Author |
Message |
Svetlana Gurevich (svetlaga)
New member Username: svetlaga
Post Number: 1 Registered: 02-2005
| Posted on Wednesday, February 09, 2005 - 06:05 am: | |
Probably it's a trivial problem, but.... I try to solve a linear eigenvalue problem Lf=lambda*f, that is a result of a linearization of the parabolic nonlinear PDE system (actually reaction-diffusion system). The problem is, that the operator L is NOT self-adjoint, so one can expect complex eigenvalues and complex eigenfunctions f. Can I use FlexPde for this problem and if so, how can be interpreted a resut of a calculation? (re(f),im(f), re(lambda)...etc?) |
Robert G. Nelson (rgnelson)
Moderator Username: rgnelson
Post Number: 316 Registered: 06-2003
| Posted on Wednesday, February 09, 2005 - 09:29 pm: | |
FlexPDE doesn't know anything about complex numbers, and usually requires breaking the system into component equations. But in your case, that implies two eigenvalues, the real and the imaginary parts. We can't do that, either. The best I can suggest is to use a single real eigenvalue and break it into real and imaginary parts using a fixed, user specified phase angle. Then sweep the phase angle in consecutive runs. |
David (assighna)
New member Username: assighna
Post Number: 1 Registered: 03-2005
| Posted on Sunday, March 27, 2005 - 09:52 am: | |
hi everybodey I am interesting on wood drying at hight temperature. I would like ton use Luikove model, but I don't know if it is valable at hight temperature. do you have an Idea. sincerly yours |
|