Vous êtes sur la page 1sur 28

COMPUTATIONAL MODELING OF

STRUCTURES
CASE STUDY
Group 10 Maria Luisa Rosales
Rob Smeekens

TABLE OF CONTENTS
1. INPUT....................................................................................................................................................... 3
Material Properties..................................................................................................................................... 3
Element Properties..................................................................................................................................... 3
Element Dimensions............................................................................................................................... 3
Finite Element Mesh................................................................................................................................ 3
Boundary Conditions.................................................................................................................................. 4
Constraints (at the bottom).................................................................................................................... 4
Constraints (at the top)........................................................................................................................... 4
Applied Loads......................................................................................................................................... 4
Prescribed Displacements....................................................................................................................... 5
2. ANALYSIS RESULTS................................................................................................................................. 6
2.1. Structural Linear Analysis............................................................................................................. 6
Reaction Forces....................................................................................................................................... 6
Displacements (Y direction, deformed shape)......................................................................................... 6
Displacements (Z direction, deformed shape)........................................................................................ 6
Stresses SZZ (outer face, deformed shape)............................................................................................ 6
Stresses SZZ (inner face, deformed shape)............................................................................................ 7
Strains EZZ (outer face, deformed shape).............................................................................................. 7
Strains EZZ (inner face, deformed shape).............................................................................................. 7
2.2. Non-Linear Analysis........................................................................................................................ 8
2.2.1. Physical Nonlinearity........................................................................................................................ 8
Displacements (Z direction, deformed shape)........................................................................................ 9
Displacements (Y direction, deformed shape)......................................................................................... 9
Stresses SZZ (outer face, Deformed Shape)........................................................................................... 9
Stresses SZZ (inner face, Deformed Shape).........................................................................................10
Strains EZZ (outer face, Deformed Shape)........................................................................................... 10
Strains EZZ (Inner face, Deformed Shape)............................................................................................10
2.2.2. Physical + Geometrical Nonlinearity.............................................................................................. 11
1
Displacements (Z direction, deformed shape)......................................................................................12
Displacements (Y direction, deformed shape).......................................................................................12
Stresses SZZ (outer face, Deformed Shape)......................................................................................... 12
Stresses SZZ (Inner face, Deformed Shape).........................................................................................13
Strains SZZ (outer face, Deformed Shape)...........................................................................................13
Strains SZZ (Inner face, Deformed Shape)............................................................................................ 13
2.2.3. Physical + Geometrical Nonlinearity (with point loads)..................................................................14
Displacements (Z direction, deformed shape)......................................................................................15
Displacements (Y direction, deformed shape).......................................................................................15
Stresses SZZ (outer face, Deformed Shape)......................................................................................... 16
Stresses SZZ (Inner face, Deformed Shape).........................................................................................16
Strains SZZ (outer face, Deformed Shape)...........................................................................................16
Strains SZZ (Inner face, Deformed Shape)............................................................................................ 17
3. COMMENTS AND CALCULATIONS....................................................................................................... 18
3.1. Hand Calculations............................................................................................................................. 18
3.2. Comments......................................................................................................................................... 21
3.3 Conclusions........................................................................................................................................ 24

2
1. INPUT
Dimension (Quadrilateral)
Assumed 2D (Plane
Stress Field Stress)

Fig. 1.2 CQ40S Element

ELEMENT DIMENSIONS
As shown in the model Sketch.
Fig. 1.1 Model Sketch

The objective is to model a compression test that was done on FINITE ELEMENT MESH
an aluminium extrusion with an imperfection on its center.
In order to have the same mesh size for all the
element, we chose a size of 5 mm.
MATERIAL PROPERTIES
Name Aluminium
Type Isotropic
Elastic Modulus 70000 N/mm2
Poissons Ratio 0.33
Model Type Von Mises
Initial Yield 200 N/mm2
Stress

ELEMENT PROPERTIES
Name ext
Data Name ext
Type Curved Shell
(Hyperbolic)
DIANA Element CQ40S
Type
Nodes 8
Degrees of 5x 8 Nodes = 40 D.O.F.
Freedom (Displacements and
Rotations)
Integration 2 x 2 x 3 = 12
Scheme Integration
points
Type of Numerically
Integration Integrated
Type of 2th order
Interpolation (quadratical)
Shape 3D (Curved) Fig. 1.3 Finite Element Mesh Size = 5 mm
Dimension
Topological 2D
3
4
BOUNDARY CONDITIONS Doing a preliminary linear analysis with a random
load we can interpolate the load required to get
CONSTRAINTS (AT THE BOTTOM) such displacement. From this analysis we get a
distributed load qa=0.038 N/mm2
For surfaces normal to the X Axis: UX, UZ, RX, RY,
RZ for each node. It is important to take into account that in this
For surfaces normal to the Y Axis: UY, UZ, RX, RY, preliminary analysis we have to set a constraint
RZ for each node. UZ for the top part of the extrusion. This is
because the prescribed displacements that are
going to be applied in the analysis constraint the
extrusion in this zone for this direction (putting a
constraint in Z is the same as putting a
displacement with a value of zero in the model,
which is the initial step for the nonlinear analysis)

Fig. 1.4 Constraints at the bottom of the element

CONSTRAINTS (AT THE TOP)


For surfaces normal to the X Axis: UX, UZ, RX, RY,
RZ for each node.
For surfaces normal to the Y Axis: UY, UZ, RX, RY,
RZ for each node.

Fig. 1.6 Applied distributed load

Fig. 1.5 Constraints at the top of the element

APPLIED LOADS
To simulate an imperfection at the center of the
element, we apply a distributed outward area
load qa over an area of 40x40 mm in order to
have a displacement of 0.1 mm.

5
Fig. 1.7 Resulting Displacement for qa=0.038 N/mm2

6
PRESCRIBED DISPLACEMENTS Fig. 1.8 Prescribed Displacement

Using elastic and plastic theory we obtain an


approximate value of 0.674 mm for the maximum
elastic capacity (see calculations in part 3). In
order to be able to observe the plastic behavior of
the element after reaching these limit, we apply a
prescribed displacement of 2 mm on each node
all over the top of the element.

7
2. ANALYSIS RESULTS
2.1. STRUCTURAL LINEAR ANALYSIS
Since there are no load steps on the linear
analysis, we apply both the imperfection load qa
and the prescribed displacement in the same
Load Case.

REACTION FORCES
Total Reaction Force on top= 121737.7

Fig. 2.1.1 Reaction Forces FB(Y) ; N

DISPLACEMENTS (Y DIRECTION , DEFORMED


SHAPE )

8
Fig. 2.1.2 Displacements DyY(V) ; mm Fig. 2.1.4 Stresses SZZ Outer Face ; N/mm2

Max Stress SXX= 790 N/mm2 (>>> Yield Strenght)

STRESSES SZZ (INNER FACE, DEFORMED


SHAPE )

DISPLACEMENTS (Z DIRECTION, DEFORMED


SHAPE )

Fig. 2.1.5 Stresses SZZ Inner Face ; N/mm2

Max Stress SXX= 696 N/mm2 (>>> Yield Strenght)

Fig. 2.1.3 Displacements DyZ(V) ; mm


STRAINS EZZ ( OUTER FACE, DEFORMED
SHAPE )

STRESSES SZZ (OUTER FACE, DEFORMED


SHAPE )

Fig. 2.1.6 Strains EZZ Outer Face ; N/mm2

9
Fig. 2.1.7 Strains EZZ Outer Face ; N/mm2

STRAINS EZZ ( INNER FACE, DEFORMED


SHAPE )

10
2.2. NON-LINEAR ANALYSIS

2.2.1. PHYSICAL NONLINEARITY

In order to simulate the imperfection in the 60000


behavior of the extrusion, we set up two load
50000
cases and subsequently two execute blocks: one
for the applied load qa and other for the 40000
prescribed displacement. In the non-linear 30000
Reaction Force N
analysis, we first apply the distributed load qa at 20000
once and then we apply the prescribed
10000
displacement using load steps.
0
Material additional Properties for Non-linear 0 0.5 1 1.5 2 2.5
Analysis
Displacement mm
Class: Steel
Material Model: Von Mises and Tresca Plasticity
Hardening Function: No hardening
Fig. 2.2.1.1 Force vs. Displacement plot Physical
Nonlinearity
Iteration and Solution Properties (For
applied load execute block)
Max Force 51900 N at =0.90 mm
Equilibrium iteration method: Newton-
Raphson In order to show properly the behaviour of the
Type: Regular extrusion in the different steps of the analysis, we
First Tangent: Tangencial show the output of it in the following steps:
Convergence Norm: Energy Based
Convergence Tolerance: 0.0001 Step 3 (Elastic Behaviour) =0.40 mm
Number of Steps: 1 Step 10 (Elastic Behaviour, just before
Step Size: 1 yielding) =0.75 mm
Max. Number of Iterations: 10 Step 25 (Yielding, Peak Load) =0.9 mm
Step 40 (Yielding, Stable load) =1.05 mm
Iteration and Solution Properties (For Step 60 (Yielding, Stable load) =1.25 mm
prescribed displacement execute block)

Equilibrium iteration method: Newton-


Raphson
Type: Regular
First Tangent: Tangencial
Convergence Norm: Energy Based
Convergence Tolerance: 0.0001
Number of Steps: 134
Step Size: 0.10 (3). 0.05(1), 0.005(130)
Max. Number of Iterations: 10

Fig. 2.2.1.2 Location of characteristic steps


DISPLACEMENTS (Z DIRECTION, DEFORMED SHAPE)

Fig. 2.2.1.3 Displacements DyZ(V) for steps 3,10,25,40,60 ; mm

DISPLACEMENTS (Y DIRECTION , DEFORMED SHAPE)

Fig. 2.2.1.4 Displacements DyY(V) for steps 3,10,25,40,60 ; mm

STRESSES SZZ (INNER FACE, DEFORMED SHAPE)

Fig. 2.2.1.5 Stresses SZZ Outer Face for steps 3,10,25,40,60; N/mm 2

STRESSES SZZ (OUTER FACE, DEFORMED SHAPE)


Fig. 2.2.1.6 Stresses SZZ Inner Face for steps 3,10,25,40,60; N/mm 2

STRAINS EZZ (INNER FACE, DEFORMED SHAPE)

Fig. 2.2.1.7 Strains EZZ Outer Face for steps 3,10,25,40,60; N/mm 2

STRAINS EZZ (OUTER FACE, DEFORMED SHAPE)

Fig. 2.2.1.8 Strains EZZ Inner Face for steps 3,10,25,40,60; N/mm 2

2.2.2. PHYSICAL + GEOMETRICAL NONLINEARITY


In order to model the buckling of the section that 50000
is shown in the experiment, we add geometrical
nonlinear effects to the analysis. 40000

30000
Material additional Properties for Non-linear
Analysis Reaction Force N 20000

Class: Steel 10000


Material Model: Von Mises and Tresca Plasticity
0
Hardening Function: No hardening
0 0.5 1 1.5 2 2.5
Iteration and Solution Properties (For Displacement mm
applied load execute block)

Equilibrium iteration method: Newton-


Raphson Fig. 2.2.2.1. Force vs. Displacement plot Physical +
Type: Regular Geometrical Nonlinearity
First Tangent: Tangencial
Convergence Norm: Energy Based Max Force = 43520 N at =0.73 mm (Step 8)
Convergence Tolerance: 0.0001
In order to show properly the behaviour of the
Number of Steps: 1
Step Size: 1 extrusion in the different steps of the analysis, we
Max. Number of Iterations: 10 show the output of it in the following steps:
Step 2 (Elastic Behaviour) =0.20 mm
Step 4 (Elastic Behaviour, just before
Iteration and Solution Properties (For yielding) =0.60 mm
prescribed displacement execute block)
Step 8 (Yielding, Peak Load) =0.73 mm
Step 20 (Yielding, decreasing Load)
Equilibrium iteration method: Newton-
Raphson =0.85 mm
Type: Regular Step 60 (Yielding, decreasing load) =1.25
First Tangent: Tangencial mm
Convergence Norm: Displacement Based
Convergence Tolerance: 0.01 50000
Number of Steps: 134
Step Size: 0.10 (3). 0.05(1), 0.005(130) 40000
Max. Number of Iterations: 100
30000
Geometrical Settings Reaction Force N 20000
Type of geometrical nonlinearity: Total
Lagrange 10000

0
0 0.5 1 1.5 2 2.5
Displacement mm

Fig. 2.2.2.2. Location of characteristic steps

DISPLACEMENTS (Z DIRECTION, DEFORMED SHAPE)


Fig. 2.2.2.3 Displacements DyZ(V) for steps 2,4,8,20,60 ; mm

DISPLACEMENTS (Y DIRECTION , DEFORMED SHAPE)

Fig. 2.2.2.4 Displacements DyY(V) for steps 2,4,8,20,60; mm

STRESSES SZZ (INNER FACE, DEFORMED SHAPE)

Fig. 2.2.2.5 Stresses SZZ Outer Face for steps 2,4,8,20,60; N/mm 2

STRESSES SZZ (OUTER FACE, DEFORMED SHAPE)


Fig. 2.2.2.6 Stresses SZZ Inner Face for steps 2,4,8,20,60; N/mm 2

STRAINS SZZ (INNER FACE, DEFORMED SHAPE)

Fig. 2.2.2.7 Strains EZZ Outer Face for steps 2,4,8,20,60

STRAINS SZZ (OUTER FACE, DEFORMED SHAPE)

Fig. 2.2.2.8 Strains EZZ Inner Face for 2,4,8,20,60


2.2.3. PHYSICAL + GEOMETRICAL NONLINEARITY (WITH POINT LOADS)

In order to find the effects of the distribution of


load that simulate the imperfection, and also to
have a better possibility to check the experiment
and the modeling by hand calculation, we apply a
point load instead of distrubted load.

To simulate an imperfection at the center of the


element, we apply a point load in the center of
each flange (instead of applying it directly in the
web) in order to have a displacement of 0.1 mm.

Doing a preliminary linear analysis with a random


load we can interpolate the load required to get
such displacement. From this analysis we get a
point load of 260 N on each flange. Fig. 2.2.3.2. Resulting displacement for point load of 260 N
on each flange

The non linear analysis is done using the same


properties for the Material, Iteration and Solution
and Geometrical Effects as in the previous
analysis

50000

40000

30000
Reaction Force N
20000

Fig. 2.2.3.1. Applied point load of 260 N on each flange 10000

0
0 0.5 1 1.5 2 2.5
Displacement mm

Fig. 2.2.3.3. Force vs. Displacement plot Physical


Nonlinearity + Geometrical Nonlinearity with point loads

Max Force = 43825 at =0.73 mm (Step 8)

In order to show properly the behaviour of the extrusion in the different steps of the analysis, we show the
output of it in the following steps:

Step 2 (Elastic Behaviour) =0.20 mm


Step 4 (Elastic Behaviour, just before yielding) =0.60 mm
Step 8 (Yielding, Peak Load) =0.73 mm
Step 20 (Yielding, decreasing Load) =0.85 mm
Step 60 (Yielding, decreasing load) =1.25 mm
50000

40000

30000
Reaction Force N
20000

10000

0
0 0.5 1 1.5 2 2.5
Displacement mm

Fig. 2.2.3.4. Location of characteristic steps

DISPLACEMENTS (Z DIRECTION, DEFORMED SHAPE)

Fig. 2.2.3.5 Displacements DyZ(V) for steps 2,4,8,20,60 ; mm

DISPLACEMENTS (Y DIRECTION , DEFORMED SHAPE)

Fig. 2.2.3.6 Displacements DyY(V) for steps 2,4,8,20,60; mm

STRESSES SZZ (INNER FACE, DEFORMED SHAPE)


Fig. 2.2.3.7 Stresses SZZ Outer Face for steps 2,4,8,20,60; N/mm 2

STRESSES SZZ (OUTER FACE, DEFORMED SHAPE)

Fig. 2.2.3.8 Stresses SZZ Inner Face for steps 2,4,8,20,60; N/mm 2

STRAINS SZZ (INNER FACE, DEFORMED SHAPE)

Fig. 2.2.3.8 Strains EZZ Outer Face for steps 2,4,8,20,60

STRAINS SZZ (OUTER FACE, DEFORMED SHAPE)


Fig. 2.2.3.9 Strains EZZ Inner Face for steps 2,4,8,20,60
3. COMMENTS AND CALCULATIONS
3.1. HAND CALCULATIONS
In order to perform the respective hand We idealize this situation as a column with length L=300 clamped
on top and bottom, with an applied horizontal load on its center and subsequent prescribed displacements.

From the geometrical properties of the section we can calculate the Elastic Moment and the Plastic
moment. The elastic moment is the moment due to the applied force that causes a displacement of 0.1
mm, which can be calculated with elastic theory. After that we calculate the elastic moment using
equilibrium.

192EI
F= 3
L

FL
M el=
8

Elastic cross section


b 80 mm
h 25 mm
L 300 mm
t 2 mm
E 70000 N/mm2
Atot 252 mm2
fy;d 200 N/mm2

z0 5,56 mm
zb 19,44 mm +
h 25,00 mm

I1 3216 mm4
I2 10020 mm4 +
I 13236 mm4

M0;el 24707 Nmm 1/8FL


b;1 -36,28 N/mm2 M*zb/I
b;2 10,39 N/mm2 M*zo/I

u 0,1 mm
u*192EI/
Fhor 659 N L3

Table 3.1.1. Elastic cross section calculations

On the other hand, the plastic moment can be calculated using plastic theory. We can also calculate the
applied force necessary to reach this plastic moment, as this happens when all the fibers of the section
reach the yield strenght of the material.

F pl =f y A

M pl=f yW pl
Plastic cross section
(Full)
z0 1,58 mm *A/b
zb 23,43 mm h-zo
h 25 mm zo+zb
S1 99 mm3 *b*zo2
A1 92 mm
z1 11,93 mm
S2 1097 mm3
A2 34 mm
A1 + A2 126 mm
z2 0,21 mm
S3 7 mm3
S1+S2+S
Wpl;u;d 1204 mm3
3

24071 Wpl;u;d *
Mpl;u;d Nmm
0 fu;d
Fpl;u;d 50400 N A * fu;d

Table 3.1.2. Plastic cross section calculations

This is the situation that is modeled in the first analysis (just physical nonlinearity). However, due to the
results of the modeling compared to the ones of the experiment (see section 3.2.), we also need to include
geometrical nonlinearity effects.

In order to do this, we gradually increase the elastic moment and the supports by increasing it with the
effect of the load eccentricity due to the displacement in the center of the extrusion, until it reaches the
yield strenght (elastic limit).

After that, we find a force F vert that generates a moment at the supports that is equal to the plastic
moment of a reduced cross section. Such cross section is the cross section of the extrusion, excluding the
area bearing the axial force of the force F vert and the area bearing the shear force due to the horizontal load
applied in the center of the extrusion.

Plastic cross section


(Reduced)
Fvert 47570 N
Vd 329 N *Fhor
AVd 2,85 mm2 Vd*31/2/fy;d
Atot/2 126,00 mm2
AF/2 118,93 mm2 *F/fy;d
mm2/sid
Anet 5,65 Atot/2-AF/2-Avd/2
e
h1 0,07 mm
a1 1,54 mm
S1 9 mm3 Anet*a1
h2 1,41 mm
a2 22,72 mm
S2 128 mm3 Anet*a2
Wpl;u;d 137 mm3 Sl+S2
Mpl;u;d 27401 Nmm Wpl;u;d * fy;d
Table 3.1.3. Reduced Plastic cross section calculations

A 1= A 2

Fig. 3.1.1. Reduced plastic cross section

We can calculate the buckling load using eulers formula with a buckling length L buc=0.5L=150 mm (under
the assumption of double clamping).

2 EI
Fbuc =
Lbuc2

We calculate the horizontal displacement in the center of the extrusion with:

n
uend = u
n1 0

With n=Fbuc/Fvert.

Elastic Plastic
N,
40641 40641 40641 40641 40641 4757 4757 4755 4755 4755
Fbuc 406417 Lbuc = 150 2EI/Lbuc2
7 7 7 7 7 0 0 9 9 9
mm
4756 4756 4755 4753 4719
Fvert 0 10000 20000 30000 39631 47570 N
7 3 9 7 5
n 1 40,64 20,32 13,55 10,26 8,54 Fbuc/Fvert
n/(n-
1 1,03 1,05 1,08 1,11 1,13
1)
uo 0 0,100 0,100 0,100 0,100 0,100 0,100 0,100 0,100 0,100 0,100 mm
uend 0 0,103 0,105 0,108 0,111 0,113 0,115 0,117 0,119 0,130 0,300 mm n/(n-1)*uo
2470 2470 2470 2470 2470
Mo 24707 24707 24707 24707 24707 24707 Nmm
7 7 7 7 7
2744 2749 2753 2779 3178 Mo+*uend*F
Mend 24707 25732 26811 27946 29098 27401 Nmm
2 0 7 7 6 vert

Fvert 0 -40 -79 -119 -157 N/mm2 Fvert/Atot


M;end -36 -38 -39 -41 -43 N/mm2 + Mend*zb/I
max -36 -77 -119 -160 -200 -200 -200 -200 -200 -200 -200 N/mm2
wvert;top 0 0,170 0,340 0,510 0,674 0,809 mm Fvert*L/(EA)

Table 3.1.3. Geometrical nonlinearity calculations

Load/displacement top
60000
47570
40000 39631
Load Fvert [N] 30000
20000 20000
10000
0 0
0.000 0.200 0.400 0.600 0.800 1.000
wvert [mm]

Fig. 3.1.2. Load vs Displacement on top of the extrusion

Geometric non-linearity
60000

50000
47195
40000

30000
Load Fvert [N]
20000

10000

0
0.1000.1500.2000.2500.3000.350

uhor mid section [mm]

Fig. 3.1.3. Load vs horizontal Displacement of the extrusion

With this calculations we find the limit for the elastic capacity at F=39631 N with a prescribed
displacement of 0.674 mm, and a limit for the plastic capacity at F=47570 N with a prescribed
displacement of 0.809 mm.

3.2. COMMENTS
First we are going to compare the results of the linear analysis and the first two nonlinear analysis, against
the hand calculations for the Geometrical nonlinearity.
60000

50000

40000
Physical Nonlinearity
30000 Phyisical +
Reaction Force N
Geometrical
20000 Nonlinearity
Linear
10000
Hand Calculations
0
0 1 2

Displacement mm

Fig. 3.2.1. Comparison between analysis

Fig. 3.2.2. Experiment Results

From this we can see that geometrical nonlinearity must be used in order to obtain an output for the
extrusion that is closer to its real behaviour. Without the geometrical nonlinearity the extrusion can keep
deforming undefinitely over a constant load, which is not the case of the experiment since the extrusion
clearly reaches a failure after which the load decreases under a increasing displacement.

Taking into account the geometrical effects, we get a curve that is very close to the one obtained in the
experiment, which confirms that the selection of the boundary conditions was appropiate for this model.
The maximum load is basically the same to the one obtained in the experiment, but the softening part of
the curve is slightly lower than the one from the experiment (in the experiment the load tends to stabilize
at 30 KN while in the model it keeps going down).

It must be taken into account as well that the deformed shape of the model is very similar to the final
shape of the extrusion after the experiment, which also confirms that the model output is very close to the
real behaviour of the extrusion.
From the hand calculations we obtain a plastic limit that is about 4 KN higher (47,57 vs. 43,83 kN = ~ 9
%) then the one derived from the experiment and from the model. This can be declared with the fact that
the n/(n-1) method is a elastic method to calculate the second order displacements according to buckling
and bending. In the quasi-linear part of the experiment, the strains (and displacements) are not linear any
more with the loads and the load-displacement line will become curved and more flat.

To evaluate the effects of the imperfection in the behaviour of the extrusion, we compare the results
between the model with an distributed load (applied in the web) that generates a displacement of 0.1 mm,
and a point load (applied in the flanges) that generates the same displacement.

It must be taken into account that with the area load we get a 0.1 mm displacement just in the mid point
of the web while with the point load we get a displacement of 0.1 in a line over the center of the web (see
output above). Hence the total load applied in the first case (0.038x40x40=608 N) is much lower than the
one applied in the second case (260x2=520 N).

50000

40000

30000
Reaction Force N
20000 Area Loading
Point Loading
10000

0
0 0.5 1 1.5 2
Displacement mm

Fig. 3.2.2. Area Loading vs. Point loading comparison

From this graph we conclude that, even though both applied loads are different, the behaviour of the
extrusion tends to be similar. The slight difference between the maximum load can be explained with the
difference between the values of the two applied loads. But apart from that, the imperfection does not
seem to have a big influence in the capacity and behaviour of the extrusion. This can be confirmed by
running a new analysis without applying any loads to simulate imperfections.

50000

40000

30000

Reaction Force N Area Loading


20000
Point Loading
10000 No imperfection

0
0 1 2
Displacement mm
Fig. 3.2.2. Comparison, with and without imperfection

3.3 CONCLUSIONS
-This is a satisfactory modeling of the experiment because the maximum load and the prescribed
displacement corresponding to that load are very similar between the model and the experiment. There is
a slight difference in the behaviour of the model after reaching that peak (softening) compared to the
experiment, which may be due to other material and element properties that cannot be taken into account
for the modeling. But in terms of getting the actual capacity of the extrusion, the experiment was properly
modeled.

-To perform the hand calculations, it is correct to idealize this experiment as a double clamped column.
However, it is not possible to take into account other factors that greatly affect the behaviour of the
element (such as local stiffness) which is why the results vary between the calculations and the model (for
example, the load needed to get a 0.1 mm dislacement varies from 659 N for the hand calculations to 520
N for the model). But doing hand calculations is still a good approach to understand the behaviour of the
extrusion and to detect possible mistakes done in the modeling.

-From Figure 3.2.2. we can conclude that, even though both applied loads are different, the behaviour of
the extrusion tends to be similar. The slight difference between the maximum load can be explained with
the difference between the values of the two loads. But apart from that, the imperfection does not seem to
have a big influence in the behaviour of the extrusion. This can be confirmed by running a new analysis
without applying any loads to simulate imperfections. We can say that due to the short length of the
extrusion and its geometrical properties, the geometrical nonlinear effects due to the imperfection are not
relevant in the whole behavior of the extrusion. However, geometrical nonlinear effects due to other
factors (i.e. local buckling of the web and flanges) must be taken into account in order to be able to model
the experiment properly, even though these effects cannot be hand calculated.

-In order for the nonlinear analysis to converge (especially when taking into account geometrical nonlinear
effects) we have to select the right analysis and iteration parameters. For example, using smaller load
steps and increasing the maximum number of iterations can increase significantly the computation time,
but must be done due to the complex behavior of the extrusion in the nonlinear range.

Vous aimerez peut-être aussi