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.
Why boundary values are changing with time?
Posted 5 mag 2012, 04:03 GMT-4 2 Replies
Please login with a confirmed email address before reporting spam
Hello,
I am using COMSOL to solve colloid transport problem using convection-diffusion equation.
My domain is a pipe (2D axisymmetric). In addition to convection and diffusion in the flow direction, there is a term equivalent to convection normal to the flow (migration) very close near to the wall. This normal convection velocity which acts only very near to the wall is a rapidly changing one with radial distance with two peaks. Due to the normal convection term, the concentration near the peaks is around two orders of magnitude greater than the concentration far away from the wall. I use Dirichlet boundary condition (concentration) at the inlet boundary and convective flux boundary condition at the outlet boundary with non-penetration B.C. at the wall. But I am confused with the results. The boundary concentration (i.e. at z=0) should always remain same as my boundary condition which is constant. But when I plot the concentration at z=0 (inlet boundary) near to the wall, I am getting peaks similar to what is happening inside the pipe due to the PDE. Also the concentration at z=0 near the wall varies with time. Why is this happening? Is this because COMSOL solves the PDE at the boundary also? I have not used any weak form pde at the boundary.
Also the results show that there is flux normal to the flow at the inlet boundary which should not be possible from the dirichlet B.C. It also shows that the total flux at the inlet boundary (z=0) is less than the flux leaving the pipe at the outlet. There is no decay of mass (no reaction) or source inside the pipe. Hence the total flux entering and leaving the pipe should be the same at steady-state.
It will be of great help if the moderators can help me in this regard.
Regards,
Seetha
I am using COMSOL to solve colloid transport problem using convection-diffusion equation.
My domain is a pipe (2D axisymmetric). In addition to convection and diffusion in the flow direction, there is a term equivalent to convection normal to the flow (migration) very close near to the wall. This normal convection velocity which acts only very near to the wall is a rapidly changing one with radial distance with two peaks. Due to the normal convection term, the concentration near the peaks is around two orders of magnitude greater than the concentration far away from the wall. I use Dirichlet boundary condition (concentration) at the inlet boundary and convective flux boundary condition at the outlet boundary with non-penetration B.C. at the wall. But I am confused with the results. The boundary concentration (i.e. at z=0) should always remain same as my boundary condition which is constant. But when I plot the concentration at z=0 (inlet boundary) near to the wall, I am getting peaks similar to what is happening inside the pipe due to the PDE. Also the concentration at z=0 near the wall varies with time. Why is this happening? Is this because COMSOL solves the PDE at the boundary also? I have not used any weak form pde at the boundary.
Also the results show that there is flux normal to the flow at the inlet boundary which should not be possible from the dirichlet B.C. It also shows that the total flux at the inlet boundary (z=0) is less than the flux leaving the pipe at the outlet. There is no decay of mass (no reaction) or source inside the pipe. Hence the total flux entering and leaving the pipe should be the same at steady-state.
It will be of great help if the moderators can help me in this regard.
Regards,
Seetha
2 Replies Last Post 5 mag 2012, 07:12 GMT-4
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hello
first of all there are many "non moderators" out here that can certainly give you more info than me, this site is for all of us, also to answer ;)
This said, my first suspiction would be on the mesh density particularly along the Boundaries, try a locla refinement, or perhaps "boundary mesh" the latter if you know that the derivative is somewhat normal to the boundary
But there are several other issues thta could be significant too, sometimes you might trigger numerical oscillations, there is a full capter somewhere in the doc about this, and several threads already
--
Good luck
Ivar
first of all there are many "non moderators" out here that can certainly give you more info than me, this site is for all of us, also to answer ;)
This said, my first suspiction would be on the mesh density particularly along the Boundaries, try a locla refinement, or perhaps "boundary mesh" the latter if you know that the derivative is somewhat normal to the boundary
But there are several other issues thta could be significant too, sometimes you might trigger numerical oscillations, there is a full capter somewhere in the doc about this, and several threads already
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Thank you Ivar for your reply.
I refined the mesh near the boundary, but still I am getting the same results.
What else can be the reason? If I do not consider the migration velocity, then the boundary condition is not changing and the results are correct.
Can anybody please give some suggestions?
Regards,
Seetha
I refined the mesh near the boundary, but still I am getting the same results.
What else can be the reason? If I do not consider the migration velocity, then the boundary condition is not changing and the results are correct.
Can anybody please give some suggestions?
Regards,
Seetha