Anda di halaman 1dari 5

Introduction of Meshfree Methods and Implementation of

Element Free Galerkin (EFG) Method to Beam Problem

Someshwar S. Pandey1, Paresh K. Kasundra2 & Sachin D. Daxini3

B.H. Gardi College of Engineering & Technology Rajkot,
Babariya Institute of Technology Vadodara,
E-mail : someshwar1584@gmail.com1,, sachin_daxini@rediffmail.com3

Abstract Numerical simulation using computers has of possible boundary conditions with a set of arbitrarily
increasingly become a very important approach for solving distributed nodes (or particles) without using any mesh
problems in engineering and science. It plays a valuable that provides the connectivity of these nodes or
role in providing tests and examinations for theories, particles. [7]
offering insights to complex physics, and assisting in the
interpretation and even the discovery of new phenomena. One important goal of the initial research is to
Grid or mesh based numerical methods such as FDM, modify the internal structure of the grid-based FDM and
CFD, FEM despite of great success, suffer from difficulties FEM to become more adaptive, versatile and robust.
in some aspects, which limit their applications in many Much effort is concentrated on problems to which the
complex problems. The major difficulties are inherited
from the use of grid or mesh. A recent strong interest is
conventional FDM and FEM are difficult to apply, such
focused on the next generation computational methods as problems with free surface, deformable boundary,
meshfree methods, which are expected to be superior to moving interface (for FDM), large deformation (for
conventional grid-based FDM and FEM in many FEM), complex mesh generation, mesh adaptivity, and
applications. The Element Free Galerkin (EFG) method is multi-scale resolution (for both FDM and FEM).
a meshless method because only a set of nodes and a
description of models boundary are required to generate The finite element methods are well established and
the discrete equations. In this paper the EFG method is powerful computational / simulation techniques which
applied to 2-D beam problem and results are compared are used for modeling and analysis of physical
with the analytical solution by using Timoshenko Beam phenomena in different fields of engineering and applied
Theory. The step by step algorithm for EFG MATLAB sciences. It has successfully been applied for a large
program is also provided inside. number of engineering applications, for example solid
Keywords FEM, EFG, MATLAB , FDM, CFD. mechanics, structure mechanics, electro magnetism,
geo-mechanics, bio mechanics, aerodynamics and so on,
I. INTRODUCTION but it is with some shortcomings as mentioned below:
1. For modeling large deformation problems,
Simulation of different engineering applications has considerable loss in accuracy arises when the
for many years been great importance in solving the elements in mesh become extremely skewed or
problem. This is done by solving partial differential compressed.
equation with initial values and boundary conditions.
2. The use of mesh in modeling these problems creates
The Finite Element Method (FEM) has been the
difficulties in the treatment of discontinuities which
standard tool for this kind of calculations. But under the
do not coincide with the original mesh lines.
last fifteen years a new mesh free method has been
under extensive research. [1] 3. Creation of a mesh for the problem domain
consumes major time of analyst using FEM
A recent strong interest is focused on the packages. It becomes a major component of the cost
development of the next generation of computational of a simulation project. The concern is more the
methods meshfree methods, which are expected to be manpower time, and less the computer time.
superior to the conventional grid-based FDM and FEM
4. It is very difficult to simulate the breakage of
for many applications. The key idea of the meshfree
material into a large number of fragments as FEM
methods is to provide accurate and stable numerical
is essentially based on continuum mechanics, in
solutions for integral equations or PDEs with all kinds
which the elements formulated cannot be broken.

ISSN : 2319 3182, Volume-2, Issue-3, 2013

International Journal on Theoretical and Applied Research in Mechanical Engineering (IJTARME)

5. Discontinuous secondary variables like stresses The principal attraction of mesh free methods is the
across element boundaries (due to piecewise possibility of simplifying adaptivity and problems with
nature of assumed shape functions), Post moving boundaries and discontinuities, such as phase
processing techniques are required to achieve changes or cracks. In crack growth problems, for
smooth stress distribution in structural problems. example, nodes can be added around a crack tip to
To overcome the above difficulties mesh free capture the stress intensity factors with the desired
methods have been developed which do not require a accuracy; this nodal refinement can be moved with a
mesh to discretize the problem and any connectivity propagating crack through a background arrangement of
between nodes. Mesh free methods are having following nodes associated with the global geometry. Adaptive
advantages as compared to finite element method: meshing for a large verity of problems including linear
and non linear stress analyses can be effectively treated
1. There is no need to provide in advance any
by these methods in a simple manner.
information about the relationship of the nodes, so
it provides flexibility in adding and deleting nodes The generalized step by step procedure for mesh
whenever and wherever needed. free methods is depicted in the below figure:
2. Since there is no need to create a mesh, and the
nodes can be created by a computer in a fully
automated manner, it saves a lot of human effort.
3. The approximate solution is constructed entirely in
terms of a set of nodes.
4. They can easily handle very large deformation
because connectivity can change with time. [2]
In spite of having above advantages over FEM
mesh free methods are not without disadvantages,
following are the drawbacks of Mesh free methods:
1. Since mesh free shape functions are rational
functions, therefore, it requires highly accurate
integration scheme to be applied.
2. It is to be noted that treatment of essential boundary
conditions is not straight forward as in mesh based
methods since the shape functions of mesh free
methods are not interpolants but approximants,
which implies that they do not satisfy the kronecker Fig.1.2 : Flow Chart Of Mesh Free Method Solution
delta property. Procedure [2]
There are various mesh free methods are developed
now a days to analyze well posed problems. Mesh free II. ELEMENT FREE GALERKIN METHOD
methods can be classified as below:
This meshfree method was developed in by the
group of Prof. Ted Belytschko based on the idea of
Lancaster and Salkauskas and probably motivated by
the purpose to model arbitrary crack propagation
without computational expensive re-meshing. In
Element Free Galerkin (EFG) we use the moving least
square (MLS) method for constructing the shape
functions. Moving least square method was first
proposed by Lancaster and Salkauskas (1981), as an
interpolation method. It was used in element free
methods by Belytschko et al. (1994), with use of
Lagrange multiplier to invoke essential boundary. [3]
The Boundary representation in meshfree methods is
done only by the arbitrary distribution of the nodes,
which may or may not be uniform. Fig. 1 shows the
boundary representation for meshfree methods only by
Fig. 1.1 : Classification of mesh free methods

ISSN : 2319 3182, Volume-2, Issue-3, 2013

International Journal on Theoretical and Applied Research in Mechanical Engineering (IJTARME)

the nodes and for FEM using the nodes & elements. The
field variables of interest are specified on these nodes.

STEP-5: The nodal discrete equations are obtained
using the constrained Galerkin weak form using the
Lagrange multiplier. Following are the governing
matrices for boundary condition, stiffness, force and
displacement respectively.



Fig. 2.1 : Boundry representations in mesh free method and (6)

fem [3]



STEP-1: Discretize the beam into number of nodes Where,

distributed uniformly over the domain = (0, 1). The
selection of the number of nodes is random and the
convergence study has to be performed to arrive at the
optimum number of nodes, depending upon the (8)
accuracy and tolerance requirement of the solution. The
discretized bar is represented by fig. 2.2

These relations are for plane stresses. In which a comma
designates a partial derivative with respect to the
indicated spatial variable; E and are Youngs modulus
Fig. 2.2 : Mesh Free Representation of Beam and Poisons ratio respectively. [6]
STEP-2: The integration cells are constructed for four STEP-6 The above discrete nodal equations are
point gauss quadrature point and linear basis function assembled into global matrix:
for 2-D beam can be given by,

(1) (11)
STEP-3: The shape function is constructed by the STEP-7 The essential boundary condition is imposed.
MLS approximation method The specified displacement at fixed end is equal to zero,
hence q=0
STEP-8 The global matrix is solved to obtain the nodal
STEP-4: The approximate solution is obtained using displacement parameters.
shape function as:

ISSN : 2319 3182, Volume-2, Issue-3, 2013

International Journal on Theoretical and Applied Research in Mechanical Engineering (IJTARME)

The solution code using the MATLAB language is 4. Determine the domain of influence for each
developed and the resulting data and plots are recorded node in the mesh.
for further analysis.
5. Set up quadrature cells in the domain.
EFG APPLIED TO 2-D BEAM PROBLEM 6. Set up gauss points, weights and jacobian for
each cell
Consider a beam of length L = 48m subjected to
parabolic traction at the free end as shown in below 7. Loop over gauss points
figure 2.3 The beam has characteristics height D=12m
7.a Determine nodes in the neighborhood of the
and is considered to be of unit depth and is assumed to
gauss points
be in a state of plane stress with P= 1000N, = 0.3 and
E= 3.0 x 107 7.b Determine weights, shape function and shape
function derivatives
7.c Assemble B matrix
7.d Add contributions to K matrix
8. Determine nodes on traction boundary and
essential boundary.
9. Set up gauss points along traction boundary and
essential boundary
Fig.2.3 : Two Dimensional Beam Problem
10. Integrate forces along traction boundary to
The exact analytical solution for the displacement form f factor
of Timoshenko beam is given by the following
equations: [5] 11. Integrate lagrange multipliers along essential
boundary to form the G matrix and q vector
12. Enforce essential boundary conditions using
lagrange multipliers.
13. Solve for nodal parameters ui
14. Loop over gauss points
14.a Determine exact and analytical stresses at
quadrature points
14.b Assemble contributions to L2 error norm
15. Print total error in strain energy
We shall use this solution to examine the quality of the
EFG solution by comparing it to the exact solution for Table 2.1 Flow Chart Of 2-D EFG Program
both the displacements and strains along the bar. The for Beam [4]
program has been written in MATLAB, a language
which allows matrix manipulations to be performed COMPARISON OF EXACT ANALYTICAL
easily. The language is very similar to FORTRAN, and SOLUTION WITH EFG MATLAB SOLUTION AND
the program can be adapted easily. The program has FEM SOLUTION:
written with a minimum of generality, and it is aimed at
solving the specific problem outlined previously. The
step by step general algorithm of program is given EFG FEM
1. Define the physical dimensions and material INCH IN INCH in INCH in Y
properties. Y direction direction

2. Define the plane stress D matrix 0 0 0 0

4 -0.0002 -0.0002 -0.0003
3. Set up the nodal coordinates for a uniform
mesh. 8 -0.0004 -0.0004 -0.0007

ISSN : 2319 3182, Volume-2, Issue-3, 2013

International Journal on Theoretical and Applied Research in Mechanical Engineering (IJTARME)

12 -0.0008 -0.0008 -0.0012 III. CONCLUSION

16 -0.0014 -0.0014 -0.0016 The details of the Element Free Galerkin (EFG)
20 -0.0021 -0.0021 -0.0023 method and its numerical implementation have been
presented. The method has been applied on benchmark
24 -0.0028 -0.0028 -0.0023
problems of structures widely used in mechanical field
28 -0.0037 -0.0037 -0.0041 and it is observed that EFG method is quite promising
32 -0.0046 -0.0047 -0.0054 in the performance as the results calculated from
36 -0.0057 -0.0057 -0.0067 analytical solution and proposed mesh free method are
quite same. Thus, EFG is validated for its effectiveness
40 -0.0067 -0.0067 -0.0075 in solving structural problems.
44 -0.0078 -0.0078 -0.0081
48 -0.0089 -0.0089 -0.0091 IV. REFERANCES
[1] David V. Hutton, Fundamentals of finite
Table 2.2 Comparison of Results element analysis, Mc-Graw Hill, 2004
[2] G.R. Liu, Mesh Free Methods: Moving beyond
the finite element method, CRC Press, 2003
[3] J.S.Kushawaha, Significance of Dimensionless
Size of Support Domain
In Element Free Galerkin Method
[4] Victor Petersson, An Implementation of Mesh
Free Methods for Mechanical Problems at Large
[5] Timoshenko, S.P. and J.N.Goodier (1970).
Theory of Elasticity(Third edi.). New York:
McGraw Hill
[6] John Dolbow and Ted Belytschko, An
Introduction to Programming the Meshless
Element Free Galerkin Method
Fig. 2.4 : Comparison of Results
[7] G.R. Liu and M.B.Liu, Smoothed Particle
The comparison of above results shows good
Hydrodynamics a mesh free particle method
agreement between EFG results and exact analytical
solution but there is slight variation in results obtained
by FEM in Pro/Mechanica

ISSN : 2319 3182, Volume-2, Issue-3, 2013