Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.
Frequency domain analysis in general PDE form
Posted Sep 22, 2011, 7:04 p.m. EDT 3 Replies
Please login with a confirmed email address before reporting spam
Hi all,
As you may know, the PDE(general) module has the form of (a*d^2(u)/dt^2+d*du/dt+nabla.nabla(u)=f ). This format provides the possibility for implementing our own equation in time domain studies. I want to use the general form PDE for a frequency domain. So the equation will not have the derivations of time anymore. For example, for the case of a pressure acoustics, the general equation is ((1/c^2) * d^2(p)/dt^2-nabla.nabla(p)=0 ) . Considering the pressure as harmonic p=p*exp(-i*omega*t), we can write the frequency domain equation for the above as:
-(1/c^2)*(omega^2)*p-nabla.nabla(p)=0)
The only way for implementing it in General PDE is puting the (+(1/c^2)*(omega^2)*p) as f.
Here the question is that:
What is the proper variable to put as "omega" in the equations that Comsol can understand it as the "frequency" and give the right solution? I know "freq" but I am not sure if it works and if I can get right results from Frequency Domain studies.
In general: How to implement the frequency domain study with the general PDE form?
As you may know, the PDE(general) module has the form of (a*d^2(u)/dt^2+d*du/dt+nabla.nabla(u)=f ). This format provides the possibility for implementing our own equation in time domain studies. I want to use the general form PDE for a frequency domain. So the equation will not have the derivations of time anymore. For example, for the case of a pressure acoustics, the general equation is ((1/c^2) * d^2(p)/dt^2-nabla.nabla(p)=0 ) . Considering the pressure as harmonic p=p*exp(-i*omega*t), we can write the frequency domain equation for the above as:
-(1/c^2)*(omega^2)*p-nabla.nabla(p)=0)
The only way for implementing it in General PDE is puting the (+(1/c^2)*(omega^2)*p) as f.
Here the question is that:
What is the proper variable to put as "omega" in the equations that Comsol can understand it as the "frequency" and give the right solution? I know "freq" but I am not sure if it works and if I can get right results from Frequency Domain studies.
In general: How to implement the frequency domain study with the general PDE form?
3 Replies Last Post Mar 25, 2013, 8:19 p.m. EDT
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
what I do in such cases (and for your issue I would have to go through the full procedure, cannot give you "the right tip" like that ;) is to index my pdf files and search for i.e. "frequency variable", which points to, among others, to p115 of COMSOLMultiphysicsUsersGuide.pdf (v4.2) where it talks about "freq"
Now, COMSOL often defines also other variables such as "omega" and "iomega", my way for those are in complement to the above remark to set up a simple i.e solid physics model and study the underlaying "equations" (turn on the COMSOl equation sub nodes in the preferences). Then I see most of COMSOLinternals formulas, and I can derive from that, mostly, how to set-up my own PDE case
--
Good luck
Ivar
what I do in such cases (and for your issue I would have to go through the full procedure, cannot give you "the right tip" like that ;) is to index my pdf files and search for i.e. "frequency variable", which points to, among others, to p115 of COMSOLMultiphysicsUsersGuide.pdf (v4.2) where it talks about "freq"
Now, COMSOL often defines also other variables such as "omega" and "iomega", my way for those are in complement to the above remark to set up a simple i.e solid physics model and study the underlaying "equations" (turn on the COMSOl equation sub nodes in the preferences). Then I see most of COMSOLinternals formulas, and I can derive from that, mostly, how to set-up my own PDE case
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Thanks Ivar.
I don't understand why below model implemented in PDE (general) and frequency domain study doesn't give any result:
A very simple model of 1D case with variable as "u" and the frequency variable as "freq" (Which seems to be a predefiened parameter for frequency domain studies (I am not sure). the displacement (u=1) at one end is applied as input in a frequency changed from (For example freq=100 to 1000).
The 1D wave equation in in frequency domain was used:
-(1/c^2)*(freq^2)*u-nabla.nabla(u)=0
Which was implemented in Comsol PDE (general) as:
gamma= - nabla(u)
and
f= (1/c^2)*(freq^2)*u;
I dont understand why the model does not provide any result. It seems it can not solve it. For all the frequencies the obtained displacement is zero and also the solution time shown is zero. Where have I been wrong? Should I use another parameters instead of "freq"? The comsol doesnt recognize "omega" or "iomega".
I appreciate your help.
Masoud
I don't understand why below model implemented in PDE (general) and frequency domain study doesn't give any result:
A very simple model of 1D case with variable as "u" and the frequency variable as "freq" (Which seems to be a predefiened parameter for frequency domain studies (I am not sure). the displacement (u=1) at one end is applied as input in a frequency changed from (For example freq=100 to 1000).
The 1D wave equation in in frequency domain was used:
-(1/c^2)*(freq^2)*u-nabla.nabla(u)=0
Which was implemented in Comsol PDE (general) as:
gamma= - nabla(u)
and
f= (1/c^2)*(freq^2)*u;
I dont understand why the model does not provide any result. It seems it can not solve it. For all the frequencies the obtained displacement is zero and also the solution time shown is zero. Where have I been wrong? Should I use another parameters instead of "freq"? The comsol doesnt recognize "omega" or "iomega".
I appreciate your help.
Masoud
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hello Masoud,
I have the same question now, do you figure out the solution?
Thank you
Dan
I have the same question now, do you figure out the solution?
Thank you
Dan