solidworks simulation tutorial REVISED.pdf
Embed Size (px)
Transcript of solidworks simulation tutorial REVISED.pdf
- 1 -
SolidWorks Simulation 14 June 2011
Sam Ettinger, HMC 2012
By completing this tutorial, you will learn to conduct finite-element analysis (FEA) tests on
SolidWorks models using the Simulation add-on.
Finite-element analysis (FEA) is useful in predicting a model’s response to various influences
such as forces, torques, periodic excitations, and heat. FEA is used to analyze large or
complicated models where analytical solutions are not possible. FEA software breaks the model
into thousands of small tetrahedral elements and solves numerically for each one individually.
Some of the leading commercial FEA tools include COMSOL, Ansys, and SolidWorks
Simulation. This tutorial covers SolidWorks Simulation because it is a comfortable environment
for those who already know 3D modeling with SolidWorks. SolidWorks Simulation is primarily
applicable to mechanical and thermal models. COMSOL specializes in multiphysics problems
involving interaction between mechanical, thermal, and electrical behavior. Ansys also addresses
mechanical and thermal simulations and has advanced capabilities required in certain fields.
- 2 -
This tutorial will cover three of the simulation studies available in SolidWorks Simulation:
Static analysis, for identifying stresses caused by static loading
Frequency analysis, for identifying resonant frequencies and associated mode shapes
Thermal analysis, for identifying heat flow through a model
Mastering these three gives you the tools and experience necessary to make use of any of the
remaining simulation studies. We will use the same model, a wine glass, for each study.
Before you begin
On two occasions I have been asked, “Pray, Mr. Babbage, if you put into
the machine wrong figures, will the right answers come out?”...I am not
able rightly to apprehend the kind of confusion of ideas that could provoke
such a question.
- Charles Babbage, Passages from the Life of a Philosopher, 1864
Since the very beginning of computing, users have been plagued by bad outputs as a
result of bad inputs. FEA is particularly prone to such problems, generating pretty
pictures that often have no bearing on reality. As a general rule, if you don’t know
what to expect, the results you get are probably incorrect and certainly unusable.
Some common reasons for error include making invalid assumptions, setting incorrect
boundary conditions, setting incorrect material properties, and general numerical
errors. As an example, let’s consider the stress analysis of a typical model. One can
assume a linear stress-strain relationship before the yield strength of the material is
reached. The model produces incorrect results as stress or strain increase to the point
of nonlinearity on the stress-strain curve.
Setting appropriate boundary conditions can be more difficult than one might first
expect, and it is easy to overlook boundaries (such as the initial temperature of an
object). This can result in nonsensical default values being used for the simulation.
Obtaining accurate physical parameters for your materials can also be difficult,
especially if you are using nonstandard or unusual materials.
FEA inherently discretizes the object being studied. The number of elements used
presents a tradeoff between runtime and accuracy.
Before you trust your FEA results, you should plan for a significant validation process.
If possible, the best place to begin the validation process is with a simple model that
can be solved analytically. Check that the FEA simulation produces comparable
results. For example, before looking at bending in an array of bolted-together I-beams,
compare the FEA results for bending in a single I-beam to analytical results. Be sure
you are using the appropriate material parameters. Another important model validation
technique is to compare the FEA results to an actual physical prototype using simple
stimulus, such as an impulse. Beware of making claims about the FEA model results
that you cannot independently support with other analysis or measurement.
- 3 -
Getting the model, defining the material
SolidWorks is, first and foremost, a 3-D modeling tool. Look in the tutorials folder for a
SolidWorks model named “glass.SLDPRT”. When you open it, it should show a wine glass, as
in Figure 1. This is the model we are going to study in this tutorial. Save a copy to your Charlie
If you wish to practice your modeling, part of a technical drawing of this wine glass is included
in Appendix A. Units are in millimeters. This should be enough information to draw your own.
Figure 1. The wine glass model.
To use the model of the glass in Simulation, we must specify the material that the model is made
of. Can you guess what material we want the model to be? That’s right, glass! You can specify a
material in the Feature Manager design tree on the left-side panel. There should be an icon
named “Material <not specified>,” as in Figure 2. Right-click this and choose “Edit material” to
be brought to the Materials window, shown in Figure 3. Glass is found in the folder SolidWorks
Materials > Other Non-metals. Click “Glass,” then click “Apply” and “Close.” Your model
should change from opaque grey to transparent grey, as in Figure 4.
- 4 -
Figure 2. How to change a model's material.
Figure 3. Materials window.
- 5 -
Figure 4. The glass, now shown as glass.
Opening Simulation for the first time
By default, SolidWorks Simulation does not open when SolidWorks does. We can change this by
going to Tools > Add-Ins, shown in Figure 5.
Figure 5. How to start Simulation.
The Add-Ins window, replicated in Figure 6, pops up. Check the box to the left of “SolidWorks
Simulation” to enable Simulation in this instance of SolidWorks. If you want Simulation to be
enabled every time you start SolidWorks, check the box to the right of “SolidWorks Simulation”
as well. If this is your first time using Simulation, you may be asked to agree to an end-user
- 6 -
Figure 6. The add-ins window.
If all goes right, there should be a new tab named “Simulation” in the upper-left, next to
“Features,” “Sketch,” “Evaluate” and the like. Now we can begin our first study!
As mentioned above, static analysis computes the effects of static loading on a model. It can
display stresses, strains, displacement, and the factor of safety at each segment of a model. In
Simulation, one has to specify the location and magnitude of each load, as well as to specify
where and how the model is supported. Identifying where stresses are highest/lowest quickly
shows the designer where a model can be improved by adding support or by removing excess
We are going to investigate what happens when a 5 kg load is placed on the lip of a glass set
upright on a table. Let’s assume that placing the glass on a table can be best approximated by a
perfectly fixed support on the entire bottom face of the wine glass base. Furthermore, let’s
assume the 5 kg load is best approximated by a 50 N force pushing against the lip of the glass
with uniform distribution. In models of large objects, it is often advisable to include gravity in
the simulation, but that is not necessary for our 10-cm-tall wine glass.
To create a static study, click the Simulation tab in the upper-left. There should be a button
labeled “Study Advisor.” Click the arrow just beneath it and choose “New Study,” as in Figure 7.
Here you can see all the types of studies available in Simulation. Click “Static,” name the study
something memorable, and click the green check mark.
- 7 -
Figure 7. Starting a new study.
Below the normal display pane on the left, a static study pane should open. We can use this or
the Simulation tab along the top of the screen to specify our fixtures and loads.
To set up the fixtures on the model, either right-click “Fixtures” in the static study pane or click
the arrow beneath “Fixtures Advisor” in the Simulation tab. Choose “Fixed Geometry” as the
fixture type for this study. This is shown in Figure 8. You can also have supports such as pins,
rollers, or hinges, if your model requires it. For now, though, the fixed geometry suffices. When
you click “Fixed Geometry,” the Fixture pane opens on the left. Select the bottom face of the
base of the glass and press the green check mark. You can select multiple faces at a time, if you
wish, but for this example only one face is fixed.
Figure 8. Adding a fixture.
To specify our 50 N load, right-click “External Loads” in the static study pane or click the arrow
beneath “External Loads Advisor” in the Simulation tab. As you can see, there are lots of
- 8 -
possible options for loading, but we just care about the simplest one, “Force.” Click that to be
taken to the Force/Torque pane. Click the lip of the glass (you may have to zoom in a ways to
make sure you are selecting the whole face, not just one edge) to select it as the face being
loaded. Rather than having the load normal to the lip, let’s specify the load as directly down (as it
would be in reality). To do this: in the Force/Torque pane, change the direction of the force from
“Normal” to “Selected direction,” then choose the Top Plane in the design tree just to the right of
the Force/Torque pane. The design tree and the lip of the glass are shown in Figure 9. By default,
the design tree is condensed to just show the name of the part; click the boxed + to the left in
order to expand it.
Figure 9. Selecting a face and a direction to apply force.
The design tree is the tree in the upper-left corner of the figure.
- 9 -
In the Force section of the Force/Torque pane, we have to specify the magnitude and direction of
our load. Click the button marked “Normal to Plane” and specify 50 N as the force. We want the
force aimed down, so check “Reverse direction” as well. These are shown in Figure 10. Finally,
click the green check mark.
Figure 10. Specifying force and direction in the Force/Torque pane.
With the forces and fixtures specified, we can run a finite element analysis now! SolidWorks
needs to break the model into small tetrahedral units, which together are called a mesh. Smaller
meshes (as in meshes with smaller individual units) produce more precise results but require
additional computing time. Large meshes run quickly but may produce wildly inaccurate results,
especially around sharp edges. It is common to use a mesh with varying element sizes: smaller
units around the areas of interest in a model, such as potential failure points, and larger units
where precise results are less valuable.
In the static study pane, right-click “Mesh” and choose “Create Mesh.” Accept the default mesh
size and check OK. This will create a uniformly sized mesh over your entire model, which
should look something like Figure 11. If you ever need a non-uniform mesh, you can do so by
right-clicking “Mesh” and choosing “Apply Mesh Control” instead.
- 10 -
Figure 11. Meshed glass.
Begin the static study by clicking “Run” in the Simulation tab. You will see that even this simple
problem consumes significant memory and time. If all goes well, a folder named “Results” will
appear in the static study pane. Right-click the folder and choose “Define Stress Plot,” then
accept the default settings that appear. This will show you the von Mises stresses from the 50 N
load by coloring the mesh. You should see the greatest stress occurring at the joint between the
stem and the bowl, shown in Figure 12. Is this reasonable?
- 11 -
Figure 12. Stress plot close-up.
To see the interior of the bowl better, we can slice the model in half. While looking at the stress
results, click “Plot Tools” in the Simulation tab and open “Section Clipping.” Set it to cut along
the front plane and check OK. Figure 13 shows an elegant cutaway view and an easy way of
observing the interior of the model!
- 12 -
Figure 13. Stress plot cutaway.
Create a new result plot to show displacement. Right-click “Results” again and choose “Define
Displacement Plot.” In the settings pane that appears, set “Deformed Shape” to True Scale and
check OK. By default the displacement is measured in URES (“resultant displacement”—U is
commonly used to abbreviate displacement), which is a simple measure of displacement
magnitude. Measuring displacement along the X, Y, or Z axes is also an option here, though we
will stick with URES, like in Figure 14.
- 13 -
Figure 14. Displacement plot.
The actual displacement caused by the 50 N load is very small, on the order of microns, and will
not be visible on this model. To make deformations visible, we need to change the displacement
settings. Right-click the actual Displacement plot in the results folder of the static study pane and
choose “Edit Definition.” Set the Deformed Shape to “Automatic.” This plot tool greatly
exaggerates the displacements, multiplying them by a factor of 5000 or so, so that they are
visible. Now we see the bowl squashed, the stem shortened, and the connection between stem
and bowl beginning to cave as in Figure 15. Again, the view can be improved by going into Plot
Tools and setting up Section Clipping, as in Figure 16.
- 14 -
Figure 15. Exaggerated displacement.
Figure 16. Exaggerated displacement, cutaway.
- 15 -
What to turn in
A screenshot of your model, showing the cutaway stress plot of the model.
A screenshot of your model, showing the deformation plot, with displacement
Frequency analysis is done to models to identify resonant frequencies and mode shapes. The
wine glass is a good model to run frequency testing on because at least one of its resonant
frequencies yields a mode shape that can shatter the bowl while leaving the stem intact. This is a
great party trick, assuming your party includes a powerful speaker and a tone generator:
http://youtu.be/17tqXgvCN0E. We are going to find that resonant frequency using finite-element
Create a new study by clicking the arrow under “Study Advisor” and selecting “New Study.”
This time, choose the “Frequency” option and rename the study to something relevant.
A frequency analysis pane will appear on the left side of the screen. This is very similar to the
static pane from before; it lets you add fixtures (which are obligatory) and external loads (which
are optional). Add the same fixtures from the static study; do not add any external loads. This is
all that is necessary to begin the frequency analysis!
Run the analysis just as in the static study. When it finishes, right click “Results” in the
frequency study pane. The results of the study that we care about most are the mode shapes, so
click “Define Mode Shape/Displacement Plot.” A new pane, shown in Figure 17, should open to
control which mode shape to show.
- 16 -
Figure 17. Mode Shape/Displacement pane.
The “Plot Step” section lets you specify which mode shape to visualize, as well as its associated
resonant frequency. Since the wine glass has plenty of contours and such, there will be several
resonances with unrelated mode shapes. We care about finding a mode shape that will result in
the glass bowl shattering without damaging the stem or base. The best way to investigate this is
to go through the mode shapes one-by-one. Beginning with mode shape 1, check OK. The result
is shown in Figure 18.
- 17 -
Figure 18. Still image from mode shape 1.
The mode shape may be easier to visualize if animated. In the Simulation tab, click Plot Tools >
Animate. In a few seconds, SolidWorks will show a looping animation of the glass bending
along its first mode shape (deformation is exaggerated and the period is increased greatly). You
can rotate the model now, just as you would if it were inanimate. Now you can see that the stem
of the glass is bending back and forth, which means this is not the desired mode shape. Time to
try the other ones!
Cancel the animation with the red X in the animation pane. Right-click the displacement study in
the Results folder of the frequency study pane and select “Edit definition.” Change the mode
shapes in the “Plot Step” section. Mode shape 2 has a resonant frequency that is less than half of
one Hertz away from mode shape 1, which strongly suggests the resulting deformation will be
the same as before, but along a different plane. The Animation function confirms this. Mode
shape 3 looks like a plausible bowl-shattering shape, as do shapes 4 and 5 (which have nearly
identical resonant frequencies). In fact, any one of these three will do the trick and shatter the
To see a list of the resonant frequencies, right-click “Results” in the analysis pane and choose
“List Resonant Frequencies.” If you are not interested in the specifics of the model’s mode
shapes, this is a quick way to quantify resonances. You can save the results as a .csv (comma
separated values) file, easily viewed in Microsoft Excel or in MATLAB.
- 18 -
What to turn in
Complete the following table:
Mode No. Frequency (Hz) Plain-language description of mode shape motion
Thermal studies are very different from structural analyses such as the static and frequency
analyses above. Still, the process has many similarities. Thermal analysis in SolidWorks can be
used to approximate heat flow through an object before or after the system reaches a steady-state
condition. Instead of fixtures and loads, we specify the initial temperature of faces and any heat
flux in/out of the system.
Obtaining good thermal analysis of a model is generally more involved than obtaining good
static analysis. Wherever the thing being modeled contacts a fluid at a different temperature—the
air surrounding it, for example, or flowing water inside a pipe—we have to specify a convection
coefficient to define how heat flows in our simulation. Finding an accurate convection coefficient
is not trivial; it requires thermodynamic calculations that vary from system to system, and it is
probably accompanied by making numerous assumptions about the model to simplify behavior.
Because the convection coefficient has a good deal of uncertainty, your finite element results
will also have corresponding uncertainty. It may be prudent to explore your results over a range
of possible convection coefficients to understand your sensitivity to the unknown parameter.
In our thermal analysis, we are going to investigate something straightforward: the temperature
profile of an empty wine glass being held by the stem or underneath the bowl. We’ll do this with
both steady-state conditions, which produce one plot showing the temperature throughout the
system after an infinite amount of time has passed, and transient conditions, which produces
several plots and shows a timeline of how heat moves and how temperature changes in the
- 19 -
Conducting a good thermal analysis on the first try requires a great deal of foresight. Before we
begin this study, consider the following:
Transient analysis takes a very long time relative to the static and frequency studies
detailed in the sections above. Fortunately, our glass is radially symmetric! For thermal
analyses, we can look at a small sliver of the glass and obtain the same results as with a
whole glass, but in a fraction of the time. There’s no limit to how small a wedge we can
use from this model, but for the sake of visibility we will make ours a 15-degree arc.
Unlike the previous studies, this thermal analysis will be done on an assembly rather than
a single part. In thermal analysis, initial conditions must be defined on entire bodies,
which is possible in assemblies but not on parts. So, after we save our wedge as a new
part, we will have to create an assembly for it as well.
o Nota bene: For any analysis in SolidWorks simulation, each individual part can
only be of one material. When performing a study on an assembly with multiple
parts, even if they are of the same material, it is critical to define the nature of the
contacts between them. This is done in the study pane, under “Connections.”
We will model the glass being held in two ways: fingers pinching the stem, and an entire
hand cradling the bowl1. We will represent these in our thermal model as segments of
faces fixed at 310 K (approximately human body temperature). It is important that we
only select segments of the model faces—for example, it would be silly to make the
entire exterior of the bowl 310 K if the hand is only supporting a portion near the bottom.
To make portions of faces selectable, we will have to create split lines in our part before
moving into assembly and analysis. This process will be covered below.
Duplicate your wine glass model in a new save file. Be sure the new file has the same units of
measurement as the original! Then, reduce the model from a full circle to a 15-degree arc, like in
Figure 19. To do this, right click on the Revolve feature. Above the pop-up menu, click on the
Edit Feature icon. Change the revolve parameters to 15 degrees.
1 Cradling the glass in this fashion is frowned upon in most wine societies. This method is the norm when holding a
snifter of brandy, however.
- 20 -
Figure 19. Just a little sliver.
Now we are going to add the split lines. These lines are what permit the selection of fractions of
a face. To start the split lines, make a sketch consisting of three lines, spaced as in Figure 20.
- 21 -
Figure 20. Spacing of split line sketches.
These are not the split lines, but they are necessary to making them. In the “Features” tab, go to
Curves > Split Line. In the split line pane that appears, choose “Projection” as the type of split,
select your new sketch as the sketch to project, and select the exterior of the bowl and stem as the
faces to split. This is shown in Figure 21. When you check OK, split lines should manifest on the
- 22 -
Figure 21. Generating split lines.
The next step is to make an assembly from this part. Save the model and click File > Make
Assembly from Part. A new window will open with a “Begin Assembly” pane. Check OK to
accept the default settings and go to your assembly. Save it immediately.
Start a thermal study via the Simulation tab, as before. In the thermal study pane, right-click the
study name and select “Properties” to open the Thermal properties window, as in Figure 22. In
this window, you should see “Steady state” selected instead of “Transient.” Keep it set to
“Steady state” for now; we will perform transient analysis later.
- 23 -
Figure 22. Thermal properties.
Instead of loads and fixtures, there are now just thermal loads to specify. For steady state
analysis, we do not have to define our initial conditions, but we do have to define where and how
heat flows through the system. Let us start by putting the 310 K human finger analogue at the
center of the stem. In the study pane, right-click “Thermal Loads,” click “Temperature,” and then
select the middle segment of the stem (in between the two split lines). In the temperature pane,
change temperature from 0 K to 310 K.
Now that we have to specified how heat enters the system, it is necessary to specify how it exits.
Let us assume that all heat is lost via convection into the surrounding air, which is at
approximately 293 K. Right-click “Thermal Loads” again, but choose “Convection.” Select only
the faces that would be exposed to air on a whole, 360-degree model of a glass. This means not
selecting the face already set to 310 K or the “imaginary” faces created by dissecting the original
model. Defining heat flow on imaginary faces is a common mistake. A helpful reminder is
shown in Figure 23.
- 24 -
Figure 23. DO NOT SELECT THAT FACE
After selecting the appropriate faces, change the settings in the convection pane. The convection
coefficient of room-temperature air is approximately 8 W/m2/K (see Appendix B). As mentioned
above, we are assuming that the ambient temperature (or “bulk ambient temperature,” since there
is much more air than there is wine glass) is 293 K. Check OK.
That’s all there is to specify for steady-state analysis. Create a mesh and run the simulation as in
the previous studies. Right-click “Results” in the study pane and select “Define Thermal Plot.”
The options here are to display temperature, temperature gradient, or heat flux. The latter two
make for useful analysis, but right now choose “Temperature” and check OK to see the
temperature profile of the glass, as in Figure 24. Right-click the result and select “Chart
Options,” then check both “Show min annotation” and “Show max annotation” to show which
nodes are hottest and coldest (of course, in this setup, there are multiple nodes of equal
temperature). Take a screenshot with annotations for the turn-ins section.
2 DO NOT SELECT THAT FACE.
- 25 -
Figure 24. Temperature profile for stem-held glass.
Not surprisingly, the center of the stem is at 310 K and the fringes of the model are at 293 K. The
body heat entering the stem hardly registers an effect on the bowl. It’s no wonder this is the
preferred way to hold a glass of chilled wine.
To create a temperature profile for holding the bowl instead of the stem, we have several options:
create a new thermal study, edit the pre-existing thermal loads, or create new thermal loads and
suppress the old ones. The first alternative is excessive, and the second alternative means
eliminating some settings we will use again, for transient analysis. It’s best to suppress them (by
selecting both thermal loads, right-clicking, and choosing “Suppress”). Create a new temperature
load on the bowl (below the split line) and a new convection load on the appropriate faces. The
numerical settings should be the same as before. Don’t hesitate to refer back to Figure 23 for
guidance. Run the study and observe the results, as before. They should resemble those in Figure
25. Again, take a screenshot of your results with annotations.
- 26 -
Figure 25. Temperature plot for bowl-cradled glass.
Now, switch to transient analysis. Right click the study name as in Figure 22 and switch the
settings to “Transient.” Pick a reasonable observation time and time increment—say, 120
seconds total time with 6-second time steps.
In addition to the surface temperature loads that we specified for steady-state analysis, we now
have to include initial conditions. Keep the temperature and convective loads that you have
already entered for the bowl-cradled glass. Add a new temperature load. Now, in the temperature
pane, “Initial temperature” should be a selectable type. Rather than selecting faces, we want to
select the entire part body, so click and drag around the part. We need to select the entire part in
this step because, if we select only the exposed faces, then SolidWorks will assume that the
interior of the model has an initial temperature of absolute zero. Set the initial temperature to 280
K, so we can observe it slowly increase to air temperature.
Run the simulation. This may take a while. The screen should display the first time step, the
temperature of the system after just 6 seconds. Right-click the result. Clicking “Edit Definition”
permits changing which time step is displayed, and clicking “Animate” cycles through the
frames (like in the frequency study).
Right-click the result again and choose “Probe,” then click anywhere on the model. In the probe
result pane, you should see nodes appear in the Results section with each click, like in Figure 26.
- 27 -
Figure 26. Probes along the stem.
Click on any row in the table of results and then click “Response,” below, to see a plot of
that node’s temperature with respect to time. Save the plot for a node near the top of the stem.
Feel free to experiment with other scenarios if you like.
What to turn in
Screenshots for both of your steady-state temperature studies, including max and min
A transient graph showing temperature vs. time for a node near the top of the stem.
Design verification (Optional, HIGHLY recommended)
Now, let us run a beam-deflection test on a glass pane. This is a simple enough problem that we
can solve it by hand. Whether or not the numerical results compare to simulation will tell us how
much faith to put in the other glass studies.
For this problem, we are going to find the maximum vertical displacement due to gravity on a
simply-supported glass pane, 1 m by 0.1 m and 6 mm thick.
- 28 -
Beam deflection equations
Gravity is equivalent to a uniformly distributed load across the beam. For a simply-supported
beam such as ours, the maximum vertical displacement, , is at the center of the beam and is
Where is the load distribution, is the beam length, is Young’s modulus, and is the second
moment of inertia.
To find the load distribution , we must know the weight of the glass pane. Assume the density
of the glass is approximately 2500 kg/m3. Then, the mass of the pane is 1.5 kg and the weight (on
earth) is about 14.7 N. Since our beam is one meter long, that makes the load distribution 14.7
Young’s modulus varies depending on the content of glass. As defined in SolidWorks, the glass
material has E = 68.9 GPa.
The second moment of inertia of a beam varies depending on its cross-section. We are using a
beam with a rectangular cross-section, width 0.1 m and height 6 mm. For this shape, ⁄ , or 1.8*10
Plugging all of these into our deflection equation gives an expected max displacement of 1.54
mm. Now, let’s see what SolidWorks gives.
Start a new part and draw a base to the dimensions specified above (1 m square, 6 mm thick). Set
glass as the part material, as in Figure 2. Start a new static study.
On solid parts such as this, there are no fixtures analogous to a simply-supported arrangement.
There are some kludges involving cylindrical faces and hinges, but the best thing to do is to tell
SolidWorks to treat the pane as a beam. Right-click the part name in the study pane and select
“Treat as beam,” as in Figure 27. Treating elements as beams is a very useful way of reducing
complexity in assemblies of linkages, though it works well for individual elements too.
- 29 -
Figure 27. How to treat a solid model as a beam.
A new option in the study pane should open, named “Joint Group.” These joints are how
SolidWorks identifies which parts of the model are the ends of the beam (and, therefore, where
our supports will go). Right-click “Joint Group” and select “Edit” to open the Edit Joints pane.
Click “calculate” to generate joints based on default settings. If it works, the default location for
these joints will be mid-way along the short faces. Sometimes, the joint calculation fails for
unknown reasons. The only known solution right now is to try on a different computer (now you
know why this section is optional).
Still here? Neat. Check OK to continue. Two spheres will appear, representing the joint
locations. This is shown in Figure 28.
Figure 28. Glass pane with joints, represented by spheres.
Now, right-click “Fixtures” and add a Fixed Geometry feature. For beams, there is a new option
in the Fixture pane, which is to alternate between “Fixed geometry” and “Immovable (No
translation.” The former, as we know, prevents translation and rotation, whereas the latter merely
prevents translation. Change the fixture type to “Immovable” and apply this to both joints.
- 30 -
For a load, we could use a force set to 14.7 N (which is the weight we calculated earlier) but it is
more convenient to choose the gravity option. Right-click “External loads” and select “Gravity.”
Apply this to the model, and it is ready to mesh and run.
Once you have meshed and run the simulation, create a new displacement plot and observe the
maximum displacement in the Y direction, as in Figure 29. It should be around 1.52 mm, which
is very close to what we expected from the math. Hooray!
Figure 29. Beam displacement.
What to turn in
Nothing for this section!
- 31 -
Appendix A: Technical drawing
- 32 -
Appendix B: References
A great resource for learning how to use Simulation is Engineering Analysis with SolidWorks
Simulation 2010 by Paul M. Kurowski, from SDC Publications. A copy can be checked out from
The website Engineering Toolbox was used to establish a plausible coefficient of convection for
The dimensions of the wine glass model are based on a “tasting glass” with ISO specifications
obtained via http://www.diwinetaste.com/dwt/en2002113.php.
Beam deflection equations were based on information from: Jennifer Stroud Rossman and Clive
L. Dym, An Introduction to Engineering Mechanics: A Continuum Approach, CRC Press, Boca
Rotan, Florida, 2008