breve intro a e.f
TRANSCRIPT
Chapter 1
Introduction to Finite Element
Analysis
1.1 Introduction
Finite element analysis is a numerical method is a numerical method that is used
in the analysis of complex mechanical and structural problems. It is used to
obtain numerical solutions and can be applied over a wide range of objective
functions and over a wide range of loading conditions. The method can be used
for analysis of static problems and dynamic problems and can even be used to
analyse linear systems and non-linear systems. A linear system is a type of
system where a linear relation-ship exists between the force and deflection and
these systems do not take plastic deformation. A non-linear system is a type of
system that takes plastic deformation into account and can allow testing all the
way up to the point(s) of fracture.
Finite element analysis is not restricted to mechanical and structural analysis.
In the methods most basic form, it can be applied to problems that involve heat
loss, fluid flow and even electric potential. Due to this wide range of application,
the method is very popular. The method is used in the analysis of many complex
two-dimensional problems, but can also be applied for problems involving three-
dimensions. As problems become more complex with finite element analysis, they
can take longer to solve. Thankfully, computers can be used to solve complex
1
1. Introduction to Finite Element Analysis
finite element problems and a wide range of finite element analysis software exists
today (LUSAS, ABAQUS, ANSYS).
Figure 1.1: Computer finite element analysis of an I-Beam
The method works by taking a problem and modelling it into smaller elements
through the use of nodes and elements. Physical and geometric properties are
assigned to the elements and loads and displacements are applied to the nodes.
From that, a finite element analysis is carried out in an attempt to earn numerical
values. The best way to see finite element analysis in action, is to see how it is
applied and used to solve simple basic problems.
1.2 Worked Example 1
The following worked example will be used to develop a one degree of freedom
finite element. This simple problem will be solved and then the results will
be compared to a theoretical solution. Consider a cantilever L. It is axially
restrained at one end and subjected to a axial load W at the other end. The
member has a continuous cross-sectional area of A and a Youngs Modulus of E.
The theoretical stress of the beam, σ, can be obtained using
σ =Force
Area(1.1)
2
1. Introduction to Finite Element Analysis
Figure 1.2: Axially loaded member
The load applied is W and the cross-sectional area is A. Therefore the equa-
tion becomes
σ =W
A(1.2)
The theoretical strain of the beam, ǫ, can be calculated using
E =σ
ǫ(1.3)
Rearranging
ǫ =σ
E(1.4)
Substituting the theoretical stress obtained previously, equation 1.1, into
equation 1.4 to give
ǫ =W
EA(1.5)
With the theoretical results obtained, the finite element method can now be
applied to the cantilever beam. The results obtained from the finite element
method can then be compared to the theoretical results obtained for stress and
strain.
1.2.1 Step 1 - Creating an element
In order to apply the finite element method, the problem shown in Figure 1.1
must be modeled as a finite element. As the beam is a simple cantilever, with
a uniform cross sectional area, we can model this as one single element. This
is achieved by first applying a node to the fixed end of the beam and a node
3
1. Introduction to Finite Element Analysis
at the free end of the beam. An element can be created between the nodes, as
demonstrated in Figure 1.3.
Figure 1.3: Cantilever beam as an element
The element has a length of L and has two nodes, nodes 1 and 2, at either
end of the beam. Each node has one degree of freedom, since the original problem
is such that any point on the member has one degree of freedom, u. The finite
element model is then used to obtain a solution.
1.2.2 Step 2 - Nodal displacements
Now that a finite element model has been established, the nodal displacements
are now taken into consideration, however, the restraints of the cantilever beam
are ignored for the timed being to create and understand the general relationship
for displacement along the cantilever beam.
Figure 1.4: Axially loaded member
If a horizontal load is applied to the element, node 1 will displace to 1 ′ and
node 2 will displace to 2 ′. The displacement of node 1 to 1 ′ can be considered
as u1 and the displacement of node 2 to 2 ′ can be considered as u2. This is
illustrated in Figure 1.4. These displacements can be written as the displacement
vector, u.
u =
[
u1
u2
]
(1.6)
4
1. Introduction to Finite Element Analysis
The displacement along the member is required. Since there are two nodal
displacements, a function with two unknowns can be used (usually, but not al-
ways, a complete polynomial is used). Thus, the displacement function can be
written as.
u = a1 + a2x (1.7)
Where u is the displacement at x along element 1 2, a1 and a2 are unknown
constants and x is a co-ordinate along the beam. Using the displacement function,
we can determine a1 and a2 in terms of the displacements u1 and u2.
At node 1, x = 0. Substituting into the displacement function, 1.7,
u1 = a1 + a2(0) (1.8)
Therefore
u1 = a1 (1.9)
Now consider the other node. At node 2, x = L. Substituting into the
displacement function,
u2 = a1 + a2L (1.10)
Previously, the coefficient a1 was determined. Therefore,
u2 = u1 + a2L (1.11)
Rearranging it terms of the constant a2,
L =u2 − u1
L(1.12)
With both a1 and a2 determined, they can be substituted back into the dis-
placement function, to give
u = u1 +u2 − u1
Lx (1.13)
5
1. Introduction to Finite Element Analysis
1.2.3 Step 3 - Strain
With the displacement function obtained, the strain can be determined along the
member. The strain in the member, ǫ, can be calculated using
ǫ =Change in Length
Original Length(1.14)
Figure 1.5 illustrates the strain along the member
Figure 1.5: Strain along the member
From Figure 1.5, the original length of the beam is denoted as dx. Node 1 is
displaced by u and node 2 is displaced by u+ ∂u
∂xdx. Therefore,
Change in Length = u+∂u
∂xdx− u (1.15)
This simplifies to
Change in Length =∂u
∂xdx (1.16)
With expressions for the change in length and the original length, an expres-
sion for the strain can be derived.
ǫ =Change in Length
Original Length=
∂u
∂xdx
dx(1.17)
This simplifies to
ǫ =∂u
∂x(1.18)
Previously, an expression for displacement, u, was derived to give equation
1.13. By differentiating this with respect to x, we obtain
6
1. Introduction to Finite Element Analysis
∂u
∂x= (0) +
u2 − u1
L(1) (1.19)
∂u
∂x=
u2 − u1
L(1.20)
Therefore,
ǫ =u2 − u1
L(1.21)
This expression can be expressed in matrix format
ǫ =[
−1L
1L
]
[
u1
u2
]
(1.22)
Consider the the matrix
[
−1L
1L
]
This is the strain-displacement matrix and can be written as [B]. Now con-
sider the vector
[
u1
u2
]
This is the displacement vector and can be written as u. Therefore, the
expression for strain can be written as
ǫ = [B] u (1.23)
In this worked example, we are dealing with a one dimensional problem, there
is only one strain. Later we will be looking at two dimensional and three di-
mensional situations where there can be 3 or 6 strains at any point in a body,
respectively.
7
1. Introduction to Finite Element Analysis
1.2.4 Step 4 - Stress
With an expression for the strain, the stress in the element can now be determined.
The expression to obtain stress is given as
σ = [D] ǫ (1.24)
[D] is the elasticity matrix and the notation is used by convention. As we
are dealing with a one dimensional problem, only one stress will be obtained.
However, σ, can contain a number of stresses depending on the problem. In
this case, as there is only one stress, the matrix [D] only contains one term the
modulus of elasticity E.
σ = E ǫ (1.25)
Previously, an expression for strain was derived. Substituting that into ex-
pression 1.25 yields,
σ = E [B] u (1.26)
1.2.5 Step 5 - Nodal Loads
The nodal loads now need to be related to the internal stresses.
Figure 1.6: Nodal loads
This yields the nodal force vector
P =
[
P1
P2
]
(1.27)
However, rather than apply just to the particular loading shown above a
general relationship will be derived. This can be done with the aid of virtual
work.
8
1. Introduction to Finite Element Analysis
1.2.6 Step 6 - Application of Virtual Work
Virtual work can be used to relate the external forces acting on the beam with
the internal forces acting from within the beam.
External Work Done = Internal Work Done (1.28)
Given a set of external applied forces, P , we can apply a set of virtual dis-
placements, u. Therefore, the external work applied can be expressed as
External Work Done = uT P (1.29)
As for the internal work, the internal stresses, σ, equilibrate P and the internal
strains, ǫ, are compatible with u. From this, the internal work can be calculated
using
Internal Work Done = ǫT σ Per Unit Vol (1.30)
This becomes
Internal Work Done =
∫
V
ǫT σ dV (1.31)
Now, consider the expression for strain that was derived previously.
ǫ = [B] u (1.32)
By taking the transpose of the strain, we obtain
ǫT = [B]T uT (1.33)
This can then be substituted into expression 1.31 to give
Internal Work Done =
∫
V
uT [B]T σ dV (1.34)
This expression can be simplified further by using the expression for strain,
expression 1.26, that was derived previously. This gives
9
1. Introduction to Finite Element Analysis
Internal Work Done =
∫
V
uT [B]T E [B] u dV (1.35)
We can now apply equilibrium.
External Work Done = Internal Work Done (1.36)
uT P =
∫
V
uT [B]T E [B] u dV (1.37)
uT is not a function of the geometry of the member. Therefore, this can be
rearranged to obtain
uT P = uT
∫
V
[B]T E [B] dV u (1.38)
Let∫
V
[B]T E [B] dV = [k] (1.39)
This is the stiffness matrix. Therefore
uT P = uT [k] u (1.40)
This can be simplified further to obtain the final expression
P = [k] u (1.41)
Where P is the nodal force vector, [k] is the element stiffness matrix and u
is the nodal dispalcement vector. As mentioned previously, the important point
to note is that the derivation is completely general. No assumptions have been
made regarding the loads, element type, material properties, etc.
1.2.7 Step 7 - Stiffness Matrix
In order to solve the problem, we mus first derive the stiffness matrix. For our 2
noded element, consider the stiffness matrix expression
10
1. Introduction to Finite Element Analysis
Let
[k] =
∫
V
[B]T E [B] dV (1.42)
Rearranging
[k] = [B]TE[B]
∫
V
dV (1.43)
Now, consider
∫
V
dV (1.44)
This is the equivalent of
∫ ∫ ∫
dxdydz (1.45)
By integrating with respect to y and z, we obtain
∫
L
O
dxyz (1.46)
yz is the equivalent of the area, A. Therefore
∫
L
O
dxA = A
∫
L
O
dx = AL (1.47)
Substituting
[k] = [B]TE[B]AL (1.48)
Now consider the strain-displacement matrix [B]. From previous, it was found
[B] =[
−1L
1L
]
By taking the transpose of the strain-displacement matrix, we obtain
[B]T =
[
−1L
1L
]
Substituting
11
1. Introduction to Finite Element Analysis
[k] =
[
−1L
1L
]
E[
−1L
1L
]
AL (1.49)
Multiplying through, we obtain
[k] =
[
EA
L−
EA
L
−EA
L
EA
L
]
(1.50)
This can then be used to solve the original problem.
1.2.8 Step 8 - Solution
Using what has been derived in the previous steps, we can now return to the
original problem.
Figure 1.7: Original Problem
Figure 1.8: Finite Element
This problem can be solved using
P = [k]u (1.51)
From previous, the nodal load vector is
P =
[
P1
P2
]
(1.52)
12
1. Introduction to Finite Element Analysis
The displacement vector is
u =
[
u1
u2
]
(1.53)
The stiffness matrix is
[k] =
[
EA
L−
EA
L
−EA
L
EA
L
]
(1.54)
Substituting these into equation 1.51, we obtain
[
P1
P2
]
=
[
EA
L−
EA
L
−EA
L
EA
L
][
u1
u2
]
(1.55)
We can now apply the beams properties, known displacements and known
loads to expression 1.55. In this case
u1 = 0 (Node 1 is restrained and cannot be displaced)
u2 = Unknown
P1 = Unknown
P2 = W
E = E
A = A
L = L
13
1. Introduction to Finite Element Analysis
Therefore
[
P1
W
]
=
[
EA
L−
EA
L
−EA
L
EA
L
][
0
u2
]
(1.56)
Multiplying the matrices, we obtain
P1 = −EA
Lu2 (1.57)
W =EA
Lu2 (1.58)
Rearranging expression 1.58, we obtain an expression for u2
u2 =WL
EA(1.59)
This can then be used with expression 1.57 to find the unknown nodal force.
P1 = −EA
L
(
WL
EA
)
= −W (1.60)
u2 can also be used to obtain the strain within the beam. Using expression
1.2.3
ǫ =[
−1L
1L
]
[
0WL
EAu2
]
(1.61)
Multiplying through, the expression for strain becomes
ǫ =W
EA(1.62)
Using the strain, the stress can be obtained.
σ = [D]W
EA(1.63)
As mentioned previously, [D] contains only one value, E, as we are dealing
with a one dimensional problem. Therefore, the stress becomes
σ =W
A(1.64)
14
1. Introduction to Finite Element Analysis
Using basic finite element method theory, it is clear to see that the results
obtained through the finite element method agree with the theoretical results
that were calculated earlier.
15
1. Introduction to Finite Element Analysis
1.3 Worked Example 2
Now consider the following problem.
Figure 1.9: Axially loaded member with two loads applied
We can use finite element analysis to obtain the stresses and strains within
this beam. However, this beam now has two applied loads at different lengths of
the beam - one at the end of the beam and one half way across the length of the
beam. In worked example 1, we were able to solve the problem by modelling the
entire beam with a single two-noded element. However, if we were to do that for
this case, we would neglect the load applied half way across the beam. In order
to solve this, we must divide the structure with two two-noded elements. Each
element has a length of L
2.
16
1. Introduction to Finite Element Analysis
Figure 1.10: The original problem and the equivalent finite element model
From this we can establish a relationship between the loads and displacements
using
P = [k] u (1.65)
However, we cannot apply this directly to the entire member. We must first
apply the relationship to the individual elements and then use them to create an
overall relationship for the beam.
1.3.1 Element 1
Consider element 1 and apply the unique nodal loads to the element.
We can obtain the the nodal force vector, the displacement vector and the
stiffness matrix to form a relationship using 1.65. The nodal vector for element
1 is given as
P (1) =
[
P(1)1
P(1)2
]
(1.66)
The displacement vector for element 1 is
17
1. Introduction to Finite Element Analysis
Figure 1.11: Element 1
u(1) =
[
u(1)1
u(1)2
]
(1.67)
Considering the element length is L
2, the cross sectional area is A and Young’s
modulus is E, the stiffness matrix is for element 1 is
[k(1)] =
[
2EA
L−
2EA
L
−2EA
L
2EA
L
]
(1.68)
Therefore, using 1.65, the relationship for element 1 is
[
P(1)1
P(1)2
]
=
[
2EA
L−
2EA
L
−2EA
L
2EA
L
] [
u(1)1
u(1)2
]
(1.69)
1.3.2 Element 2
Now consider element 2 and apply the unique nodal loads to the element.
The nodal vector for element 2 is given as
P (2) =
[
P(2)2
P(2)3
]
(1.70)
The displacement vector for element 2 is
18
1. Introduction to Finite Element Analysis
Figure 1.12: Element 2
u(2) =
[
u(2)2
u(2)3
]
(1.71)
Considering the element length is L
2, the cross sectional area is A
2and Young’s
modulus is E, the stiffness matrix is for element 2 is
[k(2)] =
[
EA
L−
EA
L
−EA
L
EA
L
]
(1.72)
The relationship for element 2 is
[
P(2)2
P(2)3
]
=
[
EA
L−
EA
L
−EA
L
EA
L
] [
u(2)2
u(2)3
]
(1.73)
1.3.3 Overall Relationship
We can now create an overall relationship for the member. However, before we
can do this, we must first apply compatibility. We must first obtain the structure
load vector, structure displacement vector and structure stiffness matrix, which
will then be used to obtain an overall relationship for the member.
F = [K] δ (1.74)
Where
19
1. Introduction to Finite Element Analysis
F = Structure load vector
[K] = Structure stiffness matrix
δ = Structure displacement vector
20
1. Introduction to Finite Element Analysis
Figure 1.13: Compatibility
Consider the forces. We can apply compatibility to create relationships for
the forces that affect each node. Therefore,
F1 = P(1)1
F2 = P(1)2 + P
(2)2
F3 = P(2)3
Therefore, the structure load vector becomes
F =
F1
F2
F3
(1.75)
Considering the displacements and applying compatibility, we obtain
δ1 = u(1)1
21
1. Introduction to Finite Element Analysis
δ2 = u(1)2 = P
(2)2
δ3 = u(2)3
Therefore, the structure displacement vector becomes
δ =
δ1
δ2
δ3
(1.76)
Now consider the stiffness matrix. The stiffness matrix from element 1 can be
combined with the stiffness matrix from element 2 to give the structure stiffness
matrix.
[K] =
2EA
L−
2EA
L0
−2EA
L
(
2EA
L+ EA
L
)
−EA
L
0 −EA
L
EA
L
(1.77)
This simplifies to give
[K] =EA
L
2 −2 0
−2 3 −1
0 −1 1
(1.78)
Therefore, the overall relationship for the member is
F1
F2
F3
=
EA
L
2 −2 0
−2 3 −1
0 −1 1
δ1
δ2
δ3
(1.79)
22
1. Introduction to Finite Element Analysis
1.3.4 Solution
As a relationship between the loads and the displacements has been derived for
the member, we can now solve the problem. We can apply the known loads and
displacements that act on the member. By inspection, the forces are
F1 = Unknown
F2 = 2W
F3 = W
By inspection, the displacements are
δ1 = 0
δ2 = Unknown
δ3 = Unknown
Using these, expression 1.79 becomes
F1
2W
W
=
EA
L
2 −2 0
−2 3 −1
0 −1 1
0
δ2
δ3
(1.80)
Multiplying through, we obtain
F1 = −2EAδ2
L(1.81)
2W =3EAδ2
L−
2EAδ3
L(1.82)
23
1. Introduction to Finite Element Analysis
W = −EAδ2
L+
EAδ3
L(1.83)
Solving gives gives the unknown force and the unknown displacements
F1 = −3W (1.84)
δ2 =3WL
2EA(1.85)
δ3 =5WL
2EA(1.86)
These values can then be used to obtain expressions for stress and strains for
the two elements.
For element 1, the strain obtained is
ǫ(1) =3W
EA(1.87)
And the stress for element 1 is
σ(1) =3W
A(1.88)
For element 2, the strain obtained is
ǫ(2) =2W
EA(1.89)
And the stress for element 2 is
σ(2) =2W
A(1.90)
24
1. Introduction to Finite Element Analysis
1.4 Worked Example 3
So far, we have dealt with problems where point loads have been applied to
beams. Consider the following problem.
Figure 1.14: Member with UDL applied
We now have a cantilever beam of length L where a load of w per unit length
has been applied. Young’s modulus, E, and the cross-sectional area, A, are
uniform through out the beam. We can still use finite element analysis to calculate
the stresses and strains within the member.
1.4.1 Theoretical results
We can first use basic theory to obtain results.
The theoretical stress of the beam, σ, can be obtained using
σ =Force
Area(1.91)
The load applied is w over length L and the cross-sectional area isA. Therefore
σ =wL
A(1.92)
The theoretical strain of the beam, ǫ, is calculated using
ǫ =σ
E(1.93)
25
1. Introduction to Finite Element Analysis
This becomes
ǫ =wL
EA(1.94)
Figure 1.15: Theoretical stress along the member
If we plot the theoretical stress along the member, as shown in figure 1.15, the
maximum stress occurs at the fixed end of the beam. Moving down the beam, the
stress decreases linearly. At the end of the beam, no stress exists on the beam.
Finite element analysis will then be used to obtain the stresses and strains
and the results obtained will be compared to the theoretical results.
1.4.2 Modelling with a single two-noded element
This problem is similar to worked example 1. Therefore, we will model this using
one two-noded element.
As this is modeled using a single two-noded element, it is clear to see that the
element force vector, element stiffness matrix and element displacement vector
are the same as the structure load vector, structure stiffness matrix and structure
26
1. Introduction to Finite Element Analysis
Figure 1.16: Two-noded element
displacement matrix respectively. Therefore, we can easily create a relationship
between the forces and displacements that will occur on the member.
[
F1
F2
]
=EA
L
[
1 −1
−1 1
] [
δ1
δ2
]
(1.95)
We can then apply the known displacements to the relationship. By inspec-
tion, the displacements are
δ1 = 0
δ2 = Unknown
We can also apply the known forces to the relationship. However, as we are
dealing with a distributed load, we must take greater care. First consider node
2. Whilst the distributed load acts on node 2, not all of it does. Therefore, the
force acting on node 2 is
27
1. Introduction to Finite Element Analysis
F2 =wL
2
Now consider node 1. As with node 2, not all of the distributed load acts
on this node. However, we must also take into consideration that there is an
unknown reaction at node 1. Therefore, the forces that act on node 1 can be
expressed as
F1 = R1 +wL
2
Where R1 is the unknown reaction. Using these, 1.95 becomes
[
F1
wL
2
]
=EA
L
[
1 −1
−1 1
] [
0
δ2
]
(1.96)
Solving gives
F1 = −EA
Lδ2 (1.97)
δ2 =wL2
2EA(1.98)
Substituting 1.98 into 1.97 gives
F1 = −wL
2(1.99)
This can then be used to obtain the unknown reaction, R1
F1 = R1 +wL
2(1.100)
Therefore
R1 = −wL (1.101)
The stress and strain for the member can now be obtained. Using expression
1.98, the strain is given as
28
1. Introduction to Finite Element Analysis
ǫ =wL
2EA(1.102)
And the stress is given as
σ =wL
2A(1.103)
Figure 1.17 illustrates the results obtained, plotted against the theoretical
results.
Figure 1.17: Single two noded element stress vs. theoretical stress
Whilst the unknown reaction obtained through finite element analysis agrees
with the theoretical result, the results for the stress and strain obtained through
finite element do not agree with the theoretical stress and strain. We can see that
the results obtained through theory show maximum stress occurring at the fixed
end of the beam and the stress decreases linearly through the beam. The results
obtained through finite element analysis show that the expressions for stress and
strain are uniform throughout the beam, which is not the case, although the finite
element results do accurately predict the stress and strain at the midpoints of
the beam. However, these are artifacts of the analysis and may not necessarily
be true.
Whilst the results obtained do not agree with theory, it does not mean that
29
1. Introduction to Finite Element Analysis
we cannot use finite element analysis with problems where a distributed load
is applied. It means that we must model our problem much more carefully to
improve the accuracy of the analysis. Now, we will try to model the problem
using two, two-noded elements.
1.4.3 Modelling with two, two-noded elements
We will now model the problem using two two-noded elements, each with lengthL
2.
Figure 1.18: Two two-noded finite element model
As in worked example 2, we must first determine the element load vector,
element displacement vector and element stiffness matrix and create the load-
30
1. Introduction to Finite Element Analysis
displacement relationship for each element before establishing an overall load-
displacement relationship for the structure. For element 1, the element load
vector is
P (1) =
[
P(1)1
P(1)2
]
(1.104)
The displacement vector for element 1 is
u(1) =
[
u(1)1
u(1)2
]
(1.105)
Considering the element length is L
2, the cross sectional area is A and Young’s
modulus is E, the stiffness matrix is for element 1 is
[k(1)] =
[
2EA
L−
2EA
L
−2EA
L
2EA
L
]
(1.106)
Therefore, the load-displacement relationship for element 1 is
[
P(1)1
P(1)2
]
=
[
2EA
L−
2EA
L
−2EA
L
2EA
L
] [
u(1)1
u(1)2
]
(1.107)
As for element 2, the nodal vector for element 2 is
P (2) =
[
P(2)2
P(2)3
]
(1.108)
The displacement vector for element 2 is
u(2) =
[
u(2)2
u(2)3
]
(1.109)
Considering the element length is L
2, the cross sectional area is A and Young’s
modulus is E, the stiffness matrix is for element 2 is
[k(2)] =
[
2EA
L−
2EA
L
−2EA
L
2EA
L
]
(1.110)
31
1. Introduction to Finite Element Analysis
The relationship for element 2 is
[
P(2)2
P(2)3
]
=
[
2EA
L−
2EA
L
−2EA
L
2EA
L
] [
u(2)2
u(2)3
]
(1.111)
In order to create an overall load-displacement relationship, we must first
apply compatibility.
Figure 1.19: Compatibility
Compatibility for the forces
F1 = P(1)1
F2 = P(1)2 + P
(2)2
F3 = P(2)3
The structure load vector becomes
32
1. Introduction to Finite Element Analysis
F =
F1
F2
F3
(1.112)
Displacement compatibility
δ1 = u(1)1
δ2 = u(1)2 = P
(2)2
δ3 = u(2)3
Structure displacement vector
δ =
δ1
δ2
δ3
(1.113)
Now consider the stiffness matrix. The stiffness matrix from element 1 can be
combined with the stiffness matrix from element 2 to give the structure stiffness
matrix.
[K] =
2EA
L−
2EA
L0
−2EA
L
(
2EA
L+ 2EA
L
)
−2EA
L
0 −2EA
L
2EA
L
(1.114)
This gives
[K] =2EA
L
1 −1 0
−1 2 −1
0 −1 1
(1.115)
Therefore, the overall relationship for the structure is
33
1. Introduction to Finite Element Analysis
F1
F2
F3
=
2EA
L
1 −1 0
−1 2 −1
0 −1 1
δ1
δ2
δ3
(1.116)
We can then solve this by applying known loads and forces. By inspection,
the known applied loads at each node are
F1 = R1 +wL
4
F2 =wL
2
F3 =wL
4
The known displacements are
δ1 = 0
δ2 = Unknown
δ3 = Unknown
Therefore
F1
wL
2
FwL4
=
2EA
L
1 −1 0
−1 2 −1
0 −1 1
0
δ2
δ3
(1.117)
Solving this gives
F1 = −6wL
8(1.118)
34
1. Introduction to Finite Element Analysis
δ2 =3wL2
8EA(1.119)
δ3 =wL2
2EA(1.120)
These are then used to obtain the unknown reaction, R1.
R1 = −wL (1.121)
The strains and stresses can then be calculated for each element. For element
1, the strain is calculated as
ǫ(1) =3wL
4EA(1.122)
And the stress in element 1 is
σ(1) =3wL
4A(1.123)
For element 2, the strain is calculated as
ǫ(2) =wL
4EA(1.124)
And the stress is
σ(2) =wL
4A(1.125)
Now, withe reference to figure 1.20, we can compare the results obtained with
the two two-noded elements, to the theoretical results and the single two-noded
element.
35
1. Introduction to Finite Element Analysis
Figure 1.20: Two two-noded finite element model
By dividing the structure into two elements, the accuracy of the solution
has improved. However, it still does not entirely match the theoretical results.
Therefore it can be said that by splitting the model up with more elements,
the solution approaches the theoretical solution. However, by doing so, this will
also increase the size of the structure stiffness matrix. Increasing the size of the
stiffness matrix also increases the complexity of the problem and will take more
time to evaluate, hence the need for a computer.
36