Scribd
Upload
42 views

Shape Functions and Numerical Integration

The document discusses shape functions and numerical integration in computational fluid dynamics (CFD) finite volume method (FVM) simulations. It describes the properties and forms of shape functions for one-dimensional and two-dimensional elements, including linear and quadratic elements. It also discusses numerical integration techniques like Gaussian quadrature that are commonly used to evaluate integrals of shape functions over element domains that arise in CFD simulations.

Uploaded by

sanjay
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
42 views

Shape Functions and Numerical Integration

The document discusses shape functions and numerical integration in computational fluid dynamics (CFD) finite volume method (FVM) simulations. It describes the properties and forms of shape functions for one-dimensional and two-dimensional elements, including linear and quadratic elements. It also discusses numerical integration techniques like Gaussian quadrature that are commonly used to evaluate integrals of shape functions over element domains that arise in CFD simulations.

Uploaded by

sanjay
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

COMPUTATIONAL FLUID DYNAMICS: FVM: Interpolation Schemes

Lecture 22
SHAPE FUNCTIONS AND NUMERICAL
INTEGRATION

22.1 SHAPE FUNCTIONS


Shape functions in finite element analysis depend on the dimensionality of the problem and
type of elements used for discretization of the problem domain. These must also satisfy
continuity requirements depending on the underlying PDE and form (strong or weak) of finite
element formulation. Normally these are chosen as polynomial functions. In addition, the
shape functions must satisfy/possess the following properties:
 Partition of unity:
n

N
k 1
k 1 (22.1)

 Kronecker-delta property:
N i (x k )   ik (22.2)

where n is number of nodes in an element, and xk represents the coordinate of kth node.

22.2 SHAPE FUNCTIONS FOR ONE DIMENSIONAL ELEMENTS


Shape functions for a two-node (linear) one dimensional element are given by
1 1
N1  (1   ), N 2  (1   ) (22.3)
2 2

where ξ is the local coordinate defined as   2( x  xc ) / l in which l is length of the element


and xc represents its mid-point. Here, nodes 1 and 2 correspond to   1 and   1
respectively. Similarly, shape functions for a quadratic (three-node) element are given by

1 1
N1   (  1), N 2  (1   2 ) N3   (1   ) (22.4)
2 2
where nodes 1, 2 and 3 correspond to   1 ,   0 and   1 respectively.

22.2 SHAPE FUNCTIONS FOR TWO DIMENSIONAL ELEMENTS


In two-dimensions, one can opt for rectangular or triangular elements. In each case, shape
functions are expressed in corresponding natural coordinates. These shape functions can also
be used to define the physical coordinates of any point as follows:

n n
x( , )   N ( , ) x , y ( , )   N ( , ) y (22.5)
 1  1

Dr K M Singh, Indian Institute of Technology Roorkee NPTEL 22.1


COMPUTATIONAL FLUID DYNAMICS: FVM: Interpolation Schemes

where n denotes the number of nodes in the element. Elements of preceding type for which
the shape functions can be used to represent the geometry as well as function approximation
are called iso-parametric elements.

22.2.1 Rectangular Elements


Shape functions for rectangular elements are expressed in terms of natural coordinates ξ and
η defined as
2( x  xc ) 2( y  yc )
 ,  (22.6)
lx ly
where (xc , yc) is the centroid of the element, and lx, ly represent its extents in x and y-
direction. Two-dimensional bilinear rectangular elements are given by
1
N  (1   )(1   ),   1,.., 4 (22.7)
4
where ξ and η are natural coordinates, and coordinates of the four nodes in natural
coordinates are (-1,-1), (1,-1), (1, 1) and (-1, 1). For further detailed discussion of different
types of rectangular element employed in finite element analysis, see Zienkiewicz, Taylor
and Zhu (2005).

22.2.1 Triangular Elements


Shape functions for triangular elements are expressed in terms of area coordinates 1 ,  2 and
 3 defined as
a  b x  c1 y a  b x  c2 y
1  1 1 , 2  2 2 , 3  1  (1   2 ) (22.8)
2 2
where Δ is the area of the triangle with vertices (x1,y1), (x2,y2), (x3,y3), and coefficients ak, bk
and ck are given by
a1  x2 y3  x3 y2 , b1  y2  y3 , c1  x3  x2
(22.9)
a2  x3 y1  x1 y3 , b2  y3  y1 , c2  x1  x3
In terms of area coordinates, shape functions for a linear triangular element are defined as
N1  1 , N 2   2 , N 3  3 (22.10)
Shape functions for a quadratic triangular element are given by
 Corner nodes (numbered as 1, 2, 3)
N1  1 (21  1), N 2   2 (2 2  1), N 3  3 (23  1)
 Mid-side nodes (numbered as 4,5,6)
N 4  41 2 , N5  4 23 , N 6  431 (22.11)
For further details of two-dimensional rectangular and triangular elements, see Zienkiewicz,
Taylor and Zhu (2005). You can also find shape functions for three dimensional elements, for
example, prism or tetrahedral elements in this reference.

22.2 EVALUATION OF INTEGRALS


Formation of discrete finite element system (20.12) requires evaluation of domain integrals
given in Eq. (20.14). Except for the simplest of element geometries, these integrals cannot be
evaluated analytically. Hence, numerical integration is the only alternative. For line elements
(1-D), quadrilateral elements (in two dimensions) and brick elements (in three dimensions),
Gaussian quadrature is most commonly employed (Zienkiewicz, Taylor and Zhu, 2005). For
example, a typical integral for a two dimensional quadrilateral element can be evaluated as

Dr K M Singh, Indian Institute of Technology Roorkee NPTEL 22.2


COMPUTATIONAL FLUID DYNAMICS: FVM: Interpolation Schemes

1 1
I  G(x)d   
e
1 1
G (x( , )) | J | d d

m n
(22.4)
  G (i , j ) wi w j , G (i , j )  G (x( , )) | J |
i 1 j 1

where ξ and η are natural coordinates, | J | is the determinant of the Jacobian matrix J given
by
 x x 
   
J  (22.5)
 y y 
   
 
i , j are Gaussian quadrature abscissa, and wi and wj are corresponding weights. Special
quadrature formulae are available for triangle and tetrahedral elements. For further details
and guidelines for evaluation of finite element integrals, see Zienkiewicz, Taylor and Zhu
(2005).

REFERENCES/FURTHER READING
Chung, T. J. (2010). Computational Fluid Dynamics. 2nd Ed., Cambridge University Press,
Cambridge, UK.
Muralidhar, K. and Sundararajan, T. (2003). Computational Fluid Dynamics and Heat
Transfer, Narosa Publishing House.
Reddy, J. N. (2005). An Introduction to the Finite Element Method. 3rd Ed., McGraw Hill,
New York.
Zienkiewicz, O. C., Taylor, R. L., Zhu, J. Z. (2005). The Finite Element Method: Its Basis
and Fundamentals, 6th Ed., Butterworth-Heinemann (Elsevier).

Dr K M Singh, Indian Institute of Technology Roorkee NPTEL 22.3

You might also like