Note: This discussion is about an older version of the COMSOL Multiphysics® software. The information provided may be out of date.

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.

Valveless Pump Boundary Load Convergence Problems

Please login with a confirmed email address before reporting spam

Hello everyone,

I've been trying to simulate a valveless MEMs pump. In this pump, a vibrating membrane causes fluid to flow through a nozzle and a diffuser, which due to changing resistances to flow, induces a net fluid flow over time in one direction. I've been following the "Peristaltic Pump Tutorial" since it's very similar to what I am trying to do.

I've been using the Fluid Structure Interaction Module to simulate the pump. Previously, I was have having convergence issues. I've been able to eliminate most problems by changing from a standard mesh to a quadrilateral mapped mesh, and by specifying the time steps and using the Intermediate time dependent solver. I unfortunately needed to specify my membrane displacement using a prescribed displacement condition applied at the center of the membrane.(Pump2DIteration1)

I am now trying to fix the membrane displacement, and apply the load on the membrane as a solid mechanics boundary load rather than a prescribed mesh displacement. My applied load follows a negative cosine equation with a positive offset (meaning it starts at 0 N/m and peaks at some value, before returning back to 0 N/m). I'm currently using 200,000 N/m as the peak load due to a static analysis showing that this provides a similar membrane displacement as the prescribed displacement condition. (Pump2DIteration2)

Errors are indicating that the mesh is inverting near the center of the interface between the fluid boundary and the solid boundary (the membrane). I'm not sure what is causing this issue at this point.

Am I applying my load in the wrong way? Is it possible that I am causing the membrane to vibrate due to lack of dampening conditions? Is my membrane compressing and flipping itself inside out due to an improper application of load?

I've include my simulation files, I'm using COMSOL 4.4 for this. I've omitted any solutions due to their large size.

Thanks,
Paul


4 Replies Last Post 8 mai 2014, 08:39 UTC−4

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 7 mai 2014, 11:02 UTC−4
So I've figure out how to fix this problem. After looking through the forums and some COMSOL tutorials given at universities, it seems the problem stems from the solver that COMSOL was using. COMSOL by default was using a segregated solver. From what I read, a segregated solver works well only for weakly coupled physics and linear problems. Unfortunately my problem is highly non-linear and strongly coupled. As such, changed the solver to a one-way coupled time dependent solver, solving for the deformation of the membrane first, and then the resultant fluid flow second. The physics of the fluid does not affect the membrane (which is an adequate assumption for my purposes). No problems solving and it runs significantly faster.

-Paul
So I've figure out how to fix this problem. After looking through the forums and some COMSOL tutorials given at universities, it seems the problem stems from the solver that COMSOL was using. COMSOL by default was using a segregated solver. From what I read, a segregated solver works well only for weakly coupled physics and linear problems. Unfortunately my problem is highly non-linear and strongly coupled. As such, changed the solver to a one-way coupled time dependent solver, solving for the deformation of the membrane first, and then the resultant fluid flow second. The physics of the fluid does not affect the membrane (which is an adequate assumption for my purposes). No problems solving and it runs significantly faster. -Paul

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 7 mai 2014, 12:15 UTC−4
Hi Mr. Paul,

Thank you so much for sharing your experience.
So, do you use Fluid-Structure Interaction in your model only?
Will a stationary coupled solver work for this problem?
Is it possible to upload your modified model, so everyone can understand better?

Best regards,
Marwan
Hi Mr. Paul, Thank you so much for sharing your experience. So, do you use Fluid-Structure Interaction in your model only? Will a stationary coupled solver work for this problem? Is it possible to upload your modified model, so everyone can understand better? Best regards, Marwan

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 7 mai 2014, 13:41 UTC−4
Hi Marwan,
I'm using only FSI in my model, but I would incorporate Non-linear Structural Mechanics if we had the license for the module. The membrane of the pump experiences large deformation, which would be more accurately described with a hyper-elastic material (rather than a linear elastic material). So in the mean time, I'm forcing linear strain and material incompressibility.

A stationary solver would not work because the deformation of the membrane is related to the force applied to it, which changes over time. I have a time dependent function which I apply the force on the membrane with. I do use a stationary solver to solve for initial conditions though.

I'll definitely upload the new simulation files when I can. But at this point I'm going to be using the simulation as part of some research I plan to publish. I'll upload the files after I've submitted the article.

Hope this help!
Paul
Hi Marwan, I'm using only FSI in my model, but I would incorporate Non-linear Structural Mechanics if we had the license for the module. The membrane of the pump experiences large deformation, which would be more accurately described with a hyper-elastic material (rather than a linear elastic material). So in the mean time, I'm forcing linear strain and material incompressibility. A stationary solver would not work because the deformation of the membrane is related to the force applied to it, which changes over time. I have a time dependent function which I apply the force on the membrane with. I do use a stationary solver to solve for initial conditions though. I'll definitely upload the new simulation files when I can. But at this point I'm going to be using the simulation as part of some research I plan to publish. I'll upload the files after I've submitted the article. Hope this help! Paul

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 8 mai 2014, 08:39 UTC−4
Thank you so much Mr. Paul. All the best for you.
Thank you so much Mr. Paul. All the best for you.

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.