Written by: Shaun Bentley, Application Engineer
SOLIDWORKS Simulation Premium includes two complex analysis types: Nonlinear and Dynamic. Nonlinear analyses can be the most challenging type of analysis to run. After years of experimenting and running my own nonlinear studies, I’ve made a list of the most common tricks that helped get me through some of the trickiest simulations. This is not a comprehensive list (to write such a list would take many pages), but hopefully should give you some general starting points to kick off from.
Trick #1: Start Simple
Many nonlinear studies take hours to run and might fail for no apparent reason with an error message like that shown below:
Frequently, scaling down your model can help with troubleshooting. This means that you might model simple blocks (or even a 2D simplification) that might emulate some of the behavior you are looking for. Studies that take only a minute or two to run can make it a lot easier to anticipate the type convergence problems you’ll bump into with a more sophisticated model. Also, it’s often overlooked that little successes of getting a simplified model to run can help to motivate you to continue to scale the nonlinear mountain.
Some typical simplifications:
- Turn off friction
- Remove some sources of nonlinearity such as a nonlinear material or no penetration contact
- Use a coarse draft quality mesh
- Overapply fixtures to keep things stable
- Exclude parts/bodies
- Model with shell or beam elements
- Take advantage of symmetry or pseudo-symmetry
Trick #2: Try Nonlinear Dynamic
You might think that you’ll have better luck with Nonlinear-STATIC but, oftentimes, Nonlinear-DYNAMIC will be more forgiving.
You’ll likely want to enter damping coefficients. In many cases starting with numbers like those shown below will be enough to keep things relatively stable.
Think of the alpha coefficient as describing a viscous fluid that surrounds your model and you are keying in how thick the fluid is. You may need to experiment with this value a bit. I’ve found that on some problems it helps if I enter values as large as 100000 for the alpha. In my experience, the Beta coefficient doesn’t usually help as much, and you might just enter 0 for it.
Trick #3: Adjust Advanced Settings
When I am troubleshooting simulations, here are the settings that I regularly use.
The numbers I boxed in green, you will frequently need to adjust usually by making them smaller. If you considered running Nonlinear-Dynamic, you may need to also try specifying a larger “End time” since the default of 1s can generate undesirable vibrations (applying a load is 1s can be too fast) leading to potential convergence issues. Showing intermediate results can be handy to turn on, but it makes the simulations run a bit slower, so I’ve disabled it here.
Trick #4: Improve Contact
For problems using no penetration contacts, some of the settings that improve my chances are shown in the image below:
If one side of the contact pair includes more complex geometry than the other side, Set1 should contain the more complex geometry and a refined mesh can help in these areas. Also, notice that I rarely use friction while I’m troubleshooting. Surface to surface gives me more accurate contact and I’ve noticed fewer problems with getting it to work in my experience.