Hi all,
I am trying to build a piezoelectric energy harvester in comsol so I coupled the electrical circuit interface with the piezoelectric device multiphysics. i was trying to use the frequency domain modal analysis so I can superimpose all the eigenmodes that are solved in the first step,
But in the first step, which is the eigenfrequency step, I tick all the four ticks in the study node, there is a orange triangle at the side of the electrical circuit node so I suppose it wont be considered in the eigensolver. However, when it is running it takes so long and there is no progressing, nor any error. Is there anyone helping with it?
Boyue Chen Could be related to your having a source in the electrical circuit.
René Christensen Hi, thanks for the answer. Actually I dont use a source in the electrical circuit except for the one from the 'terminal' node from the electrostatic physics. I waited for long enough time and it is finally solved. I can generate electric potential result but the voltage across the resistance is zero, which I suppose suggest that the electrical circuit physics is not included in both eigenvalue solver and the following modal solver.
René Christensen Hi, I found that when I enable the electrical circuit physics, the electrical potential result from electricstatics interface becomes very werid. I wonder do you have any experience it this?
Boyue Chen Looking at the Electric Circuit alone, the eigen- studies are not available. So that is probably your fundamental problem. For a domain problem the eigenvalues relate to some variation of the degree of freedom over the domain (or rather the mesh). But for the 0D electric circuit the eigenvalues would relate to e.g. currents in the individual components. And so perhaps you can simply not do what you are trying to do.
René Christensen Hi，many thanks for your answer. Can I understand your answer as the following.
The modal solver is the superimpostion of the eigenmodes I required in the eigenvalue step, and the electrical circuit physics actually has nothing to do with any degree of freedom in my meshing and contains no knowledge in terms of 'modes'. In this case it is not considered at all in modal solver?
My best regards.
Boyue Chen Perhaps, but I am not totally sure. It just seems to not be an option. In reality, the circuit would affect the modal response, but I dont know how to include the circuit in the simulation... Perhaps you can lump the entire circuit into one equivalent component that you apply in a point, or something like that.
René Christensen Hi, I found an interesting phenomenon, if in the electrical circuit setting, equation form, I chose 'frequency domain' instead of 'study controlled', I am able to incorporate electrical circuit in the eigen solver. However, if I use the eigenvalue solution as the pre-step in a frequency domain modal analysis including the electrical circuit, very weird voltage-frequency relationship is witnessed.
Hmm, yeah I am not sure modal superposition works for this case...
René Christensen It seems that it didn't work. So I decided to use float potential in electrostatics to predict the open circuit voltage and deal with the resistor analytically. Hope it gives me better result.
- Badges
- Science topic