Graphics Systems SolidNotes

The right part. The wrong stress.

When analyzing a SolidWorks assembly within COSMOSWorks, the available results to plot are initially displayed for the entire assembly.  But I'd like to tell you about a trick to change that. And it might be the most important COSMOSWorks trick you learn this week.

Here in the image on the left, we see a simple, two-piece assembly and we're plotting the normal stresses in the Z. Note that the maximum stresses occur within the piece to the left. Let's call that the "collar."  The second piece- we'll call that the pin shuttle- comes in to contact due to the loading and we're interested in the stess concentrations and maximum stress there.  What do you do??

First, you must hide all components you are not interested in. Secondly, right-click the chart results in question and modify the Chart Options. You will be then presented with a checkbox option to show the Min/Max range on the shown part only.

After this, you may need to re-edit the definition of the plot but afterwards, the plot will correctly display the stresses, displacements, or strain for the desired part only.

You can see the modifications in the image to the right. And now you know how to tweak your plots from assembly analyses to correctly show results for specific component!

COSMOS FloWorks user? Wanna grab the EV of 08's SP4? Read here first.

So you say you've been waiting and waiting and waiting for SP4 of SolidWorks 2008.  And you've seen the download for the EV (early visibility) release on your Customer Portal.  But now you're wondering if you should download it.

If you're a COSMOS FloWorks (SolidWorks' Computational Fluid Dynamics software) user you might want to think twice. Why?  Because Flo studies are not backwards compatible to the previous Service Packs.

Straight from the horse's mouth.... or better, the SolidWorks Knowledge Base, see Solution Article S-016109. Users who've upgraded to SP4 will be greeted with a message "Wrong file format" but SolidWorks will still load. The answer to this article: "COSMOSFloWorks is not backward compatible. Upgrading the machine with COSMOSFloWorks SP2 to SP4 will correct the "Wrong File Format" error message. Unfortunately there is no work around other than recreating the model using the SP2 version or upgrading to SP4."

This means that if the company is going to upgrade its users, all the users ought to upgrade, even if only one or two are FloWorks users.  This should cause the CAD administrator and the user community to think twice before haphazardly downloading and installing new service packs. As always, if you have a question or concern, please call your reseller first and let them offer their input and help guide your decision. (This is also a good time to 1. make sure your organization has backups and restore points and 2. keep an eye on those individuals who are always the first to grab new releases and install them independantly from the rest of the community)

Interesting and Insightful COSMOSWorks answers, Part III

I'd like to address some of the many COSMOSWorks questions we here at Graphics Systems see come through during training or everyday Technical Support.  And we thought that, instead of sharing the answers with lone individuals or just those who took training classes, we'd publish them "for the greater good."  To begin with, I'll deal with a couple of mesh-related questions.

1. Can the use of mesh control still cause distortion in an element? Great question! The use of Mesh Controls is meant to do just the opposite: fix distorted or failing elements. However, given the right conditions, one could still create an undesirable mesh with poor aspect ratio.  We'll deal with aspect ratios below.  In general, your use of mesh controls is a considerably safe bet.
2. What is a good guideline for aspect ratio check on my mesh? Aspect ratios, which we'll cover more in a future post, are a mathematical check for amount of element distortion (how an element gets mapped along curvilinear geometry) within a mesh. Ideally, every element of a mesh will have an aspect ratio of 1. But that's just not realistic given models that look more like reality than tinker toys. And so it is suggested that if there are approximately 5% of all elements diverging from a ratio of 1, you should be OK. But having very distorted elements isn't neccessarily a bad thing because they may be in areas of your model where accurate results aren't needed. To check this, after a mesh has been created we can create mesh plots of both aspect ratios and Jacobian checks. We'll cover more on this sub-topic as well in a future post.
3. With a p-adaptive mesh, what is a good threshold for the "energy norm error?" There are no hard and fast rules for this but, from the kind folks over at COSMOS, it is recommended to target within 10%.  Take care, however, that (like above) certain areas of your model may be of less interest than others, and refining the mesh in areas of importance is a good way to go.

Lastly, there are two other questions I'd like to deal with. the first involves the use of a "pin connector." Instead of modeling and including a real pin between components, if one isn't interested in the displacements and stresses within the pin itself, it can be excluded and replaced with a special connector. However, this connector creates special bonded pairs and so no gap can ever exist between the pin and the bore. In order to have and use a gap, a pin must be modeled and included within the study. This means it will have to be meshed and solved along with the rest of the components. You also must create a No Penetration contact condition between the pin and the components it connects.

Finally, for studies with beam elements (versus shell or solids), the user can retrieve the forces on the beams with COSMOSWork 2008 by simply right-mouse clicking the Results folder and selecting "List Beam Forces." Also, if you're new to 2008, check out the other new plots you can now create with studies that use beam elements.

Thanks for reading!

Interesting and Insightful COSMOSWorks answers, Part II

In Part I of our multi-partseries on COSMOSWorks Q-n-A, we covered three questions pertaining to Frequency Studies.  This time around, I'd like to handle a few more questions involving Buckling studies and "soft springs."

1. What can COSMOSM or GeoStar handle for Buckling that COSMOSWorks can't? The good news is that all products can handle both linear and non-linear buckling scenarios. The user must choose the correct material model for each case.
2. Can Slenderness Ratios and Euler Numbers be retrieved from the results of a Buckling study? COSMOSWorks does not report back these values but they can be calculated by hand.
3. When "soft springs" are used to stabilize a model, can the reaction forces on these "springs" be retrieved?  No; the soft springs are additional qualities to the stiffness calcs being performed. But with all static studies ran with COSMOSWorks, the user can retrieve and list the reaction forces and moments for selected items or the entire model. This is a great, quick check to ensure that your model is balanced and truly.... static. (Right-mouse click the Results folder and select List Reaction Forces)

In our next installment, we'll wrap things up with questions involving meshing, pin connectors, and beam elements.

Interesting and Insightful COSMOSWorks answers, Part I

We here at Graphics Systems regularly come across COSMOSWorks related questions whether it be during training or everyday Technical Support.  So for the next couple of posts, I will attempt to provide Interesting and Insightful answers to a few of the more interesting questions.

We'll begin with three questions related to Frequency Studies.  Frequency studies provide the user with the natural frequency response of a system and with these frequencies, engineers can design around them... hopefully avoiding the dreaded RESONANCE FREQUENCY with such components as motors.

1. Which equations to the Direct Sparse and FFEPlus solvers use to find eigenvalues?  The answer also applies to Bucking studies (but that's another post). Both solvers use the following equation to calculate resultant structural stiffness:

   [K_E + λ_iK_S][φ_i] = 0; K's of E is the elastic stiffness matrix, K's of S is the stress stiffness matrix, lambda of i are eigenvalues and phi of i are your eigenvectors.

2. Are modal participation values calculated? Mass participation factors can be retrieved from the study by right-mouse clicking the Results folder for the study in question.

3. Are the units for gravity calculated in "G's" or m/v? With all studies, unless gravity is a boundary condition set-up by the analyst, gravity is not calculated in. Should gravity be included as a boundary condition, the system of units to describe gravity are the same for any other load or restrain: English, Metric, SI. (I prefer m/s^2 personally) [on the first pass, I had the wrong units; sorry, brain fart]

Thanks for the questions, folks. And keep them coming.  In the next few posts, I'll deal more with Buckling, some items pertaining to mesh quality, and even one or two questions regarding "soft springs."  Stay tuned!

Shell Mesh: When should I use Thick vs Thin Elements

Let's consider a cantilever beam, which can be modeled as a solid or as a shell.

The actual deformation of the real beam (not mesh) is the sum of the shear deformation and the bending deformation as seen below:

When modeled as a solid it is recommend to have a minimum of 3 elements thickness in all places so that the shear deformation can be calculated accurately.  If this is not possible because of the size of the model (there are too many degrees of freedom) we use shell elements.  This is the reason we have two types of shell elements.  Thick shell elements consider the shear forces (for thicker plating) and thin shell elements do not (for this sheet metal) because they are negligable and would just waste calculation time.

As a rule, thick shell elements should be used when the thickness to span ratio is greater than 5%.  That's a very simple rule as long as you understand one thing... How exactly do we define span?  Thickness and 5% should be perfectly clear but the span can be confusing.  To explain this let's again consider the same beam.  In this case the span is the length but not the width.  But to be a little more clear lets consider this piece of sheet metal:

In this case we first need to assume that the verticle section is rigid compared to the horizontal section.  The sheet metal may be very long and thin but because of how this load is applied, if you need to conseder the deformation of the flange you should use thick elements because in this case the span is the length of the flange not the length of the part.

It's hard to put a concise definition on span but in general I'd define span as the distance between where the load is applied and the place bending starts.

COSMOS - I tried using Shell elements but I get a Singular Matrix Error.

Lets take this model for example.  The best way to overcome this error (to make sure that the nodes at the intersection of Vplate and Hplate get merged together) is to split the top surface of the Hplate (Where Vplate is touching the Hplate) into two surfaces based upon the common area between the Hplate and the Vplate.  Once the surface has been split, select the appropriate shell surface (in our case, there will be two surfaces from the Hplate and one vertical surface from the Vplate). The meshed model's nodes are merged at the intersection of the Vplate and the Hplate.

COSMOS - What are shell elements and why should I use them?

What are Shell elements?

Shell elements in COSMOS are triangular thin shell elements with membrane and bending capabilities. Unlike in thick shell elements, the shear deformation effect is neglected for thin shell elements. Six degrees of freedom per node (three translations and three rotations) are considered for structural analysis. Only one degree of freedom per node, representing the temperature, is used for thermal analysis. In both thick and thin shell elements, a constant thickness is assumed throughout the element.

Why use Shell elements?

You may mesh any solid model with tetrahedral solid elements. However, meshing thin models with solid elements results in generating a large number of elements since you have to use a small element size. Using a larger element size can deteriorate the quality of the mesh and lead to inaccurate results. Although you can use mesh controls to reduce the number of elements, shell meshing is the natural choice for sheet metal and thin parts.