What are the limitati8ons / accuracy of Optimetrices in HFSS?
时间：03-26
整理：3721RD
点击：
If I simulate a design in HFSS, I uses parametize the model, so I can change anything easily. This make is possible to use the optimetics feature, whereby one does different simulations of the basis of one or more parameters changing.
However, I'm concerned about the accuracy of this. Consider for example a complex vertical antenna, which has a simple whip at the top, of length whip_length. If the length of that whip is varied, optimetics allows one to calculate the response of the antenna. But as whip_length is varied, there is no guarantee the original mesh is valid. As far as I know, a new mesh is not generated each time. Suddenly an area which had a small change in the magnitude of E when the whip was some length, suddenly becomes an area where the magnitude of E is changing rapidly.
Am I correct in assuming a new mesh is not generated for each different value of the parameter? Is there any way to cause a new mesh to be generated for each different value of the parameter? I apprecite that for small variations in the parameter, the original mesh may be ok, but there is nothing stopping one making huge changes, where I doubt it is sensible to reuse the same mesh.
However, I'm concerned about the accuracy of this. Consider for example a complex vertical antenna, which has a simple whip at the top, of length whip_length. If the length of that whip is varied, optimetics allows one to calculate the response of the antenna. But as whip_length is varied, there is no guarantee the original mesh is valid. As far as I know, a new mesh is not generated each time. Suddenly an area which had a small change in the magnitude of E when the whip was some length, suddenly becomes an area where the magnitude of E is changing rapidly.
Am I correct in assuming a new mesh is not generated for each different value of the parameter? Is there any way to cause a new mesh to be generated for each different value of the parameter? I apprecite that for small variations in the parameter, the original mesh may be ok, but there is nothing stopping one making huge changes, where I doubt it is sensible to reuse the same mesh.
During an adaptive analysis, HFSS refines the mesh iteratively in the areas of highest error.
The fact that HFSS does create a new mesh is the bread and butter of their business. Each variation will have an adaptively converged mesh, and this can be seen in the profile or via field plots.
Have Fun
Thank you. So I was wrong in my initial assumption that a new mesh was not created. Thank you. That means I have a bit more faith in the optometrics than I otherwise did.