Title: The Finite Element Method
1The Finite Element Method
- Dr. Abdul Razzaq Touqan
- Department of Civil Engineering
2Introduction to the Finite Element Method
- Objectives
- Students will be introduced to FE
3Finite Element Method Defined
- Problems are governed by differential or integral
equations. - These equations provide an exact, closed-form
solution (analytical). - Complexities in geometry, properties and in the
boundary conditions prevent ability to obtain
exact solutions. Also 2D and 3D elements are
governed by partial differential equations, no
closed form solution exists for such elements..
4Finite Element Method Defined (cont.)
- Complex regions are discretized (anatomized)
into simple shapes called elements. The
continuum has infinite number of DOFs.
Discretized model has finite number of DOFs. - The elements are either 1D, 2D or 3D.
- Properties and relationships are analogized over
the elements in terms of unknown values at nodes. - A set of linear/nonlinear equations are obtained
by linking the individual elements and
considering effects of loads and boundary
conditions.
5Finite Element Method Defined (cont.)
- Number of equations is large ? requires digital
computers otherwise worthless. - Advances in computers and software has brought
the FEM within reach for engineers and students.
6Simplified example of concept of FE finding the
perimeter of a circle
- 1. Analytical integral equation of infinitely
small arcs - 2. Analogical approximate arc into straight
lines and then perimeter is between (4v2)r and
8r! How? - 3. Anatomical (Finite element) combine 1 and 2
discretize circle into finite number of divisions
and approximate curved arcs by straight lines
p(2rsin?/2)n, ? 2p/n
7Origins of the Finite Element Method
- Basic concepts evolved over a period of 150 or
more years. - The term finite element was first coined by
Clough in 1960. - The first book on the FEM by Zienkiewicz and
Chung was published in 1967. - In the late 1960s and early 1970s, FEM was
applied to a wide variety of engineering problems.
8Origins of the Finite Element Method (cont.)
- 1970s marked advances in mathematical treatments
- Most FEM software originated in the 1970s and
1980s. - FEM is one of the most important developments in
computational methods in the 20th century. - In few decades, the method has evolved to cover
many scientific and technological areas (solid
and fluid mechanics, chemical reactions,
electromagnetics, biomechanics, heat transfer and
acoustics ).
9Advantages of the Finite Element Method
- Handle complex geometry heart and power of it.
- Handle complex restraints indeterminate
structures -
- Handle complex loading point, pressure, inertia
- Handle nonhomogeneous material bodies every
element assigned a different material properties.
10Advantages of the Finite Element Method (cont.)
- Handle nonisotropic materials bodies
orthotropic, anisotropic - Handle special material effects temperature
dependent properties, plasticity, creep,
swelling, - Handle complex analysis types vibration,
nonlinear.. - Model special geometric effects large
displacements and large rotations
11Disadvantages of the Finite Element Method
- Handle specific problems, thus no general
closed-form solution - Approximate the mathematical model (the source of
so-called inherited errors.) - Need experience and judgment to construct a good
finite element model. - A computer and reliable FEM software are
essential. - Large amount of Input and output data to prepare
and interpret.
12Disadvantages of the Finite Element Method (cont)
- Numerical problems
- Computers carry a finite number of significant
digits? round off and error accumulation. -
- Susceptible to user-introduced modeling errors
- Poor choice of element types.
- Distorted elements.
- Geometry not adequately modeled.
- Certain effects not automatically included
- Buckling, large deflections and rotations,
material and geometric nonlinearities
13homework
- Pick a small problem in engineering. Then provide
analytical, analogical and anatomical (FE)
solutions for it. Pick it up from a library
reference if not able to visualize one.
14End of introduction to FE
Let Learning Continue
15- Finite Element Analysis of Framed Structures
16General Structural Problem
- For the following structure, we want to compute
- Bending moments, shear forces, axial forces
- Bending, shear and axial stresses and strains
- Deflections
17Finite Element Analysis (FEA), The basic idea
- Complex structures idealized through mathematical
models are too complicated to derive
relationships between applied loads, deflections
and internal stresses. - Hence discretize (anatomize) into many individual
finite elements of simpler form, e.g. a beams or
columns - Determine the relationship between load,
displacement, stresses and strains within a
finite element (analogize) - Assemble elements to satisfy equilibrium and
compatibility (analytical solution).
18(No Transcript)
19Modelling Idealization
20Analysis of bar elements
- Analytical solution review mechanics of
materials
21Analysis of bar elements (cont.) analytical
solution
- but
- for constant A
-
- Homogeneous sol. particular solution
22Analysis of bar elements (cont.)analytical
solution example
- -analyze the following structure
23Analysis of bar elements (cont.)analytical
solution example
24Analysis of bar elements (cont.)analytical
solution example
25For a finite element, we need to derive the
relationship between
- External Loads
- Deflections/deformations
- Internal stresses and strains
26 General basic steps for finite element
method applied here for bar elements
- Input
- Problem to be solved
- Physics of problem
- Mathematical model
- Processing
- Propose theory
- Formulate equations
- Solve equations
Output 1. Verify compatibility 2. Verify
equilibrium 3. Verify stress-strain relations
27(No Transcript)
28Input
- Problem analysis of a bar subjected to axial
loads - Physics bar subjected to axial
stress\deformation - Model define material, geometry and loading
29Processing 1. propose theory for solution
- a. Select element type axial, 2-node
- b. Select a displacement function
- (1) u generic displacement
302.1 propose theory for solution b. Select
element type and displacement functions
- b. con express u(x) in terms of nodal
displacements using boundary conditions. - (2)
312.1 propose theory for solution b. Select
element type and displacement functions
322.1 propose theory for solution
- c. Derive strain and stress displacement
relationships
332.2 formulate equations a. Derive the element
stiffness matrix and equations using principle of
Virtual Work (general derivation)
- (fj joint, fb equivalent joint)
loads
342.2 formulate equations a. Derive the element
stiffness matrix and equations
for uniform axial load b
352.2 formulate equations
- b. Transfer equations from local to global axes
(later) - c. Assemble element equations to obtain global
equations and introduce boundary conditions
362.3 Solve equations provide details
- a. Solve for the unknown DOF
- b. Solve for reactions
- c. Solve for element strains and stresses
373. Output
- Verify compatibility same as analytical
- Verify equilibrium reaction3X515KN ok
- Verify stress strain relationship
- -stress is constant unlike true solution which is
linear
38Ways of trial improvements?
- Use more subdivisions
- Use more terms, three node element (homework)
- c. Replace terms of homogenous with higher
order terms - Check each idea from above, propose more ideas,
carry out conclusions!
39End of FE for bars
Let Learning Continue
40- Ways of trial improvements?
41 General basic steps for FE methodreminder
- Input
- Problem to be solved
- Physics of problem
- Mathematical model
- Processing
- Propose theory
- Formulate equations
- Solve equations
Output 1. Verify compatibility 2. Verify
equilibrium 3. Verify stress-strain relationships
42Ways of trial improvements?
- Use more subdivisions
- Use more terms, three node element (homework)
- c. Replace terms of homogenous with other
terms - Check each idea from above, propose more ideas,
carry out conclusions!
43Ways of trial improvements?a.Use more
subdivisions
- To compare with previous solution
- Input, no change
- Processing
- Propose theory for solution
- A. Select element type no change, two elements
- B. Displacement function no change
- C. Stress-strain/displacement relations no
change - Formulate equations
- Element stiffness and equivalent nodal loads
change L3m to L1.5m - Local to global no change
- Assemble next
442.2 formulate equations
- c. Assemble the element equations and introduce
boundary conditions
452.3 Solve equations
- a. Solve for the unknown DOF
- b. Solve for reactions
462.3 Solve equations
- c. Solve for element strains and stresses
473. Output
- Verify compatibility same as analytical
- Verify equilibrium reaction3X515KN ok
- Verify stress strain relationship more
subdivision produces more accuracy.
48Ways of trial improvements? C. Replace some
terms with higher order terms
- To compare with original solution
- Input, no change
- Processing
- Propose theory for solution
- a. Discretize No change
- b. Element type and displacement functions
49Processing2.1 Propose theory
- c. Stress-strain/displacement relations
502.2 Formulate equations
- Derive element stiffness and equivalent joint
loads
512.2 formulate equations
- b. Transfer equations from local to global no
change - c. Assemble element equations and introduce
boundary conditions
522.3 Solve equations
- a. Solve for the unknown DOF
- b. Solve for reactions
532.3 Solve equations
- c. Solve for element strains and stresses
543. Output
- Verify compatibility different than analytical
- Verify equilibrium reaction3X515KN ok
- Verify stress strain relationship
- -stress is linear, however it is completely wrong
in values and slope
55(No Transcript)
56Sources of Error in the FEM (cont.)
- Three sources modeling (idealization
formulation and system), discretization and
solution (numerical) errors
- 1. Modeling errors result from
- -Either the use of elements or systems that
don't precisely describe the behavior of the
physical problem, - -Elements which are used to model physical
problems for which they are not suited are
sometimes referred to as ill-conditioned or
mathematically unsuitable elements. - -Using 1D or 2D instead of 3D systems
57Sources of Error in the FEM (cont.) 1. Modeling
error
- Example a FE with displacements varying in a
linear manner will produce no element modeling
error if used for linearly varying physical
problem, but would create a significant error if
used to represent a quadratic or cubic varying
displacement field.
58Sources of Error in the FEM
- 2. Discretization error can be related to
modeling the boundary shape, the boundary
conditions, etc.
Error due to poor geometry representation.
error effectively eliminated
59Sources of Error in the FEM
- 3. Solution (numerical error) includes truncation
errors and round off errors. Numerical error is
therefore a problem mainly concerning the FEM
buyers and developers. - The user can also contribute to the numerical
accuracy, for example, by specifying a physical
quantity, say Youngs modulus, E, to an
inadequate number of decimal places.
60Conclusions
- Discuss further ideas
- Continuous within and between elements
(compatibility) Co enforce displacement
continuity across common boundaries (example bar
element). C1 enforce both displacement and its
first derivative continuous across common
boundaries (example beam element). - Complete for equilibrium
- Rigid body motion for real
- Also discuss acceptance criterion (verification)
61SAP acceptance criterion (documentation\verificati
on\methodology)
62End of ways of improvements
Let Learning Continue
63- Theory of plane and space trusses
64Local versus global coordinates
- Element axes are not all the same.
- So there is a need for a coordinate
transformation
65Coordinate transformations from local to global
66Coordinate transformations from local to global
67Coordinate transformations from local to global
- Transformation matrix from global to local T is
- Notice for an orthogonal matrix, the inverse
transformation from local to global is quite easy
because
68Coordinate transformations from local to global
- -The element stiffness matrix in local
coordinates is - -since truss member connects two nodes, the
transformation
69Coordinate transformations from local to global
- -element equation in local coordinate
- -transform to global
- -multiply both sides by T-1
70Global stiffness matrix
71Plane truss example
- Analyze the truss shown, given
72Plane truss example (continues) Member connects
nodes 1?j
73Plane truss example (continues)
74Plane truss example (continues)
75Stress equation for PT
76Space truss
- Better approximate reality
- Analogize torsion in RC structures
- Analogize arches and domes.
77Space truss
78Space truss
79Space truss global stiffness
80Example
- Analyze the truss shown, given
81ST example (solution) Member connects nodes 1?j
82ST example (solution)
83ST example (solution)
84ST example (solution)
85ST example (solution)
- -equation of stress becomes
86ST example (solution)
87End of PT and ST analysis
Let Learning Continue
88 891. Condensation
- Reducing any number of DOFs up-to 1
DOF-equation. - Example solve
- First step partition into number of DOFs that
are needed to be removed, assume A, and those
which are needed to remain, assume B.
90Condensation of matrices (continues)
91Example
92Example (continue)
- Hw use matrix condensation to solve (hint
divide 2X2)
932.Symmetry
- We only need to deal with half the structure
around each symmetric plane. If two planes exist,
we work with ¼. - Example analyze if
- E20X103 MPa,
94Symmetry (continues)Notice 2 planes of symmetry
- -member 1 connects 1 to 5
95Symmetry (continues) Finding stresses
96Find reactions
97End of ways of reducing DOF's
Let Learning Continue
98Finite element for beams
99Analysis of beam elements-forces
- -Analytical solution
- review mechanics of materials
100Analysis of beam elements-deformations
101Analysis of beam elements
particular solution Homogeneous
solution
102Analysis of beam elements (cont.)analytical
solution example
103Analysis of beam elements (cont.)analytical
solution example (continued)
104 General basic steps for finite element method
beam elements
- Input
- Problem to be solved
- Physics of problem
- Mathematical model
- Processing
- Propose theory
- Formulate equations
- Solve equations
Output 1. Verify compatibility 2. Verify
equilibrium 3. Verify stress-strain relationships
105Input
- Problem analysis of the beam in the previous
example - Physics shear and moment deformations
- Mathematical model material, geometry and
loading
106Processing 2.1 propose theory for solution
- a. Select element type beam, 1 element
- b. Select displacement functions
- v generic displacement
1072.1 propose theory for solution b. Select
displacement functions
108Shape Functions
1092.1 propose theory for solution c. Derive strain
and stress displacement relationships
1102.2 formulate equations
- Derive element stiffness using principle of
Virtual Work
1112.2 formulate equations a. Derive the element
stiffness matrix and equations
1122.2 formulate equations a. Derive the element
stiffness matrix and equations
- For a uniform downward load p0
- b. Transfer from local to global axes (No need)
- c. Assemble the element equations to obtain the
global equations and introduce boundary
conditions
113Example analytical solution
114Example FE solution structure one element
SIMILAR CONCLUSIONS AS BEFORE!
115Example FE solution discretize to two elements
116Example FE solution discretize to two elements
117Example FE solution discretize to two elements
118FE versus analytical
119End of fe for beams
Let Learning Continue
120- Finite Element Analysis of framed structures
121Plane framesprinciple of superposition
122Plane framesequivalent joint loads
123Plane framestransformations
124Plane framesglobal stiffness
125Example 1
- Given E18 GPa,
- A 0.05m2, I .0003m4
- Required analyze frame
- Solution
- Member 1 2 ? 1
126Example 1(continues)
127Example 1 (continues)
128Example 1(continues)
129Example 1 (continues)
Member 2 3 ? 2
130Example 2
- Example- Repeat previous example neglecting
axial deformations - Member 1
- Member 2
131End of FE for frames
Let Learning Continue
132Finite Element Analysis of grid structures
- Dr. Abdul Razzaq Touqan
- Department of Civil Engineering
133Analysis of torsional elements
- Analytical solution review mechanics of
materials
134Analysis of torsional elements (cont.)analytical
solution
- but
- Homogeneous solution particular solution
135Analysis of torsional elements (cont.)analytical
solution example
- -analyze the following structure
136Analysis of torsional elements (cont.)analytical
solution example
137(No Transcript)
138(No Transcript)
139 General basic steps for finite element method
applied here for grid elements
- Input
- Problem to be solved
- Physics of problem
- Mathematical model
- Processing
- Propose theory
- Formulate equations
- Solve equations
Output 1. Verify compatibility 2. Verify
equilibrium 3. Verify stress-strain relationships
140Input
- Problem to be solved analysis of a shaft
subjected to torsion loads - Understand physics shaft will be subjected to
shear stresses and torsional deformations - Mathematical model define material, geometry and
loading
141Processing 1. propose theory for solution
- a. Select element type torsional element
(2-node) - b. Select displacement functions
- (1) f generic displacement
1422.1 propose theory for solution
- b. continues express u(x) in terms of nodal
displacements using boundary conditions
1432.1 propose theory for solution
c. Derive strain and stress displacement
relationships
1442.1 propose theory for solution
c. Derive strain and stress displacement
relationships
1452.2 formulate equations
- a.Derive the element stiffness matrix and
equations using principle of Virtual Work
(general derivation) - J Polar moment of inertia for torsional members
1462.2 formulate equationsa.Derive the element
stiffness matrix and equations
1472.2 formulate equations
- b. Transfer equations from local to global axes
(later) - c. Assemble the element equations to obtain the
global equations and introduce boundary
conditions
1482.3 Solve equations
- a. Solve for the unknown DOF
- b. Solve for reactions
- c. Solve for resultant element strains and
stresses
1493. Output
- Verify compatibility same as analytical
- Verify equilibrium reactiontL ok
- Verify stress strain relationship
- -resultant torque is constant unlike true
solution which is linear
150Gridprinciple of superposition
151Gridequivalent joint loads
152Gridtransformations
153Gridglobal stiffness
154Example 1
Given Required analyze grid Solution
155Example 1(continues)
Member 2
156Example 1 (continues)
157Example 1 (continues)
158End of FE for grids
Let Learning Continue
159Discussion class
160Kjj
161Kkj
162Kkk
163Equivalent joint loads
164Transformation matrix
165End of FE for space frame
Let Learning Continue
166Space-frame versus Shell
1671D Space-frame
1682D -Shell
169Methodology
- Understand exact 1D
- Perform analogical solutions between 1D and nD
models - Build up experience with 3D models
170Example
- Cantilever beam 3m span, 0.2m width by 0.3m depth
made of concrete with weight density 2.5t/m and
E2.5X10 t/m2. Find end span deflection and
maximum stress at fixed end using - Exact solution
- Finite element solution
- 1D space frame model
- 2D shell model
- 3D solid model
171End of FE for 2D and 3D
Let Learning Continue
172 173Definition
174Structural modeling assumptions
- Elements
- 1D
- 2D
- 3D
- Structures
- 1D structure with 1D elements
- 2D structure with 1D, 2D elements
- 3D structure with 1D, 2D, 3D elements
175Loading assumptions
1761D, 2D or 3D modeling 1D slab-beam-column
1771D, 2D or 3D modeling 1D slab-beam-column/
continued
1781D, 2D or 3D modeling 2D plane frame
1791D, 2D or 3D modeling 3D space frame with
slabs/walls
180Laws versus theories
- Structural Analysis laws
- constitutive (stress-strain) relationships
- equilibrium equations essential
- compatibility equations optional/present
challenge - Structural Analysis theories
- Based on assumptions
- Assumptions based on available knowledge
- Available knowledge is constrained with available
tools like hand calculators and personal
computers - Computer programs are based on assumptions on
which the theoretical basis of the software was
developed.
181Structural analysis system approach
- Input
- 1.Goal
- 2.Given
- 3.Create a mathematical model
- Processing
- 1.Propose a theory
- 2.Formulate equations
- 3.Solve the equations
- Output
- 1.Verify Compatibility
- 2. Verify equilibrium.
- 3. Verify stress-strain relationships.
182Structural analysis system approach input
- Input
- 1.Goal (ref creation versus man made),
provide - support system (science)
- functional system (facts)
- 2.Given (ref global versus local), understand
physics of the system, i.e. specify system
requirements - strength, stiffness and stability
- function, freedom and form
- 3.Create a mathematical model (ref codes of
practice), specify structural system that
provide - satisfaction
- fairness
-
183Structural analysis system approach processing
- Processing
- 1.Propose a theory reduce assumptions deviating
model from reality\ref 3D nonlinear dynamic
probabilistic soil-soil GIS interaction - 2.Formulate equations according to state of
knowledge and available tools \ref analytical,
anatomical, analogical - 3.Solve the equations analyze the structure\
ref verify analysis laws
184Structural analysis system approach output
- Output
- 1.Verify Compatibility
- 2. Verify equilibrium.
- 3. Verify stress-strain relationships.
185example
- A single storey RC slab-beam factory structure
shown next page - Fixed foundations, 4 spans 5m bays in x and a
single 8m span in y, 6m elevation - E24GPa, µ0.2, ?2.5t/m3
- Slab 25cm thickness, drop beams 30cmX80cm,
columns 30X60cm - superimposed loads5kN/m2, live load9kN/m2
- Due to cracking of elements, assume modifiers
for gross inertia - Beam 0.35
- Column 0.7
- One way slab (0.35, 0.035)
186(No Transcript)
187Analogical 1D analysis slab model
1881D analysis slab analysis
1891D analysis slab analysis
- Wd(.2524.55)11.125KN/m
- Wl9KN/m
- Wu1.211.1251.6927.75KN/m
1901D analysis slab analysis, values of bending
moment KN.m
1911D analysis slab analysis, values of reactions
in KN
1921D analysis beam and column analysis,
- Assume simply supported beam
- Beam C, Mu(1291.20.3.824.5)82 /81079
- Beam B, Mu(1591.20.3.824.5)82 /81319
- Beam A, Mu(54.521.20.3.824.5)82 /8492
- Column reactions
- Beam C, Ru (1290.3.824.5)8 /2540
- Beam B, Ru(1590.3.824.5)8 /2660
- Beam A, Ru (54.50.3.824.5)8 /2242
193Gravity equilibrium checks
- D
- Slab20X8X(0.25X24.55)1780KN
- Beams(5X82X20)X.8X.3X24.5470KN
- Columns10X6X.3X.6X24.5264.6KN
- Sum2514.6KN
- L
- R 20X8X91440KN
194Gravity equilibrium checks
195(No Transcript)
196(No Transcript)
197Homework
- Analyze and design a one story reinforced
concrete structure (entertainment hall) made of
one way solid slab sitting on drop beams
supported on six square columns 50cm dimensions.
The superimposed loads are 300kg/m2, and the live
load equal to 400kg/m2 - Analyze using analogical (local practice
slab-beam-column load path) - Analyze using anatomical finite element 3D model
(more actual representation) - Compare analogical with anatomical
198(No Transcript)
199End of structural modelings
Let Learning Continue