Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.
Calculation of dependent variables at boundary conditions
Posted 2017年6月27日 GMT-4 09:01 1 Reply
Please login with a confirmed email address before reporting spam
Hello everyone,
I have a model where I am simulating heat transfer in soils.
The question is:
As I have the dependent variable T also as a boundary condition, for instance, convective heat in the surface is q=h(Text-T), how can I be sure that Comsol does sufficient computation for a precise value, or yet if Comsol could perform less computation to save time? Has it something to do with the number of iterations?
Thanks in advance
I have a model where I am simulating heat transfer in soils.
The question is:
As I have the dependent variable T also as a boundary condition, for instance, convective heat in the surface is q=h(Text-T), how can I be sure that Comsol does sufficient computation for a precise value, or yet if Comsol could perform less computation to save time? Has it something to do with the number of iterations?
Thanks in advance
1 Reply Last Post 2017年6月28日 GMT-4 01:56
Please login with a confirmed email address before reporting spam
Posted:
7 years ago
Hi,
The fact that the dependent variable is part of the boundary condition does not make any fundamental changes to the problem. h*Text ends up in the right hand side, and h*T in the stiffness matrix.
Your example (convection) is a linear boundary condition, so there would not be any iterations anyway. For a radiation condition q=a*(Text^4-T^4), the problem would become nonlinear, and then you, as for any nonlinear problem, will have an accuracy which is controlled by the tolerance settings.
Regards,
Henrik
The fact that the dependent variable is part of the boundary condition does not make any fundamental changes to the problem. h*Text ends up in the right hand side, and h*T in the stiffness matrix.
Your example (convection) is a linear boundary condition, so there would not be any iterations anyway. For a radiation condition q=a*(Text^4-T^4), the problem would become nonlinear, and then you, as for any nonlinear problem, will have an accuracy which is controlled by the tolerance settings.
Regards,
Henrik