| Author | Message | |||
     abdulwhab (tolana) New member Username: tolana Post Number: 1 Registered: 03-2009 |
hope that someone will help me with this, here is the problem statment. Use FlexPDE to solve for the temperature distribution in a wall of thickness b extending from 0 to 10b in the y direction, and from - to + in the direction perpendicular to the x and y directions. The surfaces at x= +/- b/2 and the top surface at y=10b are held at temperature To, whereas the bottom surface of the wall at y=0 is maintained at temperature T1. and here what I did.
| |||
| Robert G. Nelson (rgnelson) Moderator Username: rgnelson Post Number: 1228 Registered: 06-2003 |
You haven't put any boundary conditions on the problem. You need to use the VALUE(temp)=T0 and VALUE(temp)=T1 commands to place the boundary conditions. However, adjacent boundaries with different VALUE conditions are a physical impossibility. It implies an infinite gradient, therefore an infinite flux at the corner point. But go ahead and do it. Just realize that the corner is unphysical. | |||
| abdulwhab (tolana) New member Username: tolana Post Number: 2 Registered: 03-2009 |
Thaks for the response, it was very helpfull. I have attached the file to make sure that I got it write.
| |||
| abdulwhab (tolana) Junior Member Username: tolana Post Number: 3 Registered: 03-2009 |
I also have assumed my own geometry and solved for it, I am just wondering if this goemetery applicable or not.
| |||
| Marek Nelson (mgnelson) Moderator Username: mgnelson Post Number: 116 Registered: 07-2007 |
Your script "my_1" has the boundary conditions that you wanted. The script "2" has a problem with the region definition. You have made a "bow-tie" not a rectangle. The FlexPDE gridder can handle this (you get two triangles with their tips touching), but now you have two boundaries crossing, one with a VALUE(temp)=70 the other is -15. Check your region definition. Your last point is probably in error. | |||
| abdulwhab (tolana) Member Username: tolana Post Number: 4 Registered: 03-2009 |
Thank you very much for being generous in answering my question. One last question, if the hight of the wall is infinity, then would FlexPDE handle the problem and if not what are my alternatives. Note: I tride to change the boundry condition for y to be a huge number, but it did not work it out. | |||
| Robert G. Nelson (rgnelson) Moderator Username: rgnelson Post Number: 1232 Registered: 06-2003 |
1. Computers cannot compute with infinities, and it would be hard to draw the picture on your screen. 2. Use a height that is great enough that a variation in the height makes negligible effect on the solution at the bottom. 3. Use a less huge number. |
