Author |
Message |
Eugene B. Postnikov (postnikov)
New member Username: postnikov
Post Number: 2 Registered: 06-2004
| Posted on Friday, June 02, 2006 - 08:10 am: | |
Hello, I've found the following trouble: I try to solve the system of two equations (see .pde file in the attachment) but FlexPDE5 shows an exponetial growth instead of a solution with two maxima (But FlexPDE4 solves this probem correctly, see .gif file in the attachment). If I use 1D geometry, the error is the same. Could you help? Thank you in advance, Eugene B. Postnikov 0.7 K | MOrlet.pde "" | | |
Robert G. Nelson (rgnelson)
Moderator Username: rgnelson
Post Number: 622 Registered: 06-2003
| Posted on Friday, June 02, 2006 - 02:40 pm: | |
Your script does not seem to be downloadable. Please send it directly to support@pdesolutions.com
|
Robert G. Nelson (rgnelson)
Moderator Username: rgnelson
Post Number: 624 Registered: 06-2003
| Posted on Saturday, June 03, 2006 - 11:31 pm: | |
It appears that changes made in the symbolic processing of equations to support moving meshes in version 5 have introduced an error in symbolic handling of equation coefficients that are explicitly time dependent. It is not sufficient to merely hide the time in a parameter definition, because definitions are expanded during processing. I will fix this in the next maintenance release, but for now there is a workaround you can use: 1) define a GLOBAL VARIABLE, for example, K 2) give the GLOBAL VARIABLE the EQUATION K: K=t 3) use K as the diffusivity in your existing equations. This process hides the time dependence of K from the equation analyzer.
|
|