Modeling of structures

Tutku Üngör

Istanbul Aydin University, Faculty of Engineering

Corresponding author: tutkungor@gmail.com

Abstract:

Modeling of structures is very common application which using Finite Element Method todays
engineering. Before using this process, we must choose the size and shape of finite element
such as bar, beam, area, solid, membrane etc. If we want more exact results, we must conduct
sensivity anaylsis otherwise we will get results which are very distant from reality. Meshing is
very important in this. Holes and corners are critical areas in meshing process and bad
meshing will result in bad response or unrealible results for desing. We must check geometry
and symmetry for a good mesh before starting. The geometry checking refers to free edges,
small fillets and small holes. In this paper a basketball basket analysis was done to check how
the mesh size affects result. The basket was excited with a static force and with a dynamic
force simulating slam dunk. In paper change of natural period, support forces and deformation
were observed under the change of mesh.

Tutku ÜNGÖR 1
Content:

1. Explanation of Finite Element Method 3

2. Finite Element Types 3
2.1 Bar Elements 3
2.2 Beam Elements 4
2.3 Area Elements 5
2.4 Membrane Elements 5
2.5 Solid Elements 6
3. How the Size and Shape of Finite Element Selected? 6
4. Sensivity Analysis 7
5. Meshing Process 8
5.1 Critical Areas in Meshing Process 9
5.2 Some Examples for Bad Meshing and Good Meshing 9
5.3 How to Start Mesh? 10
6. Validation and Verification 11
6.1 Validation 11
6.2 Verification 11
6.3 Difference of Validation and Verification 11
7. How Materiels are Modelled? 12
8. Errors 17
8.1 Determination of Numarical Errors 17
8.1.1 Characteristic of Convergence in Finite Element 17
Method
8.1.2 Unsmoothed Stress Map 17
8.2 Modeling Errors 18
8.2.1 Formulation 18
8.2.2 The Mesh 18
8.2.3 Double Elements and Double Nodes 19
8.2.4 Cross Section Orientation and End Releases 20
8.2.5 Symmetries and Antisymmetries 20
8.3 Software Errors 21
9. Compare of softwares for finite element method 21
SAP 2000 Exercise 23
References 34

Tutku ÜNGÖR 2
1 Explanation of the finite element method

The finite element method is well described in [1] and [2]. The finite element method (FEM) is

a numerical method for solving problems of engineering and mathematical physics. Typical

problem areas of interest include heat transfer, structural analysis, fluid flow, mass transport

and electromagnetic potential. The basic principles of the finite element method are simple.

The first step in the finite element solution procedure is to divide the domain into elements,

and this process is called discretization. The elements’ distribution is called the mesh. The

elements are connected at points called nodes.

After the region is discretized, the governing equations for each element must be established

for the required physics. Material properties, such as thermal conductivity for thermal analysis,

should be available. The elements’ equations are assembled to obtain the global equation for

the mesh, which describes the behavior of the body.

2 Finite element types

Depending on the problem, the elements can have different shapes, such as bars or beams,

areas, or solids. The following is described in [3].

2.1 Bar Elements

Bar elements are long and slender, have 2 nodes (Figure 1) and can be oriented anywhere

in 3D space. Bar elements transmit forces which allow translation only and not rotation. Bars

Tutku ÜNGÖR 3
are normally used to model towers, bridges, and buildings. A general cross section area is

usually assumed. They are usually used for linear elastic structural analysis in engineering

practice.

Figure 1. Bar element

2.2 Beam elements

Beam elements are long and slender, have three nodes, and can be oriented anywhere in 3D

space (Figure 2). Beam elements are 6 degrees of freedom (DOF) elements allowing both

translation and rotation at each end node. That is the primary difference between beam and

truss elements. The i j nodes define element geometry; the K node defines the cross-sectional

orientation. This is how you differentiate between the strong and weak axis of bending for a

beam. A constant cross section area is assumed.

Figure 2. Beam Element [5]

Tutku ÜNGÖR 4
2.3 Area elements

Planar or 2D elements have 3 or 4 nodes (Figure 3). Usual degrees of freedom Y and Z

translation and are normally created in the YZ plane. In solving, we are preferring 4 node

elements more than three-node elements. A three-node element can be a problem to solve

because of their shape. They are used for plane stress or plane strain analyses. Common

applications include axisymmetric bodies of revolution such as missile radomes, radial seals,

etc. and long sections with constant cross-sectional area such as a dam. Plane stress implies

no stress normal to the cross section defined - strain is allowed - suitable to model the 2D

cross section of a body of revolution.

Figure 3. Area elements

2.4 Membrane elements

Membrane elements are three or four node 2D elements that can be oriented anywhere in 3D

space. They can be used to model thin membrane like materials like fabric, thin metal shells,

etc. These elements will not support or transmit a moment load or stress normal to the surface.

They support only translational DOF not rotational and in-plane loading. The thickness of the

membrane must be small relative to its length or width. Membrane thickness is defined as a

fixed parameter which can be varied. The geometry is drawn at the midplane with zero

thickness shown, like a plate element.

Tutku ÜNGÖR 5
2.5 Solid elements

Elements defined in three dimensions (Figure 4), tend to be more used in advanced structural

analysis because the behaviour of mass structures tend to be non-linear. There are some

elastic applications. The volume elements are used to model three-dimensional bodies. Shells

are special elements. These elements used to model gears in Mechanical Engineering.

Figure 4. Solid element

3 How the size and shape of finite elements is selected?

In [4] it is well described how the shape of finite elements is selected. Firstly, we should define

to which kind of applications that we need. The solid elements are used to model three-

dimensional bodies. The line elements can be trusses, spring, links, or beams. The area

elements that could be rectangle or triangle are used to model two-dimensional structural

elements for stress analysis for a plate, or fins for thermal analysis.

Problems with parts of low thickness in relation to the other dimensions, thin-walled structures.

Model it as a shell and can be define thickness (SHELL elements). Thickness is important

parameter here. When we solving to Shell elements, we can include transverse shear

deformation in plate-bending behavior. In this case, thickness will be parameter. Thick – plate

formulations are best for applications when Shear deformation to be important when shell

thickness is greater than approximately 1/5 to 1/10 of the span of plate-bending curvature. Also

Tutku ÜNGÖR 6
shearing can be important in locations of bending-stress concentrations, which occur near

sudden changes in thickness or support conditions, and near openings or re-entrant corners

Thick-plate formulation is also recommended in general because it tends to be more accurate,

though slightly stiffer, even for thin-plate bending problems in which shear deformation is truly

negligible. However, the accuracy of thick-plate formulation is sensitive to mesh distortion and

large aspect ratios, and therefore should not be used in such cases when shear deformation

is known to be small.

4 Sensitivity analysis

Sensitivity analysis is where the effects of differenet values for features or parameters are

investigated. This can prove very useful in promoting better understanding of the behaviour of

system being modelled. For example, If the results are not sensitive to the valuse of a

particular parameter this can help to validate the model and provide information which can

help in the design. If the results are sensitive to a parameter, then speacial validation action

may be needed. The need for sensitivity analysis is a function of the degree of uncertainty

about the behaviour. When working on an unsual system, sensitivity analysis may be

essential. The following issues are relevant to sensitivity analysis; [5]

 A useful strategy is to work from a reference model. Changing one variable at a time

and reverting to the reference model after each change. As one gains understanding

it may be better to change the reference model.

 Make comprasions with indicative parameters. Parameters which lend to exemplify the

behaviour.

 In the report for sensitivity analysis, show the corrosponding resulst as close to each

other participate

Tutku ÜNGÖR 7
 5 Meshing process

This Explanation well described in literature [5], [6] and [7]. Per definition, a mesh is a network

that is formed of cells and points. It can have almost any shape in any size and is used to

solve Partial Differential Equations (PDE). Each cell of the mesh represents an individual

solution of the equation which, when combined for the whole network, results in a solution for

the entire mesh. Solving, which we need one, the entire object without dividing it into smaller

pieces can be impossible because of the complexity that is within the object.

Increasing the number of elements order will lead to significant increase in the computational

time and requires memory to solve the problem. Holes, corners and angles can make it

extremely difficult for solvers to obtain a solution.

Mesh Layouts for a Cantilever Bracket (Figure 5);

 For Figure 5a), is a regular mesh without refinement, which would tend not to give

good acccuracy for local stress

 For Figure 5b), shows a local mesh refinement where triangles are used to diffuse the

mesh outwards from the load

 For Figure 5c), does not have triangular elements of figure b and is inadmissible

because the node at point X is not connected to the element on its left

Figure 5. Different Meshes

Tutku ÜNGÖR 8
5.1 Critical Areas in meshing process

Critical areas are locations where high stress locations will occur. Dense meshing and

structured mesh is recommended in these regions. Areas away from the critical area are

general areas. Geometry simplification and coarse mesh in general areas are recommended

(to reduce the total DOFs and solution time).

Critical Area General Area

Minimum 12 elemtents around the hole 4 to 6 elements

Minimum 3 elements in fillet Suppress small fillet, 1 element for large fillet

Figure 6. Critical Areas in Meshing

5.2 Some examples for bad meshing and good meshing

This chapter provides examples that describes bad and good meshing. A tipical type of mesh

can be seen on the Figure 7. All pieces are very regular and straight on Figure 7a. Meshing

frequency is more where near holes as well. That makes it that good mesh on process. Figure

7 b is bad meshing because lines and pieces are not regular. Meshing frequency is not often

near holes, holes is part of area which is hard to solve more than flat area in solving. That

makes it that bad mesh.

Tutku ÜNGÖR 9
 a) b)

Figure 7. Example of a mesh: a) Good Mesh, b) Bad Mesh

5.3 How to start mesh?

Method on how to start or generate a mesh is provided in [6]. A short example of

staring a mesh is provided next.

1) Create a geometry

2) Geometry Check

The geometry should be carefully checked for;

 Free Edges

 Scarlines

 Duplicate Surface

 Small Fillets

 Small Holes

3) Symmetry Check

4) Selection of type of element

5) Joint Modelling

Tutku ÜNGÖR 10
6 Validation and verification

6.1 Validation

Validation is consideration of whether or not a process is suited to its purpose [5]. Validation

is the process of evaluating the final product to check whether the software meets the

customer expectations and requirements. It is a dynamic mechanism of validating and testing

the actual product. This can answer the following question,

 Am I building the right product?

 Am I accessing the right data (in terms of the data required to satisfy the requirement)

6.2 Verification

The process of evaluating software to determine whether the products of a given development

phase satisfy the conditions imposed at the start of that phase. [8] Verification is done at the

starting of the development process. It includes reviews and meetings, walk-throughs,

inspection, etc. to evaluate documents, plans, code, requirements and specifications. It is a

relatively objective process. Verification will help to determine whether the software is of high

quality, but it will not ensure that the system is useful. Verification is concerned with whether

the system is well-engineered and error-free. It can be high – quality instead of useful and

thus It will not be for our process.

6.3 Difference of validation and verification

This chapter will show to difference of Validation and Verification on Table 1. This table is well
described in literature [10].

Tutku ÜNGÖR 11
 Table 1. Verification and Validation Comparison

Verification Validation
Verification is a static practice of verifying Validation is a dynamic mechanism of validating

documents, design, code and program. and testing the actual product.

It does not involve executing the code. It always involves executing the code.

It is human based checking of documents and files. It is computer-based execution of program.

Validation uses methods like black box

Verification uses methods like inspections, reviews,
(functional) testing, gray box testing, and white
walkthroughs, and Desk-checking.
box (structural) testing.

Verification is to check whether the software Validation is to check whether software meets the

conforms to specifications. customer expectations and requirements.

It can catch errors that validation cannot catch. It is It can catch errors that verification cannot catch. It

low level exercise. is High Level Exercise.

Target is requirements specification, application

Target is actual product-a unit, a module, a bent of
and software architecture, high level, complete
integrated modules, and effective final product.
design, and database design.

7 How materials are modeled?

Procedues of materials modeling are well described in [2]. The following steps show how

materials can be modeled in a step-by-step procedure using program ANSYS,

First Step, Main Menu > Preferences (Figure a)

The user selects Main Menu > Preprocessor > Element type >Add/Edit/Delete (Figure b) Then

the user selects Select solid – Select triangle 6node 2 (Figure c)

Options (Figure d)

Tutku ÜNGÖR 12
 a) b)

c) d)

Figure 8. Four steps in modeling material using program ANSYS: a) Preference for GUI
filtering, b) Element Types, c) Library of element types, d) Element types

The users should follow following step;

Main Menu > Preprocessor > Real Constants > Add/Edit/Delete (Figure b)

Then click to Add (Figure c)

Tutku ÜNGÖR 13
 a) b) c)

Figure 9. Next three steps in modeling materials using program ANSYS: a) Plane element
type options, b) Real Constants, c) Element Type for Plane

The users should follow following steps;

Type 0.005 in Thickness (Figure a)

Main Menu > Preprocessor > Material Props > Material Models (Figure b)

For the material properties, the modulus of elasticity and Poisson’s ratio are required

to solve the problem.

a) b)

Figure 10. Next two steps in modeling materials using program ANSYS: a) Real Constant
Set, b) Define Material Model Behavior

Tutku ÜNGÖR 14
The users should follow following steps;

Main Menu > Preprocessor > Modeling > Create > Area > Rectangle > By 2 Corners (Figure

a)

Then, type those values to openned window;

Type 0 in WP X

Type 0 in WP Y

Type 0.25 in Width

Type 0.25 in Height

Main Menu > Preprocessor > Meshing > Mesh Tool (Figure b)

Then,

Select Smart Size

Set Smart Size to 1

Mesh

Pick All

Main Menu > Solution > Defi ne Load > Apply > Structural > Pressure > On Lines (Figure d)

a) b)

Tutku ÜNGÖR 15
 c) d)
Figure 11. Next four steps in modeling materials using program ANSYS: a) Rectangle by 2
corners, b) Mesh Tool, c) Mesh, d) Apply Pressure on Lines

The users should follow following steps;

Main Menu > Solution > Solve > Current LS (Figure a)

ANSYS graphics shows the plate before and after deformation (Figure b)

a) b)

Figure 12. Last two steps in modeling materials using program ANSYS: o) Current LS,

p) Deformation

Tutku ÜNGÖR 16
8) Errors

This following decribtions are well described in literature [9] There are essentially three main

possible sources of problems an insufficient density of the mesh and consquently the

insufficient precision with which the solution was calculated; errors arising during the course

of modelling; and errors in the software program.

8.1 Determination of numerical errors

8.1.1 Characteristic of convergence in finite element method

A first possible source of errors lies in the insufficient precision, but the leading and most

frequent one is related to the insufficient density of the mesh. If we increase the precision as

density of the mesh, we will get results as many close as real.

a) b)

Figure 13. Characteristic of Convergence Errors in Finite Element Method a) Table of error

difference as depend element grid, b) Showing Stress Maps of error difference [9]

8.1.2 Unsmoothed stress map

An extremly fruitful method for evaluating the degree of refinement of an analysis at least in

cases dealing with continua is the analysis intrelement jumps, both by direct means (with maps

of the jumps) and by indirect means (unsmoothed maps). We have seen that he presence of

significant intrelement jumps where we instead expect the stresses to be continuous is a direct

measure of how imprecise an analysis is. Although It is not possible to establish a direct

relationship between the error and the intrelement jumps, certainly the order of

magnitude is an important indicator.

Tutku ÜNGÖR 17
 Figure 14. Table of difference values as depend frequency of mesh

a) b)

Figure 15. Unsmoothed Stress Map: a) 6x6, b) 48x48 [9]

8.2 Modelling errors

8.2.1 Formulation

One typical source of errors in results is an error in formulation. This occurs when a more

general problem is related back to a theroretical framework in which important parts are lost.

The fact is that only in certaian circumstances is it possible to relate the dominate part of the

stress of a given problem back to what is obtained by means of a calculation performed

according to one of simplifying theories. In the great majority of cases, this can not be done.

8.2.2 The Mesh

The most typical and frequent causes of errors in a model are related to problems with the

mesh. It is always good practice to check the mesh carefully in order to avoid problems of

disconnection or violations of congruity that might compromise the validity of the entire

analysis. The use of shrunken reprasantations can facilitate checking the mesh to a

considarible degree, since this makes it easier to identify elements that are facing and

unconnected.

Tutku ÜNGÖR 18
 Figure 16. Erroneous mesh of a truss (Element Numbering) [9]

8.2.3 Double elements and double nodes

Double elements and double nodes warrant a separate discussion. These rather quite

common errors are rather deceptive because they are not quite evident. Double nodes can

lead to disconnections that are not immediately discernible and thus, as a consequence, to

mechanism. In general, the meshing programs are equipped with specific tools for checking

that are intended to prevent the generation of double nodes, which tend to be generated

almost at every turn.

Figure 17. Double Elements and Double Nodes [9]

Tutku ÜNGÖR 19
8.2.4 Cross section orientation and end releases

Structures where there are beam elements, it is necessary to check the orientation of the

elements carefully, because an incorrect orientation is certain to lead to incorrect result. The

represantation of the local axes or the solid represantation of the elements is helpful in

checking that all of the elements of the model are correctly oriented. Expect for this particular

function, this kind of “Solid“ represantation is of no use to the analsyt, who will find it to be

condiradictory and misleading in beam and truss models which are “Wire Frame“ by

defination.

Figure 18. “Solid“ represantation of a beam and truss model [9]

8.2.5 Symmetries and Antisymmetries

If there are symmetries or antisymmetries that are not taken advantage of during the modelling

phase, these must be used to validate the structural solution. The presence of unjustified

asymmetries in the solution will naturally lead us to consider the model unreiable and look fort

he causes of the asymmetry. It should be recalled that to call an answer symmetrical (or

ansymmetrical) there has to be symmetry in:

1) The formulation of the elements

2) The geometry

3) The mesh

4) The load

5) The constraints and end releases

6) The applied actions

Tutku ÜNGÖR 20
8.3 Software errors

Software errors are perpahs the most deceptive of all errors that the analyst must deal with.

This error is not possible to prove definitively and with absolute certainly that any computer

programs is error – free.

9 Compare of softwares for finite element method

This chapter will show 5 different software comparing on Table 2. This table is well described

in [10].

Table 2. Comparing of Several Software [10]

NX Autodest
Solidworks Abaqus /
Ansys Simcenter Simulation
Simulation CAE
3D Mechanical

Graphical

Geometry Yes Yes Yes No Yes

Modeler

Graphical

Manuel Yes No Yes No Yes

Meshing

Cad Import Yes Yes Yes Yes Yes

Units Aware Yes No No No No

Linear Static Yes Yes Yes Yes Yes

Nonlinear –

Large Yes Yes Yes Yes Yes

Displacement

Nonlinear –
Yes Yes Yes Yes Yes
Contact

Tutku ÜNGÖR 21
 Linear
Yes No Yes Yes Yes
Buckling

Heat
Yes Yes Yes Yes Yes
Transfer

This following describtions are well described in literature [10]

Meanings of headings;

 Graphical geometry modeler: That is mean that to build a model defined geometry like

a CAD program.

 Graphical manual meshing: Uses tools such as extruding, revolving and laying

structured meshes on simple shapes to to build a finite element mesh directly.

 CAD import: This is meaning that ability to read geometry from a CAD program.

 Units aware: Aware of units that mostly using for engineering process such as

milimeter or inch or KN for force.

 Linear static: This is a very common type of analyis that can find stresses in structures

that undergo small displacements and whose displacement is a linear function of load.

 Nonlinear - large displacements: This is a geometric nonlinearity which allows

displacements to be large compared to the dimensions of the structure.

 Nonlinear – contact: This can be detected when two-part touch each other. Thus, this

is applying a reaction force or constraint to prevent them occupying the same space.

 Linear buckling: This can help us on that find the critical buckling loads of certain types

of structure without doing a time consuming non-linear analysis.

 Heat transfer: This is behavior of heat dynmaic as usual using mechanical analysis.

Tutku ÜNGÖR 22
SAP 2000 Exercice

This chapter shows a numerical model of a basketball basket (Figure 19.) made using program

SAP2000, version 18.2.0. This basketball basket has six frame elements and one shell

element.

a) b)

Figure 19. Basketball basket: a) side view, b) front view

(http://www.enterpan.co.rs/index_files/image447.jpg)

The bearing structure and the panel (Figure 20.) were made of steel of class S275 (E = 210000

N/mm2, γs = 76,97 kN/m3). The beam was made using hollow steel cross section with outside

deptht of 200 mm, outside width of 150 mm and wall thickness of 7 mm. The diagonal elements

carrying the beam were made using hollow circular cross sections with outside diameter of

100 mm and wall thickness of 5 mm. The panel was made using thin shell element that was 3

mm thick.

Tutku ÜNGÖR 23
 a) b)

Figure 19. 3D numerical model of basketball basket: a) extruded model, b) in standard view

Figure 21. shows numerical model with node labels and Figure 22. shows numerical model

with frame element labels. In nodes 1, 6 and 7 a support was defined. The support has three

degress of freedom restrained: translation in direction 1 (red arrow), translation in direction 2

(greed arrow) and translation in direction 3 (blue arrow).

Tutku ÜNGÖR 24
 a) b)

Figure 21. Numerical model: a) node labels, b) node local axees

Figure 22. Frame element labels

Tutku ÜNGÖR 25
Table 3. contains coordinates of the nodes of the basketball basket.

Table 3. Coordinates of the nodes of the basketball basket

Joint X (m) Y (m) Z (m)

1 0,9 3,6 0
2 0,9 3 2,5
6 0,04 2,4 0
7 1,76 2,4 0
9 0,9 0 3,5
10 0,9 1,8 2,9
11 0 0 2,9
12 1,8 0 2,9
13 1,8 0 4,1
14 0 0 4,1
15 0,9 0 2,9
16 0 0 3,5
17 0,9 0 4,1
18 1,8 0 3,5
19 0,45 0 2,9
20 0,45 0 3,2
21 0 0 3,2
22 0,45 0 3,5
23 0,9 0 3,2
24 0,45 0 3,8
25 0 0 3,8
26 0,45 0 4,1
27 0,9 0 3,8
28 1,35 0 2,9
29 1,35 0 3,2
30 1,35 0 3,5
31 1,8 0 3,2
32 1,35 0 3,8
33 1,35 0 4,1
34 1,8 0 3,8
35 0,9 1,35 3,05
36 0,9 0,9 3,2
37 0,9 0,45 3,35
38 0,9 2,4 2,7
39 0,9 3,3 1,25
40 0,47 2,7 1,25
41 1,33 2,7 1,25
42 0,47 2,1 1,45
43 1,33 2,1 1,45

Tutku ÜNGÖR 26
The basketball basket will be loaded in joint 9 (Figure 23.) with a dynamic force that changes

during time as described by Figure 24. It is assumed that a 100 kg basketball player will make

a slamdunk, so the force of 1 kN will be applied at the joint 9 and assigned to the function

described in Figure 24. In software a discrete time step of 0,01 s and damping of 5 % were

used.

Figure 23. Numerical model node labels (front view)

1,5

1
coefficient

0,5

0
0 0,2 0,4 0,6 0,8 1 1,2 1,4 1,6 1,8 2
t (s)

Figure 24. The change of force simulating a slamdunk

Tutku ÜNGÖR 27
The first calculatinon was made for numerical model (Figure 19.) that contains 39 points, 16

frame elements and 16 area elements (reference model). To check whether the model is

sensitive to mesh density several tests were made as describe in Table 4.

Table 4. shows that finer mesh of the area elements affects modal analysis but not the static

displacement due to the self-weight of the structure. Changing the number of frame elements

does not affect the behavior of the model.

Table 4. Changing result on different mesh density

Test Description T1 (s) dv,9 (mm)

0 (ref. model) 16 frame elements and 16 area elements 0,67744 -0,55161

1 16 frame elements and 64 area elements 0,68936 -0,55161

2 16 frame elements and 256 area elements 0,72232 -0,55161

3 32 frame elements and 16 area elements 0,67744 -0,55161

4 64 frame elements and 16 area elements 0,67744 -0,55161

To verify the model, the translations of nodes 1, 6 and 7 were checked. All the three nodes

are modelled correctly since there are no translations (Table 5.). In Table 5. U1 referes to

translation (U) in direction 1 (red local axis as shown in Figure 21. b) and R2 referes to rotation

(R) around axis 2 (green local axis as shown in Figure 21. b).

Tutku ÜNGÖR 28
 Table 5. Joint displacements for self-weight

U1 U2 U3 R1 R2 R3
Joint
m m m Radians Radians Radians
1 0 0 0 0,000056 -1,344E-18 -1,731E-17
2 -1,375E-17 -0,000044 0,000005175 -0,000011 5,728E-17 -3,245E-18
6 0 0 0 0,000017 0,000038 0,000007823
7 0 0 0 0,000017 -0,000038 -0,000007823
9 3,627E-16 -0,00023 -0,000552 0,000353 1,871E-16 1,632E-16
10 3,574E-17 -0,000057 -0,00003 0,000135 9,711E-17 6,402E-17
11 3,229E-07 -0,000018 -0,000553 0,000353 -7,901E-07 1,706E-16
12 -3,229E-07 -0,000018 -0,000553 0,000353 7,901E-07 1,723E-16
13 3,229E-07 -0,000442 -0,000553 0,000353 7,901E-07 1,672E-16
14 -3,229E-07 -0,000442 -0,000553 0,000353 -7,901E-07 1,728E-16
15 1,475E-16 -0,000018 -0,000552 0,000353 2,001E-16 1,688E-16
16 3,631E-16 -0,00023 -0,000553 0,000353 -1,783E-07 1,733E-16
17 5,786E-16 -0,000442 -0,000552 0,000353 2,005E-16 1,712E-16
18 3,629E-16 -0,00023 -0,000553 0,000353 1,783E-07 1,697E-16
19 2,422E-07 -0,000018 -0,000552 0,000353 -5,458E-07 1,69E-16
20 1,205E-07 -0,000124 -0,000552 0,000353 -5,321E-07 1,696E-16
21 1,087E-07 -0,000124 -0,000553 0,000353 -7,631E-07 1,719E-16
22 3,63E-16 -0,00023 -0,000552 0,000353 -9,095E-07 1,734E-16
23 2,554E-16 -0,000124 -0,000552 0,000353 5,209E-16 1,7E-16
24 -1,205E-07 -0,000336 -0,000552 0,000353 -5,321E-07 1,74E-16
25 -1,087E-07 -0,000336 -0,000553 0,000353 -7,631E-07 1,732E-16
26 -2,422E-07 -0,000442 -0,000552 0,000353 -5,458E-07 1,727E-16
27 4,704E-16 -0,000336 -0,000552 0,000353 5,226E-16 1,713E-16
28 -2,422E-07 -0,000018 -0,000552 0,000353 5,458E-07 1,717E-16
29 -1,205E-07 -0,000124 -0,000552 0,000353 5,321E-07 1,712E-16
30 3,629E-16 -0,00023 -0,000552 0,000353 9,095E-07 1,697E-16
31 -1,087E-07 -0,000124 -0,000553 0,000353 7,631E-07 1,719E-16
32 1,205E-07 -0,000336 -0,000552 0,000353 5,321E-07 1,67E-16
33 2,422E-07 -0,000442 -0,000552 0,000353 5,458E-07 1,677E-16
34 1,087E-07 -0,000336 -0,000553 0,000353 7,631E-07 1,687E-16
35 9,242E-17 -0,000086 -0,000118 0,00024 1,261E-16 1,083E-16
36 1,701E-16 -0,000128 -0,000244 0,000307 1,508E-16 1,396E-16
37 2,624E-16 -0,000177 -0,000393 0,000343 1,711E-16 1,579E-16
38 -1,087E-18 -0,000043 0,000008824 0,000021 7,142E-17 1,307E-17
39 -6,873E-18 -0,000051 -0,000004806 0,000013 -1,344E-18 -1,731E-17
40 0,000022 -0,000016 -0,000006809 0,000023 -0,000008902 0,000003736
41 -0,000022 -0,000016 -0,000006809 0,000023 0,000008902 -0,000003736
42 0,000018 -0,000023 -0,00002 0,000015 -0,000021 0,000014
43 -0,000018 -0,000023 -0,00002 0,000015 0,000021 -0,000014

Tutku ÜNGÖR 29
Figure 25. shows reactions due to self-weight of the structure. The sum of the vertical reaction

is equal to 6,632 kN what is equal to the total self weight of the structure. From this 0,7 t of

steel is needed to build this kind of structure. If we assume that 1 kg of steel costs 1 EUR,

approximate price of the structure is 700 EUR, not including connections, foundation, and

labor.

Figure 25. Nodal reactions due to self-weight in KN

Figure 26. shows vibrations of the panel due to slamdunk of a 100 kg basketball player. From

the figure it was possible to learn that there are no vibrations after 4,5 s. The largest vertical

displacement was 0,02 mm in the moment when the player grabed the hoop. After the player

released the hoop the largest deformation was 0,013 mm.

Tutku ÜNGÖR 30
 0,015

0,01

0,005
displacement (mm)
0
0 1 2 3 4 5 6
-0,005

-0,01

-0,015

-0,02

-0,025
time (s)

Figure 26. Vibraions of the structure due to slamdunk

Joint 1, joint 6 and joint 7 have reactions on the during time of slamdunk. Joint 6 and joint 7

are same reactions because of their placement on same coordinate y axes. When basketball

player grabbed the hoop, joint 7 reaction was 290 N and joint 1 reaction was 175 N as highest

values. Joint 7 and joint 1 reactions were oriented same direction until basketball player

released the hoop. Both of reactions direction changed reciprocally and severity changed

diminish (Figure 27.).

300

200
vertical reaction (N)

100

0
0 1 2 3 4 5 6
-100

Joint 1 Joint 7
-200
time (s)

Figure 27. Change of reaction in joint 1 and 7 due to slamdunk

Tutku ÜNGÖR 31
In the short moment when the player grabed the hoop, both reactions in joint 1 and 7 were

oriented in the same direction until the player was hanging on the hoop, when the reaction in

joint 1 changed the direction. After the player released the hoop the reactions in joint 1 and 7

were mutually changing direction.

1200 1200
Joint 1
1000 1000
Joint 7
vertical reaction (N)

slamdunk force (N)

800 800
slamdunk
600 600

400 400

200 200

0 0
0,8 1 1,2 1,4
-200 -200
time (s)

Figure 28. Change of reaction joint 7 and 1 over slamdunk force

Figure 29. shows meanings of S11, S12 and S22. S11 is horizontal stretching as well as S22

is vertical stretching and S12 is cutting of shell element on stress map analysis. S is shortly

for “Strain” and numbers are according to coordinate plane.

Figure 29. Meanings of stress map shortcut [11]

Tutku ÜNGÖR 32
Figure 30 shows strain maps of panel under static force. In the first strain map (S11), Panel

stretching was 120 N/mm2 as highest. Stress increased from middle of panel to each side on

z axes.

a) b)

c) d)

Figure 30. Models on strain analysis: a) Strain model, b) S11 strain map (N/mm2), c) S22

strain map (N/mm2), d) S22 strain model (N/mm2)

Tutku ÜNGÖR 33
References

[1] https://en.wikipedia.org/wiki/Finite_element_method (Avaiable on 30th of June 2018)

[2] Alawadhi, E.M., 2010. Finite Element Simulations Using ANSYS. CRC Press, Taylor &

Francis Group

[3] http://fea-cae-engineering.com/fea-cae-engineering/element_types.htm (Avaiable on 30th

of June 2018)

[4]https://www.researchgate.net/post/Which_element_type_should_analyst_use_for_Finite_

Element_Analysis_simulation (Avaiable on 30th of June 2018)

[5] MacLeod, I.A. 2005 Model Structural Analysis. Thomas Telford Publishing, Thomas

Telford Ltd

[6] https://altairuniversity.com/wp-content/uploads/2014/02/meshing.pdf (Avaiable on 30th of

June 2018)

[7] https://www.simscale.com/docs/content/simwiki/preprocessing/whatisamesh.html

(Avaiable on 30th of June 2018)

[8] http://toolsqa.com/software-testing/difference-between-verification-and-validation/

(Avaiable on 30th of June 2018)

[9] Rugarli, P. 2010 Structural Analysis with Finite Elements. Thomas Telford Limited

[10] http://feacompare.com/ (Avaiable on 30th of June 2018)

[11] CSI Analysis Reference Manuel, SAP 2000, 2016

Tutku ÜNGÖR 34