Discussion Forum

I am working on modeling the interaction of microwave plane wave pulses (1 GHz) with spherical geometries simulating a human head. Using the RF module, we built a layered spherical model, and after some iterations, were able to successfully get an electric field distribution for multiple lossy dielectric layers. Next, instead of building the geometry in COMSOL, we imported a mesh based on a segmented MRI image. After remeshing, we were able to get the model to run; however, COMSOL appears to default to assuming perfect conductors in the outer materials (skin, then bone) and not following the over-ride of the perfect conductor with the dielectric properties for the layers. We are using an impedance boundary condition. Why is the model continuing to treat the boundaries as perfect conductors? We are losing all of the field at the surface. In the spherical model, we used the same materials, same parameters, etc. The only difference is we went to an imported geometry. What are we missing??