Te ZZZZZZZ ZZZZZZZZ
Te ZZZZZZZ ZZZZZZZZ
Te ZZZZZZZ ZZZZZZZZ
TRAKYA UNIVERSITY
ENGINEERING FACULTY
MECHANICAL ENGINEERING
PROJECT- 2
May 2018
EDİRNE
ABSTRACT
Truss is one of the important components for a structure and need to be designed in such
a way that they have enough strength and rigidity to satisfy the strength and serviceability
limitation. Element analysis has been reported to be suitable tool to analysis the truss.
This is because finite element analysis is cheaper and easier to conduct compared to full
scale test. This report presents the development of truss finite element analysis software
using ANSYS.
In this Project, the planar truss systems, finite element method and Ansys software have
been introduced. Axial force, strain and stresses have been analyzed via Ansys. The
results have been compared with that results in which thesis have analytical solution.
Then observed the impact of the analysis results by increasing the number of nodes. As
the number of nodes increased, I realized that the results of the analysis were closer to
reality.
Key words:
Finite element method is one of the main methods used in solving engineering problems.
Many structural analysis programs operate using this finite element method. Using
ANSYS, which is one of these software, an engineering problem in real life is solved and
a precaution is taken. In this study, we aim to analyze an example of an engineering
problem.
My thesis advisor who gives me the best guidance by supporting me at every step of this
project Prof. Dr. Metin Aydoğdu I offer my sincere thanks to my teacher.
CONTENTS
ABSTRACT...................................................................................................................... 2
FOREWORD AND THANKS ......................................................................................... 3
LIST OF FIGURES .......................................................................................................... 6
LIST OF TABLES ............................................................................................................ 7
1. ABOUT FINITE ELEMENT METHOD ..................................................................... 8
1.1. A HISTORY OF FINITE ELEMENT ANALYSIS .............................................. 8
1.2. WHAT IS FEA? ................................................................................................... 10
2. ANSYS ....................................................................................................................... 13
2.1. WHAT IS ANSYS? ............................................................................................. 13
2.2. BASIC PROGRAM STRUCTURE ..................................................................... 14
2.2.1. Preprocessor .................................................................................................. 15
2.2.2 Solution processor .......................................................................................... 16
2.2.3 Postprocessor .................................................................................................. 16
3. TURUSSES ................................................................................................................ 17
3.1. ASSUMPTIONS FOR DESIGN.......................................................................... 18
3.1.1. All loadings are applied at the joints. ............................................................ 18
3.2.2. The members are joined together by smooth pins ......................................... 18
3.2. SIMPLE TRUSS .................................................................................................. 20
4. MATRIX ANALYSIS OF TRUSSES ........................................................................ 21
4.1. FINITE ELEMENT FORMULATION ............................................................... 21
5. MANUAL SOLUTION .............................................................................................. 28
5.1. PREPROCESSING PHASE ................................................................................ 29
5.1.1. Discretize the problem into nodes and elements. .......................................... 29
5.1.2. Assume a solution that approximates the behavior of an element. ............... 29
5.1.3. Develop equations for elements. ................................................................... 30
5.1.4. Assemble elements. The global stiffness matrix is obtained by assembling, or
adding together, the individual elements’ matrices: ................................................ 37
5.1.5. Apply the boundary conditions and loads. .................................................... 38
5.2. SOLUTION PHASE ............................................................................................ 39
5.3. POSTPROCESSING PHASE .............................................................................. 40
5.3.1. Obtain other in formation. ............................................................................. 40
6. ANSYS SOLUTİON .................................................................................................. 44
6.1. INTRODUCTION TO ANALYSIS .................................................................... 45
6.1.1. Geometry ....................................................................................................... 45
6.1.2 Material........................................................................................................... 47
6.1.4. Element type .................................................................................................. 48
6.1.5. Loads ............................................................................................................. 49
6.1.6. Solution.......................................................................................................... 49
6.1.7 Results ............................................................................................................ 49
7. COMPARISON OF RESULTS .................................................................................. 52
7.1. COMPARISON OF ANALYTICAL SOLUTION AND ANSYS ANALYSIS
RESULTS ................................................................................................................... 52
7.1.1 Displacement of the nodes.............................................................................. 52
7.1.2 Reaction forces in support .............................................................................. 52
7.1.3 Stresses in each element ................................................................................. 53
7.2. COMPARISON OF ANSYS RESULTS ACCORDING TO NUMBER OF
NODES ....................................................................................................................... 53
8. CONCLUSION ........................................................................................................... 55
REFERENCES ............................................................................................................... 56
RESUME ........................................................................................................................ 57
LIST OF FIGURES
TABLE 5.1 The relationship between the elements and their corresponding nodes ...... 29
Press Apply to create the first four nodes. Press OK to create the last nodes and close
the dialog box.................................................................................................................. 46
Table 6.1 Coordinate of nodes ........................................................................................ 46
Table 6.2 Element’s Nodes ............................................................................................. 47
Table 7.1 Comparison of nodes displacement ................................................................ 52
Table 7.2 Comparison of Reaction Forces...................................................................... 52
Table 7.3 Comparison of stresses ................................................................................... 53
Table 7.4 Comparison of ANSYS displacement results according to number of nodes. 54
1. ABOUT FINITE ELEMENT METHOD
The central activity of engineering, as distinguished from science, is the design of new
devices, processes and systems. Design has always been, and will continue to be, one of
the most important facets of the field of engineering. The finite element method, one tool
used for design, is a numerical method that is applied to real world problems involving
complicated phenomenon to solve engineering problems. An examination of the finite
element process first requires a look back at the history of the method to see how we got
to modern day finite element analysis. A fundamental historical perspective can also help
increase the users understanding of the finite element tool. The method of representing a
domain as a collection of discrete parts can bet raced back as far as ancient
mathematicians in their estimations of the value of π to the accuracy of nearly 40
significant figures by representing a circle as a polygon of a very large but finite number
of sides. In 1851, Schell back discretized a surface into finite right triangles and wrote a
finite difference expression to solve for the total area in an attempt to determine the
surface of minimum area bound by a closed curve in space. Later, in 1906, researchers in
Germany discovered that a body having several bars in a regular pattern behaves like an
isotropic elastic body. These examples show early applications in which bodies were
discretized into “finite elements” to solve problems. The modern development of the
finite element method began in 1940s in the field of aircraft structural engineering.
Aircraft fuselages and wings were treated as assemblies of strings, skins, and shear panels,
which were represented by a framework of one-dimensional bars and beams to solve for
stresses and bending over the continuous solid. This work, done by Hrennikoff in 1941
and McHenry in 1943, was known as the framework method. The framework method is
largely regarded as a precursor to FEA as it works well for framed structures but cannot
be applied to oddly shaped members, nor does it discretize a body into smaller pieces but
rather substitutes members of a different type. The modern finite element method can be
further traced through the works on Courant, who determined the torsional rigidity of a
hollow shaft by dividing the cross section using piecewise shape functions into triangular
subunits and then solving a stress function over each triangle at certain net-points, or
nodes as they are now known in 1947, Levy developed the flexibility method, and
published aworkin1953 suggesting the stiffness method as an alternative for use in
statically redundant aircraft structures. His stiffness method, although successful, was too
bulky to solve by hand, and did not become popular until the development of the
highspeed computer Engineers debate as the which of the works of Hrennikoff, Courant,
or even Levy present early forms, or rather precursors to the finite element method, but
each of their works have certain key features of the modern day method. The formal
presentation of the method came about later in the 1950s and early 1960s, starting with
Turner, Robinson, and Argyris, who each worked with two dimensional elements,
including three-node triangular elements, and used these elements to derive stiffness
matrices for the procedure now known as the direct stiffness method.
The phrase “finite element” was not used until 1960 when Clough used both three and
four node elements to solve plane stress analysis problems. In the years thereafter, many
new techniques and elements were developed in 1961, Melosh developed a flat,
rectangular-plate bending-element stiffness matrix. A similar curved-shell version for
axisymmetric shells and pressure vessels was developed in 1963 by Grafton and Strome.
The finite element method was extended to three-dimensional problems by using
tetrahedral elements in the early 1960s. Later, additional three-dimensional elements were
used and methods for axisymmetric solids were developed in 1965. It must be noted that
all off the a fore mentioned works dealt with small scale strains and displacements, elastic
behavior, and static conditions. However, works on large deflection and thermal analysis,
nonlinear materials, buckling, distributed-mass systems, and visco-elasticity were also
performed in the 1960s. Another big development in the finite element method was the
use of variational formulation to set up problems to allow for the solving of non-structural
application and field problems. This allowed for the use of the finite element method to
solve fluid flow, heat conduction, and torsion problems for shafts. The method was further
expanded when Szabo and Lee developed the weighted residual method in 1969 and
Zienkiewicz and Parekh used it to solve transient field problems in 1970. These works in
the 1960s outlined the basis for many of the methods and techniques that are core to the
modern finite element method. However, many of the applications of these methods were
limited by computing power, as equations with tens of thousands of degrees of freedom
are not solvable by hand. Therefore, the development of finite element methods is linked
closely with the development of computing power. General finite element computer
programs began appearing in the late 1960sand early 1970s. In the late 1970s, computer
graphics had advanced enough to advent the use of finite element software for actual
design, rather than simply completed design verification or structural failure analysis.
More about the use of computer FEA software in the design cycle will be discussed later.
With modern advances in finite element software, graphics, and computing power, it is
sometimes easy to forget what is behind these powerful software programs. look back at
the history and fundamentals of the finite element method can greatly help an engineer
understand the logic behind the tool, which greatly increases its power, accuracy, and
reliability. This fundamental understanding of the finite element method and how it
makes the tool more effective will be evident in the following studies.
The acronym FEA stands for finite element analysis. At its basis, FEA is a numerical
method used to solve engineering field problems by dividing a domain into several
smaller finite subdomains, which each act as individual elements over which algebraic
equations are applied and an approximate solution is given using the finite difference
method. The results from each finite element are then reassembled and different types of
analysis can be run to solve any number of complicated engineering problems using this
method and a powerful solver.
The basic method by which a problem is solved using FEA can be broken down into
several steps. First, the problem must be identified and classified. There are several
different types of analysis that can be performed. Selecting the correct analysis
for the correct problem is important. Next, a simplified mathematical model should be
derived from which to build the basic physical concepts of the analysis. Preliminary
analysis is then performed, in which a solution is obtained to help ballpark the result
sought after from the FEA study. The next step is to actually perform the finite element
analysis, which is almost always done with the aid of a computer.
The final step is to check the results. It is important to first note if the results “look”
correct, if they make sense, and if they are similar to the preliminary analysis performed.
It may also be necessary to check the results against other solution forms, or against a
physical model. It must also be noted that the FEA process is a very iterative one. Rarely
is the first FEA study the final one, and revisions are often needed after interpreting the
results of a study. The actual FEA portion of the solution method described above can be
broken down into eight steps, many of which are typically performed by a computer. The
first step is to break the system into “finite elements” by discretizing the system into
several elements formed by associated nodes. There are a variety of different types and
shapes of elements available for use, depending on the geometry of the system. The
second step is to select a displacement function to fit each element. The strain-
displacement and stress-strain relationships are then selected in the third step, which is
dependent on the material properties of each element. Fourth, the stiffness matrix
equations are derived by the computer, and fifth, they are assembled into a global matrix
with all elements and boundary conditions incorporated. Sixth, the equations are solved
simultaneously to determine the displacement at each node. Consequently, the seventh
step is to then calculate the stress and strain for each element using the relationships
determined before the equation assembly. Finally, the eighth step is to interpret the
results, a vague but important task that will be discussed in more detail later. This method
is a general one, and there are multiple ways that it is implemented in different FEA
computer programs. On a larger scale, the method can be broken down into preprocessing,
numerical analysis, and post-processing. Preprocessing refers to all the data input by the
user related to the geometry, material properties, element type sand mesh, loads, supports,
boundary conditions, etc. The numerical solution step is then performed by a computer.
This step was broken down into parts above and consists of the processor combining the
equations into matrix form and then solving simultaneously. Finally, post-processing
occurs, in which the results are displayed in graphical form. The information is
automatically generated once requested by the user. The use of FEA provides several
advantages over other solution and modeling methods. First, by dividing a system into
small parts, systems with complex geometries or several materials can be easily
represented in one model. This provides a huge advantage over traditional analysis
methods and allows engineers to solve unique problems which would have required much
simplification years ago. The size of the elements can also be varied throughout the model
to increase or decrease resolution in areas that need to be more closely examined or are
not important to the result. Additionally, the method can handle general load conditions
and an unlimited number and types of boundary conditions.
The method can also handle a large variety of problems, including heat transfer, stress
analysis, dynamic problems, and nonlinear problems.
2. ANSYS
ANSYS can import CAD data and also enables to build a geometry with its
"preprocessing" abilities. Similarly, in the same preprocessor, finite element model (a.k.a.
mesh) which is required for computation is generated. After defining loadings and
carrying out analyses, results can be viewed as numerical and graphical.
ANSYS can carry out advanced engineering analyses quickly, safely and practically by
its variety of contact algorithms, time-based loading features and nonlinear material
models.
Treatment of engineering problems basically contains three main parts: create a model,
solve the problem and analyses the results. ANSYS, like many other FE-programs, is also
divided into three main parts (processors) which are called preprocessor, solution
processor, post processor. Other software may contain only the preprocessing part or only
the postprocessing part. During the analysis you will communicate with ANSYS via a
Graphical User Interface (GUI), which is described below and seen in Figure 2.2
• Toolbar
Push buttons to commonly used commands.
• Main Menu
Here you can find the processors used when analysing your problem.
• Graphics window
In the graphics window your model is displayed: geometry, elements, visualisation of
results and so forth.
• Input Window
You can type commands in the input window.
2.2.1. Preprocessor
Within the preprocessor the model is set up. It includes a number of steps and usually in
the following order:
•Build Geometry
Depending on whether the problem geometry is one, two or three dimensional, the
geometry consists of creating lines, areas or volumes. These geometries can then, if
necessary, be used to create other geometries by the use of boolean operations. The key
idea when building the geometry like this is to simplify the generation of the element
mesh. Hence, this step is optional but most often used. Nodes and elements can however
be created from coordinates only.
• Define materials
The problem is discretized with nodal points. The nodes are connected to form finite
elements, which together form the material volume. Depending on the problem and the
assumptions that are made, the element type has to be determined. Common element types
are truss, beam, plate, shell and solid elements. Each element type may contain several
subtypes, e.g. 2D 4-noded solid, 3D 20-noded solid elements. Therefore, care has to be
taken when the element type is chosen. The element mesh can in ANSYS be created in
several ways. The most common way is that it is automatically created, however more or
less controlled. For example, you can specify a certain number of elements in a specific
area, or you can force the mesh generator to maintain a specific element size within an
area. Certain element shapes or sizes are not recommended and if these limits are violated,
a warning will be generated in ANSYS. It is up to the user to create a mesh which is able
to generate results with a sufficient degree of accuracy.
Here you solve the problem by gathering all specified information about the problem:
•apply loads
Boundary conditions are usually applied on nodes or elements. The prescribed quantity
can for example be force, traction, displacement, moment, rotation. The loads may in
ANSYS also be edited from the preprocessor.
•obtain solution
The solution to the problem can be obtained if the whole problem is defined.
2.2.3 Postprocessor
A truss is a structure composed of slender members joined together at their end points.
The members commonly used in construction consist of wooden struts or metal bars. In
particular, planar trusses lie in a single plane and are often used to support roofs and
bridges. The truss shown in Fig. 6.1-a is an example of a typical roof-supporting truss. In
this figure, the roof load is transmitted to the truss at the joints by means of a series of
purlins. Since this loading acts in the same plane as the truss, Fig.6.1.-b, the analysis of
the forces developed in the truss members will be two-dimensional.
In the case of a bridge, such as shown Fig.3.1-a, the load on the deck is first transmitted
to stringers, then to floor beams, and finally to the joints of the two supporting side trusses.
Like the roof truss, the bridge truss loading is also coplanar, Fig.3.1-b
To design both the members and the connections of a truss, it is necessary first to
determine the force developed in each member when the truss is subjected a given
loading. To do this we will make two important assumptions:
In most situations, such as for bridge and roof trusses, this assumption is true. Frequently
the weight of the members is neglected because the force supported by each member is
usually much larger than its weight. However, if the weight is to be included in the
analysis, it is generally satisfactory to apply it as a vertical force, with half of its
magnitude applied at each end of the member.
The joint connections are usually formed by bolting or welding the ends of the members
to a common plate, called a gusset plate, as shown in Figure3.3, or by simply passing a
large bolt or pin through each or the members. Figure3.4. We can assume these
connections act as pins provided the center lines of the joining members are concurrent.
Figure3.3 Connections bolting or welding to a common plate.
Figure3.4 Connections simply passing a large bolt or pin through each or the members
Because of these to assumptions, each truss member will act as a two-force member, and
therefore the force acting at each end of the member will be directed along the axis of the
member. If the force tends to elongate the member, it is a tensile force(T).
Figure 3.4-a; Whereas if it tends to shorten the member, it is a compressive force(C)
Figure 3.4-b; In the actual design of a truss it is important to state whether the nature of
the force is tensile or compressive. Often, compression members must be made thicker
than tension members because of the buckling or column effect that occurs when a
member is in compression.
Figure 3.5. Tension and compression
If three members are pin connected at their ends they form a triangular truss that will be
rigid. Attaching two more members and connecting these members to a new joint D forms
a larger truss, Figure 3.6
This procedure can be repeated as many times as desired to form an even larger truss. If
a truss can be constructed by expanding the basic triangular truss in this way, it is called
simple truss
4. MATRIX ANALYSIS OF TRUSSES
Let us consider the deflection of a single member when it is subjected to force F, as shown
in Figure 4.1 The forthcoming derivation of the stiffness coefficient is identical to the
analysis of a centrally loaded member that was presented in Section 1.4. As a review and
for the sake of continuity and convenience, the steps to derive the elements’ equivalent
stiffness coefficients are presented here again. Recall that the average stresses in any two-
force member are given by
F
σ= (4.1)
A
ΔL
ε= (4.2)
L
Over the elastic region, the stress and strain are related by Hooke’s Law,
σ = Eε (4.3)
AE
F = ( )ΔL (4.4)
L
Figure 4.1. Example of statically determinate and statically indeterminate problems.
Note that Eq. (2.4) is similar to the equation of a linear spring, F = kx. Therefore, a
centrally loaded member of uniform cross section may be modeled as a spring with an
equivalent stiffness of
AE
keq = (4.5)
L
A relatively small balcony truss with five nodes and six elements is shown in Figure 4.3.
From this truss, consider isolating a member with an arbitrary orientation. Let us select
element (5).
In general, two frames of reference will be required to describe truss problems: a global
coordinate system and a local frame of reference. We choose a fixed global coordinate
system, X Y (1) to represent the location of each joint (node) and to keep track of the
orientation of each member (element), using angles such as 6; (2) to apply the constraints
and the applied loads in terms of their respective global components; and (3) to represent
the solution that is, the displacement of each joint in global directions. We will also need
a local, or an elemental, coordinate system to describe the two-force member behavior of
individual members (elements). The relationship between the local (element) descriptions
and the global descriptions is shown in Figure 4.4.
The global displacements are related to the local displacements according to the equations
Uix = uixcosθ - uiysinθ
Where
{U} and {u} represent the displacements of nodes i and j with respect to the global X Y
and the local x y frame of references, respectively. [T] is the transformation matrix that
allows for the transfer of local deformations to their respective global values. In a similar
way, the local and global forces may be related according to the equations
where
FiX
FiY
{F}={ }
FjX
FjY
are components of forces acting at nodes i and j with respect to global coordinates and
fix
fiy
{f}={ }
fjx
fjy
represent the local components of the forces at nodes i and j.
A general relationship between the local and the global properties was derived in the
preceding steps. However, we need to keep in mind that the displacements and the forces
in the local y-direction are zero.
This fact is simply because under the two-force assumption, the members can only be
stretched or shortened along their longitudinal axis (local x-axis). Of course, this fact also
holds true for the internal forces that act only in the local x-direction. We do not initially
set these terms equal to zero in order to maintain a general matrix description that will
make the derivation of the element stiffness matrix easier. This process will become clear
when we set the y-components of the displacements and forces equal to zero. The local
internal forces and displacements are related through the stiffness matrix
fix k 0 −k 0 uix
fiy 0 0 0 0 uiy
fjx = [ ] {u } (2.10)
−k 0 k 0 jx
f
{ jy } 0 0 0 0 u jy
AE
where k = keq = ,and using matrix form we can write
L
After substituting for {f} and {u} in term is of {F} and {U}, we have
Substituting for values of the [T], [K], [T]-1, and {U} matrices in Eq. (2.14) and
multiplying, we are left with
Equations (2.15) express the relationship between the applied forces, the element stiffness
matrix [K](e) , and the global deflection of the nodes of an arbitrary element. The stiffness
matrix [K](e) for any member (element) of the truss is
The next few steps involve assembling, or connecting, the elemental stiffness matrices,
applying boundary conditions and loads, solving for displacements, and obtaining other
information, such as normal stresses.
5. MANUAL SOLUTION
Consider the balcony truss in Figure 5.1, shown here with dimensions. We are interested
in determining the deflection of each joint under the loading shown in the figure. All
members are made from structural steel with a modulus of elasticity of E = 29e6 lb/in and
a cross-sectional area of 2 in. We are also interested in calculating average stresses in
each member. First, we will solve this problem manually. Later, once we learn how to
use ANSYS, we will revisit this problem and solve it using ANSYS.
There are seven steps involved in any finite element analysis Here, these steps are
discussed again to emphasize the three phases (preprocessing, solution, and
postprocessing) associated with the analysis of truss problems.
5.1. PREPROCESSING PHASE
Each truss member is considered an element, and each joint connecting members is a
node. Therefore, the given truss can be modeled with five nodes and six elements. Consult
Table 5.1 while following the solution.
TABLE 5.1 The relationship between the elements and their corresponding nodes
As discussed in Section 2.2, We will model the elastic behavior of each element as a
spring with an equivalent stiffness of k as given by Eq. (2.5). All elements have the same
length, cross-sectional area, and modulus of elasticity, the equivalent stiffness constant
for these elements (members) is
𝑙𝑏
𝐴𝐸 (2 𝑖𝑛2 )(29𝑥106 )
𝑖𝑛2 𝑙𝑏
k= = = 604166.6 𝑖𝑛
𝐿 96 𝑖𝑛
For elements (1), (2), and (5), the local and the global coordinate systems are aligned,
which means that θ = 0. This relationship is shown in Figure 2.6. Using Eq. (2.16), we
find that the stiffness matrices are
Figure 5.2 The orientation of the local coordinates with respect to the global coordinates
for elements (1), (2), and (5).
cos20 sin0cos0 −cos2 0 −sin0cos0
2
[K](e) = 6.041𝑥105 𝑙𝑏 sin0cos0 sin 0 −sin0cos0 −sin20
𝑖𝑛2 −cos2 0 −sin0cos0 cos2 0 sin0cos0
2
[−sin0cos0 −sin 0 sin0cos0 sin20 ]
1 0 −1 0 𝑈1𝑋
[K](1) = 6.041𝑥105 2 [ 0
𝑙𝑏 0 0 0] 𝑈1𝑌
𝑖𝑛 −1 0 1 0 𝑈2𝑋
0 0 0 0 {𝑈2𝑌 }
and the position of element (1)’s stiffness matrix in the global matrix is
𝑈1𝑋
1 0 −1 0 0 0 0 0 0 0 𝑈
1𝑌
0 0 0 0 0 0 0 0 0 0
−1 0 1 0 0 0 0 0 0 0 𝑈 2𝑋
0 0 0 0 0 0 0 0 0 0 𝑈2𝑌
[K](1G) = 6.041𝑥105 𝑙𝑏 0 0 0 0 0 0 0 0 0 0 𝑈3𝑋
𝑖𝑛2 0 0 0 0 0 0 0 0 0 0 𝑈3𝑌
0 0 0 0 0 0 0 0 0 0 𝑈4𝑋
0 0 0 0 0 0 0 0 0 0 𝑈4𝑌
0 0 0 0 0 0 0 0 0 0 𝑈
[ 0 0 0 0 0 0 0 0 0 0] {𝑈5𝑋 }
5𝑌
Note that the nodal displacement matrix is shown alongside element (1)’s position in the
global matrix to aid us in observing the location of element (1)’s stiffness matrix in the
global matrix. Similarly, the stiffness matrix for element (2) is
1 0 −1 0 𝑈2𝑋
[K](2) = 6.041𝑥105 𝑙𝑏
[ 0 0 0 0] 𝑈2𝑌
𝑖𝑛2 −1 0 1 0 𝑈3𝑋
0 0 0 0 {𝑈3𝑌 }
and its position in the global matrix is
𝑈1𝑋
0 0 0 0 0 0 0 0 0 0 𝑈
1𝑌
0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 𝑈 2𝑋
0 0 1 0 −1 0 0 0 0 0 𝑈2𝑌
[K](2G) = 6.041𝑥105 𝑙𝑏 0 0 0 0 0 0 0 0 0 0 𝑈3𝑋
𝑖𝑛2 0 0 −1 0 1 0 0 0 0 0 𝑈3𝑌
0 0 0 0 0 0 0 0 0 0 𝑈4𝑋
0 0 0 0 0 0 0 0 0 0 𝑈4𝑌
0 0 0 0 0 0 0 0 0 0 𝑈
[0 0 0 0 0 0 0 0 0 0] {𝑈5𝑋 }
5𝑌
1 0 −1 0 𝑈4𝑋
0 0 0 0 𝑈4𝑌
[K](5) = 6.041𝑥105 [ ]{ }
−1 0 1 0 𝑈5𝑋
0 0 0 0 𝑈5𝑌
0 0 0 0 0 0 0 0 0 0 𝑈1𝑋
0 0 0 0 0 0 0 0 0 0 𝑈1𝑌
0 0 0 0 0 0 0 0 0 0 𝑈2𝑋
0 0 0 0 0 0 0 0 0 0 𝑈2𝑌
0 0 0 0 0 0 0 0 0 0 𝑈3𝑋
[K](5G) =6.041𝑥105
0 0 0 0 0 0 0 0 0 0 𝑈3𝑌
0 0 0 0 0 0 1 0 −1 0 𝑈4𝑋
0 0 0 0 0 0 0 0 0 0 𝑈4𝑌
0 0 0 0 0 0 −1 0 1 0 𝑈5𝑋
[0 0 0 0 0 0 0 0 0 0] {𝑈5𝑌 }
Figure 5.3 The orientation of the local coordinates with respect to the global coordinates
for elements (4), and (6).
For element (4), (6) the orientation of the local coordinate system with respect to the
global coordinates is shown in Figure 5.2. Thus, for element (4), (6) θ = 60, which leads
to the stiffness matrix
[K](4G) =
0.25 0.433 0 0 0 0 −0.25 −0.433 0 0 𝑈1𝑋
0.433 0.75 0 0 0 0 −0.433 −0.75 0 0 𝑈1𝑌
0 0 0 0 0 0 0 0 0 0 𝑈2𝑋
0 0 0 0 0 0 0 0 0 0 𝑈2𝑌
𝑙𝑏 0 0 0 0 0 0 0 0 0 0 𝑈3𝑋
6.041𝑥105 𝑖𝑛2 0 0 0 0 0 0 0 0 0 0 𝑈3𝑌
−0.25 −0.433 0 0 0 0 0.25 0.433 0 0 𝑈4𝑋
−0.433 −0.75 0 0 0 0 0.433 0.75 0 0 𝑈4𝑌
0 0 0 0 0 0 0 0 0 0 𝑈5𝑋
[ 0 0 0 0 0 0 0 0 0 0] {𝑈5𝑌 }
[K](6G) =
0 0 0 0 0 0 0 0 0 0 𝑈1𝑋
0 0 0 0 0 0 0 0 0 0 𝑈1𝑌
0 0 0.25 0.433 0 0 0 0 −0.25 −0.433 𝑈2𝑋
0 0 0.433 0.75 0 0 0 0 −0.433 −0.75 𝑈2𝑌
0 0 0 0 0 0 0 0 0 0 𝑈3𝑋
6.041𝑥105
0 0 0 0 0 0 0 0 0 0 𝑈3𝑌
0 0 0 0 0 0 0 0 0 0 𝑈4𝑋
0 0 0 0 0 0 0 0 0 0 𝑈4𝑌
0 0 −0.25 −0.433 0 0 0 0 0.25 0.433 𝑈5𝑋
[0 0 −0.433 −0.75 0 0 0 0 0.433 0.75 ] {𝑈5𝑌 }
Figure 5.4 The orientation of the local coordinates with respect to the global coordinates
for elements (3), and (7).
For element (3), (7) the orientation of the local coordinate system with respect to the
global coordinates is shown in Figure 2.8. Thus, for element (2), θ = 125, yielding the
stiffness matrix
[K](3) =
cos 2 120 sin120cos120 −cos26 120 −sin120cos120
6.041𝑥105 [ sin120cos120 sin2 120 −sin120cos120 −sin2 120 ]
−cos 2 120 −sin120cos120 cos 2 120 sin120cos120
−sin120cos120 −sin2 120 sin120cos120 sin2 120
[K](3G) =
0 0 0 0 0 0 0 0 0 0 𝑈1𝑋
0 0 0 0 0 0 0 0 0 0 𝑈1𝑌
0 0 0.25 −0.433 0 0 −0.25 0.433 0 0 𝑈2𝑋
0 0 −0.433 0.75 0 0 0.433 −0.75 0 0 𝑈2𝑌
0 0 0 0 0 0 0 0 0 0 𝑈3𝑋
6.041𝑥105
0 0 0 0 0 0 0 0 0 0 𝑈3𝑌
0 0 −0.25 0.433 0 0 0.25 −0.433 0 0 𝑈4𝑋
0 0 0.433 −0.75 0 0 −0.433 0.75 0 0 𝑈4𝑌
0 0 0 0 0 0 0 0 0 0 𝑈5𝑋
[0 0 0 0 0 0 0 0 0 0] {𝑈5𝑌 }
[K](7G)
0 0 0 0 0 0 0 0 0 0 𝑈1𝑋
0 0 0 0 0 0 0 0 0 0 𝑈1𝑌
0 0 0 0 0 0 0 0 0 0 𝑈2𝑋
0 0 0 0 0 0 0 0 0 0 𝑈2𝑌
0 0 0 0 0.25 −0.433 0 0 −0.25 0.433 𝑈3𝑋
=6.041𝑥105
0 0 0 0 −0.433 0.75 0 0 0.433 −0.75 𝑈3𝑌
0 0 0 0 0 0 0 0 0 0 𝑈4𝑋
0 0 0 0 0 0 0 0 0 0 𝑈4𝑌
0 0 0 0 −0.25 0.433 0 0 0.25 −0.433 𝑈5𝑋
[0 0 0 0 0.433 −0.75 0 0 −0.433 0.75 ] {𝑈5𝑌 }
It is worth noting again that the nodal displacements associated with each element are
shown next to each element’s stiffness matrix. This practice makes it easier to connect
(assemble) the individual stiffness matrices into the global stiffness matrix for the truss.
5.1.4. Assemble elements. The global stiffness matrix is obtained by assembling, or
adding together, the individual elements’ matrices:
[K](G)=
[K](G) =
The following boundary conditions apply to this problem: nodes 1 and 3 are fixed; which
implies that U1X = 0, U1Y = O, U3X = 0, and U3Y= 0. Incorporating these conditions into
the global stiffness matrix and applying the external loads at nodes 2, 4 and 5 such that
F2Y= -1000lb , F4X= 1500 lb and F5x = - 1250lb and F5y= - 2165.063lb results in a set of
linear equations that must be solved simultaneously:
0
0
0
−1000
0
=
0
1500
0
−1250
{−2165.063}
Because U = 0, U= 0,U= 0, and U = 0,we can eliminate the first, second, fifth, and sixth
rows and columns from our calculation such that we need only solve a 6 X 6 matrix:
𝑈1𝑋 0
𝑈1𝑌 0
𝑈2𝑋 −0.000416
𝑈2𝑌 −0.000705
𝑈3𝑋 0
= in
𝑈3𝑌 0
𝑈4𝑋 0.00233
𝑈4𝑌 −0.00340
𝑈5𝑋 −0.00192
{𝑈5𝑌 } {−0.000604}
Recognize that the displacements of the nodes are given with respect to the global
coordinate system.
5.3. POSTPROCESSING PHASE
{R}=[K]{U}-{F}
such that
𝑅1𝑋
𝑅1𝑌
𝑅2𝑋
𝑅2𝑌
𝑅3𝑋
=105
𝑅3𝑌
𝑅4𝑋
𝑅4𝑌
𝑅5𝑋
{𝑅5𝑌 }
7.551 2.615 −6.041 0 0 0 −1.51 −2.615 0 0
2.615 4.53 0 0 0 0 −2.615 −4.53 0 0
−6.41 0 15.102 0 −6.041 0 −1.51 2.615 −1.51 −2.615
0 0 0 9.06 0 0 2.615 −4.53 −2.615 −4.53
0 0 −6.041 0 7.551 −2.615 0 0 −1.51 2.615
0 0 0 0 −2.615 4.53 0 0 2.615 −4.53
−1.51 −2.615 −1.51 2.615 0 0 9.061 0 −6.041 0
−2.615 −4.53 2.615 −4.53 0 0 0 9.06 0 0
0 0 −1.51 −2.615 −1.51 2.615 −6.041 0 9.061 0
[ 0 0 −2.615 −4.53 2.615 −4.53 0 0 0 9.06 ]
0 0
0 0
−0.000416 0
−0.000705 −1000
0 0
-
0 0
0.00233 1500
−0.00340 0
−0.00192 −1250
{−0.000604} {−2165.063}
Performing matrix operations, yields the reaction results
𝑅1𝑋 788.66
𝑅1𝑌 933
𝑅2𝑋 0
𝑅2𝑌 0
𝑅3𝑋 −1038.7
= lb
𝑅3𝑌 2232
𝑅4𝑋 0
𝑅4𝑌 0
𝑅5𝑋 0
{𝑅5𝑌 } { 0 }
Internal Forces and Normal Stresses Now let us compute internal forces, and the average
normal stresses, in each member. The member internal forces fix and fjx, which are equal
and opposite in direction, are
Note that the sum of fix and fjx is zero regardless of which representation of Figure 5.5 we
select. However, for the sake of Consistency in the forthcoming derivation, we will use
the second representation so that fix and fjx are given in the positive local x-direction. In
order to use Eq. (2.17) to compute the internal force in a given element, we must know
the displacements of the element’s end nodes, uix and ujx with respect to the local
coordinate system, x, y.
{U} = [T]-1{u}
{u} = [T]-1{U}
Once the internal force in each member is computed, the normal stress in each member
can be determined from the equation
internal force f
σ= =
area A
AE
f k( uix − ujx ) L
( uix − ujx ) uix − ujx
σ= = = = E( )
A A A L
As an example, let us compute the internal force and the normal stress in element (5). For
element (5), θ = 0, U4X =0.00233 in, U4Y = -0.00340 in, U5X = -0.00192 in, and
U5Y = -0.000604 in. First, we solve for local displacements of nodes 4 and 5 from the
relation
Consider the balcony truss in Figure 2.4, shown here with dimensions. We are interest-
ed in determining the deflection of each joint under the loading shown in the figure. All
members are made of structural steel with a modulus of elasticity of E = 29e6 lb/in and
a cross-sectional area of 2in2. We are also interested in calculating average stresses in
each member. First, we will solve this problem manually. Later, once we learn how to
use ANSYS, we will revisit this problem and solve it using ANSYS.
Your model can be saved in a database by specifying your working directory (the folder
where you want your ANSYS files to be saved) and a job name (every problem has a job
name, for example truss).
6.1.1. Geometry
We will now draw the structure shown in Figure 2 by first defining nodes and then
drawing lines between them. A visible working plane often makes the creation of the
geometry easier. Therefore:
Change to “Grid and Triad” (triad meaning a triplet of base vectors) and set the minimum
and maximum to 0 and 2 respectively, OK.
We will now define nodes at the joints of the truss, see Table 6.2 for the location of the
nodes.
ANSYS Main menu: Preprocessor > Modeling > Create > Nodes > In Active CS
Press Apply to create the first four nodes. Press OK to create the last nodes and close
the dialog box.
We will now create elements between the nodes, see Figure 6.4
ANSYS Main menu: Preprocessor > Modeling > Create > Element > Auto Numbered
> Thru Nodes
Press Apply to create the first five lines. Press OK to create the last line and close the
dialog box.
Table 6.2 Element’s Nodes
6.1.2 Material
We assume that the structural steel linearly elastic. Define the material model and the
material constants, see Figure 6.4
ANSYS Main menu: Preprocessor > Material Props > Material Models
Figure 6.5 Define Material Properties
Enter 29e6 for Ex (Young’s modulus) and save your database: ANSYS Toolbar:
SAVE_DB
The element type to use is called link1. Add this element from the library:
ANSYS Main menu: Preprocessor > Element type > Add/Edit/Delete > Add…
The cross-sectional area of the elements in the truss structure also has to be defined. This
is accomplished with a so called real constant set:
Choose the element type link1, OK. Fill in the value of the cross-sectional area in the box.
6.1.5. Loads
ANSYS Main menu: Solution > Define Loads > Apply > Structural > Displacement > On
Nodes
ANSYS Main menu: Solution > Define Loads > Apply > Structural > Force/Moment >On
Nodes Choose and apply forces
6.1.6. Solution
6.1.7 Results
ANSYS Main menu: General Postproc > Read Results > First Set Now there are several
results to study.
ANSYS Main menu: General Postproc > Plot Results > Deformed Shape
The undeformed and deformed shape should be similar to what is shown in Figure 6.7
Figure 6.7 Deformed shape
ANSYS Utility menu: List > Results > Nodal solution ...
In the dialog box select Nodal solution, DOF solution and x and y displacement
components, click Apply.
ANSYS Utility menu: List > Results > Element solution > Stress >x-component of stress
One of the most important issues of finite element analysis is whether the size of the
elements used and the number of elements are enough for that analysis. Finite element
density is an important factor used to control the accuracy of analysis. (element type and
shape also affect the accuracy of the analysis). If we assume that there is no singularity
region in the model, high mesh density will give high accuracy results. But its causes long
working times and consumes a lot of memory. In my study, I realized that the analysis
results were more accurate as the number of nodes increased.
REFERENCES
4. Charles Head Norris, John Benson Wilbur, Senol Utku, 1977, Elementary Structural
Analysis, third edition, McGraw – Hill.
RESUME
He was born on 19 June 1993 in Bursa. After completing his primary and secondary
education again, he completed his high school education at Tophane Industrial
Occupational in Bursa / Osmangazi. Later, he completed the Department of Machinery
program in Ordu University. In 2015, the Department of Mechanical Engineering of
Trakya University started and is continuing.