A First Course in the Finite Element Method, 6th Edition Logan

Types of Residual Methods

• Collocation Method
• Requires that the error or residual function, R, be forced
to zero at as many points as there are unknown
coefficients
• Subdomain Method
• Requires that the integral of the error or residual function
over some selected subintervals to be set to zero
• Least Squares Method
• Requires the integral of the error function squared to be
minimized with respect to each of the unknown
coefficients
• Galerkin’s Method
• Requires the error to be orthogonal to some weighting
functions

1
A First Course in the Finite Element Method, 6th Edition Logan

Example – four different weighted residual methods

2
A First Course in the Finite Element Method, 6th Edition Logan

Exact solution

3
A First Course in the Finite Element Method, 6th Edition Logan

Pick an approximated function

4
A First Course in the Finite Element Method, 6th Edition Logan

5
A First Course in the Finite Element Method, 6th Edition Logan

6
A First Course in the Finite Element Method, 6th Edition Logan

7
A First Course in the Finite Element Method, 6th Edition Logan

Galerkin’s method

8
A First Course in the Finite Element Method, 6th Edition Logan

Galerkin’s method

9
A First Course in the Finite Element Method, 6th Edition Logan

Chapter 3 review

10
A First Course in the Finite Element Method, 6th Edition Logan

Summary – p1
General differential equation for axial bar problems

Exact solution

11
A First Course in the Finite Element Method, 6th Edition Logan

Steps for FEA of a truss problem

Step 1: Discretize and Select the Truss Element

Step 2: Select a Displacement Function
Step 3: Define the Element Relationships
Step 4: Derive the Element Stiffness Matrix
Equations (methods in three general categories)
Step 5: Assemble the Element Equations
Step 6: Solve for the Nodal Displacements
Step 7: Solve for the Element Forces

12
A First Course in the Finite Element Method, 6th Edition Logan

Summary – p2
Introduce linear approximation (nodal displacements are linked to
the entire element by the linear shape [interpolation] functions):

Two formulations:
1) x – global coordinate system
2) 𝜉 – local coordinate system

13
 A First Course in the Finite Element Method, 6th Edition Logan

Summary – p3
Introduce quadratic approximation (nodal displacements are linked to
the entire element by the quadratic shape [interpolation] functions):

u(x)
Two formulations:
1) x – global
coordinate
system
2) 𝜉 – local
coordinate
system
A First Course in the Finite Element Method, 6th Edition Logan

Hw 5 P4 – quadratic version of:

For any approximation function, you now can derive the shape
functions to interpolate for the entire element. Then the second
properties like strains and stresses can be obtained accordingly.
15
A First Course in the Finite Element Method, 6th Edition Logan

Summary – p4
Shape function properties:
1）Kronecker-delta property
2) Compatibility
3) Completeness

16
A First Course in the Finite Element Method, 6th Edition Logan

Steps for FEA of a truss problem

Step 1: Discretize and Select the Truss Element

Step 2: Select a Displacement Function
Step 3: Define the Element Relationships
Step 4: Derive the Element Stiffness Matrix
Equations (methods in three general categories)
Step 5: Assemble the Element Equations
Step 6: Solve for the Nodal Displacements
Step 7: Solve for the Element Forces

17
A First Course in the Finite Element Method, 6th Edition Logan

Summary – p5
Direct (stiffness) method:
- Displacements as the unknowns

Direct (flexibility) method (not covered):

- Forces as the unknowns

18
A First Course in the Finite Element Method, 6th Edition Logan

Summary – p6
Potential Energy Approach:

Multi-element
expressions after
meshing

19
A First Course in the Finite Element Method, 6th Edition Logan

Summary – p6
Potential Energy Approach:
since we have obtained the approximation functions: u = [N] {d},
we can put this relationship to the above equations.

For any
distributed/concentrated
loads, you now can
formulate the nodal force
vector.

20
A First Course in the Finite Element Method, 6th Edition Logan

Summary – p7
Residual methods: collocation method, subdomain method, least
squares method, Galerkin’s method
- Approximate function:

- Residual function:

Without
distributed
loads
The goal of residual methods: build equations to solve
for the coefficients
21
A First Course in the Finite Element Method, 6th Edition Logan

Summary – p8
Weighted residual methods:

22
A First Course in the Finite Element Method, 6th Edition Logan

Summary – p9
Weighted residual methods:

Integration by parts
to reduce order of
the derivatives

23
A First Course in the Finite Element Method, 6th Edition Logan

Summary – p10
Weighted residual methods:
- Galerkin’s methods (when picking the weight functions as the
follows).

24
A First Course in the Finite Element Method, 6th Edition Logan

Summary – p11
We’ve introduced two examples (without meshing) using the Bubnov-Galerkin approach

Case 1:
choosing the basis of the weighting functions the same as the
basis in the approximate function.

25
A First Course in the Finite Element Method, 6th Edition Logan

Summary – p12
We’ve introduced two examples (without meshing) using the Bubnov-Galerkin approach

Case 2:
choosing the basis of the weighting functions the same as the
basis in the approximate function.

For any second order differential

equations, you should be able to apply
similar approaches by picking the basis in
the approximated functions yourselves
then using Galerkin’s method to obtain the
approximate solution yourselves.

26
A First Course in the Finite Element Method, 6th Edition Logan

Summary – p12
Finally, we have meshed the bar and within one element, we formulate
stiffness matrix using n-order element.

27
A First Course in the Finite Element Method, 6th Edition Logan

Summary
• Developed truss equations
• Derived the stiffness matrix for a bar element
• Described the process of solving for element forces of the bar
element
• Discussed guidelines for choosing displacement functions
• Derived the global stiffness matrix for an arbitrarily oriented bar
element in the X – Y plane
• Computed stress in a bar in the X – Y plane
• Reviewed the plane truss
• Discussed the transformation matrix of a bar in three-
dimensional space

28
A First Course in the Finite Element Method, 6th Edition Logan

Summary
• Stated the global stiffness matrix of a bar in three-dimensional
space
• Reviewed the type of symmetry that occurs in structures
• Discussed reflective symmetry
• Derived the stiffness matrix using the principle of minimum
energy
• Compared the finite element solution to the exact solution for a
bar
• Discussed weighted residual methods in general
• Described types of residual methods commonly used
• Listed a computational flow chart for solving three-dimensional
truss problems

30