Upload
reduan-sagor
View
19
Download
1
Tags:
Embed Size (px)
Citation preview
ME3602 FEA Lab Report
1
Applications of Finite Element Modelling in Stress Analysis
Student Number: 1216434
1. Abstract
The stress analysis carried out in this report was done using ANSYS Workbench on a thin plate
consisting of double row of holes in it. The report demonstrates the study of localized stress or
stress concentration around the holes using nominal stress and stress concentration factor. The
analysis system used for this report was Static Structural Analysis which determined the total
deformation, maximum equivalent stress and maximum principal stress of the plate when it was
under an axial load in both positive and negative y-axis. The assumption made here is that load
applied was constant and uniformly distributed on the horizontal edges of the plate, taking into
account the effect of Poisson’s ratio. Using the Static Structural Analysis system the geometry
creation, mesh generation and refinement, applying boundary conditions and solving were carried
out. The results for maximum principal stress and number of elements were noted down in MS
Excel and convergence plot of the results were obtained. The results showed convergence after
the meshes were refined each time until no further refinement causes any variation. Both
theoretical and actual stress concentration factors, Kt and K, were determined and similar pattern
were obtained in the plot of K vs staggered angle between the rows of the holes.
2. Introduction to FEA Analysis Finite Element Modelling is an important numerical analysis technique in mechanical design where
a physical model is converted into a discrete mathematical model which to carry out mesh
generation and simulation for obtaining a certain set of results. Mechanical engineers use FEA
software such as ANSYS Workbench to solve static, structural, dynamic, linear, nonlinear, and
thermal problems. FEA analysis generally involves three stages: Pre-processor, Solver and Post-
processor. For a given static or dynamic problem under investigation, Pre-processor stage is
where the model creation and mesh generation can be carried out. The Solver stage takes over to
assemble a stiffness matrix from the mesh and proceed to solve the structural field where
displacement is first calculated and later other results such as stress, deformation and strains are
obtained. Linear problems can be solved easily using Hooke’s Law whereas non-linear problems
can be complicated to solve. The Post-processing stage is where the data for different set of
results are generated and displayed in a contour plot. Advantages of FEA analysis are: irregular
boundaries, general loads, different materials, boundary conditions, variable element size, easy
modification etc. Disadvantages of FEA include: approximate solutions, element density affects the
solution for a plate with a circular hole, poor shape quality of elements reduce accuracy of the
solution, errors in input data. [1]
ME3602 FEA Lab Report
2
3. Pre-Processing 3.1 3D-modelling of the plate
Diameter of
hole (a) /
mm
Spacing
between
holes (b) /
mm
Length of
the plate /
mm
Width of
the plate /
mm
Elastic
Modulus /
GPa
Poisson’s
ratio
Yield
Stress /
MPa
2 32 250 100 210 0.32 250
Table_1: Parameters given for the geometry.
Figure_1 shows the schematic of the semi-finite plate containing the double-row of staggered
holes subject to axial tension.
Figure_1: Plate geometry.
The distance between the rows of holes, c, was variable from 0 mm to 45 mm and also the
thickness of the plate, t, had values of 3, 5 and 7 mm. The staggered angle, 𝜃, was dependent on
the values of c and b. Stress was applied on the plate by means of an axial tensile load of 1200 N
in the y-axis. The main tasks were: to determine the peak stress around the hole, and hence using
nominal stress, to calculate the stress concentration factor, SCF for different staggered angle and
different thickness of the plate. Figure_2 shows the 3D model of the plate created, with symmetry
of axis.
Figure_2: Extruded plate with double-row of holes.
c
b
a θ
σ
σ
w
l
𝜽
c
ME3602 FEA Lab Report
3
The symmetry of axis was created and used to reduce the plate total surface area, hence
increasing the mesh elements and nodes in the half-plate to get more accurate results for better
mesh refinement. Also it is less time consuming to obtain the results for half-plate than a full plate.
3.2 Addition of material properties of the plate
The properties given were the Elastic modulus, Poisons ratio and Tensile or Yield strength. The
materials inside the ‘Engineering Data’ of the Static Structural Analysis system did not have the
exact properties which were given therefore a new material called ‘FEA_Material’ was added to the
material library and given parameters were inserted manually and used for the plate model for
meshing and simulation to make sure that peak stress and SCF calculation are solely based on the
input parameters only.
3.3 Mesh Discretization and Mesh-refinement study
Once the 3D model of the geometry is created, mesh can be generated in terms of elements and
nodes, with choice of different types of meshes such as prism mesh, tetrahedron mesh,
hexahedron mesh, quadrilateral mesh etc. The results obtained were affected by the choice of
mesh type. Four types mesh were generated, amongst which one was an ‘Auto Mesh’. The other
three were Prism mesh, Hexahedron mesh and Quadrilateral mesh. The figures below show the
different types of mesh generated.
Figure_3: Auto Mesh Figure_4: Quadrilateral Mesh
Figure_5: Tetrahedron/Prism Mesh Figure_6: Hexahedron Mesh
ME3602 FEA Lab Report
4
The peak stress, 𝜎max, will be highest around the edges and inside of the holes hence a critical
mesh refinement study was carried out using Edge sizing and Inflation of the whole plate including
the holes. The horizontal and vertical edges of the plate were chosen for Edge sizing and certain
numbers of divisions were created on these edges. This will create an uniform appearance of the
mesh over the plate. To obtain as accurate results as possible, the edges of the holes were divided
into number of elements and Inflation was done on the inside faces of the holes.
Figure_7: Inflation and edge sizing of the holes.
Figure_8: Edge sizing of the horizontal edges of the plate.
Figure_9: Edge sizing of the holes.
The method chosen was called
‘Multizone’ which allowed the
user to choose between
hexahedron or prism mesh. The
meshes were refined by
increasing the Edge sizing
values of the edges of plate
same along with increasing the
number of divisions for edges of
the holes and changing the
Inflation settings. The numbers
of layer of circles around the
holes were increased while the
maximum thickness of the
inflation was decreased. This in
return increased elements and
nodes of the mesh hence
increasing the accuracy of the
results which were obtained
later. Although a certain inflation
settings did not generate the
mesh sometimes but a different
setting with a combination of
different edge sizing created a
denser mesh around the holes.
Figures on the left show how
edge sizing and inflation was
done. To obtain a convergence
plot of the results the meshes
were refined and three refined
meshes were obtained.
ME3602 FEA Lab Report
5
3.4 Boundary Conditions
Figure_10: Forces and support on the plate.
4. Solving, Post-Processing and Discussion of Results
After applying the loads and support to the half-plate model, simulation was carried out and a
warning was obtained which indicated that the software used an iterative solver to solve the model.
Hence to get rid of that warning, a direct solver was used which did not change the results
obtained but made the solving process less time consuming. The simulation was run to obtain
results for total deformation, equivalent (von-Mises) stress, maximum principal stress and safety
factor. The simulation generated contour plots for these three results which are shown below.
4.1 Contour plots
Figure_11: Total Deformation of the half-plate.
The load of 0.6 kN was applied on
both horizontal faces of the plate in
the y-axis. To stop the plate from
experiencing rigid body motion, a
fixed support was used in one
vertex of the plate, which had
constrained the plate. There were
other types of support but they
were unable to stop the plate from
experiencing rigid body motion.
The contour plots on the left are displayed once
the solving process was carried out. The
deformation of the half-plate obtained was
expected to be as it is due to the fixed support
used at one vertex on the left vertical edge,
where it is deformed the minimum (blue area).
The deformation had a maximum value of
0.00049066 mm on the top horizontal face of
the plate due to the applied force of 0.6 kN. This
was elastic deformation which follows Hooke’s
Law meaning that if load applied was removed
from the half-plate, then the plate will return
back to original un-deformed shape. The holes
are seen to experience moderate deformation
(green area) by using the fixed support hence
they are not going to fail at this load.
ME3602 FEA Lab Report
6
Figure_12: Equivalent stress around the hole.
Figure_13: Principal stress around the hole.
4.2 Convergence Plot of the Hexahedron Mesh
Figure_14: Convergence of results for Max. Principal Stress.
0
2
4
6
0 5000 10000 15000 20000 25000
Max Principal Stress/ MPa
Elements
Convergence_Plot_Hexa_Mesh
thickness 3 mm
thickness 5 mm
thickness 7 mm
The Equivalent (von-Mises) stress was highest
on the inner surface of the holes as seen in the
contour plot, with a value of 2.7934 MPa. The
von-Mises stress is a combination of principal
stresses which can be used to check whether
the plate will withstand a given load condition or
not. The plate will fail if the maximum equivalent
stress exceeds the yield stress of the material.
The safety factor tool can be used to determine
the yield stress of the plate. The safety factor of
the plate had a maximum value of 15 MPa,
hence yield stress was calculated to be (15 x
2.7934) = 41.9 MPa. So it can be concluded
that the plate will not fail.
The Maximum Principal Stress had a value of
3.0104 MPa on the inner face of the hole which
is the peak stress due to the presence of that
hole in that area of the plate. Minimum Principal
Stress was present in the y-direction which can
be seen in the blue region of the plot. Other
regions of the plate were within a green color
band as the maximum principal stress was
moderate there due to absence of holes, with a
value of 1.3111 MPa.
ME3602 FEA Lab Report
7
For convergence of the results, Hexahedron mesh was used. Three refined hexahedron meshes
were used. The plot consists of three lines for three different thickness of the plate. Maximum
Principal stress converged at 5.10 MPa (for 3 mm), 2.99 MPa (for 5 mm) and 2.11 MPa (for 7 mm).
Therefore any further refinement of the meshes will not change the stress results. Although the
patterns for convergence plot were similar, the value of stress decreased for different thickness of
the plate.
Figure_15: Convergence of results for different mesh densities.
4.3 Analysis of stress concentration factor, K
Stress concentration factor is a dimensionless parameter that is determined using the maximum
principal stress, 𝜎max, and nominal stress for a given plate with certain thickness and number of
holes. The nominal stress is the stress calculated using the uniaxial load applied over the net area
of the plate. It is same for a certain thickness of the plate but changes when the plate gets
thicker/thinner. The value of K, either theoretical or experimental, depends on the principal stress,
nominal stress, diameter of the holes, horizontal pitch (spacing between the holes) and the
staggered angle between the holes. This can be demonstrated from Equation_1 and Equation_2
as shown below. The equations to calculate K will be different for staggered angles more than 60°.
𝑲𝒕𝒏 = 𝝈𝒎𝒂𝒙
𝝈 (𝟏 −
𝟐𝒂
𝒃𝐜𝐨𝐬 𝜽) --------------------- Equation_1 (for staggered angle < 60°)
𝑲𝒕𝒏 = 𝝈𝒎𝒂𝒙
𝝈 (𝟏 −
𝒂
𝒃) ---------------------------------Equation_2 (for staggered angle > 60°)
To calculate the theoretical K for a given staggered angle, linear interpolation method was used on
the chart of Ktn vs staggered angle for a plate with double-row of holes. Then using the values of K,
nominal stress, diameter of hole (a), horizontal pitch (b) and θ, theoretical 𝜎max can be determined
for a given staggered angle. To calculate experimental K for the plate under uniaxial load, the
value of 𝜎max will be the result obtained for the maximum principal stress.
0
1
2
3
4
5
6
0 5000 10000 15000 20000 25000 30000
Max Principal Stress/ MPa
Elements
Convergence_Plot_Different_Mesh
Hexa_Mesh
Prism_Mesh
ME3602 FEA Lab Report
8
Figure_16: Variation of stress concentration factor with staggered angle between holes.
The plot for the K (7 mm) was closest to the theoretical plot of K vs staggered angle. This means it
has the lowest value of K in comparison with plate thickness of 3 and 5 mm for a certain value of
staggered angle. The plot tells us that increasing the staggered angle between the double-row of
holes will increase the stress concentration factor steadily until the values do not change anymore.
The peak value of theoretical K was 2.7 whereas it was 2.88 for 7 mm thickness, 2.92 for 5 mm
thickness and 2.98 for 3 mm thickness of the plate. The difference in K for different plate thickness
was due to the varied maximum principal stress which decreased with increasing thickness. But as
nominal stress also decreased with increasing thickness, the ratio of maximum principal stress
over nominal stress was not significantly different hence causing less variation in values of K which
is indicated by the plots being closer to one another. The percentage difference between the
values theoretical K and K (7 mm) was the least with the difference being 1-2% and was the
highest between K and K (3 mm) with the difference being around 4-5%.
Figure_17: Variation of experimental Maximum Principal Stress with staggered angle for different
plate thickness.
2
2.2
2.4
2.6
2.8
3
3.2
0 10 20 30 40 50 60 70 80
K
Staggered Angle / degree
K variation with staggered angle
Theoretical K
K (3 mm)
K (5 mm)
K (7 mm)
0
2
4
6
0 20 40 60 80
Max Principal Stress / MPa
Staggered Angle / degree
Max Principal Stress variation with Staggered Angle
Thickness 3 mm
Thickness 5 mm
Thickness 7 mm
ME3602 FEA Lab Report
9
Figure_16 demonstrates the effect of changing staggered angle between the holes on the
maximum principal stress. The variation of results for a particular thickness of plate is very little.
The maximum principal stress was 5.10 MPa (for 3 mm), 3.01 MPa (for 5 mm) and 2.10 MPa (for 7
mm). This shows a gradual decrease in the stress value with increasing plate thickness. The
values of principal stresses obtained for this plot were experimental and were similar to the
theoretical values of maximum principal stresses. Therefore the mesh convergence was occurred
at the desired maximum principal stress for each plate thickness.
4.4 Recommendation for the best set of variable of ‘a’, ‘b’, ‘c’ and ‘t’
The value of stress concentration factor, K depends on a number of parameters such as diameter
of the holes (a), horizontal pitch between holes (b), the staggered angle, which changes by
changing the spacing between rows of holes (c) and the plate thickness (t).
To come up with an understanding of what values of these parameters will give the least value of
K, the chart in page-351 of reference [2] was used. If for example the plate under investigation for
this report was used in a mechanical product in real life, the peak stress around the holes will
determine when the structure will fail. This requires the hole to have minimum peak stress with the
least stress concentration factor value. Using the chart in reference [2], it can be seen that for
increasing ‘a/b’ the Ktn decreases. The two ways to increase ‘a/b’ will be: (i) increase ‘a’ (ii)
decrease ‘b’. Therefore chosen value of ‘a’ will be the given maximum value of 9 mm and of ‘b’ will
be the given minimum value of 12 mm.
Now to demonstrate how the value of ‘c’ affects the value of K, Figure_16 can be observed. In that
plot of maximum principal stress against staggered angle (different for different ‘c’) it can be seen
that maximum principal stress does not change significantly with increasing staggered angle. But
also, nominal stress will be same for each staggered angle therefore Equation_1 and Equation_2
are not enough to justify for choosing a certain value for staggered angle. Therefore using the
chart in reference [2] and Figure_15 it can be seen that smaller value of staggered angle will give
smaller value of K. Hence the spacing between the rows should be minimum so that staggered
angle will decrease, decreasing K, with constant ‘a’ and ‘b’.
To choose a suitable value for thickness (t), Figure_15 and Figure_16 can be analyzed, which
shows that the least value of K was obtained for the plate thickness of 7 mm. Therefore it can be
justified that increasing thickness will distribute the peak stress around the holes, hence reducing
the stress concentration factor, K.
ME3602 FEA Lab Report
10
5. Conclusions
ANSYS 15.0 and Workbench was used to numerically model and simulate the problem under
investigation in this report which was a thin plate with double-row of staggered holes. The objective
was to carry out a stress analysis on the plate when it undergoes uniaxial tensile load of 1.2 kN.
The load applied was uniformly distributed on the horizontal faces of the plate in the y-direction. To
reduce the simulation run-time and apply more mesh densities around the holes, an axis of
symmetry was applied on the full plate of 250 mm length in the y-axis so that it generates a 3D
half-plate model. The thickness of the plate was 3, 5 and 7 mm. It was chosen as a design
parameter so that later the value can be changed without going back to the geometry frequently.
The vertical spacing between the holes was also chosen to be a design parameter hence can be
changed quickly in a way similar to changing thickness. The effect of changing thickness and
staggered angle were investigated in terms of determining peak stress around the holes and
subsequently the stress concentration factor. In order to achieve this, mesh generation was carried
out. The Auto mesh was generated without defining the mesh density of the plate. The other three
mesh types generated were manually defined. To get most accurate results of peak stress around
the hole, mesh density on the inner face of holes was increased using the tools Inflation and Edge
sizing. A direct solver was used to simulate the mesh which made the simulation less time
consuming. The use of correct boundary conditions and material properties ensured that results
obtained were not biased by any manual errors. Although in the assignment brief, the support for
the plate was not mentioned, therefore different types of support were tried and finally a fixed
support was used, which took into account the effect of Poisson’s ratio and stopped the plate from
experiencing rigid body motion and being under constrained. The maximum principal stress values
were obtained for different number of elements and convergence study was carried out. The
results were converged hence further refinement of meshes did not cause any variation. The stress
concentration factor variation with staggered angle was investigated and similar patterns were
obtained for different thickness of plate that looked similar to the pattern of theoretical SCF.
Results of the numerical analysis showed that SCF was affected by staggered angle and thickness
of the plate in the way predicted by theoretical analysis.
6. References
[1] FINITE ELEMENT METHOD. FINITE ELEMENT METHOD, [Online]. Lecture 8, 1-24. Available
at:
http://www.google.co.uk/url?sa=t&rct=j&q=&esrc=s&source=web&cd=4&cad=rja&uact=8&ved=0C
DEQFjAD&url=http%3A%2F%2Fmochsafarudin.files.wordpress.com%2F2008%2F10%2Flecture8.
pdf&ei=n0S1VLDECuuf7gbDsYDIBA&usg=AFQjCNHcWQJHjOvGYiJaPEoi8gnicMMeZg&bvm=bv.
83339334,d.ZGU [Accessed 14 January 2015].
[2] Pilkey, WDP, DFP, 2008. PETERSON’S STRESS. 3rd ed. Canada: JOHN WILEY & SONS, INC.
ME3602 FEA Lab Report
11
7. ANSYS Summary Report
Geometry
TABLE 2 Model (A4) > Geometry
Object Name Geometry
State Fully Defined
Definition
Source D:\FEA_Final\FEA_2014\Hexa_Mesh\FEA_2014_files\dp0\SYS-1\DM\SYS-1.agdb
Type DesignModeler
Length Unit Meters
Element Control Program Controlled
Display Style Body Color
Bounding Box
Length X 125. mm
Length Y 100. mm
Length Z 5. mm
Object Name Solid
State Meshed
Material
Assignment FEA_Steel
Nonlinear Effects Yes
Thermal Strain Effects Yes
Statistics
Nodes 29529
Elements 4141
Mesh Metric None
Symmetry
TABLE 5 Model (A4) > Symmetry
Object Name Symmetry
State Fully Defined
TABLE 6 Model (A4) > Symmetry > Symmetry Region
Object Name Symmetry Region Symmetry Region 2
State Fully Defined
Scope
Scoping Method Named Selection
Named Selection Symmetry:XYPlane Symmetry:YZPlane
Definition
Scope Mode Automatic
Type Symmetric
Coordinate System XYPlane YZPlane
Symmetry Normal Z Axis
Suppressed No
ME3602 FEA Lab Report
12
Mesh
TABLE 7 Model (A4) > Mesh
TABLE 8 Model (A4) > Mesh > Mesh Controls
Object Name
MultiZone Edge Sizing
Edge Sizing 2
Edge Sizing 3
Edge Sizing 4
Edge Sizing 5
Edge Sizing 6
Inflation
State Fully Defined
Static Structural (A5)
TABLE 10 Model (A4) > Analysis
Object Name Static Structural (A5)
State Solved
Definition
Physics Type Structural
Analysis Type Static Structural
TABLE 11 Model (A4) > Static Structural (A5) > Analysis Settings
Object Name Analysis Settings
State Fully Defined
Solver Controls
Solver Type Direct
Weak Springs Program Controlled
TABLE 12 Model (A4) > Static Structural (A5) > Loads
Scope
Scoping Method Geometry Selection
Geometry 1 Face 1 Vertex 1 Face
Definition
Type Force Fixed Support Force
Define By Components Components
Coordinate System Global Coordinate System Global Coordinate System
X Component 0. N (ramped) 0. N (ramped)
Y Component 600. N (ramped) -600. N (ramped)
Z Component 0. N (ramped) 0. N (ramped)
Material Data
FEA_Steel
TABLE 16 FEA_Steel > Isotropic Elasticity
Temperature C Young's Modulus MPa Poisson's Ratio Bulk Modulus MPa Shear Modulus MPa
2.1e+005 0.32 1.9444e+005 79545
TABLE 17 FEA_Steel > Tensile Yield Strength
Tensile Yield Strength MPa
ME3602 FEA Lab Report
13