Mindlin - Implementation - Padhye-UEL
Mindlin - Implementation - Padhye-UEL
Implementation
Page 1 of 97
Nikhil Padhye
Subodh Kalia
2
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Contents
6 Acknowledgments 59
Page 3 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
List of Figures
1 W , L and t represent the width, length and thickness of the plate, respectively. 7
5 UEL subroutine header from Abaqus 2016 documentation, Abaqus User Sub-
11 Plots for reaction forces in Z-direction at the nodes with respect to nodal dis-
placement. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
12 Plots for reaction moments in X-direction at the nodes with respect to nodal
displacements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
displacements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
15 Plate deflection from Abaqus S4 element analysis with a scale factor of 106 . . . 57
Page 4 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Page 5 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
A BSTRACT
In this report, we present a model for plate bending based on the Mindlin plate theory for
small elastic deformations. The related field and constitutive equations, and finite element
discretization (FEM) at an element level are also presented. The implementation of the
Mindlin plate element is carried out for the UEL module of ABAQUS. The overall goal of
this technical report is to facilitate the understanding and implementation of finite element
discretization for the UEL module of Abaqus, so that users can define their own elements. The
overall procedure described in this technical report is expected to aid the implementation
of custom defined user elements for ABAQUS, a topic that is currently not discussed well
in the opinion of the authors. For example, other than official Abaqus documentation [1],
only one comprehensive documentation for UEL in Abaqus is available online [9]. The
current demonstration is carried out for small deformations, for the sake of simplicity, and
applications involving other plate/shell theories (along with large deformation or rotation
effects) will be done in the subsequent reports and publications 1 . Although benchmarking
on simple test problems is performed in this study, detailed analyses on the performance of
current plate formulation such as patch test performance, shear locking issues, etc. are not
discussed. Reader is referred to [3], [7], [4], [2] and [6] for detailed discussions on these topics,
1
The readers are encouraged to contact the authors in case they need any further clarification. Use of this
document or any content/code presented here for any commercial purposes is strictly prohibited. Readers are
welcome to cite this report if it is found useful to them in anyway.
Page 6 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Top surface
z L
t
y
Mid surface
Figure 1: W , L and t represent the width, length and thickness of the plate, respectively.
to qualify as a plate, we require that the thickness (t ) of the structure is small compared
to its length (L) and width (W ), i.e. t << L,W . This assumption simplifies the underlying
governing equilibrium equations and also allows for several simplifications on the kinematics
of deformation due to the “thin" nature of the structure at hand. The Mindlin plate theory
can account for homogeneous through thickness shear deformations (as opposed to another
The two key features of this theory are: (a) we assume by construction that σz = 0, and
(b) since this is a first order shear deformation theory; therefore, allowing σxz and σ y z to be
non-zero. We estimate σxz , σ y z from γxz , γ y z which themselves are taken to be non-zero and
constant through the thickness of the plate. Membrane effects (i.e. in plane stretching of the
plate mid-surface) are not included in this theory. Some other consequences of such ad-hoc
Page 7 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
The deformation in the Mindlin plate can be characterized by the displacement of the mid-
surface of the plate in the Z-direction (denoted by w), and rotation of the normal to the
mid-surface (denoted by θx and θ y ) about X-axis and Y-axis, respectively. Thus we require
three degrees of freedom w, θx , θ y at any point on the mid-surface of the plate to describe the
complete deformation.
Let us consider a material line L 1 L 2 , across thickness of the plate, perpendicular to the
we assume that points on the material line L 1 L 2 still remain on a straight line l 1 l 2 in the
direction with respect to the Y-axis. For small deformations, the displacement of any material
point lying on the line L 1 L 2 , relative to the point ‘O’ on the mid-surface, in the X-direction,
can be approximated as 2 :
u x = zθ y , (1)
∂u x ∂θ y
²x = =z . (2)
∂x ∂x
2
We have assumed no membrane action on the plate, i.e., no in-plane strains on the mid-surface of the
plate, therefore the mid-surface displacement does not vary spatially, nor does it plays any role in strains across
thickness as it shall merely enter as a constant sum to u x in equation 2.
Page 8 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
In the above equation, z indicates the coordinate of a material point along Z-axis in
the undeformed configuration, and for the sake of clarity we repeat that θ y is measured in
anti-clockwise direction about Y-axis when observing from positive direction to the negative
direction.
It is worth highlighting that a rather tacit assumption has been made in this formulation
(while locating a material point along the thickness direction) - regarding no change in
thickness (implying ²z = 0), which seems inconsistent with the prior assumption of σz = 0. In
opinion of the authors, it is quite important to understand the limitations of such classical
plate/shell theories (even with their more elegant extensions as made in the literature), since
they can lead to rather spurious predictions in scenarios where the underlying assumptions
Following a similar procedure one can estimate u y = −zθx . Please note the minus sign in
the estimation of u y , which was not present in the case of u x . Thus, ² y is given as
∂u y ∂θx
²y = = −z (3)
∂y ∂y
∂θx ∂θ y
· ¸
γx y = z − + . (4)
∂x ∂y
The shear strains across the thickness (γxz and γ y z ) can be calculated as
∂u x ∂u z
γxz = + , (5)
∂z ∂x
Page 9 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
N
θy
L1
‘O’
L2
z
Top surface
N
z L1
y ux = z θy
x ‘O’
L2
Figure 2: Rotation of a material line (normal to the neutral plane in the undeformed configuration),
about Y-axis in anti-clockwise direction when viewed from the positive Y-axis, and denoted by θ y .
Page 10 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
∂w(x, y)
γxz = θ y (x, y) + . (6)
∂x
∂w(x, y)
γ y z = −θx (x, y) + . (7)
∂y
Now we define a bending strain vector ²b and shear strain vector ²s as follows:
²x
²b = ² y , (8)
γx y
γxz
²s = . (9)
γy z
1£
²z = σz − ν(σ y + σx ) ,
¤
(10)
E
where E is Young’s modulus and ν is Poisson’s ratio. Under the assumption of plane stress,
i.e. σz = 0, we get
−ν
²z = (σx + σ y ). (11)
E
Page 11 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Similarly, we get
1£
²x = σx − νσ y ,
¤
(12)
E
1£
²y = σ y − νσx .
¤
(13)
E
γx y = Gσx y , (14)
E
where G = 2(1+ν) is the shear modulus. Solving for σx and σ y from equations 12 and 13,
we get
E £
σx = ² ν²
¤
x + y , (15)
1 − ν2
E £
σy = ν²x + ² y .
¤
(16)
1−ν2
Now we assemble all the (x and y) stress and strain components (and refer to them as
bending components):
σx 1 ν 0 ²x
E
σb = σ y = (17)
1−ν ν 1 ²y
2
0
1−ν
σx y 0 0 2
γx y
Page 12 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
σy z = α G γy z (19)
where α is some shear correction factor, and we choose α = 5/6 in this study. Shear stress
components in terms of shear strain components now can be assembled and written as
σxz 1 0 γxz
σs = =G . (20)
σy z γy z
0 1
Thus far we have formulated stress and strain components for both the bending and
shear separately, and this was possible due to several kinematic assumptions and small
strain setting. Next, we present the finite element discretization at an element level such
that Mindlin theory can be simulated. We once again wish to highlight that the current finite
element plate formulation is rather naive, and more effective and consistent proposals have
been made in the literature. Our goal here is to focus on the simplicity so that implementation
We will use a four-noded bilinear isoparametric element where each node has three degrees
of freedom (w, θx and θ y ). Four integration points (as per 2x2 quad integration rule) are used
Page 13 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
1 1 1 1
- , ,
3 3
Node 4 3 3
η
Node 3
(x4, y4) y (-1, 1) (1, 1)
(x3, y3)
z x x
ζ
x x
x
(-1, -1) (1, -1)
Node 1 Node 2 -
1
3
,-
1
3
1
3
,-
1
3
(x1, y1) (x2, y2)
Figure 3: Isoparametric formulation for 4-noded element using 2x2 quad integration.
to capture bending response, and single integration point (reduced integration) is used to
account for shear response. The shape functions for this isoparametric element (with ξ and η
1
N1 = (1 − ξ)(1 − η) , (21)
4
1
N2 = (1 + ξ)(1 − η) , (22)
4
1
N3 = (1 + ξ)(1 + η) , (23)
4
1
N4 = (1 − ξ)(1 + η) . (24)
4
Page 14 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
We denote the shape function vector as N (which has a size of 4x1 and suppress ξ and η
N1
N2
N = . (25)
N
3
N4
∂N1 ∂N1
∂ξ ∂η −(1 − η) −(1 − ξ)
∂N2 ∂N2
1 (1 − η) −(1 + ξ)
∂ξ ∂η
dN = = . (26)
∂N3 ∂N3
4
(1 + η) (1 + ξ)
∂ξ ∂η
∂N4 ∂N4
∂ξ ∂η
−(1 + η) (1 − ξ)
The derivatives of the shape functions (Ni ) with respect to the isoparametric coordinates can
be written as
Page 15 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
∂Ni ∂x ∂y ∂Ni
∂ξ
∂ξ ∂ξ
∂x
= . (29)
∂Ni ∂x ∂y ∂Ni
∂η ∂η ∂η ∂y
∂x ∂y
∂ξ ∂ξ
J = (30)
∂x ∂y
∂η ∂η
where x and y are spatial coordinates of any point within the element. Using equation
∂Ni ∂Ni
∂x
∂ξ
= J −1 (31)
∂Ni ∂Ni
∂y ∂η
Now we can assemble a matrix which contains the spatial derivatives of all the shape
∂N1 ∂N2 ∂N3 ∂N4 ∂N1 ∂N2 ∂N3 ∂N4
∂x ∂x ∂x ∂x ∂ξ ∂ξ ∂ξ ∂ξ
−1
= J = J −1 d N > . (32)
∂N1 ∂N2 ∂N3 ∂N4 ∂N1 ∂N2 ∂N3 ∂N4
∂y ∂y ∂y ∂y ∂η ∂η ∂η ∂η
We see from equation 31, that the derivatives of the shape functions with the x, y spatial
coordinates can be obtained from the derivates of the shape functions with respect to the
Page 16 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
isoparametric coordinates (ξ, η) if the Jacobian is known. Next we derive more explicit
Given the spatial nodal coordinates (x i , y i ), we can use the shape functions (Ni ) to map
4
X
x= Ni x i , (33)
i =1
4
X
y= Ni y i . (34)
i =1
Page 17 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
x1 y1
∂x ∂y ∂N1 ∂N2 ∂N3 ∂N4
∂ξ ∂ξ
∂ξ ∂ξ ∂ξ ∂ξ
x2 y2
J = = . (39)
x y3
3
∂x ∂y ∂N1 ∂N2 ∂N3 ∂N4
∂η ∂η ∂η ∂η ∂η ∂η
.
x4 y4
Using equation 39 we can calculate J from the known quantities on the right hand side.
J = dNT X , (40)
where d N T and X are the first and second matrices on the right hand side of equation
39.
If the nodal point displacements are given then the corresponding displacement field values
within the element can be obtained using the shape functions. The displacement values at
the integration points of the elements allow us to calculate strain and stress values at those
integration points and thereby leading to calculation of stiffness matrix through numerical
integration.
Page 18 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
n
X
w= Ni w i , (41)
i =1
n
θx = Ni θxi ,
X
(42)
i =1
n
θy = Ni θ yi .
X
(43)
i =1
∂ Pn
∂x i =1 Ni (ξ, η)θ yi
²b = z
∂ Pn . (44)
− ∂y i =1 Ni (ξ, η)θxi
∂ ∂
Pn Pn
− ∂x i =1 Ni (ξ, η)θxi + ∂y i =1 Ni (ξ, η)θ yi
Page 19 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
∂N1 θ y1 ∂N2 θ y2 ∂N3 θ y3 ∂N4 θ y4
∂x
+ ∂x
+ ∂x
+ ∂x
²b = z − ∂N∂y
1 θx1
− ∂N∂y
2 θx2
− ∂N∂y
3 θx3
− ∂N∂y
4 θx4
. (45)
∂N1 θ y1 ∂N2 θ y2 ∂N3 θ y3 ∂N4 θ y4
− ∂N∂x
1 θx1
− ∂N∂x
2 θx2
− ∂N∂x
3 θx3
− ∂N∂x
4 θx4
+ ∂y
+ ∂y
+ ∂y
+ ∂y
We can split the matrix on the right hand side of equation 45 as product of a matrix B b
Page 20 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
w1
θx1
θ y1
w2
∂N1 ∂N2 ∂N3 ∂N4
0 0 0 0 0 0 0 0
∂x ∂x ∂x ∂x
θx2
θ y2
²b = z ∂N
0 − ∂N 0 − ∂N 0 − ∂N
3 , (46)
0 − ∂y1 0 2
0 0 4
0
∂y ∂y ∂y
w3
θx3
0 − ∂N ∂N1
0 − ∂N ∂N2
0 − ∂N ∂N3
0 − ∂N ∂N4
1 2 3 4
∂x ∂y ∂x ∂y ∂x ∂y ∂x ∂y
θ y3
w4
θx4
θ y4
i.e.,
²b = z B b d . (47)
∂Ni ∂Ni
The spatial derivatives of the shape functions ∂x and ∂y , in equation 46, can be obtained
from equation 31. A similar procedure can be adopted to compute the components of the
Page 21 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
γxz
Pn ∂ Pn
i =1 Ni θ yi + ∂x i =1 Ni w i
²s = = . (48)
Pn ∂ Pn
γy z i =1 Ni θxi
− + ∂y i =1 Ni w i
∂N1 w 1
∂x
+ N1 θ y1 + ∂N∂x2 w2
+ N2 θ y2 + ∂N∂x3 w3
+ N3 θ y3 + ∂N∂x4 w4
+ N4 θ y4
²s = . (49)
∂N1 w 1 ∂N2 w 2 ∂N3 w 3 ∂N4 w 4
θ θ θ θ
∂y − N 1 x1 + ∂y − N 2 x2 + ∂y − N 3 x3 + ∂y − N 4 x4
The matrix on the right hand side in equation 49 can again be written as product of a
Page 22 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
w1
θx1
θ y1
w2
θx2
∂N1 ∂N2 ∂N3 ∂N4
∂x 0 N1 ∂x 0 N2 ∂x 0 N3 ∂x 0 N4
θ y2
²s =
, (50)
w3
∂N1 ∂N2 ∂N3 ∂N4
∂y
−N1 0 ∂y
−N2 0 ∂y
−N3 0 ∂y
−N4 0
θx3
θ y3
w4
θx4
θ y4
i.e.,
²s = B s d . (51)
∂Ni ∂Ni
As before ∂x and ∂y in equation 50 can be obtained from equation 31.
Page 23 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Since we have considered small elastic deformations only, therefore both K b and K s (stiffness
matrices corresponding to bending and shearing response, respectively) are constant. Non-
linear material behavior or finite deformation may require updating of the element stiffness
At each node we have considered a total three degrees of freedom, one translational
(w in z direction) and two rotational (θx and θ y about X and Y directions, respectively).
Corresponding to these degrees of freedom, at every node we have a work conjugate force
F z and moments M x and M y . Thus, the nodal force vector of an element consists of 12
Page 24 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
F z1
M x1
M y1
F z2
M x2
M y2
F=
.
(52)
F z3
M x3
M y3
F z4
M x4
M y4
To derive the expressions for bending stiffness and shear stiffness we decompose the
internal work done into bending and shear components. Let us consider an element which
has undergone a certain level of bending (²b ) and shear deformation (²s ). In this state, the
associated stress vectors are denoted by σb and σs , the nodal point displacement vector is
virtual displacement of δU , then according to the principle of virtual work, external virtual
Page 25 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Z Z
>
F δU = σ>
b δ²b d v + σ>
s δ²s d v. (53)
V V
δ²b = zB b δU (54)
and,
δ²s = B s δU . (55)
Z Z
>
F δUb = U >
B b> D b> B b δU z 2 d v + U > B s> D s> B s δU d v. (56)
V V
Since above equation holds good for all virtual displacements δU , thus we have
Z Z
>
F = U >
B b> D b> B b z 2 d v + U > B s> D s> B s δU d v. (57)
V V
By taking transpose of both sides in the above equation, and using F = K b U + K s U we can
Page 26 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
arrive at:
Z
Kb = B b> D b B b z 2 d v (58)
V
and
Z
Ks = B s> D s B s d v. (59)
V
Now we shall analyze the computation of both K b and K s . In equation 58 the quantities
within the volume integral are not a function of thickness, except the ‘z term’ itself, and hence
it can be integrated across the thickness from −h/2 to h/2 to yield a result in terms of area
h3
Z
Kb = B b> D b B b d Ω. (60)
12 Ω
In finite element analysis the integration of the stiffness matrix is done numerically, and
since we are using isoparametric elements, we can re-write the integral on the right hand
h3 1 1
Z Z
Kb = B b (ξ, η)> D b B b (ξ, η) |J (ξ, η)| d ξd η. (61)
12 −1 −1
Here, B b and J depend on the integration point coordinates (ξ, η). We shall use 2x2
−1 −1
quad integration rule with four integration points with coordinates of (ξ1 = p ,η = p ),
3 1 3
(ξ2 = p1 , η 2 = −1
p ), (ξ3 = p1 , η 3 = p1 ) and (ξ4 = −1
p ,η = p1 ). The right hand side of the
3 3 3 3 3 4 3
Page 27 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
h3 X 4
Kb = B b (ξi , η i )> D b B b (ξi , η i ) |J (ξi , η i )| ∆ξ∆η, (62)
12 i =1
where ∆ξ = 1 and ∆η = 1. For the sake of simplicity we express the above equation as
4
IPi
X
Kb = (63)
i =1
where,
h3
IPi = B b (ξi , η i )> D b B b (ξi , η i ) |J (ξi , η i )| ∆ξ∆η. (64)
12
Next, the integral on the right hand side of equation 59 for shear stiffness K s can be
written as
Z Z h/2
Ks = B s> D s B s d z d Ω. (65)
Ω −h/2
As all the quantities inside the integral on the right hand side of the above equation are
independent of the thickness coordinate, therefore we can integrate along that direction to
arrive at
Z
Ks = h B s> D s B s d Ω. (66)
Ω
Page 28 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Z 1 Z 1
Ks = h B s (ξ, η)> D s B s (ξ, η) |J (ξ, η)| d ξd η. (67)
−1 −1
In order avoid shear locking, we perform reduced numerical integration for the shear
K = Kb + K s . (69)
In this section we demonstrate through hand calculations how K b and K s are calculated for
an element. Since we are working under the assumption of small strain elasticity, the stiffness
matrix for each element will be the same and constant. E , ν and h are chosen as 2 × 108 , 0.3
and 0.1, respectively throughout this study. The shear correction factor is taken to be 56 .
Page 29 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
The expression for K b was derived in equation 62 based on four integration points. The
( p1 , p1 ) and ( p
−1 p1
, ). We will show the calculation of I P 1 (i = 1 in equation 64) for the first
3 3 3 3
−1 −1
integration point with ξ = p and η = p . Substituting these values in equations 25 and 26 we
3 3
obtain
1 1 1 1
- , ,
3 3
3 3
η
Node 3
(-1, 1) (1, 1)
(x3, y3)
x x
ζ
x x
x
(-1, -1) (1, -1)
Node 2 -
1
3
,-
1
3
1
3
,-
1
3
(x2, y2)
Figure 4: Four-noded element with integration points in isoparametric formulation.
0.622008
0.166666
N = , (70)
0.044658
0.166666
Page 30 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
and
−0.39433 −0.39433
0.39433 −0.10566
dN = . (71)
0.10566 0.10566
−0.10566 0.39433
If we choose the spatial nodal coordinates of the element nodes in X-Y coordinates as
(0, 0), (1, 0), (1, 1) and (0, 1) then we can construct X as
0 0
1 0
X = . (72)
1 1
0 1
0 0
−0.39433 0.39433 0.10566 −0.10566 1 0
J = (73)
−0.39433 −0.10566 0.10566 0.39433
1 1
0 1
Page 31 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
i.e.,
0.5 0
J = . (74)
0 0.5
2 0
J −1 = . (75)
0 2
Spatial derivatives of the all shape functions are denoted in a matrix X Y 0 and with aid of
equation 31 as
∂N1 ∂N1
∂x ∂y
∂N2 ∂N2
∂x ∂y
0 = d N J −> .
XY =
(76)
∂N3 ∂N3
∂x ∂y
∂N4 ∂N4
∂x ∂y
Page 32 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
based on equation 76 as
−0.78867 −0.78867
0.78867 −0.211325
XY 0 = . (77)
0.211325 0.211325
−0.211325 0.78867
Now we can assemble the elements of B b matrix (as defined by equations 46 and 47) using
0 0 −.78867 0 0 .78867 0 0 .211325 0 0 −.211325
Bb =
0 .78867 0 0 .211325 0 0 −.21135 0 0 −.78867 0
.
0 .78867 −.78867 0 −.78867 −.211325 0 −.211325 .21135 0 .211325 .78867
(78)
Now I P 1 (as defined by equation 64) can be calculated with all the known quantities.
according to equation 62. It is worth mentioning that B b and J will vary as a function of (ξ, η)
Page 33 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
0 0 0 0 0 0 0 0 0 0 0 0
0 0.824 −0.298 0 0.091 −0.022 0 −0.412 0.298 0 −0.504 0.022
0 −0.298 0.824 0 0.022 −0.504 0 0.298 −0.412 0 −0.022 0.091
0 0 0 0 0 0 0 0 0 0 0 0
0 0.091 0.022 0 0.824 0.298 0 −0.504 −0.022 0 −0.412 −0.298
0 −0.022 −0.504 0 0.298 0.824 0 0.022 0.091 0 −0.298 −0.412
× 104
Kb =
(79)
0 0 0 0 0 0 0 0 0 0 0 0
0 −0.412 0.298 0 −0.504 0.022 0 0.824 −0.298 0 0.091 −0.022
0 0.298 −0.412 0 −0.022 0.091 0 −0.298 0.824 0 0.022 −0.504
0 0 0 0 0 0 0 0 0 0 0 0
0 −0.504 −0.022 0 −0.412 −0.298 0 0.091 0.022 0 0.824 0.298
0 0.022 0.091 0 −0.298 −0.412 0 −0.022 −0.504 0 0.298 0.824
For calculation of stiffness associated with shear deformation, we use the reduced integration
approach, i.e., only one integration point with ξ = 0 and η = 0 is chosen. According to
0.25
0.25
N = , (80)
0.25
0.25
−0.25 −0.25
0.25 −0.25
dN = . (81)
0.25 0.25
−0.25 0.25
Matrix of nodal coordinates X is chosen same as in equation 72, and the Jacobian J is
Page 34 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
0 0
−0.25 0.25 0.25 −0.25 1 0
J = , (82)
−0.25 −0.25 0.25 0.25 1 1
0 1
0.5 0
J = . (83)
0 0.5
J −1 turns out to be
2 0
J −1 = . (84)
0 2
to equation 76 as
−0.5 −0.5
0.5 −0.5
XY 0 = . (85)
0.5 0.5
−0.5 0.5
Using the components of X Y 0 from equation 85, we can now construct the B s matrix
Page 35 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
−0.5 0 0.25 0.5 0 0.25 0.5 0 0.25 −0.5 0 0.25
Bs = .
−0.5 −0.25 0 −0.5 −0.25 0 0.5 −0.25 0 0.5 −0.25 0
Finally, the shear stiffness matrix K s can be calculated from equation 68 and we get
0.385 0.096 −0.096 0 0.096 −0.096 −0.385 0.096 −0.096 0 0.096 −0.096
0.096 0.048 0 0.096 0.048 0 −0.096 0.048 0 −0.096 0.048 0
−0.096 0 0.048 0.096 0 0.048 0.096 0 0.048 −0.096 0 0.048
0 0.096 0.096 0.385 0.096 0.096 0 0.096 0.096 −0.385 0.096 0.096
0.096 0.048 0 0.096 0.048 0 −0.096 0.048 0 −0.096 0.048 0
−0.096 0 0.048 0.096 0 0.048 0.096 0 0.048 −0.096 0 0.048
× 107 .
Ks =
(86)
−0.385 −0.096 0.096 0 −0.096 0.096 0.385 −0.096 0.096 0 −0.096 0.096
0.096 0.048 0 0.096 0.048 0 −0.096 0.048 0 −0.096 0.048 0
−0.096 0 0.048 0.096 0 0.048 0.096 0 0.048 −0.096 0 0.048
0 −0.096 −0.096 −0.385 −0.096 −0.096 0 −0.096 −0.096 0.385 −0.096 −0.096
0.096 0.048 0 0.096 0.048 0 −0.096 0.048 0 −0.096 0.048 0
−0.096 0 0.048 0.096 0 0.048 0.096 0 0.048 −0.096 0 0.048
The total stiffness matrix for the element is then obtained by summing the bending and
0.385 0.096 −0.096 0 0.096 −0.096 −0.385 0.096 −0.096 0 0.096 −0.096
0.096 0.048 −0.298 0.096 0.048 −0.022 −0.096 0.047 0.298 −0.096 0.047 0.022
−0.096 −0.298 0.048 0.096 0.022 0.047 0.096 0.298 0.047 −0.096 −0.022 0.048
0 0.096 0.096 0.385 0.096 0.096 0 0.096 0.096 −0.385 0.096 0.096
0.096 0.048 0.022 0.096 0.048 0.298 −0.096 0.047 −0.022 −0.096 0.047 −0.298
−0.096 −0.022 0.047 0.096 0.298 0.048 0.096 0.022 0.048 −0.096 −0.298 0.047
× 107 .
K =
(87)
−0.385 −0.096 0.096 0 −0.096 0.096 0.385 −0.096 0.096 0 −0.096 0.096
0.096 0.047 0.298 0.096 0.047 0.022 −0.096 0.048 −0.298 −0.096 0.048 −0.022
−0.096 0.298 0.047 0.096 −0.022 0.048 0.096 −0.298 0.048 −0.096 0.022 0.047
0 −0.096 −0.096 −0.385 −0.096 −0.096 0 −0.096 −0.096 0.385 −0.096 −0.096
0.096 0.047 −0.022 0.096 0.047 −0.298 −0.096 0.048 0.022 −0.096 0.048 0.298
−0.096 0.022 0.048 0.096 −0.298 0.047 0.096 −0.022 0.047 −0.096 0.298 0.048
Page 36 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Here we discuss consequences of the assumptions made in the Mindlin theory and associated
First we assumed that transverse normal stress σz = 0, by construction, through the plate
thickness. On the other hand we allowed for non-zero transverse shear stresses through
thickness. To investigate this let us study the equilibrium equation in the Z direction:
∂σzx ∂σz y
+ = 0. (89)
∂x ∂y
Substitute σzx = Gγzx and σz y = Gγz y in equation 89, where G is the shear modulus, to
arrive at
∂γzx ∂γz y
+ = 0. (90)
∂x ∂y
∂w
γzx = + θy ,
∂x
and
Page 37 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
∂w
γz y = − θx .
∂y
∂ ∂w ∂ ∂w
· ¸ · ¸
+ θy + − θx = 0 (91)
∂x ∂x ∂y ∂y
∂2 w ∂θ y ∂2 w ∂θx
+ + − = 0. (92)
∂x 2 ∂x ∂y 2 ∂y
∂2 w
For small deformation, we can approximate ∂x 2
= K x (curvature in X-direction) and
∂2 w
∂y 2
= K y (curvature in Y-direction), and therefore the equation 92 becomes
∂θx ∂θ y
Kx + K y = − . (93)
∂y ∂x
∂θx ²y ∂θ y ²x
=− , = . (94)
∂y z ∂x z
The above condition may not be exactly satisfied in general throughout the thickness and
Page 38 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
As mentioned in the beginning, the Mindlin theory does not take into account any mem-
brane strains, i.e. if we assume a case of homogeneous biaxial straining, with ²x = ² y = δ > 0,
then at z = 0 equation 95 is not satisfied. Lastly, we had assumed σz = 0, but this in turn
Page 39 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
UEL is a programming interface provided with Abaqus Standard using which one can define
regards to the UEL. In order to work with UEL, one must start by using the default UEL
header from the Abaqus User Subroutines Reference Guide, Section 1.1.28 in Abaqus 2016
documentation, as shown in Figure 5. The user specified program is contained under this
header and the end of the UEL is marked by the ‘END’ syntax.
Figure 5: UEL subroutine header from Abaqus 2016 documentation, Abaqus User Subroutines Refer-
ence Guide, Section 1.1.28.
Conceptually it is important to understand the overall operation of the ‘UEL’ block. Figure
6 shows a schematic describing the ‘input’ and ‘output’ flow from the UEL subroutine, that is
Page 40 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
COORDS RHS
Unew
dUold UEL AMATRX
State variables Updated state variables
called for every element. For a current time increment, Abaqus provides the incremental and
final nodal point displacements. State variables at the start of that time increment are also
provided. UEL is then required to return the updated element stiffness matrix, internal force
and state variables all at the end of the current time step. Next we discuss in greater detail,
the incoming quantifies provided by Abaqus into the UEL and associated outgoing quantities
the degree of freedom values at each node at the end of the current time increment. This is
incremental degree of freedom values for all the nodes in the previous increment. This is
Page 41 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
COORD I N AT E S needs to be defined in the UEL section of the Abaqus input file which sets
SV ARS: is a vector which stores the state variables of an element and needs to be passed
back to Abaqus after being updated. Typically stress and strains at all integration points
AM AT R X : is the stiffness matrix for an element. For our case, small deformation and linear
elasticity, the stiffness matrix is constant, however, if there are non-linear effects then the
R H S: is the internal force vector for the element which needs to be returned at the end of
the current increment. Since, we are solving a linear elasticity problem, for simplicity, the
R H S = −AM AT R X ×U (96)
The overall implementation of the UEL is summarized in form of a pseudo code as shown in
Algorithm 1. The pseudo code is based on the theory presented earlier and is self-explanatory.
Page 42 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
We also provide the actual UEL written in Fortran corresponding to this pseudo code subse-
quently.
Page 43 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
1 CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC
2 CCCCCCC THIS ROUTINE IS WRITTEN BY SUBODH KALIA, NIKHIL PADHYE CCCCCCCCC
3 CCCCCCC FOR ABAQUS 2016. THE DISTRIBUTION OF THIS CODE WITHOUT PRIOR CCC
4 CCCCCCC PERMISSION OR ANY KIND OF COMMERICAL USE IS STRICTLY CCCCCCCCCCC
5 CCCCCCC PROHIBITED. PLEASE CITE THE WORK IF YOU WANT TO USE IT FOR CCCCC
6 CCCCCCC YOUR RESEARCH. CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC
7 CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC
8 SUBROUTINE UEL(RHS,AMATRX,SVARS,ENERGY,NDOFEL,NRHS,NSVARS,
9 1 PROPS,NPROPS,COORDS,MCRD,NNODE,U,DU,V,A,JTYPE,TIME,DTIME,
10 2 KSTEP,KINC,JELEM,PARAMS,NDLOAD,JDLTYP,ADLMAG,PREDEF,NPREDF,
11 3 LFLAGS,MLVARX,DDLMAG,MDLOAD,PNEWDT,JPROPS,NJPROP,PERIOD)
12 C
13 INCLUDE 'ABA_PARAM.INC'
14 C
15 DIMENSION RHS(MLVARX,*),AMATRX(NDOFEL,NDOFEL),PROPS(*),
16 1 SVARS(*),ENERGY(8),COORDS(MCRD,NNODE),U(NDOFEL),
17 2 DU(MLVARX,*),V(NDOFEL),A(NDOFEL),TIME(2),PARAMS(*),
18 3 JDLTYP(MDLOAD,*),ADLMAG(MDLOAD,*),DDLMAG(MDLOAD,*),
19 4 PREDEF(2,NPREDF,NNODE),LFLAGS(*),JPROPS(*)
20
21 Double precision, Dimension(12, 1) :: FORCE, V_TEMP
22 Double precision, Dimension(12, 1) :: U_CONV, DU_CONV
23 Double precision, Dimension(4, 1) :: N
24 Double precision, Dimension(4, 2) :: DN, IP_COORDS, COORDS_T
25 Double precision, Dimension(4, 2) :: XY_DERIVATIVES
26 Double precision, Dimension(2, 4) :: XY_DERIVATIVES_T
27 Double precision, Dimension(2, 4) :: DN_TRANSPOSE
28 Double precision, Dimension(2, 2) :: J, INV_J
29 Double precision, Dimension(3, 12) :: M_B_MAT_BENDING, M_TEMP
30 Double precision, Dimension(12, 12) :: M_STIFFNESS, M_TEMP_STIFF
31 Double precision, Dimension(12, 12) :: M_STIFF
32 Double precision, Dimension(12, 3) :: M_B_MAT_BENDING_TRANSPOSE
33 Double precision, Dimension(3, 1) :: V_STRAIN_INCREMENT_BENDING
34 Double precision, Dimension(3, 1) :: V_STRESS_INCREMENT_BENDING
35 Double precision, Dimension(3, 3) :: M_D_BENDING
36 Double precision, Dimension(2, 12) :: M_B_MAT_SHEAR
37 Double precision, Dimension(2, 1) :: V_STRAIN_INCREMENT_SHEAR
38 Double precision, Dimension(2, 1) :: V_STRESS_INCREMENT_SHEAR
39 Double precision, Dimension(12, 2) :: M_B_MAT_SHEAR_TRANSPOSE
40 Double precision, Dimension(2, 2) :: M_D_SHEAR
41
42 Double precision :: CHI, ETA, S_E, S_NU, S_INERTIA, S_THICKNESS
43 Double precision :: S_J_DETERMINANT, S_KAPPA
44
45 INTEGER :: NDOFEL, NSVARS, NPROPS, MCRD, NNODE, K1, K2, K3, NUM_IP
46 INTEGER :: NDOF, NUM_SVAR_PER_IP, POSITION_POINTER
47
48 PARAMETER(NUM_IP=4, NUM_SVAR_PER_IP=10)
49
50 COORDS_T(1,1) = COORDS(1,1)
51 COORDS_T(1,2) = COORDS(2,1)
52
53 COORDS_T(2,1) = COORDS(1,2)
54 COORDS_T(2,2) = COORDS(2,2)
55
56 COORDS_T(3,1) = COORDS(1,3)
57 COORDS_T(3,2) = COORDS(2,3)
58
59 COORDS_T(4,1) = COORDS(1,4)
60 COORDS_T(4,2) = COORDS(2,4)
61
62 DO K1 = 1, 12
63 U_CONV(K1, 1) = U(K1)
64 DU_CONV(K1, 1) = DU(K1,1)
65 END DO
66
67 IF (LFLAGS(3).EQ.4) THEN
68 DO K1=1, 12
69 DO K2=1, 12
Page 45 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Page 46 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Page 47 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Page 48 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
277 C EPSILON_YZ
278 SVARS(POSITION_POINTER+9) = SVARS(POSITION_POINTER+9)+
279 1 V_STRAIN_INCREMENT_SHEAR(2, 1)
280
281 END DO
282
283 CCCCCCCCCCCCCCCCCCCCCCCC CALCULATE FORCE CCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC
284 CALL GENERIC_MATRIX_MULTIPLY(M_STIFFNESS, 12, 12,
285 1 U_CONV, 12, 1, FORCE)
286
287 CCCCCCCCCCCCCCCCCCCCCCCCCCC RETURN TO ABAQUS CCCCCCCCCCCCCCCCCCCCCCCCCCC
288 DO K1 = 1, 12
289 RHS(K1, 1) = -FORCE(K1, 1)
290 END DO
291
292 DO K1 = 1, 12
293 DO K2 = 1, 12
294 AMATRX(K1, K2) = M_STIFFNESS(K1, K2)
295 END DO
296 END DO
297
298 999 CONTINUE
299
300
301 RETURN
302 END
Page 49 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
E LEMENT
In this section, we compare the results of plate bending from the Abaqus with the default S4
element, Matlab implementation of Mindlin plate from [5], and our developed 4-noded UEL
subroutine.
A single element of length and width of 2 units each and a thickness of 0.1 units is
considered. Boundary conditions for this element are shown in Figure 7, where we have
displacement of 0.1 in the positive Z-direction. Thus, this is a well defined displacement
Page 50 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
driven boundary value problem for which we seek to find the solution. It is important to note
that Abaqus S4 element has a total of 6 degrees of freedom at each node, therefore we simply
Figure 8, 9 and 10 shows the reaction forces (RF3) in Z-direction, and reaction moments
RM1 and RM2 in X- and Y-directions, respectively from Abaqus S4 element analysis.
Results from the referenced Matlab code, ABAQUS UEL implementation of Mindlin plate,
and Abaqus S4 element analysis were gathered and compared for the nodal point reaction
forces and moments and shown in Figures 11, 12 and 13. Our UEL implementation yields
results exactly similar to those predicted by the referenced Matlab code. In general there
is a good agreement with Abaqus S4 element analysis, however, there is deviation in the
reaction forces at nodes 2 and 3. A detailed investigation reveals that reaction moments for
the UEL implementation and Matlab code were same in magnitude at four nodes, whereas
Page 51 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Page 52 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
in case of Abaqus S4 element the reaction moments in both directions varied within the
element. A simple explanation for this observation stems from the fact that in the Mindlin
plate element (UEL or Matlab implementation) the transverse shear strains are assumed to
be constant through the thickness, and interpolated based on the nodal point rotations and
displacements. The bending strains are interpolated based on nodal point rotations only.
In the current case, we have set the rotations and displacements at all the nodes equal to
zero except for the imposed displacement in the vertical direction at node 4. This causes the
vanishing of the bending strains all together and linearly varying (constant through thickness)
transverse shear strains. Thus the reaction moments at the nodes in the Mindlin element
appear to simply maintain the global moment-equilibrium and have the same magnitude in
both X- and Y-directions, at all nodes, due to symmetric shape of the element. On the other
hand the S4 element of Abaqus is more sophisticated and picks up the variation of bending
Exact matching of the results between our UEL implementation and referenced Matlab
code, and overall consistency with respect to the Abaqus S4 element (foregoing certain
Page 53 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
0 10,000
Abaqus S4 element
Reaction force in z-direction at node
2
-200,000 -30,000
-250,000 -40,000
-300,000 -50,000
-350,000 -60,000
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
Displacement Displacement
(a) (b)
10,000 400,000
Abaqus S4 element
Reaction force in z-direction at node
4
-30,000
150,000
-40,000
100,000
-50,000 50,000
-60,000 0
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
Displacement Displacement
(c) (d)
Figure 11: Plots for reaction forces in Z-direction at the nodes with respect to nodal displacement.
0 0
Abaqus S4 element Abaqus S4 element
Reaction moment in x-direction at
Reaction moment in x-direction at
-80,000
-100,000
-100,000
-120,000
-120,000
-140,000
-140,000 -160,000
-160,000 -180,000
-180,000 -200,000
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
Displacement Displacement
(a) (b)
0 0
Abaqus S4 element Abaqus S4 element
Reaction moment in x-direction at
Reaction moment in x-direction at
node 4
-80,000
-100,000
-100,000
-120,000
-120,000
-140,000
-140,000 -160,000
-160,000 -180,000
-180,000 -200,000
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
Displacement Displacement
(c) (d)
Figure 12: Plots for reaction moments in X-direction at the nodes with respect to nodal displacements.
Page 54 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
180,000 180,000
Abaqus S4 element Abaqus S4 element
120,000 120,000
node 2
node 1
100,000 100,000
80,000 80,000
60,000 60,000
40,000 40,000
20,000 20,000
0 0
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
Displacement Displacement
(a) (b)
200,000 200,000
Abaqus S4 element Abaqus S4 element
Reaction moment in y-direction at
180,000 180,000
UEL 4-noded element UEL 4-noded element
160,000 MATLAB 4-noded element 160,000 MATLAB 4-noded element
140,000 140,000
120,000 120,000
node 3
node 4
100,000 100,000
80,000 80,000
60,000 60,000
40,000 40,000
20,000 20,000
0 0
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
Displacement Displacement
(c) (d)
Figure 13: Plots of reaction moments in Y-direction at the nodes with respect to nodal displacements.
Page 55 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Now we shall compare the results from Abaqus, our UEL implementation and the Matlab
code on a laterally loaded plate, which is clamped at the edges, and made up of multiple
elements. 400 elements in a (20 × 20) grid along X and Y directions are employed. The overall
dimensions of the plate in X-, Y- and Z-directions are again taken as 2 units, 2 units and 0.1
units, respectively. Maximum lateral pressure of 1 unit per unit area is applied3 .
central node. This is shown in Figure 14. With our implemented UEL subroutine,a similar
in Z-direction, Figure 15. For Abaqus and UEL results, we use a scale factor of 106 to give a
consistent visual comparison with Matlab generated displacements. These results reveal an
excellent agreement.
vary the lateral pressure to 1 unit. The obtained results confirm that our implementation is
3
This is achieved by applying an equivalent level of concentrated force of 0.01 units over all the element
nodes.
Page 56 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
#10 -6
0
-0.2
-0.4
Z-axis
-0.6
-0.8
-1
-1.2
0
0.5 2
1 1.5
1
1.5
0.5
X-axis 2 0 Y-axis
Figure 15: Plate deflection from Abaqus S4 element analysis with a scale factor of 106 .
Page 57 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Figure 16: Plate deflection from UEL with a scale factor of 106 .
Abaqus S4 element
-1E-07
UEL 4-noded element
Maximum displacement
-5E-07
-7E-07
-9E-07
-1.1E-06
-1.3E-06
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Lateral pressure
Figure 17: Comparison of deflections for a laterally loaded plate clamped at the edges.
Page 58 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
6 A CKNOWLEDGMENTS
Nikhil Padhye appreciates past technical discussions with his advisor Professor David M.
Parks for motivating the usage of Abaqus in finite element studies during the course of his
Page 59 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
R EFERENCES
[1] Ver ABAQUS. 6.14 documentation. Dassault Systemes Simulia Corporation, 2014.
[2] F Auricchio and RL Taylor. A shear deformable plate element with an exact thin limit.
[3] Klaus-Jürgen Bathe and Eduardo N Dvorkin. A four-node plate bending element based
[4] Jean-Louis Batoz and Mabrouk Ben Tahar. Evaluation of a new quadrilateral thin
18(11):1655–1677, 1982.
[5] Ferreira. MATLAB Codes for Finite Element Analysis - Solids and Structures. Solid mechan-
[6] Bo Häggblad and Klaus-Jürgen Bathe. Specifications of boundary conditions for reiss-
ner/mindlin plate bending finite elements. International Journal for Numerical Methods
[7] Thomas JR Hughes and TEi Tezduyar. Finite elements based upon mindlin plate theory
[8] Nikhil Padhye. Sub-Tg, solid-state, plasticity-induced bonding of polymeric films and
Page 60 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
[9] Erik Saether and Alexander Tessler. User-defined subroutine for implementation of
higher-order shell element in abaqus. Technical report, ARMY RESEARCH LAB ADELPHI
MD, 1993.
Page 61 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Page 62 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Page 63 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Page 64 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
70
71
72 stiffness*U
Page 65 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Page 66 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Page 67 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Page 68 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Page 69 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Page 70 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Page 71 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Page 72 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Page 73 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Page 74 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Page 75 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Page 76 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Page 77 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Page 78 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
898 ** LOADS
899 **
900 ** Name: Load-1 Type: Concentrated force
901 *Cload
902 all_nodes, 3, -0.01
903 **
904 ** OUTPUT REQUESTS
905 **
906 *Restart, write, frequency=0
907 **
908 ** FIELD OUTPUT: F-Output-2
909 **
910 *Output, field
911 *Node Output
912 CF, RF, RM, RT, TF, VF
913 *Element Output, directions=YES
914 BF, CENTMAG, CENTRIFMAG, CORIOMAG, ESF1, GRAV, HP, NFORC, NFORCSO, P, ROTAMAG, SF,
TRNOR, TRSHR
915 **
916 ** FIELD OUTPUT: F-Output-1
917 **
918 *Output, field, variable=PRESELECT
919 **
920 ** HISTORY OUTPUT: H-Output-2
921 **
922 *Output, history
923 *Element Output
924 IRF1, IRF2, IRF3, IRM1, IRM2, IRM3
925 **
926 ** HISTORY OUTPUT: H-Output-1
927 **
928 *Output, history, variable=PRESELECT
929 *End Step
930
Page 79 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Page 80 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
69 _PickedSet13, 4, 4
70 _PickedSet13, 5, 5
71 **
72 ** OUTPUT REQUESTS
73 **
74 *Restart, write, frequency=0
75 **
76 ** FIELD OUTPUT: F-Output-2
77 **
78 *Output, field
79 *Node Output
80 CF, RF, RM, RT, TF, VF
81 *Element Output, directions=YES
82 BF, CENTMAG, CENTRIFMAG, CORIOMAG, ESF1, GRAV, HP, NFORC, NFORCSO, P, ROTAMAG, SF,
TRNOR, TRSHR
83 **
84 ** FIELD OUTPUT: F-Output-1
85 **
86 *Output, field, variable=PRESELECT
87 **
88 ** HISTORY OUTPUT: H-Output-2
89 **
90 *Output, history
91 *Element Output
92 IRF1, IRF2, IRF3, IRM1, IRM2, IRM3
93 **
94 ** HISTORY OUTPUT: H-Output-1
95 **
96 *Output, history, variable=PRESELECT
97 *End Step
98 *STEP
99 *MATRIX GENERATE, STIFFNESS
100 *MATRIX OUTPUT, STIFFNESS, FORMAT=COORDINATE
101 *END STEP
102
Page 81 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Page 82 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
70 f=[jj; displacements'];
71 fprintf('node U\n')
72 fprintf('%3d %12.8f\n',f)
73
74 % deformed shape
75 figure
76 plot3(xx,yy,displacements(1:numberNodes),'.')
77 xlabel('X-axis') % x-axis label
78 ylabel('Y-axis') % y-axis label
79 zlabel('Z-axis') % y-axis label
80 format long
81 D1=E*thickness^3/12/(1-poisson^2);
82 min(displacements(1:numberNodes))
Page 83 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Page 84 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Page 85 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Page 86 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Page 87 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Page 88 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Page 89 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Page 90 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Page 91 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Page 92 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Page 93 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Page 94 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Page 95 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Page 96 of 97
Nikhil Padhye, Subodh Kalia Implementation of Mindlin plate element
Page 97 of 97