Instability in simulation

Log Out | Topics | Search Moderators | Register | Edit Profile

Author Message   Bernd Vosseler (bernd) Member Username: bernd Post Number: 9 Registered: 02-2006 Posted on Tuesday, January 30, 2007 - 12:12 pm:    Dear Robert, something strange happened in my model: I simulate the surface temperature of a metallized polymer sheet, that moves below a fixed laser spot. The result seems to approach something sensible in the beginning, but after 80µs, there occur hot spots, where I would expect none. The effect is dependent of the film thickness (dicke), if dicke is 4 or smaller, it happens, if dicke is five or greater, it behaves proper (at least for the 100 µs i am watsching). Culd you help me? Best regards, Bernd Ti_Heat.pde (3.3 k)   Robert G. Nelson (rgnelson) Moderator Username: rgnelson Post Number: 753 Registered: 06-2003 Posted on Tuesday, January 30, 2007 - 05:01 pm:    Your boundary condition on the incoming Y-edge is unstable. Presumably, the incoming material is at temperature Ta. With a natural BC there, any upward bounce of the boundary slope implies hotter material flowing in. This is self-reinforcing, and leads to an instability. Use value(Temp)=Ta, and the trouble goes away. You might also want to use a somewhat smaller ERRLIM, to clamp down on the interior oscillations.