Learning Forum

[agarwa20]

Hello,

I was running stress analysis in Ansys Mechanical. For running multiple cases I created a parametric variation. I am using the Nissan Leaf motor model (shown below) I obtained from the Ansys learning hub. The thickness of the I and V shape magnets are the parameters. I am considering stresses only on the hole edges. 

Based on a response surface based DoE, only three designs were successfully evaluated. For the three design points which were successfully evaluated I manually ran the stress analysis for those cases. However, I observed that the results were different from the DoE based stress. 

I have also checked the case where stress is checked for the entire rotor body. However, I observed that the DoE based iterations gave much higher stress values than the ones I obtained with manual iterations for the same design points. In order to further check this, I took the rotational velocity as a parameter instead of the width of the magnets. In this case the DoE based results matched with the manually checked results. Am I making a mistake here? Is it feasible to perform a geometric parametric study in Ansys Mechanical?

Thank you.

[peteroznewman]

Which geometry editor are you using to have a width parameter on the magnet?

In my experience, DesignModeler is safer for shape changes than SpaceClaim.

[agarwa20]

Hello, the parameter for width is defined in Ansys Maxwell 2D. I am transferring the geometry from Maxwell 2D to Mechanical in Workbench.

[peteroznewman]

Sorry, I don't know anything about Maxwell goemetry engine and I don't have that software installed. I hope someone else will reply.