The optimization calculates the von Mises stress and shows these stress values in the postprocessing as well.
...
The stress goal is dependent on the resolution levels. The optimization works with runs on different resolution levels and switches between them to achieve the best results. The first itertions are calculated on a coarse resolution level. With rising number of iterations the resolution gets finer. This means that the first iterations calculate much faster than the last iterations but aren’t as detailed as the last. With the increasing resolution, the size of the output data and , the calculation time as well as the surface quality increase in the same way.
...
Furthermore the stress goal For each level a different stress percent goal is defined which is dependent on the steps of the stress goal set for the optimization. On lower resolution levels . The the stress percent goal changes between the different levels. This is a percentage value at which the target stress is reachedgoal varies on purpose from the requested stress goal to realize a fast design generation. On the highest resolution level the requested stress goal will reach 100% an thus satisfy the defined stress goal . On lower resolution levels the stress goal varies on purpose from the requested stress goalas the geometry allows it.
The stress calculation is from an FE point of view an estimation. That’s why we recommend a FE analysis using Apex Structures/Nastran to verify the results regarding all boundary conditions. For the reanlysis the intersected model should be used. For more information regarding intersection have a look here. By the means of the reanalysis dynamic & other boundary conditions can be considered and checked.