wang J-S, Qin Q-H. Saint-Venant decay analysis of FGPM laminates and
[28] Ray MC, Mallik N. Finite element analysis of smart structures containing dissimilar piezoelectric laminates. Mech Mater 2007;39:1053–65.
piezoelectric fiber-reinforced composite actuator. AIAA J 2004;42:1398–405. [62] Li XY, Ding HJ, Chen WQ. Three-dimensional analytical solution for functionally
[29] Samal MK, Seshu P, Parashar SK, von Wagner U, Hagedorn P, Dutta BK, et al. graded magneto–electro-elastic circular plates subjected to uniform load. Compos
Nonlinear behaviour of piezoceramics under weak electric fields Part-I: 3-D finite Struct 2008;83:381–90.
element formulation. Int J Solids Struct 2006;43:1422–36. [63] Liew KM, He XQ, Ray T. On the use of computational intelligence in the optimal
[30] Saravanos DA, Heyliger PR, Hopkins DA. Layerwise mechanics and finite element
shape control of functionally graded smart plates. Comput Methods Appl Mech
for the dynamic analysis of piezoelectric composite plates. Int J Solids Struct
Engrg 2004;193:4475–92.
1997;34:359–78.
[64] Lu P, Lee H, Lu C. Exact solutions for simply supported functionally graded
[31] Sheikh AH, Topdar P, Halder S. An appropriate FE model for through-thickness
piezoelectric laminates by Stroh-like formalism. Compos Struct 2006;72:352–63.
variation of displacement and potential in thin/moderately thick smart laminates.
[65] Lu P, Lee H, Lu C. An exact solution for functionally graded piezoelectric
Compos Struct 2001;51:401–9.
laminates in cylindrical bending. Int J Mech Sci 2005;47:437–58.
[32] Shu X. Free vibration of laminated piezoelectric composite plates based on an
[66] Ootao Y, Tanigawa Y. Three-dimensional transient piezothermoelasticity in
accurate theory. Compos Struct 2005;67:375–82.
functionally graded rectangular plate bonded to a piezoelectric plate. Int J Solids
[33] Zhang BN, Qing C, Zhang JQ. A sub-layer model for a thick piezoelectric patch
Struct 2000;37:4377–401.
bonded on elastic substrate. Acta Mech 2004;170:163–86.
[67] Pan E, Han F. Exact solution for functionally graded and layered
[34] Heyliger P. Static behavior of laminated elastic/piezoelectric plates. AIAA J
magneto-electro-elastic plates. Internat J Engrg Sci 2005;43:321–39.
1994;32:2481–4.
[68] Pan E, Han F. Green’s functions for transversely isotropic piezoelec-
[35] Heyliger P. A note on the static behavior of simply-supported laminated
tric functionally graded multilayered half spaces. Int J Solids Struct
piezoelectric cylinders. Int J Solids Struct 1997;34:3781–94.
2005;42:3207–33.
[36] Heyliger P. Exact solutions for simply supported laminated piezoelectric plates.
[69] Ray MC, Sachade HM. Finite element analysis of smart functionally graded plates.
J Appl Mech 1997;64:299–306.
Int J Solids Struct 2006;43:5468–84.
[37] Chen C-Q, Shen Y-P, Wang X-M. Exact solution of orthotropic cylindrical
[70] Zhong Z, Shang ET. Three-dimensional exact analysis of a simply-supported
shell with piezoelectric layers under cylindrical bending. Int J Solids Struct
functionally gradient piezoelectric plate. Int J Solids Struct 2003;40:5335–52.
1996;33:4481–94.
[71] Brischetto S, Carrera E. Refined 2D models for the analysis of functionally graded
[38] Dumir PC, Dube GP, Kapuria S. Exact piezoelastic solution of simply-supported
piezoelectric plates. J Intell Mater Syst Struct 2009;20:1783–97.
orthotropic circular cylindrical panel in cylindrical bending. Int J Solids Struct
1997;34:685–702. [72] Wu X-H, Chen C, Shen Y-P, Tian X-G. A high order theory for functionally graded
[39] Correia IFP, Mota Soares CM, Mota Soares CA, Herskovits J. Active control of piezoelectric shells. Int J Solids Struct 2002;39:5325–44.
axisymmetric shells with piezoelectric layers: A mixed laminated theory with a [73] Wu X-H, Shen Y-P, Chen C. An exact solution for functionally graded
high order displacement field. Comput Struct 2002;80:2265–75. piezothermoelastic cylindrical shell as sensors or actuators. Mater Lett
[40] Ganesan N, Kadoli R. Semianalytical finite element analysis of piezothermoelastic 2003;57:3532–42.
shells of revolution. Comput Struct 2005;83:1305–19. [74] Duc ND, Quan TQ, Luat VD. Nonlinear dynamic analysis and vibration
[41] Kapuria S, Sengupta S, Dumir PC. Assessment of shell theories for hybrid of shear deformable piezoelectric FGM double curved shallow shells under
piezoelectric shell under electromechanical load. Int J Mech Sci 1998;40:461–77. damping-thermo-electro-mechanical loads. Compos Struct 2015;125:29–40.
[42] Oh IK, Lee I. Supersonic flutter suppression of piezolaminated cylindrical panels [75] Liew KM, He XQ, Kitipornchai S. Finite element method for the feedback control
based on multifield layerwise theory. J Sound Vib 2006;291:1186–201. of FGM shells in the frequency domain via piezoelectric sensors and actuators.
[43] Wang HM, Ding HJ, Chen YM. Dynamic solution of a multilayered orthotropic Comput Methods Appl Mech Engrg 2004;193:257–73.
piezoelectric hollow cylinder for axisymmetric plane strain problems. Int J Solids [76] Mellouli H, Kalleli S, Mallek H, Said LB, Ayadi B, Dammak F. Electromechanical
Struct 2005;42:85–102. behavior of piezolaminated shell structures with imperfect functionally graded
[44] Wu Y-C, Heyliger P. Free vibration of layered piezoelectric spherical caps. J porous materials using an improved solid-shell element. Comput Math Appl
Sound Vib 2001;245:527–44. 2024;155:1–13.
[45] Zhu J-Q, Chen C, Shen Y-P. Three dimensional analysis of the dynamic stability [77] Ninh DG, Hoang VNV, Huy VL. A new structure study: Vibrational analyses
of piezoelectric circular cylindrical shells. Eur J Mech A Solids 2003;22:401–11. of FGM convex-concave shells subjected to electro-thermal-mechanical loads
[46] Carrera E, Brischetto S. Reissner mixed theorem applied to static analysis of surrounded by Pasternak foundation. Eur J Mech A Solids 2021;86:104168, 1-37.
piezoelectric shells. J Intell Mater Syst Struct 2007;18:1083–107. [78] Yang S, Mahjouri H, Jamalpoor A. Underwater temperature-dependent sound
[47] Carrera E, Brischetto S. Piezoelectric shell theories with a priori continuous scattering and acoustic radiation force issues of a functionally graded sandwich
transverse electromechanical variables. J Mech Mater Struct 2007;2:377–99. spherical shell integrated with piezoelectric layers. Ocean Eng 2024;294:116730,
[48] Vel SS, Batra RC. Three-dimensional analytical solution for hybrid multilayered 1-15.
piezoelectric plates. J Appl Mech 2000;67:558–67. [79] Gao X, Wang Q, Liu X, Zhong R, Qin B, Xu H. An electro-mechanical
[49] Vel SS, Batra RC. Cylindrical bending of laminated plates with distributed and shell model for predicting stochastic vibration behavior of smart piezoelec-
segmented piezoelectric actuators/sensors. AIAA J 2000;38:857–67. tric structure utilizing the Spectral-Tchebyche method. Eng Anal Bound Elem
[50] Vel SS, Batra RC. Generalized plane strain thermopiezoelectric analysis of 2023;155:584–98.
multilayered plates. J Therm Stress 2003;26:353–77. [80] Alibeigloo A, Chen WQ. Elasticity solution for an FGM cylindrical panel
[51] Kulikov GM, Plotnikova SV. A sampling surfaces method and its application to integrated with piezoelectric layers. Eur J Mech A Solids 2010;29:714–23.
three-dimensional exact solutions for piezoelectric laminated shells. Int J Solids [81] Kulikov GM, Plotnikova SV. Exact electroelastic analysis of functionally graded
Struct 2013;50:1930–43. piezoelectric shells. Int J Solids Struct 2014;51:13–25.

28
S. Brischetto and D. Cesare International Journal of Mechanical Sciences 280 (2024) 109620

[82] Brischetto S. A closed-form 3D shell solution for multilayered structures subjected [85] Brischetto S, Cesare D, Torre R. A layer-wise coupled thermo-elastic shell model
to different load combinations. Aerosp Sci Technol 2017;70:29–46. for three-dimensional stress analysis of functionally graded material structures.
[83] Brischetto S. A 3D layer-wise model for the correct imposition of transverse Technologies 2023;11:35, 1-28.
shear/normal load conditions in FGM shells. Int J Mech Sci 2018;136:50–66. [86] Brischetto S, Cesare D. Three-dimensional vibration analysis of multilayered
[84] Brischetto S, Torre R, Cesare D. Three dimensional coupling between elastic composite and functionally graded piezoelectric plates and shells. Compos Struct
and thermal fields in the static analysis of multilayered composite shells. 2024;346:118413.
CMES-Comput Model Eng Sci 2023;136:2551–91.

29