Discussion Forum

Hello dear colleagues,

I'm doing electromagnetic simulations to find the eigenmodes of microwave cavities. After I create the model, I implement a sphere for air domain. And finally, on top of that, i create a spherical shell to define a perfectly match layer. When i need an intense mesh for my model, DOF increases too much and i don't want to spend my computational power on air domain.

My question is that, what if I omit the sphere for air domain and just go on with perfectly match layer? Would this approach decrease the accuracy of my model? Because, it is so tempting since it decreases the DOF pretty much and allow me to reach the results around similar freqeuencies??

-I tried to define a separate lower mesh for air domain, but when my model includes a very small feature, with implementation of the air sphere, I always face a narrow edge/face problem. I have been thinking on this subject quite a while, any helpful advice is much appreciated-

Thanks everyone for your kind responses in advance.

mehmet