| Author | Message | ||||
     Manu fantuzzi (mfantu) New member Username: mfantu Post Number: 1 Registered: 05-2005 |
Hi, Is it possible to use discontinious variables on a surface (between 2 layers) with contact and jump because when I specify contact(T)=1/R*jump(T) about a surface, Flex PDE tell me there is an error. Thank U | ||||
| Robert G. Nelson (rgnelson) Moderator Username: rgnelson Post Number: 368 Registered: 06-2003 |
It would help if you gave me a little information: like how you tried to do it, and what the diagnostic was. I ran some experiments and was unable to find a combination that reported an error. Unfortunately, there does seem to be a malfunciton in the solution process, and the Contact boundary between layers is not imposed. I will look into this. But if you got a diagnostic, you must have specified the boundary condition wrong. | ||||
| Manu fantuzzi (mfantu) New member Username: mfantu Post Number: 2 Registered: 05-2005 |
Hello, In fact, the error message was an other thing. But now, when I run my program on FlexPDE 4.2.14 , It seems that Flex PDE doesn't consider the "contact(T)=1/R*jump(T)" because there is no discontinuity on the graphics. I join you my program so that you can help me. As you can see, I'm trying to put a discontinuity of Temp between the 2 layers, on the surface"middle". Thank you very much for your help
| ||||
| Robert G. Nelson (rgnelson) Moderator Username: rgnelson Post Number: 369 Registered: 06-2003 |
There is an error in version 4 that does not apply contact boundary conditions between layers in 3D. I will investigate and repair it. | ||||
| Manu fantuzzi (mfantu) Junior Member Username: mfantu Post Number: 3 Registered: 05-2005 |
Do tou manage to repair the mistake in the software ? Thank you very much | ||||
| Robert G. Nelson (rgnelson) Moderator Username: rgnelson Post Number: 371 Registered: 06-2003 |
Patience, I have a lot on my plate. | ||||
| Robert G. Nelson (rgnelson) Moderator Username: rgnelson Post Number: 374 Registered: 06-2003 |
The error in handling 3D contact resistance boundary conditions has been corrected in FlexPDE version 4.2.15. There is still trouble in applying mesh refinement at such boundaries in 3D. So, for the present, mesh refinement must be disabled in 3D problems with contact boundary conditions. | ||||
| gcb (gcb) Junior Member Username: gcb Post Number: 3 Registered: 12-2008 |
Hi, In my equations,miu is discontinious,how can I defint the value?Can you help me?
how can I deal whit it? | ||||
| Robert G. Nelson (rgnelson) Moderator Username: rgnelson Post Number: 1199 Registered: 06-2003 |
One way to do this is to use a TABLE to define MIU, and program a ramp into the table jumps, so that the value does not jump instantaneously. You could do the same with a cascade of RAMP functions or SWAGES, but the TABLE is probably simpler. | ||||
| gcb (gcb) Member Username: gcb Post Number: 4 Registered: 12-2008 |
Hello, In boundaries,I want to defint a variable below 1789K with value,and to defint the variable high 1789K with natural. How can I write this boundaries? | ||||
| Robert G. Nelson (rgnelson) Moderator Username: rgnelson Post Number: 1200 Registered: 06-2003 |
You can't change the class of boundary condition (Value or Natural) dynamically during the run. You can, however, simulate a Value condition with a Natural. Write your Natural to generate a large inward flux if the boundary temperature is too low, and a large outward flux if it is too high. It would look something like this: Natural(Temp) = Big*(Temp0 - Temp). The sign of Big must be chosen consistent with the way you write your PDE. You can now switch the BC at the trigger temperature: Natural(Temp) = IF(Temp > 1789) Then Flux0 Else Big*(Temp1-Temp) If Big is too big, it will force FlexPDE to use very small timesteps, so don't overdo it. Natural(Temp) = Big*(Temp0-Temp)^2 matches slope as well as value at the switch point, and may work better in some cases. |
Modele thermique