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.
Velocity boundary condition in MEMS
Posted Jan 26, 2011, 8:42 a.m. EST 3 Replies
Please login with a confirmed email address before reporting spam
Hello!
I'm having difficulties running a simulation of an microfluidic device. Since I don't know the pressure of the fluid entering the device, I was trying to use the velocity since I know how much of the fluid is pumped into the device per timeunit.
However, I get this response when trying to solve:
-"Failed to find a solution:
The relative residual (0.023) is greater than the relative tolerance.
Returned solution has not converged."
The simulation runs great when using pressure driven flow though for boundary condition.
Anyone knows how to address this issue? Are there any good equation for pressure-fluid velocity relevant to this simulation?
Cheers
/Thomas Läräng, Uppsala
I'm having difficulties running a simulation of an microfluidic device. Since I don't know the pressure of the fluid entering the device, I was trying to use the velocity since I know how much of the fluid is pumped into the device per timeunit.
However, I get this response when trying to solve:
-"Failed to find a solution:
The relative residual (0.023) is greater than the relative tolerance.
Returned solution has not converged."
The simulation runs great when using pressure driven flow though for boundary condition.
Anyone knows how to address this issue? Are there any good equation for pressure-fluid velocity relevant to this simulation?
Cheers
/Thomas Läräng, Uppsala
Attachments:
3 Replies Last Post Feb 1, 2011, 7:36 a.m. EST
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
I believe that if you use a pressure driven output, with i.e. "0" pressure you will get as solving model, and you will see the pressure drop along the system. Since this is a "gauge" pressure it's valid (so long the materials/fluids are incompressible) for any absolute pressure, no ?
Restraining input and output velocities, as the profiles changes is probably too restrictive, if you integrate the output flow you should get the sum of the two inlets to the the numerical precision.
Sometimes its also worth setting a point at a given pressure i.e. "0" to "fix" the gauge pressure, then probably it works also for the outlet velocity, try it ;)
--
Good luck
Ivar
I believe that if you use a pressure driven output, with i.e. "0" pressure you will get as solving model, and you will see the pressure drop along the system. Since this is a "gauge" pressure it's valid (so long the materials/fluids are incompressible) for any absolute pressure, no ?
Restraining input and output velocities, as the profiles changes is probably too restrictive, if you integrate the output flow you should get the sum of the two inlets to the the numerical precision.
Sometimes its also worth setting a point at a given pressure i.e. "0" to "fix" the gauge pressure, then probably it works also for the outlet velocity, try it ;)
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Thanks Ivar, it seemed to work. I'm now using "0 pressure" output, and velocity driven inputs.
However, the simulation reveal some odd results, e.g. negative values on concentrations, but also concentrations larger than the boundary conditioned concentration. This can be seen in the attached picture, strange artifacts can be seen at the inlet.
Is this a result of the "0 pressureoutput/velocity driven input"-method, and will it be solved by your recommendations of fixing the gauge pressure and then solve again, now with pressure driven input?
/Thomas
However, the simulation reveal some odd results, e.g. negative values on concentrations, but also concentrations larger than the boundary conditioned concentration. This can be seen in the attached picture, strange artifacts can be seen at the inlet.
Is this a result of the "0 pressureoutput/velocity driven input"-method, and will it be solved by your recommendations of fixing the gauge pressure and then solve again, now with pressure driven input?
/Thomas
Attachments:
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
indeed the negative concentration is a recurrent issue taken up on the forum, that is really something you should take up with "support". I suspect that some "log(C)" of the concentration approach and adapting the Physics accordingly could be a way out to handle the large "numerical, positive only, dynamic" of chemistry induced physics. But I also understand from others on the forum here that it is not that trivial ;)
You could also use the driven velocity input/output, and then only fix the pressure on one point at the outlet with a Point pressure node. remains the issue "0" gauge pressure or 1 atm or ...
As you have also concentration I'm always uncomfortable with the physics coupling: for me "0" pressure means also "0" concentration (if the formulas are 100% linked, but I believe its not the case.) Anyhow Chemistry physics is not the one I use most, so I have had no time to test it correctly and to make up my mind
--
Good luck
Ivar
indeed the negative concentration is a recurrent issue taken up on the forum, that is really something you should take up with "support". I suspect that some "log(C)" of the concentration approach and adapting the Physics accordingly could be a way out to handle the large "numerical, positive only, dynamic" of chemistry induced physics. But I also understand from others on the forum here that it is not that trivial ;)
You could also use the driven velocity input/output, and then only fix the pressure on one point at the outlet with a Point pressure node. remains the issue "0" gauge pressure or 1 atm or ...
As you have also concentration I'm always uncomfortable with the physics coupling: for me "0" pressure means also "0" concentration (if the formulas are 100% linked, but I believe its not the case.) Anyhow Chemistry physics is not the one I use most, so I have had no time to test it correctly and to make up my mind
--
Good luck
Ivar