Discussion Forum

I am attempting to use COMSOL to simulate the temperature response at a surface (1-D model) at a range of frequencies (1 Hz to 10 kHz) under the influence of a sinusoidal heat source. I am then using MATLAB to do a Fourier transform of the surface temperature response as well as the heat source. The phase delay between the heat source and the surface temperature is a function of the heat source's frequency. In theory, the heat flux from the source penetrates into the material increasingly shallow as the frequency increases, with the phase delay being influenced by the material the heat flux goes through. When two dissimilar materials have an interface with a contact resistance, the phase delay as a function of the heat source frequency behaves in a specific manner.

The potential issue I am having is not being able to accurately model this using COMSOL. The phase delay as a function of frequency curve of a validated study is shown here.

While the modelled phase delay through COMSOL data is here.

I have tried increasing mesh density, increasing the length of time taking data at a given frequency, and have triple-checked the boundary conditions to make sure they match the values used to compute the theoretical phase delay curve. Everything seems to converge on the modelled phase delay curve. I am not familiar with changing the solver configurations, etc. but I have tried a little. Does anyone have knowledge or experience with these kinds of simulations that could help me?