A polynomial layerwise model for smart complex shells

J.C. Monge a, J.L. Mantari b,c,1, M.N. Llosaa, M.A. Hinostrozab

a
Faculty of Physics, National University of San Marcos, Av. Carlos German Amezaga
375, Cercado de Lima, Lima, Perú.
b
Faculty of Mechanical Engineering, National University of Engineering, Av. Túpac
Amaru 210, Rimac, Lima, Perú.
c
Department of Sciences, Universidad de Ingeniería y Tecnología - UTEC, Jr. Medrano
Silva 165, Barranco, Lima, Perú.

Abstract
This paper presents a polynomial layerwise model in the framework of Carrera’s
Unified Formulation for the bending analysis of a magneto-electric shells with variable
radii of curvature. A parametric surface is used to model the middle surface of the shell.
Lame Parameters and Radius of Curvature are calculated by using the Differential
Geometry. The mechanical displacement and electric and magnetic scalar potential
function are written and modeled in term of Chebyshev polynomial of the First Kind.
The shell panels are exposed to different mechanical, electrical and magnetic loads.
Principle of Virtual Displacement is employed for obtaining the governing equations
which are discretized by Chebyshev-Gauss-Lobatto grid distribution and solved in
semi-analytical manner by the so-called Differential Quadrature Method. The basis
function selected is the Lagrange polynomial. A simple stress recovery technique based
on the three-dimensional (3D) equilibrium equations is introduced for the out-of-plane
shear and normal stresses, transverse electric, and magnetic induction. Close to 3D
solutions are obtained for classical shells structures. Finally, benchmarks solutions for
smart complex shells with variable radii of curvature such as parabolic and cycloidal
shells are introduced.

Keywords: Magneto-electro-elastic Shell, Principle of Virtual Displacement,

Equilibrium Equations, Differential Quadrature Method, Carrera’s Unified Formulation,
Layerwise.
______________________________________________________________________
1Corresponding Author email: [email protected]
 1. Introduction

Smart materials can convert different kind of signals to reach a desired behavior in a
structure made of a particular material. Piezoelectric and magnetostrictive materials are
considered as smart materials and are employed in different type of industries including
acoustic, vibration, control, and health monitoring. Companies related to these
disciplines adopted the concept of magneto-electric-elastic (MEE) structures and it is a
topic of relevant numerical and experimental studies in the engineering field.
In the context of numerical methods to study smart materials it is important to obtain
reliable mathematical models to accurately calculate the mechanical deformations, and
the electric and magnetic potential. Over the years, several studies for the static and
dynamic analysis of MEE beams, plates and shells were introduced to the scientific
community. For example, Chen and Lee [1] established two independent state equations
by introducing two displacements functions and two stress functions for the three-
dimensional (3D) analysis of MEE plates considering the thermal effect, body force,
free charge density and electric density current. Vinyas and Kattimani [2] employed the
3D finite element for the bending evaluation of functionally graded stepped MEE beams
subjected to thermal loads. Ragb et. al. [3] employed the Sinc Differential Quadrature
and Discrete Singular Convolution Differential Quadrature for calculating the natural
frequencies of MEE nanobeams resting on nonlinear elastic foundation. Wang et. al. [4]
used a nonlinear finite element theory and perturbation technique for investigating the
bending and buckling of magneto-thermo-elastic behavior of simply supported
ferromagnetic plates. Heyliger et al. [5] showed the through-thickness behavior of MEE
plates in cylindrical bending by considering an approximate discrete-layer
representation of the three mechanical displacements and the electric and magnetic
potential. Pan and Han [6] derived the exact solution for exponentially graded MEE
rectangular plates based on the pseudo-Stroh formalism and the propagator matrix for a
multilayered case problem. Sobhy [7] analyzed the bending of doubly curved shallow
shell reinforced by functionally graded graphene platelets surrounded by two
piezoelectromagnetic face sheets. A four-variable higher order shear deformation theory
(HSDT) was considered for the mechanical displacement assumption and a linear and
non-linear variation were assumed for electric and magnetic potential. Badri and Al-
Kayiem [8] used a first order shear deformation theory (FSDT) for the static and
dynamic analysis of simply supported MEE rectangular plate employing Navier
method. Liu [9] used the Kirchoff hypothesis for studying the bending of MEE thin
plates with different boundary conditions. The in-plane components of electric and
magnetic field were neglected, so only transverse components were considered. Vinyas
and Kattimani [10] evaluated the static response of MEE beams subjected to various
thermal loading and boundary conditions employing a finite element procedure. The
coupling between elastic, electric and magnetic and thermal properties was considered.
Pan and Heyliger [11] derived an analytical solution for the 3D free vibration of simply
supported MEE plates. Moita et al. [12] used the third order HSDT for the static and
free vibration of MEE plates. The fundamental equations were solved by the finite
element method. The magnetic and electric field were considered to vary linearly over
the thickness direction. Wang and Li [13] derived an analytical solution to study
axisymmetric problems of the transversely isotropic and linear MEE by using the
Hankel transform. Liu [14] imposed the Kirchhoff assumptions to study deformations of
fiber reinforced MEE rectangular plates. The effect of the volume-fractions on the
deformation’s variations were discussed. Pan and Wang [15] developed a 3D finite
element code for the evaluation of general and finite-size multiferroic composite under
arbitrary boundary conditions. An eight node isoparametric element was adopted.
Buchanan [16] studied the 3D free vibration of MEE infinite rectangular plate. Ebrahimi
and Jafari [17] evaluated the buckling response of MEE functionally graded plates with
different boundary conditions via a tangential exponential refined theory. The porosity
effect was considered by a modified power law distribution. Chen et. al. [18] presented
an analytical solution for the propagation of harmonic waves in an infinitely extended
MEE plate. The boundary conditions were imposed as traction-free boundary on the top
and bottom surfaces on the layered plate. Carrera et. al. [19] developed a finite element
code for studying the bending response of MEE plates. The electric and magnetic
potential were written in term of a layerwise perspective, and the mechanical
displacement were assumed as equivalent single layer or a layerwise kinematic. Daga et
al. [20] studied the transient response of MEE beams. The MEE material was bonded on
the top surface of a mild steel beam.
Milazzo [21] derived a one-dimensional model for the dynamic analysis of MEE beams.
A first order model was considered for the mechanical displacements and the magnetic
and electric potential were assumed as quasi-static. Ewolo Ngak et. al. [22] developed a
semi-analytical 3D of multilayered MEE plates under simply supported boundary
condition. A combination of pseudo-Stroh formalism and the Lagrange polynomials
was elaborated for the space and time response. Ebrahimi and Barati [23] examined the
thermal buckling of nonlocal MEE functionally graded beams under various type of
thermal loading. The mechanical displacements were modeled by a third order HSDT
and Navier type method was used to solve the differential equations. Zhang et al. [24]
employed the scaled boundary finite element method incorporated with a precise
integration technique to study the free vibration and bending of MEE laminated
composite plates. The matrix exponent method was used to solve analytically the
governing equations in term of the thickness coordinate.
Alaimo et. al. [25] formulate an isoparametric four-node finite element procedure by
using the mixed interpolation of tensorial components for the static and dynamic
analyses of MEE plates. Only the transverse component of the magnetic and electric
field were taken into account. Ebrahimi et al. [26] employed the classical shell theory
for the study of wave propagation of a cylindrical MEE. Different modes, nonlocal
parameters, length parameters, geometry, magnetic and electric field were compared.
Phoenix et. al. [27] addressed a polynomial layer-wise mixed with finite element for the
static and dynamic analysis of MEE plates. The governing equations were derived in
term of Reissner mixed variational theorem and no post processing was required for
transverse stresses, electric displacement, and magnetic induction.
Physical phenomenon can be studied from a mathematical perspective by the
formulation of partial differential equations. Sometimes analytical methods cannot be
provided due to the number of unknowns’ coefficients, boundary conditions, adopted
method, etc.; in this context, semi-analytical methods such as the Differential
Quadrature Method (DQM) appears to be an important option. This simple but
interesting method for evaluating the differential operators was implemented by
Bellman and Casti [28, 29]. The main idea of the method is calculating the approximate
numerical derivative of a certain function by a weighted linear sum of the function
evaluated in a particular grid distribution. The principal fundaments, applications and
historic review can be found in classical textbooks [30, 31] and review articles [32, 33].
Overall, after a literature survey, it is important to remark that there are very limited
contributions to applied classical shells [45-47], applied complex shells [48-51] or smart
shells with refined theories [34]. This paper aims to contribute to the scientific literature
with a quasi-exact solution for complex smart shell structures that may be used
potentially in many ocean engineering applications.
The achieve the objective of this paper, the bending response of MEE complex shells
with variable radii of curvature is evaluated by employing a two-dimensional solution.
The middle surface of the shell is modelled by a parametric equation written in
orthogonal curvilinear coordinates. The volume of the shell is generated by the normal
vector of the parametric surface. Differential Geometry is used for deriving the radius of
curvature and Lamé coefficients. The mechanical displacements, electric and magnetic
potential are written in term of Carrera’s Unified formulation (CUF) [19, 34-36] which
permits to easily evaluate different order of expansions of the theory. A layerwise
theory based on Chebyshev polynomial of the first kind to model the thickness
functions is used. The governing equations are derived based on the Principal of Virtual
Displacement (PVD). Chebyshev-Gauss-Lobatto grid distribution is used for
discretizing the parametric surface. The governing equations are solved numerically by
the DQM, and Lagrange polynomials are employed as basis functions. To achieve the
traction of transverse stresses, electrical displacement, and magnetic induction, a
recovery procedure employing 3D equilibrium equations is formulated. Two case
problems are employed for the validation of the presented model. The first case is
related to a rectangular plate subjected to a bi-sinusoidal mechanical pressure and the
second problem is a shallow shell panel subjected to electric and magnetic load [27, 43].
The solutions of these case problems are also compared with a quasi-3D solution
previously reported by the authors [44]. Then, 02 benchmark problems of smart shells
with variable radii of curvature are introduced in this paper. The first is a parabolic shell
subjected to transverse magnetic induction and the other is a cycloidal panel subjected
to a transverse electric displacement. The paper is divided in four different sections:
introduction, mathematical model, results, and conclusions.

2. Mathematical model

The bending response of MEE shells with variable radii of curvature is studied by a 2D
theory. The mechanical displacement, electric and magnetic potential are written in term
of a polynomial Layerwise. The Chebyshev polynomials of the first kind are used as
shear strain shape functions. The shells are subjected to mechanical, electrical, and
magnetic load.
The mathematical formulation starts with the establishment of the type of curvature and
the corresponding curvilinear coordinate system. Then, it is important to define the
mechanical displacement and scalar potential functions used to derive the constitutive
equations of stresses, magnetic inductions and electric displacements. Subsequently, the
Principle of Virtual Displacement is used for stablishing the governing equations.
Finally, a post-process procedure is introduced to enhance the through-the-thickness
response of out-of-plane stresses, transverse electric displacement, and magnetic
induction. The 3D equilibrium equations for mechanical stresses, diverge of electrical
displacement, and magnetic induction written in orthogonal curvilinear coordinates are
utilized for recovering the mentioned transverse variables.

2.1 Differential Geometry

The Differential Geometry technique [37-39] is used for calculating the three main
parameters: Radius of curvature, Lamé parameters, and metric coefficients which is
used for the linear-strain relations. The midsurface of the shell is described
geometrically by the use of a parametric curve r (α , β) which are written in curvilinear
orthogonal coordinates. The limits of the curvilinear coordinates are defined as:
α (α 1 ≤ α ≤ α 2) and β (β 1 ≤ β ≤ β 2). A certain point of the midsurface of the shell is defined
mathematically as:

[ ]
f 1 (α , β )
r (α , β)= f 2 ( α , β ) (1)
f 3 (α , β )

The unitary normal vector n(α , β) is used for defining the volume of the shell and the
proper definition of this vector is given by:
∂r ∂r
×
∂α ∂ β
n(α , β)=
|
∂r ∂r
×
∂α ∂ β |
(2)
The thickness of the shell is defined as orthogonal to the midsurface and its notation is
given by ζ (−h/2 ≤ ζ ≤ h /2) and the thickness of the panel is defined as h . The thickness
of the shell is parallel to the unitary normal vector. The 3D representation of the shell is
denoted as R(α , β , ζ ) and its representation is shown in Figure 1. The complete shell
structure is defined as:
R(α , β , ζ )=r (α , β)+ ζn(α , β) (3)
Lamé parameters also known as the first fundamental parameters are defined as the
square root of the dot product of the first derivative of the midsurface vector. The
principal equations are given by

A α (α , β)=
∂r ∂r
•
∂α ∂α√ , A β (α , β)=
∂r ∂r
•
∂β ∂β
.
√
The dot product is given by the following operator • . The second parameter to be
(4)

defined is the radius of curvature which are defined as the quotient of dot products and
mathematically are defined as:
∂r ∂r ∂r ∂r
− • − •
∂α ∂α ∂β ∂β
Rα (α , β)= R β (α , β)= . (5)
∂2 r ∂2 r
• n • n
∂ α2 ∂ β2
These radii of curvature define the type of shell. If a structure is a plate, the radii of
curvature is defined as infinite; if just one radius of curvature is considered as infinite,
the shell is denominated as a singly curved shell; finally, if both radius of curvature are
different of infinite, the panel is a doubly curved shell.
The metric coefficients depend on the thickness coordinates and the radius of curvature,
ζ ζ
H α =1+ H β =1+ . (6)
Rα Rβ
The complete derivation of the formulas presented above can be consulted in classical
textbooks [37-39]. It is important to known that these relations can only be applied
when the curvilinear coordinates are orthogonal to each other.

2.2 Unified formulation and constitutive relations

The CUF [34] is compact theory where different shear strain shape functions or
thickness function can be systematically studied. In the present paper, the mechanical
displacement, magnetic and electrical potential are written in term of Layerwise
approach in which each layer is evaluated separately from each other. See the
superindex k that identifies to each shell layer in Equation 7.
k k k k k
δ u ( α , β , ζ ) =Ft ( z k ) ⋅δ ut ( α , β )+ F b ( z k ) ⋅δ ub ( α , β ) + F r ( z k ) ⋅ δ ur ( α , β )=F τ ( z k ) ⋅δ u τ ( α , β ) , τ=t , b , r ; r=1 , …
k k k k k
u ( α , β , ζ )=F t ( z k ) ⋅ut ( α , β ) + Fb ( z k ) ⋅ub ( α , β ) + F r ( z k ) ⋅u r ( α , β ) =F s ( z k ) ⋅u s ( α , β ) , s=t , b , r ; r=1, … , M e
k k k k k
δ ϕ ( α , β , ζ ) =F t ( z k ) ⋅δ ϕ t ( α , β ) + F b ( z k ) ⋅ δ ϕ b ( α , β ) + Fr ( z k ) ⋅ δ ϕ r ( α , β )=F τ ( z k ) ⋅ δ ϕ τ ( α , β ) , τ=t , b , r ; r=1, …
k k k k k
ϕ ( α , β , ζ )=F t ( z k ) ⋅ϕ t ( α , β )+ F b ( z k ) ⋅ϕ b ( α , β )+ F r ( z k ) ⋅ϕ r ( α , β )=F s ( z k ) ⋅ϕ s ( α , β ) , s=t , b , r ; r=1 , … , M e
 k k k k k
δ ψ ( α , β , ζ )=F t ( z k ) ⋅δ ψ t ( α , β ) + F b ( z k ) ⋅ δ ψ b ( α , β )+ F r ( z k ) ⋅δ ψ r ( α , β )=F τ ( z k ) ⋅δ ψ τ ( α , β ) , τ=t , b ,r ; r =1 ,…
k k k k
ψ ( α , β ,ζ )=F t ( z k ) ⋅ ψ t ( α , β ) + F b ( z k ) ⋅ψ b ( α , β ) + F r ( z k ) ⋅ψ r ( α , β )
k
¿ F s (z k )⋅ψ s (α , β) , s=t , b , r ; r=1 , ... , M e (7)
In this equation, the subindexes s and τ are defined to follow the Einstein’s notation.
The virtual variation, denoted as δ , is included to easily undertint the formulation of the
Principal of Virtual Displacement. The mechanical displacements are defined as a
vector uk =[ uk v k wk ] . The order of expansion of the presented layerwise model are
denoted as M e. A multilayered shell is reported in Figure 2. Every layer of the shell has
its own local thickness h k and are calculated as h k =ζ k +1−ζ k. The thickness functions are
expressed in term of a dimensional coordinate z k (−1 ≤ z k ≤ 1) which is related to the
global thickness of the panel by the following relation:

zk=
2
ζ k +1−ζ k (
ζ−
ζ k +1+ ζ k
2 ) (8)

The thickness functions are evaluated as Chebyshev’s polynomial of the first kind Gi
which are expressed as follows:
G1=1G2=z k Gn=2 z k Gn−1−Gi−2 , i=3 , 4 ,... , M e +1

(9)

The thickness functions F τ are reported by the following relations:

1−z k 1+ z k
F 1= F r=Gr +1−Gr −1 , r=2, ... , M e F M +1 =
2 2 e

(10)

The linear strain relations [37-39] for shells with variable radii of curvature are
expressed by the following matrix:
 (11)
The electric potential [34, 35] can be written in term of the three curvilinear coordinates
and these are denoted as:

[ ]
k
−F s ∂ ϕ s

[]
k
k Aα Hα ∂ α
Eα k
k −F s ∂ ϕ s
Eβ = (12)
k A β H kβ ∂ β
E ζ
−∂ F s k
ϕs
∂ζ

The magnetic potential [19] can be also written in curvilinear coordinates and these are
expressed in an identical manner to the electric potential:

[ ][ ]
k
−F s ∂ ψ s
k
k Aα H α ∂ α
Fα k
k −F s ∂ ψ s
F = β
(13)
k A β H kβ ∂ β
F ζ
−∂ F s k
ψs
∂ζ

Consequently, the stresses are expressed as a function of the linear-strain relations,

electric, and magnetic potential by the following constitutive equation:
 (14)
where C kij, e kij, q kij are denoted as stiffness, piezoelectric and piezomagnetic coefficients.
The electric displacements Dki are given by:

(15)
where ε kii and d kii are considered as dielectric and magnetoelectric coefficients. The
magnetic induction Bki are given by:

(16)
where μkii stands for the magnetic permeability coefficients.
2.3 Principle of Virtual Displacement
The governing equations for mechanical, electrical and magnetic interactions of MEE
complex shells are derived in term of the Principle of Virtual Displacement (PVD).
When an external load is applied on a shells, an energetic variation is originated in the
inner state of the shell. This idea is expressed in a mathematical language by the
following relation:
δ L∫ ¿=δ L ¿ (17)
ext
where δ L∫ ¿¿ is the internal work, and δ Lext is the external work originated by the
applied load on the shell panel. The internal work is calculated by
δL ζ k +1 β2 α2

∫ ¿=∫ ∫∫ {δ εkαα σkαα +δ εkββ σ kββ +δ γ kαβ τ kαβ+ δ γkαζ τ kαζ +¿ δ γ kβζ τ kβζ +δ ε kζζ σ kζζ−δ Ekα Dkα −δ Ekβ Dkβ −¿δ Ekζ D kζ −δ F kα Bkα−δ F kβ Bkβ−δ Fkζ B kζ } A α A β H kα H kβ dαdβdζ ¿
ζ k β1 α1

(18)
The linear strain relations (Equation 11), electric potential (Equation 12) and magnetic
potential (Equation 13) are replaced in the internal work (Equation 18) and the
following relation is formed:

δL
( ( )) ( ( )) ( ( ) ( )) ( (
ζ k +1 β2 α2 k k k k k k k k k k
Fτ 1 ∂ δ uτ δ v τ ∂ A α δ w τ Fτ δ uτ ∂ A β 1 ∂ δ v τ δ w τ Fτ 1 ∂ δ v τ δ uτ ∂ Aα Fτ 1 ∂ δ u τ δ v τ ∂ A β Fτ 1 ∂
∫ ¿=∫ ∫∫ { k
+ +
A α ∂ α Aα A β ∂ β R α
k
σαα + k
+
A α A β ∂ α Aβ ∂ β
+
Rβ
k
σ ββ+¿ k
Aα ∂α
−
Rα ∂ β
+ k −
k
τ αβ +¿ k
ζ k β1 α1 H α H β H α H β Aβ ∂ β Rβ ∂ α H α Aα

(19)
Some integration over the thickness functions are considered for simplifying Equation
19:
N l ζ k+ 1
(N kαα , N kββ , N kαβ 1 , N kαβ 2 )=∑ ∫ F τ ( H kβ σ kαα , H kα σ kββ , H kβ τ kαβ , H kα τ kαβ ) dζ
k=1 ζ k

N l ζ k+ 1

(N k
αζ 1 ,N k
βζ 1 )=∑ ∫ F τ ( H kβ τ kαζ , H kα τ kβζ ) dζ
k=1 ζ k

N l ζ k+ 1
∂ Fτ k k k k k
(N k
αζ 2 ,N k
βζ 2 , N )=∑ ∫k
ζζ H H ( τ , τ , σ ) dζ
k=1 ζ k ∂ ζ α β αζ βζ ζζ

( H H D ) dζ
N l ζ k+ 1
∂F
(T k
αα ,T ,T )=∑ ∫ F τ H kβ D kα , F τ H kα Dkβ ,
k
ββ
k
ζζ
τ k
α
k
β
k
ζ
k=1 ζ k ∂ζ

, R ) =∑ ∫ ( F H B , F H B , H H B ) dζ
N l ζ k +1
∂F
( R kαα , Rkββ k
ζζ
∂ζ τ
k
β
k
α τ
k
α
k
β
τ k
α
k
β
k
ζ (20)
k=1 ζ k

The integration presented before are replaced in Equation 19 and after some
mathematical manipulations, the following expressions are obtained:

(21)
The applied loads on the top of the shell is expressed in terms of external virtual work:
 β2 α2

δ Lext =∫ ∫ ( δ u Nτ ( Pα ) +δ v τN ( P β ) + δ w Nτ ( Pζ ) + δ ϕ Nτ ( D ζζ ) + δψ Nτ ( B ζζ ) ) F τ|ζ =h /2 A α A β H kα|ζ =h /2 H kβ|ζ =h/ 2 dαdβ

l l l l l

β 1 α1

(22)

The external mechanical load are denoted as Pα , P β , Pζ , the transverse electric

displacement are denominated as Dζζ and magnetic induction is expressed as Bζζ . All
the loads are applied at the top of the shell.
Equation 21 is integrated by parts and it is equalized to the external virtual work
obtained by Equation 22. The governing equations for smart complex shells are
expressed below:
k k k
∂( A β N αα ) ∂( A α N αβ 2 ) ∂ A β k ∂ Aα k N αζ 1
−A α A β N αζ 2 =A α A β Pα F τ|ζ =h/ 2 H α|ζ =h / 2 H β|ζ =h /2
k k k
+ − N ββ + N αβ 1+ A α A β
∂α ∂β ∂α ∂β Rα
k k k
∂( A β N αβ 1 ) ∂( A α N ββ ) ∂ Aα k ∂ A β k N βζ 1
−A α A β N βζ 2= Aα A β P β F τ|ζ =h /2 H α|ζ =h/ 2 H β|ζ =h /2
k k k
+ − N αα + N αβ 2+ A α A β
∂α ∂β ∂β ∂α Rβ
k k k k
∂( A β N αζ 1 ) ∂( A α N βζ 1 ) N αα N ββ
− A α A β N ζζ = A α A β Pζ F τ|ζ =h/ 2 H α|ζ =h / 2 H β|ζ =h /2
k k k
+ − Aα Aβ −A α A β
∂α ∂β Rα Rβ
k k
∂( A β T αα ) ∂( A α T ββ )
− A α A β T ζζ =A α A β Dζζ F τ|ζ =h /2 H α|ζ =h/ 2 H β|ζ =h /2
k k k
+
∂α ∂β
k k
∂( A β Rαα ) ∂( A α R ββ )
− A α A β Rζζ = A α A β B ζζ F τ|ζ =h / 2 H α|ζ =h /2 H β|ζ =h/ 2 (23)
k k k
+
∂α ∂β
The boundary conditions for the borders are considered as clamped, simply supported
and free. The complete assessment of boundary conditions are described by Equations
24 to 26:
Clamped edges:
k k k
u τ ( α 0 , β)=v τ ( α 0 , β)=wτ (α 0 , β)=0 , β 0 ≤ β ≤ β 1 , τ =0,1,..., M m
k k
ϕ τ (α 0 , β)=ψ τ (α 0 , β)=0 , β0 ≤ β ≤ β1 , τ =0,1,..., Me
k k k
u τ ( α 1 , β)=v τ (α 1 , β)=w τ (α 1 , β)=0 , β 0 ≤ β ≤ β1 , τ =0,1,..., Mm
k k
ϕ τ (α 1 , β )=ψ τ (α 1 , β )=0 , β 0 ≤ β ≤ β 1 , τ =0,1,..., M e
k k k
u τ ( α , β 0)=v τ ( α , β0 )=wτ (α , β0 )=0 , α 0 ≤ α ≤ α 1 , τ =0,1,..., M m
k k
ϕ τ (α , β 0)=ψ τ (α 1 , β 0)=0 , α 0 ≤ α ≤ α 1 , τ =0,1,..., Me
k k k
u τ ( α , β 1)=v τ (α , β 1)=w τ (α , β 1)=0 , α 0 ≤ α ≤ α 1 , τ =0,1,..., Mm
k k
ϕ τ (α , β1 )=ψ τ (α , β1 )=0 , α 0 ≤ α ≤ α 1 , τ =0,1,..., M e (24)
Simply supported edges:
 k k k
N αα ( α 0 , β)=v τ (α 0 , β )=w τ (α 0 , β )=0 , β 0 ≤ β ≤ β 1 , τ =0,1,..., M m
k k
ϕ τ (α 0 , β)=ψ τ (α 0 , β)=0 , β0 ≤ β ≤ β1 , τ =0,1,..., Me
k k k
N αα ( α 1 , β )=v τ (α 1 , β)=w τ ( α 1 , β)=0 , β 0 ≤ β ≤ β 1 , τ =0,1,..., M m
k k
ϕ τ (α 1 , β )=ψ τ (α 1 , β )=0 , β 0 ≤ β ≤ β 1 , τ =0,1,..., M e
k k k
u τ ( α , β 0)=N ββ (α , β 0)=wτ ( α , β 0)=0 , α 0 ≤ α ≤ α 1 , τ =0,1,..., Mm
k k
ϕ τ (α , β 0)=ψ τ (α 1 , β 0)=0 , α 0 ≤ α ≤ α 1 , τ =0,1,..., Me
k k k
u τ ( α , β 1)=N ββ (α , β 1 )=w τ (α , β 1)=0 , α 0 ≤ α ≤ α 1 , τ =0,1,..., M m
k k
ϕ τ (α , β1 )=ψ τ (α , β1 )=0 , α 0 ≤ α ≤ α 1 , τ =0,1,..., M e (25)
Free edges:
k k k
N αα ( α 0 , β)=N αβ 1 (α 0 , β)=N ας 1 (α 0 , β)=0 , β 0 ≤ β ≤ β1
k k
T αα (α 0 , β)=Rαα (α 0 , β )=0 , β 0 ≤ β ≤ β 1 ,
k k k
N αα ( α 1 , β )=N αβ 1( α 1 , β)=N ας 1 (α 1 , β )=0 , β 0 ≤ β ≤ β 1 ,
k k
T αα (α 1 , β)=R αα (α 1 , β)=0 , β 0 ≤ β ≤ β1 ,
k k k
N αβ 2(α , β 0)=N ββ ( α , β 0)=N βζ 1 (α , β 0 )=0 , α 0 ≤ α ≤ α 1 ,
k k
T ββ (α , β 0)=R ββ (α 1 , β0 )=0 ,α 0 ≤ α ≤ α 1 ,
k k k
N αβ 2(α , β 1)=N ββ ( α , β1 )=N βζ 1 (α , β1 )=0 , α 0 ≤ α ≤ α 1 , τ =0,1,..., M m
k k
T ββ (α , β 1)=R ββ (α , β1 )=0 , α 0 ≤ α ≤ α 1 (26)
An analytical solution cannot be provided for the five governing equations due to the
dependency of the curvilinear coordinates ( α , β ) . In this context, a numerical solution
must be provided for mechanical displacements, electric and magnetic potential.

2.4 Discretization and numerical solution

The governing equations are discretized in a bidimensional domain by the classical
Chebyshev-Gauss-Lobatto distribution [32]:
(α 2−α 1) ( β 2−β 1)
α i=α 1+ ¿β i=β 1+ ¿ (27)
2 2
where N α and N βare the number of grid distribution over the curvilinear domain. DQM
[30-33] is used for solving numerically the proper discretized equations. The
approximate derivative of mechanical displacements and potential functions are
approximately calculated as linear weighted sum evaluated over the grid domain. The
partial derivatives for a certain function f (α i , β j ) are evaluated as:
 |
(m) Nα
∂ f (α , β)
(m) α =αi
=∑ ω(m )
ik f (α k , β j )i=1 , 2 ,... , N α m=1 , 2 ,... , N α −1
∂α β= β j
k=1

|
(n) Nβ
∂ f ( α , β)
(n) α =αi
=∑ ω(n)
jl f (α i , β l ) j=1 ,2 , ... , N β n=1 ,2 , ... , N β −1
∂β β= β j
l=1

| (∑ )
Nα Nβ
∂(m +n) f ( α , β )
=∑ ω (m)
ik ω(njl ) f (α k , β l ) i=1 , 2 ,... , N α , j=1 ,2 , ... , N β
∂ α (m) ∂ β(n) α=α i
β =β j k=1 l=1

n=1 ,2 , ... , N β −1, m=1 , 2 ,... , N α −1 (28)

where ω(m) and ω(n) are the weights of the coordinates α and β , respectively. The
weights can be calculated by different manners in terms of various basis functions [32,
40, 41]. The selected basis function is the Lagrange interpolation polynomial (LIP) and
has the versatility that can be applied to any grid distribution:
Nα Nβ

Lα (α j )= ∏
i=1 ,i ≠ j
(α j−α i)L β (β j )= ∏
i=1 ,i ≠ j
( β j−β i)

(29)

A recursive formula for calculating the weights for any order of differentiation was
reported by Shu [30]. With the following formulas, the weights ω(m) for a given
coordinate axis α can be obtained.
(1)
(1) L (α j)
ω =
ij , i, j=1 ,2 , ... , N α , i≠ j
(α i −α j) L(α j )
(m−1)
(m) (1) (m−1) a ij
ω ij =n(aii aij − ), i, j=1 ,2 , ... , N , i≠ j , m=2 , 3 , 4 , ... , N α−1
( α i−α j)
N
ω(m)
ii =− ∑ ω (m )
ij , i , j=1 ,2 , ... , N α , i= j , m=1 , 2, 3 , 4 ,... , N α −1 (30)
j=1 , j ≠i

The weights for the axis β are calculated in similar manner. The layerwise theory has
the following number of degrees of freedom for the stiffness matrix
5 N α N β [ N th layer ( M e +1 ) + 1 ]. Where N th layer is the number of layers of the shell panel.

2.5 Stress recovery procedure for transverse variables

A static analysis of MEE shells subjected to mechanical, electrical and magnetic loads is
accomplished. The derivative for the thickness coordinates of potential functions and
mechanical displacement is needed for the calculation of out-plane stresses, transverse
electric displacement, and magnetic induction. Layerwise perspective in conjunction
with PVD do not guarantee the continuity conditions for out-of-plane stresses,
transverse electric displacement and magnetic induction. In this context, a recovery
procedure based on 3D equilibrium equations is needed for the most precise through-
the-thickness distributions. The following procedure is inspired by the contribution of
Tornabene et al. [42] and it can be considered as an extension to an electric and
magnetic case problems. The 3D equilibrium equations are stated below:
∂ τ kαζ σ kββ −σ kαα ∂ A β k k
2 τ kαβ ∂ A α
∂ς (+
2
+
1
)
Rα H kα Rβ H kβ
τ
k
αζ = −
1 ∂ σ αα
A α A β H kβ ∂ α A α H kα ∂ α
−
1 ∂ τ αβ
−
A β H kβ ∂ β A α A β H kα ∂ β
k k k k k k

( )
∂ τ βζ 1 2 k σ αα −σ ββ ∂ A α 1 ∂ σ ββ 1 ∂ τ αβ 2 τ αβ ∂ A β
+ k
+ k
τ βζ = k
− k
− k
− k
∂ς Rα H α R β H β Aα A β H α ∂ β A β H β ∂ β Aα H α ∂ α Aα A β H β ∂ α

σ kαα σ kββ k
τ kαζ k
τ kβζ
( )
∂ σ ζζ 1 1 1 ∂ τ αζ ∂ Aβ 1 ∂ τ βζ ∂ Aα
+ k
+ k
σ ζζ = k
+ k
− k
− k
− k
− k
∂ς Rα H α R β H β R α H α R β H β A α H α ∂ α Aα A β H β ∂ α A β H β ∂ β A α A β H α ∂ β
k k k k

( )
∂ Dζ 1 1 −Dα ∂ A β 1 ∂ Dα Dβ ∂ Aα 1 ∂ Dβ
+ k
+ k
Dζ = k
− k
− k
− k
∂ς Rα Hα Rβ H β Aα Aβ H β ∂ α Aα H α ∂ α Aα Aβ Hα ∂ β Aβ H β ∂ β
k k k

( )
∂ Bζ 1 1 −B α ∂ A β 1 ∂ Bα Bβ ∂ Aα 1 ∂ Bβ
+ k
+ k
Bζ = − − −
∂ς Rα Hα Rβ Hβ A α A β H β ∂ α A α H α ∂ α Aα A β H α ∂ β A β H kβ ∂ β
k k k

(31)
Equations 31 are solved in a semi-analytical manner by the use of DQM. The in-plane
normal stresses ( σ αα , σ ββ ) and shear stress ( τ αβ ), electric displacements ( Dα , Dβ ) and
magnetic induction ( Bα , B β ) are calculated by constitutive equations described in
Equations14-16. The discretization of Equation 31 is given by Chebyshev-Gauss-
Lobatto and N ζ =21 points are used per each shell layer. The derivatives of stresses,
electric displacement, and magnetic induction in terms of ( α , β ) are evaluated by DQM.
The complete procedure for out-plane shear stresses are presented in what follows:

(32)
The number of layers of the shell panel is given by NL. This procedure guarantees the
interlaminar continuity and the correct traction conditions at the shell top and the
bottom. In the case of transverse electric displacement and magnetic induction, a similar
procedure is carried out:

(33)
The through-the-thickness out-of-plane shear stresses and transverse magnetic induction
and electric displacement are recovered by the following manner:

( )| ( )|
k k
2 ς (r )−h 2 ς (r )−h
τ̄ kαζ (r )|ij =τ kαζ (r )|ij + τ 1αζ (1) τ̄ kβζ (r )|ij = τ kβζ (r )|ij + τ 1βζ (1)
2h ij 2h ij

( )| ( )|
k k
2 ς (r )−h 2 ς (r )−h
D̄kζ (r )|ij = Dkζ (r )|ij + D1ζ (1) B̄ kζ (r )|ij =B kζ (r )|ij + B1ζ (1)
2h ij 2h ij

(34)
The results of out-of-plane shear stresses are used for recovering the out-of-plane
normal stress. The derivatives of out-of-plane shear stresses are calculated directly by
the DQM. The following procedure is adopted:
 (35)
The zero-traction condition for the shell bottom is considered in the recovery procedure
by Equation 36:

( )|
k
2 ς (r )−h
σ̄ kζζ (r)|ij = σ kζζ (r )|ij + σ 1ζζ (1) (36)
2h ij

An alternative procedure for recovering the normal strain ε ςς, transverse magnetic and
electric field (E kζ , F kζ ) can be obtained by using the out-of-plane normal stress,
transverse electric displacement and magnetic induction. The following formulas are
used for this purpose:
sm =σ̄ ςς −C 13 ε αα −C23 ε ββ s p= D̄ς −e31 ε αα −e32 ε ββ sn= B̄ ς−q31 ε αα −q 32 ε ββ
2
s11 =(−d 33 + ε 33 μ33)/ s s12=(e 33 μ33−d 33 q33)/ss13=(ε 33 q33−d 33 e 33)/s
2 2
s22=(−q 33 −C 33 μ33 )/s s23=(C33 d 33+ e33 q33 )/s s33=(−e 33 −C33 ε 33)/ s

s=−C33 d 33 −2 d33 e 33 q 33+ μ33 e33 +ε 33 q33 +C 33 ε 33 μ33 ε̄ ςς =s11 sm + s12 s p+ s 13 sn

2 2 2

Ēζ =−(s 12 s m + s22 s p+ s 23 sn ) F̄ ζ =−(s 13 sm + s 23 s p + s33 s n)

(37)
The use of recovery (ε ςς , Ekζ , F kζ ) can be easily applied to recalculation of normal
stresses (σ kαα , σ kββ ). The out-of-plane linear strain γ kαζ , γ kβζ are recovered by the stresses
k k
τ αζ , τ βζ , in the following manner:
k k k k k k k k k k
k τ̄ αζ + e15 Eα +q15 Fα k τ̄ βζ +e 24 E β + q24 F β
γ̄ αζ = k
γ̄ βζ = k
C55 C 44

(38)
These results can be used for recovering the electric displacement (Dkα , D kβ ) and
magnetic induction (Bkα , Bkβ ).

3. Results
The static response of shell panels subjected to mechanical, electric, magnetic loads are
studied. The different kind of loads are applied at the shell top. The shell is discretized
with N α =N β =13 (see Equation 27). The results are compared with other layerwise
theories and exact elasticity solutions. The validation of the Layerwise model is
established by 2 assessments: the first problem was originally formulated by Pan [43]
and it is about a three-layered square plate subjected to bisinusoidal mechanical load,
and the second assessment is about a simply supported shallow shell subjected to a
bisinusoidal electric and magnetic load. As the results are validated, this Layerwise
model can be used a referential solution for shells with variable radii of curvature. The
first case to evaluate is a parabolic shell subjected to a magnetic flux. The second
problem is related to a cycloidal panel submitted to electric flux.

3.1 Assessment 1: A three-layered plate subjected to mechanical load

A simply supported three-layered plate is submitted to a bisinusoidal mechanical
pressure. This case problem was solved by Pan [43] by the Stroh formalism and the
propagator matrix method. Phoenix et al. [27] employed a layerwise theory and the
finite element method in conjunction with Reissner mixed variation theorem to solve
this problem. Monge and Mantari [44] developed a quasi-3D solution for this problem,
by employing the equilibrium equations. The solutions were derived in term of Navier
method and then the equations in term of the thickness coordinates are solved
numerically by DQM. The geometry of the plate is given by: 0 ≤ α ≤ 1 and 0 ≤ β ≤ 1 and
the thickness is represented as h=0.3. The plate is composed of two different material
which are: barium titanate BaTi O3 (B) and cobalt ferrite CoF e 2 O 4 (F) and its
mechanical, electric, and magnetic properties are reported in Table 1. The sequence of
the layers is considered as B/F/B and F/B/F. The through-the-thickness distribution of
the panels are calculated at the coordinates α =0.75 , β=0.25 . The transverse mechanical
load is considered as:
π π
P=P o sin (¿ α )sin(¿ β), P o=1 Pa. ¿ ¿
a b
(39)
The results at the middle, top, and bottom of the plate are reported in Table 2. A good
agreement is appreciated between the quasi-3D theories and the proposed layerwise
models. It is seen that for the four order or more, the same numerical result is obtained.

3.2 Assessment 2: A shallow shell subjected to magnetic and electrical load.

A shallow shell is subjected to a magnetic and electrical load. The boundary conditions
are considered as simply supported. The results are evaluated at the coordinates
α =0.75 , β=0.25 . The geometry of the shell is described by the following parametric
surface:

( ) ( )
X =2 R α sin
α
2 Rα
cos
α −a
2 Rα ( ) ( )
Y =2 R β sin
β
2 Rβ
cos
β−b
2 Rβ

Z=−2 R sin
( 2αR ) sin ( α−a
α
α 2R )−2 R sin
α
( β
2Rβ
) sin
( 2R )
β−b
β β

Rα =R β =20 m ,a=b=1 m, h=0.15 m0 ≤ α ≤ 1 ,0 ≤ β ≤ 1

(40)
The results are compared with quasi-3D solutions proposed by Monge and Mantari [44].
The shell has a stacking sequence of B/F/B when an electric load is considered and
F/B/F when it is subjected to a magnetic load. The electric load is given by a
bisinusoidal function:
−5 2
Dζζ =D ζ sin ( πα ) sin(¿ πβ ), Dζ =1 0 C /m ¿ (41)
The through-the-thickness distribution of displacements ( u , w ), potential functions ( ϕ , ψ )
, electric displacement ( Dζ ) and magnetic induction ( Bζ ) are shown in Figure 3. A linear
response is shown in the in-plane displacement u. A constant behavior can be seen at
the middle layer of the transverse electrical displacement; this is evidenced due to
piezoelectric properties. A linear behavior is presented for the middle surface of the
electric scalar potential function. The traction conditions for the transverse electric
displacement and electric field are accomplished. The distribution for normal stresses
( σ αα , σ ζζ ) and shear stresses ( τ αβ , τ αζ ) are given in Figure 4. A certain discontinuity can be
evidenced for the normal stress σ αα ; this can be different depending on the case problem
such as functionally graded MEE. For the shear stress τ αβ, the discontinuity is less than
in the case of σ αα because of the absence of electric and magnetic load in the
calculation. The zero-traction condition are accomplished for the stresses τ αζ , σ ζζ .
Similar as in the previous condition, the shell panel is subjected to a magnetic load with
a bi-sinusoidal form:
−3
Bζζ =Bζ sin ( πα ) sin(¿ πβ), B ζ =1 0 N / Am ¿ (42)
The distribution of displacements ( u , w ), potential functions ( ϕ , ψ ), electric
displacement ( Dζ ) and magnetic inductance ( Bζ ) are displayed in Figure 5. A linear zig-
zag behavior can be seen for the magnetic inductance ( Bζ ), and the exact traction
condition at the shell top is accomplished. A nonlinear behavior is appreciated at the
middle surface in the distribution of electric displacement ( Dζ ) . A linear behavior is
observed for the middle surface of potential functions ( ϕ , ψ ). The through-the-thickness
distribution of stresses are reported in Figure 6. It can be seen that the layerwise models
can follows the quasi-3D response. The in-plane shear stress τ αβ present higher
interlaminar discontinuities due to materials than when an electric load is applied. The
numerical values of stresses ( σ αα , τ αβ , τ αζ , σ ζζ ), displacements ( u , w ), potential functions
( ϕ , ψ ), transverse electric displacement Dζ , and magnetic induction Bζ are reported in
Table 3. The present results describe a similar response when compared to the quasi-3D
referential solution and the Layerwise models.

3.2 A parabolic shell subjected to magnetic load.

A parabolic shell is subjected to a constant magnetic load at the shell top,
−2
Bζζ =1 0 N / Am. The boundary conditions are considered as clamped. The stacking
sequence as F/B/F and each layer presents the same thickness. The numerical results are
evaluated at ( 0.25 ( α 1−α 0 ) + α 0 , 0.25 ( β 1−β 0 ) + β 0 ). The geometry is given in term of the
following parametric surface:
X =α cos ( β )Y =−α sin ( β ) Z=a . α 2αε [ 0.5 , 1 ] , βε [ 0 , π /2 ] , h=0.09 , a=2

(43)
The parabolic shell is shown in Figure 7. The through-the-thickness distributions of
displacements are shown in Figure 8. The displacement v presents a linear response for
each layer. The displacements ( u , w ) present a nonlinear response at the shell top. The
potential functions distributions are shown in Figure 9. For the magnetic potential
function all the order of expansion presents almost the same results. It can be seen just a
slightly numerical difference between LD3 and LD4 and higher orders. The electrical
displacements and magnetic inductions are delivered in Figure 10. Almost a constant
zero-response is found for in-plane electric displacement and magnetic induction at the
shell top and bottom, and at the middle shell layer, respectively. The traction condition
for transverse electric displacement and magnetic induction at the shell top and bottom
are fully achieved. The through-the-thickness response of stresses is reported in Figure
11. Certain interlaminar discontinuities are seen for in-plane stresses, due to the abrupt
change of electric and magnetic properties. The traction conditions for transverse out-
plane stresses are accomplished correctly. The numerical values for mechanical and
electrical displacements, stresses, magnetic induction, and potential functions are
reported in Table 4. It is seen that there is a sort of discrepancy for transverse electric
displacement when LD4 is used, and this can be seen graphically in Figure 10.
However, it is numerically demonstrated that LD5 and LD6 present almost the same
response.

3.3 A cycloidal shell subjected to an electric load.

A cycloidal shell panel is subjected to a constant transverse electrical displacement
−8 2
Dζ =10 C /m . The lamination is B/F/B and every layer is considered as equal
thickness. The boundary conditions are considered as clamped. The numerical results
are calculated at the point ( 0.25 ( α 1−α 0 ) + α 0 , 0.25 ( β 1−β 0 ) + β 0 ). The cycloidal shell is
shown in Figure 12 and is modeled by the following parametric surface:

X =a ( 2 α + sin ( 2 α ) ) cos ( β )Y =−a ( 2 α + sin (2 α ) ) sin ( β ) Z=a ( 1−cos ( 2 α ) )

a=1.5 , π /6 ≤α ≤ π /3 , 0 ≤ β ≤ π /2.h=0.06 m .
(44)
The through-the-thickness distributions of mechanical displacement are presented in
Figure 13. It is demonstrated that the in-plane displacements have a linear response. A
nonlinear behavior is observed for the transverse displacements. Almost constant
behavior is appreciated for the displacement v at the middle of the shell. The
distributions of potential functions are shown in Figure 14. For the electric potential
function, a constant behavior is evidenced at the shell bottom and top, and a linear
behavior at middle surface. A nonlinear response is evidenced for the magnetic
potential, LD5 and LD6 presents almost the same accuracy. The electrical
displacements and magnetic inductions are delivered in Figure 15. In-plane electric
displacement is zero at the middle layer due to the absence of piezoelectric properties in
the middle layer. A zig-zag linear behavior is presented for the transverse electric
displacement. In the case of transverse magnetic induction, a linear response is
evidenced at the shell bottom and top. The through-the-thickness distribution of stresses
are given in Figure 16. It is demonstrated that higher interlaminar discontinuities are
presented for σ ββ than for σ αα . As it is well known this problem is related to the material
properties of the layers, and not a numerical problem. However, it can be redesigned to
have smoothly through-the-thickness behavior of the normal and in-plane stresses by
using reinforced MEE (with the use of magnetoelectric coefficient) or MEE functionally
graded materials. The numerical results are reported in Table 5. Results based on the
unified formulation shows good accuracy and tendency.
Overall, 02 benchmark problems for smart complex shells are introduced to the
literature. Still is important to investigate the manner to alleviate the discrepancies or
continuity challenges found in this article in order to provide a robust computational
model to study complex smart shell structures for marine, civil and aerospace
applications.

4. Conclusions
The bending response of magneto-electric-elastic complex shells with variable radii of
curvature for marine, civil and aerospace applications is analyzed by a two-dimensional
solution. Differential Geometry is employed for the calculation of Lamé parameters,
radius of curvature and metric coefficients. The mechanical displacement, electric and
magnetic potential functions are modeled in term of a layerwise theory which is based
on the concept of Chebyshev polynomials of the first kind. Principle of virtual
displacements is employed for obtaining the governing equations. Chebyshev-Gauss-
Lobatto grid distribution is used for discretizing the partial differential equations.
Differential quadrature is employed for solving the equations numerically and Lagrange
polynomials are employed as basis function. The classical three-dimensional
equilibrium equations written in curvilinear coordinates are employed for recovering
out-plane stresses, transverse electric displacement, and magnetic induction. This
method can solve several types of complex shell and explore different boundary
conditions due to its numerical formulation. The results demonstrated good accuracy
compared to 3D and layerwise solutions. Finally, benchmarks solutions are provided.

Acknowledgments

This paper was written in the context of the project: “Desarrollo de un algoritmo
autónomo y óptimo de mecánica computacional para un análisis de estructuras
complejas impresa con tecnología 3D, utilizando inteligencia artificial y algoritmos
genéticos” founded by CONCYTEC under the contract number N°060-2021-
PROCIENCIA. The authors of this manuscript appreciate the financial support from the
Peruvian Government.

References
[1] Chen W.Q., Lee K.Y., Alternative state space formulations for magnetoelectric
thermoelastic with transverse isotropy and the application to bending analysis of
nonhomogeneous plates. International journal of solids and structures 40 (2003) 5689-
5705.
[2] Vinyas M., Kattimani S.C., Static studies of stepped functionally graded magneto-
electro-elastic beam subjected to different thermal loads. Composite Structures 163
(2017) 216-237.
[3] Ragb O., Mohamed M., Matbuly M.S., Vibration analysis of magneto-electro-
thermo nanobeam resting on nonlinear elastic foundation using sinc and discrete
singular convolution differential quadrature method. Modern Applied Science 13(7)
(2019) 49-79.
[4] Wang X., Lee J.S., Zheng X., Magneto-thermo-elastic instability of ferromagnetic
plates in thermal and magnetic fields. International journal of solids and structures 40
(2003) 6125-6142.
[5] Heyliger P.R., Ramirez F., Pan E., Two-dimensional static fields in
magnetoelectroelastic laminates. Journal of intelligent material systems and structures
15 (2004) 689-709.
[6] Pan E., Han F., Exact solution for functionally graded and layered magneto-electro-
elastic plates. International journal of engineering science 43 (2005) 321-339.
[7] Sobhy M., Magneto-electro-thermal bending of FG-graphene reinforced polymer
doubly-curved shallow shells with piezomagnetic faces. Composite Structures 203
(2018) 844-860.
[8] Badri T.M., Al-Kayiem H.H., Analytical solution for simply supported and
multilayered magneto-thermo-electro-elastic plates. Asian journal of scientist research
6(2) (2013) 236-244.
[9] Liu M.F., Exact solution for the bending deformations of layered magneto-electro-
elastic laminates based on thin-plate formulation. International journal of engineering
and applied sciences 3(4) (2016) 2394-3661.
[10] Vinyas M., Kattimani S.C., A finite element based assessment of static behavior of
multiphase magneto-electro-elastic beams under different thermal loading. Structural
engineering and mechanics 62 (5) (2017) 519-535.
[11] Pan E., Heyliger P.R., Free vibrations of simply supported and multilayered
magneto-electro-elastic plates. Journal of sounds and vibrations 252(3) (2002) 429-442.
[12] Moita J.M.M., Soares C.M.M., Soares C.A.M., Analyses of magneto-electro-elastic
plates using a higher order finite element model. Composite Structures 91(4) (2009)
421-426.
[13] Wang J., Li X., Analytical solutions for the magnetoelectric effect of multilayered
magneto-electro-elastic media. Smart materials and structures 17 (2008) 045028.
[14] Liu M.F., An exact deformation analysis for the magneto-electro-elastic fiber-
reinforced thin plate. Applied Mathematical modelling 35 (2011) 2443-2461.
[15] Pan E., Wang R., Effects of geometric size and mechanical boundary conditions on
magnetoelectric coupling in multiferroic composites. Journal of physics D: Applied
physics 42 (2009) 245503.
[16] Buchanan G.R., Layered versus multiphase magneto-electro-elastic composites.
Composites: Part B 35 (2004) 413-420.
[17] Ebrahimi F., Jafari A., Buckling behavior of smart MEE-FG porous plate with
various boundary conditions based on refined theory. Advanced in material research 5
(4) (2016) 279-298.
[18] Chen J., Pan E., Chen H., Wave propagation in magneto-electro-elastic
multilayered plates. International journal of solids and structures 44 (2007) 1073-1085.
[19] Carrera E., Di Gifico M., Nali P., Brischetto S., Refined multilayered plate
elements for coupled magneto-electro-elastic analysis. Multidiscipline modeling in
materials and structures 5 (2009) 119-138.
[20] Daga A., Ganesan N., Shankar K., Behavior of magneto-electro-elastic sensors
under transient mechanical loading. Sensor and actuator A: Physics 150 (2009) 46-55.
[21] Milazzo A., A one-dimensional model for dynamic analysis of generally layered
magneto-electro-elastic beams. Journal of sound and vibration 332 (2013) 465-483.
[22] Ewolo Ngak F.P., Ntamack G.E., Azrar L., Dynamic analysis of multilayered
magnetoelectroelastic plates based on a pseudo-Stroh formalism and Lagrange
polynomials. Journal of intelligent material systems and structures 30(6) (2019) 939-
962.
[23] Ebrahimi F., Barati M.R., Buckling analysis of smart size-dependent higher order
magneto-electro-thermo-elastic functionally graded nanosize beams. Journal of
mechanics 33(1) (2017) 23-33.
[24] Zhang P., Qi C., Fang H., Ma C., Huang Y., Semi-analytical analysis of static and
dynamics responses for laminated magneto-electro-elastic plates. Composite Structures
222 (2019) 110933.
[25] Alaimo A., Milazzo A., Orlando C., A four-node MITC finite element for
magneto-electric-elastic multilayered plates. Computers and structures 129 (2013) 120-
133.
[26] Ebrahimi F., Dehghan M., Seyfi A., Eringen’s nonlocal elasticity theory for wave
propagation analysis of magneto-electro-elastic nanotubes. Advances in nano research
70(1) (2019) 1-11.
[27] Phoenix S.S., Satsangi S.K., Singh B.N., Layer-wise modelling of magneto-electro-
elastic plates. Journal of sound and vibration 324 (2009) 798-815.
[28] Bellman R., A new method for the identification of systems. Mathematical
bioscience 5 (1969) 201-204.
[29] Bellman R., Casti J., Differential quadrature and long-term integration. Journal of
mathematical analysis and applications 34(2) (1971) 235-238.
[30] Shu C., Differential Quadrature and its application in engineering: Springer:
London (2000).
[31] Wang X., Differential Quadrature and Differential Quadrature based element
methods. Butterworth-Heinemann (2015).
[32] Tornabene F., Fantuzzi N., Ubertini F., Viola E., Strong formulation finite element
based on differential quadrature: A survey. Applied mechanics: Reviews 67(2) (2015)
1-55.
[33] Bert C.W., Malik M., Differential quadrature method in computational mechanics:
a review. Applied mechanics: Reviews 49(1) (1996) 49(1) 1-28.
[34] Cinefra M., Carrera E., Valvano S., Variational kinematic Shell elements for the
analysis of electro-mechanical problems. Mechanics of advanced materials and
structures 22 (2015) 77-106.
[35] Giunta G., Koutsawa Y., Belouettar S., Analysis of three-dimensional piezo-
electric beams via unified formulation. Advanced materials research 745 (2013) 101-
118.
[36] Carrera E., Nali P., Multilayered plate elements for the analysis of multifield
problems. Finite elements in analysis and design 46(9) (2010) 732-747.
[37] Soedel W., Vibrations of shells and plates. Marcell Dekker (2004).
[38] Leissa A.W., Vibrations of shells. NASA Sp. 288 (1973).
[39] Kraus H., Thin elastic shells. John Willey & Sons: New York (1967).
[40] Monge J.C., Mantari J.L., Arciniega R.A., Computational semi-analytical method
for the 3D elasticity bending solution of laminated composite and sandwich doubly-
curved shells. Engineering Structures 221 (2020) 110938.
[41] Tornabene F., Fantuzzi N., Bacciocchi M., Strong and weak formulations based on
differential and integral quadrature methods for the free vibration analysis of composite
plates and shells: Convergence and accuracy. Engineering analysis with boundary
elements 92 (2018) 3-37.
[42] Tornabene F., Fantuzzi N., Bacciochi M., Viola E., Accurate inter-laminar
recovery for plates and doubly-curved with variable radii of curvature using layer-wise
theories. Composite Structures 124 (1) (2015) 369-393.
[43] Pan E., Exact solution for simply supported and multilayered magneto-electro-
elastic plates. Transaction of the ASME 68 (2001) 608-618.
[44] Monge J.C., Mantari J.L., Three dimensional numerical solution for the bending
study of magneto-piezo-elastic spherical and cylindrical shells. Engineering Structures
238 (2021) 112158.
[45] Shen Li, Do Kyun Kim, Ultimate strength characteristics of unstiffened cylindrical
shell in axial compression, Ocean Engineering, 243, 2022, 110253.
[46] Dechun Zhang, Peng Li, Jun Lu, Chengxiang Zhang, Yiren Yang, Stability analysis
of cylindrical shell in axial flow: A DQ-based approach and an instability prediction
formula, Ocean Engineering, 267, 2023, 113198.
[47] Wei Guan, Yongmei Zhu, Weili Wang, Fang Wang, Jian Zhang, Yu Wu, Zhong
Zheng, Experimental and numerical buckling analysis of carbon fiber composite
cylindrical shells under external pressure, Ocean Engineering, 275, 2023, 114134.
[48] Yi Ren, Yuxuan Qin, Fuzhen Pang, Hongfu Wang, Yumin Su, Haichao Li,
Investigation on the flow-induced structure noise of a submerged cone-cylinder-
hemisphere combined shell, Ocean Engineering, 270, 2023, 113657.
[49] Xiao Fu, Zhiyuan Mei, Xuefei Bai, Yi Xia, Shuo Wang, Mechanical properties and
optimal configurations of variable-curvature pressure hulls based on the equal-strength
shell theory, Ocean Engineering, 266, 3, 2022, 112938.
[50] Emad Sobhani, Amir R. Masoodi, Amir Reza Ahmadi-Pari, Free-damped vibration
analysis of Graphene Nano-Platelet nanocomposite Joined Conical-Conical-Cylindrical
Shell marine-like structures, Ocean Engineering, 261, 2022, 112163.
[51] Emad Sobhani, Vibrational characteristic simulations regarding connecting two
different semi-spheroidal shells and a full-spheroidal shell with a conical shell
categorized in underwater structures, Ocean Engineering, 276, 2023, 114252.

Table Captions
Table 1. Mechanical, electric, and magnetic properties of BaTi O3 and CoF e 2 O 4.
Table 2. Numerical results for ψ and ϕ at the top, midsurface and bottom of a
rectangular plate.
Table 3. Numerical values of mechanical displacements, potential functions, transverse
electric displacement, magnetic induction, and stresses for a shallow shell.
Table 4. Numerical values of stresses, displacements, potential functions, magnetic
induction, and electric displacements for a parabolic shell.
Table 5. Numerical values of stresses, displacements, potential functions, magnetic
induction, and electric displacements for a cycloidal shell.

Figure Captions
Figure 1. A general doubly curved shell and its curvilinear coordinate.
Figure 2. The global and local thickness coordinate of a multilayered shell.
Figure 3. Through-the-thickness distributions of displacements ( u , w ), potential
functions ( ϕ , ψ ), electric displacement ( Dζ ) and magnetic inductance ( Bζ ) of a shallow
shell subjected to electric load.
Figure 4. Through-the-thickness distributions of stresses ( σ αα , τ αβ , τ αζ , σ ζζ ) for a shallow
shell subjected to electric load.
Figure 5. Through-the-thickness distributions of displacements ( u , w ), potential
functions ( ϕ , ψ ), electric displacement ( Dζ ) and magnetic inductance ( Bζ ) of a shallow
shell subjected to magnetic load.
Figure 6. Through-the-thickness distributions of stresses ( σ αα , τ αβ , τ αζ , σ ζζ ) for a shallow
shell subjected to magnetic load.
Figure 7. A parabolic shell.
 Figure 8. Through-the-thickness of mechanical displacements for the parabolic shell.
Figure 9. Through-the-thickness of potential functions for the parabolic shell.
Figure 10. Through-the-thickness of electric displacements and magnetic inductions for
the parabolic shell.
Figure 11. Through-the-thickness of stresses for the parabolic shell.
Figure 12. A cycloidal shell.
Figure 13. Through-the-thickness of mechanical displacements for the cycloidal shell.
Figure 14. Through-the-thickness of potential functions for the cycloidal shell.
Figure 15. Through-the-thickness of electric displacements and magnetic inductions for
the cycloidal shell.
Figure 16. Through-the-thickness of stresses for the cycloidal shell.

Tables

Table 1.

Mechanica
l
Properties BaTi O3 CoF e 2 O 4 Electric Properties BaTi O3 CoF e 2 O 4 Magnetic Properties BaTi O3 CoF e 2 O 4
C 11 (GPa) 166 286 2
e 31(C/m ) -4.4 0 q 31(N /( A . m)) 0 580.3
C 22(GPa ) 166 286
2
C 12(GPa ) 77 173 e 32(C/m ) -4.4 0 q 32(N /( A . m)) 0 580.3
2
C 13(GPa ) 78 170.5 e 33 (C /m ) 18.6 0 q 33(N /( A . m)) 0 699.7
2
C 23(GPa ) 78 170.5 e 24 (C /m ) 11.6 0 q 24 (N /( A .m)) 0 550
2
C 33(GPa ) 162 269.5 e 15 (C /m ) 11.6 0 q 15(N /( A . m)) 0 550
−9 2 2 −6 2 2
C 44 (GPa) 43 45.3 p11 (1 0 C / N m ) 11.2 0.08 μ11 (10 N s /C ) 5 -590
−9 2 2 −6 2 2
C 55(GPa ) 43 45.3 p22 (1 0 C / N m ) 11.2 0.08 μ22 (1 0 N s /C ) 5 -590
−9 2 2 −6 2 2
C 66 (GPa) 44.5 56.5 p33 (1 0 C / N m ) 12.6 0.093 μ33 (1 0 N s /C ) 10 157
Table 2.
B/F/B
Phoenix
et. Al. Monge and
Variables Pan [43] [27] Mantari [44] LD3 LD4 LD5 LD6
ψ (−h/ 2 ) [ 10−6 C/ s ] -2.5270 -2.5200 -2.5316 -2.5316 -2.5315 -2.5316 -2.5315
ψ ( 0 ) [ 10−6 C / s ] -2.6000 -2.5700 -2.5974 -2.5980 -2.5973 -2.5973 -2.5973
ψ ( h/ 2 ) [ 10−6 C / s ] -2.1880 -2.1830 -2.1936 -2.1935 -2.1935 -2.1935 -2.1936
ϕ (−h/2 ) [ 1 0−4 V ] 7.8740 8.0060 7.8620 7.8323 7.8320 7.8320 7.8326
ϕ ( 0 ) [ 1 0−3 V ] 1.3590 1.3580 1.4173 1.4158 1.4158 1.4158 1.4158
ϕ ( h/2 ) [ 1 0−3 V ] 1.0530 1.0500 1.0544 1.0542 1.0542 1.0541 1.0542
F/B/F
Phoenix
et. Al. Monge and
Variable Pan [43] [27] Mantari [44] LD3 LD4 LD5 LD6
ψ (−h/ 2 ) [ 10−6 C/ s ] -1.2460 -1.1840 -1.0832 -1.0835 -1.0835 -1.0834 -1.0834
ψ ( 0 ) [ 10−6 C / s ] -1.6700 -1.6700 -1.6695 -1.6697 -1.6697 -1.6697 -1.6697
ψ ( h/ 2 ) [ 10 C / s ]
−6
-8.6920 -8.6530 -8.6504 -8.6510 -8.6510 -8.6514 -8.6515
ϕ (−h/2 ) [ 1 0−3 V ] 1.9000 1.8970 1.9068 1.9073 1.9073 1.9073 1.9073
ϕ ( 0 ) [ 1 0−3 V ] 2.2970 2.2940 2.2450 2.2456 2.2455 2.2455 2.2455
ϕ ( h/2 ) [ 1 0−3 V ] 2.1540 2.1480 2.1569 2.1574 2.1574 2.1574 2.1574

Table 3.
B/F/B
Monge and
Variable Mantari [44] LD3 LD4 LD5 LD6
u ( h/6 ) [ 1 0−9 m ] -1.8788 -1.8743 -1.8743 -1.8744 -1.8744
w ( 0 ) [1 0−8 m ] -1.5740 -1.5727 -1.5727 -1.5727 -1.5727
ϕ ( h/2 ) [ V ] 324.6497 324.6622 324.6619 324.6623 324.6620
ψ ( 0 ) [ 10−4 C /s ] 2.8148 2.7801 2.7781 2.7771 2.7786
Dζ ( h/6 ) [ 1 0−7 C /m ] -5.2772 -5.2760 -5.2764 -5.2759 -5.2763
Bζ ( 0 ) [ 10 N / Am ]
−8
-5.3306 -5.2456 -5.2462 -5.1428 -5.2511
σ αα ( h /2 ) [ Pa ] 709.0662 713.6877 713.6460 713.6726 713.5856
τ αβ (−h/2 ) [ Pa ] 401.9081 398.1719 398.1597 398.1743 398.1738
τ αζ ( 0 ) [ Pa ] 55.0585 54.3581 54.3887 54.3920 54.3914
 σ ζζ ( 0 ) [ Pa ] -28.2589 -28.2544 -28.2499 -28.2559 -28.2373
F/B/F
Monge and
Variable Mantari [44] LD3 LD4 LD5 LD6
u ( h/6 ) [ 1 0−9 m ] 1.1021 1.1013 1.1012 1.1012 1.1013
w ( 0 ) [1 0 m ]
−9
2.1719 2.1629 2.1630 2.1630 2.1630
ϕ ( h/2 ) [ 1 0−2 V ] -9.9512 -9.9976 -9.9943 -9.9973 -9.9892
ψ ( 0 ) [ 10−1 C / s ] -4.5318 -4.5318 -4.5318 -4.5318 -4.5318
Dζ ( 0 ) [ 1 0−10 C /m ] 9.2401 9.2406 9.2321 9.2359 9.2334
Bζ ( 0 ) [ 10−4 N / Am ] 5.1980 5.1980 5.1980 5.1980 5.1980
σ αα ( h /2 ) [ Pa ] -971.5696 -973.3868 -973.4845 -973.4231 -973.4073
τ αβ (−h/2 ) [ Pa ] -437.1073 -436.4174 -436.4070 -436.4303 -436.4307
τ αζ ( 0 ) [ Pa ] -8.4751 -8.2075 -8.2150 -8.2050 -8.2073
σ ζζ ( 0 ) [ Pa ] 24.0108 24.0142 24.0242 24.0193 24.0119

Table 4.
Variable LD3 LD4 LD5 LD6
u(h/ 3) [ 1 0−8 m ] 1.7372 1.7555 1.7191 1.7196
v (h/2) [ 1 0−8 m ] 2.6432 2.6433 2.6420 2.6420
w (−h /6) [ 1 0−8 m ] 2.1120 2.1139 2.1259 2.1259
ϕ (−h/6) [ V ] -3.0137 -3.0153 -3.0988 -3.0982
ψ (h /6) [ 10 C/ s ] -3.5451 -3.5451 -3.5446 -3.5446
Dα (h /6) [ 1 0−6 C /m ] -1.4025 -1.4016 -1.3545 -1.3544
D β (h /6) [ 1 0 C /m ]
−7
1.4938 1.4875 1.5232 1.5228
Dς (0) [1 0 C /m ] −8
8.9322 8.0577 8.8153 8.8380
Bα (h/ 2) [ 1 0 N / Am ]
−2
3.7873 3.7845 3.6728 3.6728
Bβ (h/2) [ 1 0 N / Am ] −2
-3.9199 -3.9212 -3.9230 -3.9230
Bζ (0) [ 10 N / Am ]
−2
1.0017 1.0017 1.0015 1.0015
σ αα (h/2) [ 1 0 Pa ] 4
-2.7083 -2.7056 -2.6897 -2.6898
σ ββ (h/ 2) [ 1 0 Pa ]
4
-1.4376 -1.4348 -1.4274 -1.4274
τ αβ (h /2) [ 10 Pa ] 3
1.2764 1.2768 1.2572 1.2572
τ αζ (0) [ 1 0 Pa ]
2
-4.7477 -4.7087 -4.8143 -4.8150
τ βζ (0) [1 0 Pa ] 3
-3.4198 -3.4140 -3.4140 -3.4141
σ ζζ (0) [ 1 0 Pa ]
3
-1.0420 -1.0398 -1.0257 -1.0257
Table 5.
Variable LD3 LD4 LD5 LD6
u(−h/2) [ 1 0−11 m ] 8.3902 8.3908 8.3886 8.3911
v (h/ 2) [ 1 0−12 m ] -2.2890 -2.2940 -2.2915 -2.2975
w (0) [1 0−11 m ] 7.4086 7.4065 7.4130 7.4107
ϕ (0) [ V ] 8.0094 8.0104 8.0093 8.0103
ψ (0) [ 10−5 C / s ] -4.2893 -4.3145 -4.2974 -4.3142
D α (−h / 2) [ 1 0−8 C/m ] -9.1405 -9.1414 -9.1360 -9.1422
D β (−h /2) [ 1 0 C/m ]
−8
-2.7070 -2.7074 -2.7066 -2.7085
D ς (0) [1 0 C / m ]
−9
-4.8646 -4.8629 -4.8667 -4.8634
Bα (0) [ 1 0 N / Am ]
−8
-6.7127 -6.7151 -6.7013 -6.7274
Bβ (0) [ 1 0 N / Am ]
−8
-7.5349 -7.5256 -7.7221 -7.2623
Bζ (h /6) [ 1 0 N / Am ]
−10
-9.1936 -9.2976 -9.1359 -9.2004
σ αα (h/ 2) [ Pa ] 15.5909 15.5944 15.5850 15.5951
σ ββ (h/2) [ Pa ] 3.2008 3.2039 3.1983 3.2040
τ αβ (h /2) [ Pa ] -0.5245 -0.5242 -0.5243 -0.5247
τ αζ (0) [ Pa ] 0.8906 0.8916 0.8914 0.8914
τ βζ (0) [1 0−2 Pa ] 3.6723 3.6703 3.6685 3.6744
σ ζζ (0) [ 1 0 Pa ]
−2
-3.9148 -3.9078 -3.9032 -3.9085

Figures
Figure 1.
Figure 2.

Figure 3
 1 1

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2
3D
0 LD3 0
LD4
-0.2 LD5 -0.2
LD6 3D
-0.4 -0.4 LD3
LD4
-0.6 -0.6 LD5
LD6
-0.8 -0.8

-1 -1
-2 0 2 4 6 8 10 -1.72 -1.7 -1.68 -1.66 -1.64 -1.62 -1.6 -1.58 -1.56 -1.54
10 -9 10 -8

1 1
3D
0.8 0.8 LD3
LD4
0.6 0.6 LD5
LD6
0.4 0.4

0.2 0.2

0 0

-0.2 -0.2

3D
-0.4 -0.4
LD3
LD4
-0.6 -0.6
LD5
LD6
-0.8 -0.8

-1 -1
0 50 100 150 200 250 300 350 0.5 1 1.5 2 2.5 3 3.5
10 -4

1 1

3D
0.8 0.8
LD3
LD4
0.6 0.6
LD5
LD6
0.4 0.4

0.2 0.2

0 0

3D
-0.2 -0.2
LD3
LD4
-0.4 -0.4
LD5
LD6
-0.6 -0.6

-0.8 -0.8

-1 -1
-5 -4.5 -4 -3.5 -3 -2.5 -2 -1.5 -1 -0.5 0 -6 -5 -4 -3 -2 -1 0 1
10 -6 10 -8
Figure 4
1 1

0.8 0.8

0.6 0.6

0.4 0.4

3D
0.2 0.2
LD3
LD4
0 0
LD5
LD6 3D
-0.2 -0.2
LD3
LD4
-0.4 -0.4
LD5
LD6
-0.6 -0.6

-0.8 -0.8

-1 -1
-2000 -1500 -1000 -500 0 500 1000 -1000 -500 0 500 1000 1500 2000 2500 3000

1 1
3D
0.8 LD3 0.8
LD4
0.6 LD5 0.6
LD6
0.4 0.4

0.2 0.2
3D
LD3
0 0 LD4
LD5
-0.2 -0.2 LD6

-0.4 -0.4

-0.6 -0.6

-0.8 -0.8

-1 -1
-150 -100 -50 0 50 100 -35 -30 -25 -20 -15 -10 -5 0 5
Figure 5
1 1

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0 0

-0.2 -0.2

-0.4 -0.4
3D 3D
-0.6 LD3 -0.6 LD3
LD4 LD4
-0.8 LD5 -0.8 LD5
LD6 LD6
-1 -1
-2 -1.5 -1 -0.5 0 0.5 1 1.5 1.95 2 2.05 2.1 2.15 2.2 2.25
10 -9 10 -9

1 1
3D 3D
0.8 LD3 0.8 LD3
LD4 LD4
0.6 LD5 0.6 LD5
LD6 LD6
0.4 0.4

0.2 0.2

0 0

-0.2 -0.2

-0.4 -0.4

-0.6 -0.6

-0.8 -0.8

-1 -1
-0.15 -0.1 -0.05 0 0.05 0.1 0.15 0.2 -2 -1.5 -1 -0.5 0 0.5 1

1 1
3D
0.8 LD3
0.8
LD4
0.6 LD5 0.6
LD6
0.4 0.4

0.2 0.2

0 0
3D
-0.2 -0.2
LD3
LD4
-0.4 -0.4
LD5
LD6
-0.6 -0.6

-0.8 -0.8

-1 -1
-2 0 2 4 6 8 10 -6 -4 -2 0 2 4 6
10 -10 10 -4
Figure 6
1 1

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0 0

3D
-0.2 -0.2
LD3
LD4
-0.4 -0.4
3D LD5
LD3 LD6
-0.6 -0.6
LD4
LD5
-0.8 -0.8
LD6
-1 -1
-1000 -800 -600 -400 -200 0 200 400 600 800 -800 -600 -400 -200 0 200 400

1 1

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0 0

3D
-0.2 -0.2
LD3
LD4
-0.4 -0.4 3D
LD5
LD6 LD3
-0.6 -0.6 LD4
LD5
-0.8 -0.8 LD6

-1 -1
-80 -60 -40 -20 0 20 40 60 80 100 -5 0 5 10 15 20 25

Figure 7.
Figure 8.
 1

0.8
LD3
LD4
0.6
LD5
LD6
0.4

0.2

-0.2

-0.4

-0.6

-0.8

-1
-2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2
10 -8

0.8

0.6

0.4

0.2

-0.2
LD3
LD4
-0.4
LD5
LD6
-0.6

-0.8

-1
-3 -2 -1 0 1 2 3
10 -8

0.8

0.6

0.4

0.2

-0.2
LD3
LD4
-0.4
LD5
LD6
-0.6

-0.8

-1
1.8 1.85 1.9 1.95 2 2.05 2.1 2.15
10 -8
Figure 9.

1
LD3
0.8 LD4
LD5
0.6 LD6

0.4

0.2

-0.2

-0.4

-0.6

-0.8

-1
-6 -5.5 -5 -4.5 -4 -3.5 -3

1
LD3
0.8 LD4
LD5
0.6 LD6

0.4

0.2

-0.2

-0.4

-0.6

-0.8

-1
-40 -35 -30 -25 -20 -15 -10 -5 0
Figure 10.
1 1

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0 0

-0.2 -0.2

-0.4 -0.4
LD3 LD3
LD4 LD4
-0.6 -0.6
LD5 LD5
LD6 LD6
-0.8 -0.8

-1 -1
-1.5 -1 -0.5 0 -5 0 5 10 15 20
10 -6 10 -8

1 1

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0 0

-0.2 -0.2

-0.4 -0.4
LD3 LD3
LD4 LD4
-0.6 -0.6
LD5 LD5
LD6 LD6
-0.8 -0.8

-1 -1
-2 0 2 4 6 8 10 12 14 -0.08 -0.06 -0.04 -0.02 0 0.02 0.04
10 -8

1 1
LD3 LD3
0.8 LD4 0.8 LD4
LD5 LD5
0.6 LD6 0.6 LD6

0.4 0.4

0.2 0.2

0 0

-0.2 -0.2

-0.4 -0.4

-0.6 -0.6

-0.8 -0.8

-1 -1
-0.05 -0.04 -0.03 -0.02 -0.01 0 0.01 -0.01 -0.005 0 0.005 0.01 0.015
Figure 11.
1 1

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0 0

-0.2 -0.2

-0.4 -0.4
LD3 LD3
-0.6 LD4 -0.6 LD4
LD5 LD5
-0.8 LD6 -0.8 LD6

-1 -1
-3 -2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 -1.5 -1 -0.5 0 0.5 1 1.5
10 4 10 4

1 1

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0 0
LD3
-0.2 -0.2 LD4
LD5
-0.4 -0.4 LD6

LD3
-0.6 -0.6
LD4
LD5
-0.8 -0.8
LD6

-1 -1
-14000 -12000 -10000 -8000 -6000 -4000 -2000 0 2000 4000 6000 -2000 -1500 -1000 -500 0 500

1 1

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0 0
LD3
LD4
-0.2 -0.2
LD5 LD3
LD6 LD4
-0.4 -0.4
LD5
LD6
-0.6 -0.6

-0.8 -0.8

-1 -1
-4000 -3500 -3000 -2500 -2000 -1500 -1000 -500 0 500 -1200 -1000 -800 -600 -400 -200 0 200 400
Figure 12.
Figure 13.

1
LD3
0.8 LD4
LD5
0.6 LD6

0.4

0.2

-0.2

-0.4

-0.6

-0.8

-1
-4 -2 0 2 4 6 8 10
10 -11

1
LD3
0.8 LD4
LD5
0.6 LD6

0.4

0.2

-0.2

-0.4

-0.6

-0.8

-1
-5 0 5 10 15 20
10 -12

0.8

0.6

0.4

0.2

-0.2
LD3
LD4
-0.4
LD5
LD6
-0.6

-0.8

-1
7.2 7.25 7.3 7.35 7.4 7.45
10 -11
Figure 14.

0.8

0.6

0.4

0.2

-0.2
LD3
LD4
-0.4
LD5
LD6
-0.6

-0.8

-1
7.4 7.6 7.8 8 8.2 8.4 8.6

0.8

0.6

0.4

0.2

-0.2
LD3
LD4
-0.4
LD5
LD6
-0.6

-0.8

-1
-4.32 -4.31 -4.3 -4.29 -4.28 -4.27 -4.26 -4.25 -4.24 -4.23 -4.22
10 -5
Figure 15.
1 1
LD3 LD3
0.8 LD4 0.8 LD4
LD5 LD5
0.6 LD6 0.6 LD6

0.4 0.4

0.2 0.2

0 0

-0.2 -0.2

-0.4 -0.4

-0.6 -0.6

-0.8 -0.8

-1 -1
-1 -0.9 -0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0 -3 -2.5 -2 -1.5 -1 -0.5 0
10 -7 10 -8

1 1
LD3 LD3
0.8 LD4
0.8
LD4
LD5 LD5
0.6 LD6 0.6
LD6

0.4 0.4

0.2 0.2

0 0

-0.2 -0.2

-0.4 -0.4

-0.6 -0.6

-0.8 -0.8

-1 -1
-10 -8 -6 -4 -2 0 2 -8 -7 -6 -5 -4 -3 -2 -1 0 1
10 -9 10 -8

1 1
LD3
0.8 LD3 0.8 LD4
LD4
LD5
0.6 LD5 0.6 LD6
LD6
0.4 0.4

0.2 0.2

0 0

-0.2 -0.2

-0.4 -0.4

-0.6 -0.6

-0.8 -0.8

-1 -1
-9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 -2 -1.5 -1 -0.5 0 0.5 1
10 -9 10 -9
Figure 16.
1 1

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0 0

-0.2 -0.2
LD3
LD4
-0.4 -0.4
LD5
LD3
LD6
-0.6 -0.6 LD4
LD5
-0.8 -0.8 LD6

-1 -1
-20 -15 -10 -5 0 5 10 15 20 -8 -6 -4 -2 0 2 4

1 1
LD3 LD3
0.8 LD4 0.8 LD4
LD5 LD5
0.6 LD6 0.6 LD6

0.4 0.4

0.2 0.2

0 0

-0.2 -0.2

-0.4 -0.4

-0.6 -0.6

-0.8 -0.8

-1 -1
-1 -0.5 0 0.5 1 1.5 2 -0.1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

1 1
LD3
0.8 LD4
0.8
LD5
0.6 LD6 0.6

0.4 0.4

0.2 0.2

0 0

-0.2 -0.2
LD3
LD4
-0.4 -0.4
LD5
LD6
-0.6 -0.6

-0.8 -0.8

-1 -1
-0.04 -0.03 -0.02 -0.01 0 0.01 0.02 0.03 0.04 -0.07 -0.06 -0.05 -0.04 -0.03 -0.02 -0.01 0 0.01