Robert Koslover
Certified Consultant
Please login with a confirmed email address before reporting spam
Posted:
10 years ago
Jun 4, 2015, 8:07 p.m. EDT
That depends. Are you trying to do this in the frequency domain or time domain? Presumably the properties of this material depend on the applied bias field, right? So: (1) if you already know exactly how that dependence works, then you should probably just define the material properties (while knowing separately what bias would yield them) instead of trying to impose a dc bias literally in the model. This would allow you to work conveniently in either freq or time domain for the rest of your problem, while holding the "dc bias" fixed. (2) Alternatively, if you don't know how the dc bias affects the material properties, but you have some kind of relationships that are subject to a statics solution in Comsol, you could employ that statics computation first, and use it to document the required relationship, thus effectively bringing you back to situation # 1. Finally, (3), If neither of those approaches appear feasible, then it seems to me that you need to use a time-domain model. In that case, you can impose any fields you want, from dc to whatever, in whatever temporal or spatial configuration or combination you decide, just so long as your mesh sizes and time steps are small enough, and you have enough computer memory and patience.
That depends. Are you trying to do this in the frequency domain or time domain? Presumably the properties of this material depend on the applied bias field, right? So: (1) if you already know exactly how that dependence works, then you should probably just define the material properties (while knowing separately what bias would yield them) instead of trying to impose a dc bias literally in the model. This would allow you to work conveniently in either freq or time domain for the rest of your problem, while holding the "dc bias" fixed. (2) Alternatively, if you don't know how the dc bias affects the material properties, but you have some kind of relationships that are subject to a statics solution in Comsol, you could employ that statics computation first, and use it to document the required relationship, thus effectively bringing you back to situation # 1. Finally, (3), If neither of those approaches appear feasible, then it seems to me that you need to use a time-domain model. In that case, you can impose any fields you want, from dc to whatever, in whatever temporal or spatial configuration or combination you decide, just so long as your mesh sizes and time steps are small enough, and you have enough computer memory and patience.
Please login with a confirmed email address before reporting spam
Posted:
9 years ago
Oct 12, 2015, 9:21 a.m. EDT
Thanks! I run it in the frequency domain. I finally used option (1) and it worked well.
Now, I need to use a solution of electrostatic field in my RF module. I would like to do it without modeling the material or the volume of the solution to an equation. What are my possibilities?
Thanks! I run it in the frequency domain. I finally used option (1) and it worked well.
Now, I need to use a solution of electrostatic field in my RF module. I would like to do it without modeling the material or the volume of the solution to an equation. What are my possibilities?