See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/319179248

MATLAB Code for Structural Analysis of 2-D Structures Subjected to Static and
Self-Weight Loading Conditions

Article · January 2017

CITATIONS READS

2 1,240

1 author:

Naman Jain
G. B. Pant University of Agriculture and Technology, Pantnagar
22 PUBLICATIONS   10 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

master thesis View project

P.hD. Thesis View project

All content following this page was uploaded by Naman Jain on 19 August 2017.

The user has requested enhancement of the downloaded file.

 International Journal of Trend in Research and Development, Volume 4(1), ISSN: 2394-9333
www.ijtrd.com
MATLAB Code for Structural Analysis of 2-D
Structures Subjected to Static and Self-Weight Loading
Conditions
Naman Jain
Department Of Mechanical Engineering, G. B. Pant University of Agriculture and Technology, Pantnagar, India

Abstract: In present investigation MATLAB code for structural functions is to use a reference element, which is a square of side 2
analysis of 2-dimension linear elastic isotropic structures (units) as represented in Figure 1.
subjected to static and self-weight loading conditions is been
presented. In this paper implementation of MATLAB code on
two structures i.e. Cantilever beam and Bolkow Blohm Beam is
done to show the application of code. The meshing of the 2-
dimension structure is done with quadrilateral 4-node element, for
analysis of self-weight loading condition the weight of an element
is equally transferred on each node in downward direction and
results obtained through MATLAB code is been compared with
ANSYS software.
Keywords— Finite element method; self-weight; MATLAB; 4-
node element
I. INTRODUCTION
Figure 1: Geometrical transformation of 4-node element
The finite element method (Hutton 2004; Chandrupatla and (Khennane 2013)
Belegunda 2004; Khennane 2013) represent is one of the most
significant achievements in the field of computational methods in To define the geometrical transformation, we will assume that the
the last century. Historically, it has its roots in the analysis of coordinates (x, y) of an arbitrary point of the parent element are
weight-critical framed aerospace structures. These framed the unknown functions defined over the domain represented by
structures were treated as an assemblage of one-dimensional the reference element in its local coordinate system (ξ, η). Notice
members, for which the exact solutions to the differential that both the variables x and y belong to the linear class of
equations for each member were well known. These solutions functions since they are continuous and their first derivatives are
were cast in the form of a matrix relationship between the forces constant equal to 1. Therefore, we start by constructing a general
and displacements at the ends of the member. Hence, the method approximation for x in terms of ξ and η.
was initially termed matrix analysis of structures. Later, it was
extended to include the analysis of continuum structures. Since x = α1 + α2ξ + α3η + α4ξη (1)
continuum structures have complex geometries, they had to be
α1
subdivided into simple components or “elements” interconnected
α2
at nodes. It was at this stage in the development of the method 𝑥 = [1 𝜉 𝜂 𝜉𝜂] α (2)
3
that the term “finite element” appeared. However, unlike framed α4
structures, closed form solutions to the differential equations
governing the behavior of continuum elements were not Then, we will transform the general approximation, equation (2),
available. Energy principles such as the theorem of virtual work to a nodal approximation by using the nodal values x1, x2, x3, and
or the principle of minimum potential energy, which were well x4 respectively at nodes 1, 2, 3, and 4. Notice also that the couple
known, combined with a piece-wise polynomial interpolation of (ξ, η) takes on the values of (−1, −1), (1, −1), (1, 1), and (−1, 1)
the unknown displacement, were used to establish the matrix respectively at nodes 1, 2, 3, and 4. It follows
relationship between the forces and the interpolated x1 = α1 − α2 − α3 + α4
displacements at the nodes numerically. In the late 1960s, when x2 = α1 + α2 − α3 − α4
the method was recognized as being equivalent to a minimization x3 = α1 + α2 + α3 + α4
process, it was reformulated in the form of weighted residuals and
variational calculus and expanded to the simulation of x4 = α1 − α2 + α3 − α4
nonstructural problems in fluids, thermomechanics and
which, when rewritten in a matrix form, yields
electromagnetics.
𝑥1 1 −1 −1 1 α1
II. LINEAR QUADRILATERAL 4-NODE ELEMENT 𝑥2 α2
1 1 −1 −1
In the quadrilateral family of elements, except for the square or 𝑥3 = 1 1 1 1 α3 (3)
the rectangle, it is impossible to construct the shape functions 𝑥4 1 −1 −1 −1 α4
directly in terms of x and y. The only way to construct these or in a more compact form as

IJTRD | Jan-Feb 2017

Available [email protected] 316
 International Journal of Trend in Research and Development, Volume 4(1), ISSN: 2394-9333
www.ijtrd.com
𝑋 = 𝐴 α (4) 𝜕
0
𝜀𝑥𝑥 𝜕𝑥
The parameters αi can be obtained easily by solving the equation 𝜀𝑦𝑦 = 0 𝜕 𝑢
(15)
(3). It can be noticed that the columns of the matrix [A] are 𝜀𝑥𝑦
𝜕𝑦 𝑣
𝜕 𝜕
actually orthogonal vectors of norm 4. Hence, the inverse of the
𝜕𝑦 𝜕𝑥
matrix [A] is obtained as
or in a more compact form as
1 −1 −1 1
𝜀 = 𝐿𝑈 (16)
𝐴 −1
= 1/4 𝐴 𝑇
= 1/4 1 1 −1 −1 (5)
1 1 1 1 Where [L] is a linear differential operator. Let us consider a finite
1 −1 −1 −1 element approximation for the unknown functions u and v. For an
and the parameters αi as element having n nodes, the unknown displacements are
interpolated using nodal approximations as
α1 1 −1 −1 1 𝑥1
α2 1 1 −1 −1 𝑥2 u = N1u1 + N2u2 +· · ·+Nnun
α3 = 1/4 1 1 1 1 𝑥3 (6) v = N1v1 + N2v2 +· · ·+Nnvn
α4 1 −1 −1 −1 𝑥4
which, when written in a matrix form, yields
Substituting for the parameters αi in equation (2) yields 𝑢1
𝑣1
1 −1 −1 1 𝑥1 𝑢2
1 1 −1 −1 𝑥2 𝑢 𝑁 0 𝑁2 0 . . . 𝑁𝑛 0 𝑣2
𝑥(ξ, η) = 1/4[1 𝜉 𝜂 𝜉𝜂] = 1 .. (17)
1 1 1 1 𝑥3 (7) 𝑣 0 𝑁1 0 𝑁2 . . . 0 𝑁𝑛
.
1 −1 −1 −1 𝑥4
𝑢𝑛
Expanding and rearranging equation (7) leads to 𝑣𝑛
or simply as
x(ξ, η) = N1(ξ, η)x1 + N2(ξ, η)x2 + N3(ξ, η)x3 + N4(ξ, η)x4 (8)
N1(ξ, η) = 0.25(1 − ξ − η + ξη) {U} = [N]{a} (18)
N2(ξ, η) = 0.25(1 + ξ − η − ξη) with {a} = {u1, v1, u2, v2, . . . , un, vn} being the vector of nodal
N3(ξ, η) = 0.25(1 + ξ + η + ξη) displacements. The number and the form of the shape functions
N4(ξ, η) = 0.25(1 − ξ + η − ξη) depend on the element used.
Following exactly the same process for the variable y, we obtain Substituting for {U} using equation (15), the strain displacement
y(ξ, η) = N1(ξ, η)y1 + N2(ξ, η)y2 + N3(ξ, η)y3 + N4(ξ, η)y4 (9) equation (16) become
Expressions (3) and (5) represent well and truly a linear {ε} = [B]{a} (19)
geometrical transformation. The center of the reference square is where
given by (ξ, η) = (0, 0). 𝜕𝑁1 𝜕 𝑁2
. . . 𝜕𝑁𝑛
0 0 0
𝜕 𝜕 𝜕
III. FINITE ELEMENT FORMULATION FOR PLANE 𝜕𝑁1 𝜕𝑁2 𝜕𝑁𝑛
𝐵 = 0 0 . . . 0
STRESS PROBLEMS 𝜕 𝜕 𝜕
𝜕𝑁1 𝜕𝑁1 𝜕𝑁2 𝜕𝑁2 𝜕𝑁𝑛 𝜕𝑁𝑛
The stress–strain relationships for plane stress given as . . .
𝜕 𝜕 𝜕 𝜕 𝜕 𝜕
1 0 (20)
𝜎𝑥𝑥 µ 𝜀𝑥𝑥
1 0 The matrix [B] is called the strain matrix; it relates the nodal
𝜎𝑦𝑦 = 𝐸 µ 𝜀𝑦𝑦 (10)
1−µ2 (1−µ) displacements to the strains. It is formed by the partial derivatives
𝜎𝑥𝑦 0 0 𝜀𝑥𝑦
2 of the shape functions Ni(x, y).
𝜎 = 𝐷 𝜀 (11)
IV. DISPLACEMENT FIELD
Whether it is a state of plane stress or plane strain, a material
The displacement field over the element is approximated as
point can only move in the directions x and y. Therefore, the two
displacement variables that play a role are u(x, y) and v(x,y). The u = N1u1 + N2u2 + N3u3 + N4u4
infinitesimal strain displacements relations for both theories are
v = N1v1 + N2v2 + N3v3 + N4v4
the same and they are given as
𝜕𝑢
or in a matrix form as
𝜀𝑥𝑥 = (12)
𝜕𝑥 𝑢1
𝜕𝑣
𝜀𝑦𝑦 = (13) 𝑣1
𝜕𝑦
𝜕𝑢 𝜕𝑣 𝑢2
𝜀𝑥𝑦 = + (14) 𝑢 𝑁 0 𝑁2 0 𝑁3 0 𝑁4 0 𝑣2
𝜕𝑦 𝜕𝑥
= 1 𝑢3 (21)
𝑣 0 𝑁1 0 𝑁2 0 𝑁3 0 𝑁4
These relations can be written in a matrix form as 𝑣3
𝑢4
𝑣4

IJTRD | Jan-Feb 2017

Available [email protected] 317
 International Journal of Trend in Research and Development, Volume 4(1), ISSN: 2394-9333
www.ijtrd.com
To evaluate the matrix [B], it is necessary to relate the partial
derivatives in the (x, y) coordinates to the local coordinates (ξ, η).
The derivative of the shape functions can be written as follows
using the chain rule:
𝜕 𝑁𝑖 𝜕𝑁𝑖 𝜕𝑥 𝜕𝑁𝑖 𝜕𝑦
= + (30)
𝜕ξ 𝜕x 𝜕ξ 𝜕𝑦 𝜕ξ
𝜕 𝑁𝑖 𝜕𝑁𝑖 𝜕𝑥 𝜕𝑁𝑖 𝜕𝑦
= + (31)
𝜕η 𝜕x 𝜕η 𝜕𝑦 𝜕η

The derivatives of the shape functions in the (x, y) system is

obtained by inversing the previous equation:
𝜕𝑁𝑖 𝜕𝑥 𝜕𝑦
𝜕x −1 𝜕ξ 𝜕ξ
𝜕𝑁𝑖 = 𝐽 𝜕𝑥 𝜕𝑦
(32)
𝜕𝑦 𝜕η 𝜕η
Figure 2: Linear 4-node quadrilateral element (Khennane 2013)
B. Stiffness Matrix
or more compactly as
{U} = [N]{a} (22) The stiffness matrix for the quadrilateral element can be derived
from the strain energy in the body, given by
The element is isoparametric, therefore the shape functions Ni(ξ,
η) also define the geometrical transformation between the 1
𝑈= 𝜎 𝑇 𝜀𝑑𝑣 (33)
2
reference and the parent element. The coordinates x and y of any 1 𝑇
𝑈= 𝑡 𝜎 𝜀𝑑𝐴 (34)
point of the parent element are given as 2 𝑒
𝜎 = 𝐷 𝐵 𝑎 (35)
x = N1x1 + N2x2 + N3x3 + N4x4 (23)
𝜀 = 𝐵 𝑎 (36)
y = N1y1 + N2y2 + N3y3 + N4y4 (24)
where te is the thickness of element e
Subsequently, we will need to express the derivatives of a
1 1 1
function in x and y coordinates in terms of its derivatives in ξ and 𝑈= 𝑡 𝑎 𝐵 𝑇
𝐷 𝐵 𝑎 𝑑𝑥𝑑𝑦 (37)
2 𝑒 −1 −1
η coordinates. This is done as follows 1 1 1 𝑇
𝑈= 𝑡 𝑎 𝐵 𝐷 𝐵 𝑎 𝑑𝑒𝑡𝐽𝑑ξ𝑑η (38)
𝜕𝑁 𝜕𝑁 𝜕𝑥 𝜕𝑁 𝜕𝑦 2 𝑒 −1 −1
= + (25) 1
= 𝑎 𝑘𝑎𝑇
(39)
𝜕ξ 𝜕x 𝜕ξ 𝜕𝑦 𝜕ξ
2
𝜕𝑁 𝜕𝑁 𝜕𝑥 𝜕𝑁 𝜕𝑦 where
= + (26) 1 1 𝑇
𝜕η 𝜕x 𝜕η 𝜕𝑦 𝜕η 𝑘 = 𝑡𝑒 −1 −1
𝐵 𝐷 𝐵 𝑑𝑒𝑡𝐽𝑑ξ𝑑η (40)
𝜕𝑁 𝜕𝑁
𝜕ξ 𝜕x
is the element stiffness matrix of dimension (8x8).
𝜕𝑁
=𝐽 𝜕𝑁 (27)
C. Numerical Integration
𝜕η 𝜕𝑦
The Gaussian quadrature approach is used for integration of
Where J is Jacobian transformation matrix which is given as
quadrilateral element. Consider the n-point approximation
𝜕𝑥 𝜕𝑦 4 𝜕𝑁 𝑖 4 𝜕𝑁𝑖
𝑖=1 𝜕ξ 𝑥𝑖 𝑖=1 𝜕ξ 𝑦𝑖 1
𝜕ξ 𝜕ξ 𝐼= 𝑓 ξ 𝑑ξ ≈ w1 f ξ1 + w2 f ξ2 + ⋯ . +wn f(ξn ) (41)
𝐽= 𝜕𝑥 𝜕𝑦
= (28a) −1
4 𝜕𝑁 𝑖 4 𝜕𝑁𝑖
𝑖=1 𝜕η 𝑥𝑖 𝑖=1 𝜕η 𝑦𝑖 Where w1,w2,….,and wn are the weights and ξ1, ξ2, …., and ξn
𝜕η 𝜕η
are the sampling points or gauss points. The idea behind Gaussian
After deriving and rearranging, the Jacobian is written in the form
quadrature is to select the n Gauss points and n weight such that
of a product of two matrices:
𝑥1 𝑦1 equation 3.51 provides an exact answer or polynomials f(ξ) of as
−(1 − η) (1 − η) (1 + η) −(1 + η) 𝑥2 𝑦2 large a degree as possible. In other words, the idea is that if the n-
𝐽 = 1/4
−(1 − ξ) −(1 + ξ) (1 + ξ) (1 − ξ) 𝑥3 𝑦3 point integration formula is exact for all polynomial. In our work
𝑥4 𝑦4 we have two variables so we used two point formula such as
1
(28b) −1
𝑓(ξ)𝑑ξ ≈ w1 f ξ1 + w2 f ξ2 (42)
A. Strain Matrix We have four parameters to choose: w1, w2, ξ1 and ξ2. We can
Substituting for the displacements u and v in equation (16) using therefore accept the formula in equation (42) to be exact for a
equation (22), the strain vector is obtained as cubic polynomial. Thus choosing f(ξ)=ao+a1 ξ+a2 ξ2+a3 ξ3 yields
1
𝜀 = 𝐵 𝑎 (29) 𝐸𝑟𝑟𝑜𝑟 = (a
−1 o
+ a1 ξ + a 2 ξ2 + a 3 ξ3 ) − w1 f ξ1 +
𝜕𝑁1 𝜕𝑁2 𝜕𝑁3 𝜕𝑁4 w2fξ2 (42)
0 0 0 0
𝜕𝑥 𝜕𝑥 𝜕𝑥 𝜕𝑥
𝜕𝑁1 𝜕𝑁2 𝜕𝑁3 𝜕𝑁4
Requiring zero error yields
𝐵 = 0 0 0 0 w1+w2=2
𝜕𝑦 𝜕𝑦 𝜕𝑥 𝜕𝑦
w1 ξ1+w2 ξ2=0
𝜕𝑁1 𝜕𝑁1 𝜕𝑁2 𝜕𝑁2 𝜕𝑁3 𝜕𝑁3 𝜕𝑁4 𝜕𝑁4
w1 ξ12+w2 ξ22=2/3
𝜕𝑦 𝜕𝑥 𝜕𝑦 𝜕𝑥 𝜕𝑥 𝜕𝑥 𝜕𝑦 𝜕𝑥 w1 ξ13+w2 ξ23=0

IJTRD | Jan-Feb 2017

Available [email protected] 318
 International Journal of Trend in Research and Development, Volume 4(1), ISSN: 2394-9333
www.ijtrd.com
These nonlinear equations have the unique solution  Nodal coordinate generation ( line 4-14)
 Nodal connectivity generation ( line 15-28)
w1 =w2 =1 -ξ1 =ξ2=1/ 3=0.57735022691…..
 Input value used in FEM ( line 29-42)
These values are directly used in the finite element analysis for 4-  Calculating stiffness matrix of body ( line 43-74)
nodes quadrilateral element.  Defining boundary condition force applied and
V. MATLAB CODE IMPLEMENTATION calculating displacement of body ( line 75-119)
 Plotting final result obtained ( line 120-169)
In present investigation a compact MATLAB code for structural
analysis of cantilever beam and Messerschmitt Bolkow Blohm 1. For structure analysis of cantilever beam under static loading
Beam subjected to static and self-weight loading has been condition as shown in Figure 4 (b) input
performed. The Matlab code (see the Appendix), is built up as a
standard topology optimization code. The main program is called FEM(32,20,1,0,0)
from the Matlab prompt by the line 2. For structure analysis of cantilever beam under self-weight
function FEM (Nx,Ny,CANTILEVER,MMB,GRAVITY) loading condition as shown in Figure 5 (b) input

Where nelx and nely are number of elements in the horizontal and FEM(32,20,0,0,11)
vertical directions, respectively. CANTILEVER and MBB is used 3. For structure analysis of MBB beam under static loading
for analysis of cantilever beam and Messerschmitt Bolkow condition as shown in Figure 6 (b) input
Blohm beam respectively. GRAVITY is used when structure is
analyzed under self weight. Other variables as well as boundary FEM(60,20,0,1,0)
conditions are defined in the MATLAB code itself and can be 4. For structure analysis of MBB beam under self-weight loading
edited if needed. The Compete MATLAB code for static and self- condition as shown in Figure 7 (b) input
weight loading conditions are described in appendix. The
description of MATLAB code (cantilever beam under static FEM(60,20,0,0,12)
loading meshing by 4-node elements) written is as follow:
F
60 mm

20 mm
20 mm

(a) (b)
Figure 3 Geometry and Boundary Conditions (a) cantilever beam (b) MBB beam
Table 1: Maximum deformation in x and y directions under point load
CANTILEVER BEAM MBB BEAM
ANSYS MATLAB ANSYS MATLAB
CANTILEVER BEAM CANTILEVER BEAM MBB BEAM MBB BEAM
UX 11.208 mm UX 30.545 mm UX 30.545 mm UX 30.545 mm
UY 27.471 mm UY 125.88 mm UY 125.88 mm UY 125.88 mm
MATERIAL PROPERTIES E =1 ν =0.3

(a) DISPLACEMENT UX DISPLACEMENT UY

IJTRD | Jan-Feb 2017

Available [email protected] 319
 International Journal of Trend in Research and Development, Volume 4(1), ISSN: 2394-9333
www.ijtrd.com

(b) DISPLACEMENT UX DISPLACEMENT UY

Figure 4: Deformation of catiliver beam under point load (a) ANSYS (b) MATLAB

(a) DISPLACEMENT UX DISPLACEMENT UY

(b) DISPLACEMENT UX DISPLACEMENT UY

Figure 5: Deformation of MBB beam under point load (a) ANSYS (b) MATLAB
Table 1: Maximum deformation in x and y directions self-weight load
CANTILEVER BEAM MBB BEAM
ANSYS MATLAB ANSYS MATLAB
CANTILEVER BEAM CANTILEVER BEAM MBB BEAM MBB BEAM
UX 16450 UX 16450 UX 2.5849×105 UX 2.5849×105
UY 54401 UY 54401 UY 9.4256×105 UY 9.4256×105
MATERIAL PROPERTIES ρ =1 g =10 E =1 ν =0.3

(a) DISPLACEMENT UX DISPLACEMENT UY

IJTRD | Jan-Feb 2017

Available [email protected] 320
 International Journal of Trend in Research and Development, Volume 4(1), ISSN: 2394-9333
www.ijtrd.com

(b) DISPLACEMENT UX DISPLACEMENT UY

Figure 6: Deformation of cantiliver beam under self-weight load (a) ANSYS (b) MATLAB

(a) DISPLACEMENT UX DISPLACEMENT UY

(b) DISPLACEMENT UX DISPLACEMENT UY

Figure 7: Deformation of MBB beam under self-weight load (a) ANSYS (b) MATLAB

CONCLUSION 10. coordinates((n+j),1)=x(i);

11. coordinates((n+j),2)=y(j);
In present investigation MATLAB code for structural analysis 12. end
of 2-D linear elastic isotropic structures i.e. cantilever beam and 13. n=n+(nely+1);
MBB beam is presented. For both the beam maximum 14. end
displacement obtained in x & y directions with MATLAB code % Input data for nodal connectivity for each element
are been compared with the ANSYS and from above result we 15. nodes=zeros(nelx*nely,4);
conclude that the accuracy of MATLAB code is same as that of 16. B=[1:nely];
ANSYS software for structural analysis of above numerical 17. t=[2:(nely+1)];
examples (cantilever beam and MBB beam). Another 18. m=[0:nely:nely*(nelx-1)];
19. n=0;
application of FEM MATLAB code is for self-weight analysis
20. for j=1:nelx
in which weight of the structure also include in structure 21. for i=1:nely
analysis of structures which also equally efficient to ANSYS 22. nodes((i+m(j)),1)=B(i)+n;
software. 23. nodes((i+m(j)),2)=B(i)+n+(nely+1);
APPENDIX 24. nodes((i+m(j)),3)=t(i)+n+(nely+1);
1. function FEM (Nx,Ny,CANTILEVER,MMB,GRAVITY) 25. nodes((i+m(j)),4)=t(i)+n;
2. nelx=Nx; 26. end
3. nely=Ny; 27. n=n+(nely+1);
% Input data for nodal coordinate values 28. end
4. coordinates=zeros((nely+1)*(nelx+1),2); % Input data
5. y=[nely:-1:0]; 29. nel = length(nodes);
6. x=[0:1:nelx]; 30. nnel=4;
7. n=0; 31. ndof=2;
8. for i=1:(nelx+1) 32. nnode = length(coordinates) ;
9. for j=1:(nely+1) 33. sdof=nnode*ndof;

IJTRD | Jan-Feb 2017

Available [email protected] 321
 International Journal of Trend in Research and Development, Volume 4(1), ISSN: 2394-9333
www.ijtrd.com
34. edof=nnel*ndof; 86. w4=[((nely+1)*(nelx)+2):((nely+1)*(nelx+1)-1)];
35. stiffness = zeros(sdof,sdof); 87. w5=[1:(nely+1)*(nelx+1)];
36. displacement=zeros(sdof,1); 88. W1=union(w1,w2);
37. force = zeros(sdof,1); 89. W2=union(w3,w4);
38. E = 2.1*10^9; 90. W3=union(W1,W2);
39. density=7850; 91. W4=setdiff(w5,W3);
40. gravity=9.81; 92. W5=setdiff(W4,W);
41. F=(density*gravity); 93. force(2*W')=-F/4;
42. nu = 0.3; % Poisson's ratio 94. force(2*W3')=-F/2;
% Computation of element matrices and vectors and their assembly 95. force(2*W5')=-F;
43. D = E/(1-nu^2)*[1 nu 0 ; nu 1 0; 0 0 (1-nu)/2] ; 96. if (GRAVITY==11)
44. Gausspointx=[-1 1 1 -1]/sqrt(3); 97. displacement((2*(nely+1)+1):sdof) =
45. Gausspointy=[1 1 -1 -1]/sqrt(3); stiffness((2*(nely+1)+1):sdof,(2*(nely+1)+1):sdof)\force((2*(nely
46. Gaussweight=[1 1 1 1]; +1)+1):sdof);
47. for iel=1:nel 98. end
48. for i=1:nnel 99. if (GRAVITY==12)
49. nd(i)=nodes(iel,i); 100. fixeddofs = union([1:2:2*(nely+1)],[2*(nelx+1)*(nely+1)]);
50. xx(i)=coordinates(nd(i),1); 101. alldofs = [1:2*(nely+1)*(nelx+1)];
51. yy(i)=coordinates(nd(i),2); 102. freedofs = setdiff(alldofs,fixeddofs);
52. end 103. displacement(freedofs,:) = stiffness(freedofs,freedofs) \
53. K = zeros(edof,edof); force(freedofs,:);
54. for int=1:4 104. displacement(fixeddofs,:)= 0;
55. xi = Gausspointx(int); 105. end
56. wtx = Gaussweight(int); 106. end
57. eta = Gausspointy(int); % solution
58. wty = Gaussweight(int) ; 107. UX = displacement(1:2:sdof) ;
% Compute isoparametric four-node Quadrilateral shape functions and 108. UY = displacement(2:2:sdof)
their derivatives 109. disp('The maximum displacement UX')
59. j=0.25*[-(1-eta) (1-eta) (1+eta) -(1+eta);... 110. max(abs(UX))
(1-xi) (1+xi) -(1+xi) -(1-xi)]; 111. disp('The maximum displacement UY')
60. J=j*[xx;yy]'; 112. max(abs(UY))
61. A=[J(2,2) -J(1,2) 0 0; 113. for iel=1:nel
0 0 -J(2,1) J(1,1); 114. for i=1:nnel
-J(2,1) J(1,1) J(2,2) -J(1,2)]*(1/det(J));
62. G=(1/4)*[-(1-eta) 0 (1-eta) 0 (1+eta) 0 -(1+eta) 0; 115. nd(i)=nodes(iel,i); X(i,iel)=coordinates(nd(i),1);
(1-xi) 0 (1+xi) 0 -(1+xi) 0 -(1-xi) 0; Y(i,iel)=coordinates(nd(i),2);
0 -(1-eta) 0 (1-eta) 0 (1+eta) 0 -(1+eta); 116. end
0 (1-xi) 0 (1+xi) 0 -(1+xi) 0 -(1-xi)]; 117. profile1(:,iel) = -UX(nd') ;
63. B=A*G; 118. profile2(:,iel) = -UY(nd') ;
64. K=K+B'*D*B*wtx*wty*det(J); 119. end
65. end 120. % Plotting the FEM mesh and profile of the given component
66. EDOF = [2*nd(1)-1; 2*nd(1); 2*nd(2)-1; 2*nd(2); 2*nd(3)-1; 121. f3 = figure ;
2*nd(3); 2*nd(4)-1; 2*nd(4)]; 122. set(f3,'name','Postprocessing','numbertitle','off') ;
67. stiffness(EDOF,EDOF)=stiffness(EDOF,EDOF)+K; 123. plot(X,Y,'k')
68. end 124. fill(X,Y,profile1)
% constrain, boundary condition and force vector 125. axis off ;
%-----------------------external force----------------------------- 126. cbar = colorbar;
69. if (CANTILEVER==1) 127. set(figure,'name','Postprocessing','numbertitle','off') ;
70. force(2*(nelx+1)*(nely+1)) = -1 ; 128. plot(X,Y,'k')
71. displacement((2*(nely+1)+1):sdof) = 129. fill(X,Y,profile2)
stiffness((2*(nely+1)+1):sdof,(2*(nely+1)+1):sdof)\force((2*(nely 130. axis off ;
+1)+1):sdof); 131. cbar = colorbar;
72. end
%-----------------------gravity force-------------------------------- References
73. if (MMB==1) [1] Chandrupatla, T. R. and Belegunda, A. D. 2004.
74. force(2,1) = -1;
Introduction to finite element in engineering, Prentice hall
75. fixeddofs = union([1:2:2*(nely+1)],[2*(nelx+1)*(nely+1)]);
76. alldofs = [1:2*(nely+1)*(nelx+1)]; upper Inc.
77. freedofs = setdiff(alldofs,fixeddofs); [2] Hutton, D. V. 2004. Fundamental of finite element analysis,
78. displacement(freedofs,:) = stiffness(freedofs,freedofs) \ The McGraw-Hill companies
force(freedofs,:); [3] Khennane, A. 2013. Introduction to finite element analysis
79. displacement(fixeddofs,:)= 0; using MATLAB and Abaqus, Taylor & Francis Group,
80. end LLC
%-----------------------gravity force------------------------------- [4] Ain, J.E 1982. Application and implementation of finite
81. if (GRAVITY>0) element methods, London:Academic.
82. W=[1 (nely+1) ((nely+1)*(nelx+1)-nely) (nely+1)*(nelx+1)];
[5] Baker, A.J. and Pepper D.W., 1988. Finite element analysis
83. w1=[2:nely];
84. w2=[(nely+2):(nely+1):(nely+1)*(nelx)]; on microcomputers. New York:McGraw-Hill.
85. w3=[(2*(nely+1)):(nely+1):(nely+1)*(nelx)];

IJTRD | Jan-Feb 2017

Available [email protected] 322
 International Journal of Trend in Research and Development, Volume 4(1), ISSN: 2394-9333
www.ijtrd.com
[6] Lepi, S.M., 1998. Practical guide to finite elements: a solid [13] Logan D.L. A First Course in the Finite Element Method
mechanics approach, Marcel Dekker. Using Algor, 2nd edn. Brooks/Cole Thompson Learning,
[7] Smith I.M. and Griffiths D.V. Programming the Finite Pacific Groove, CA, 2001.
Element Method, 2nd edn. Wiley, Chichester, U.K., 1988. [14] Mase G.E. Schaum’s Outline Series: Theory and Problems
[8] Timoshenko S. and Goodier J. Theory of Elasticity, 3rd of Continuum Mechanics. McGraw-Hill, New York, 1970.
edn. McGraw-Hill, New York, 1970. [15] McGuire M., Gallagher G.H., and Ziemian R.D. Matrix
[9] Timoshenko S. and Woinowsky-Krieger S. Theory of Plates Structural Analysis, 2nd edn. Wiley, New York, 2000.
and Shells. McGraw-Hill, New York, 1959. [16] Meek J.L. Computer Methods in Structural Analysis. E &
[10] Zienkiewicz O.C. The Finite Element Method, 3rd edn. FN SPON, London, U.K., 1991.
McGraw-Hill, York, London, 1977. [17] Reddy J.N. An Introduction to the Finite Element Method,
[11] Cook R.D. Finite Element Modeling for Stress Analysis. 3rd edn. McGraw-Hill, New York, 2006.
Wiley, New York, 1995. [18] Saada A.S. Elasticity: Theory and Applications, 2nd edn.
[12] Kwon Y.W. and Bang H. The Finite Element Method Using Krieger Publishing, Melbourne, FL, 1993.
Matlab, 2nd edn. CRC Press, London, U.K., 2000.

IJTRD | Jan-Feb 2017

Available [email protected] 323
View publication stats