Dear COMSOL community,
after reading the COMSOL documentation and this forum for a couple of weeks I am reaching out to you. I am using COMSOL V4.4.
I have been desperately trying to run my simulation model and either failing or receiving a very slow convergence (time step in the order of 1e-6). So I am hoping for some fresh ideas and help. Or just one or the other comments would be really helpful as most of you had a whole lot more experience and expertise as myself who is a master’s student of biomedical engineering.
It is a 2D FSI model and I am trying to model the pulsatile movement of the liquid that surrounds the brain and its. The pulsatile motion comes from the pulsation of the expansion of the arteries that are all over the human brain. So the model consists of the brain and the spinal cord. Both of them are surrounded by the liquid. Within the brain there is a wormlike structure which also contains some of that liquid. In the ongoing research a parameter dependency study shall be conducted.
Inlets: I do have 14 inlets , “holes” all around the brain which are supposed to model the arteries. Fluid comes at a given time dependent velocity and goes back, netto volume flow is equal to zero. Together with the area it is supposed to be a kind of Volume flow as Volume flow=velocity*area. The same idea is behind the the artery which is in the brain itself. “Volume” is added to the area and the expansion together with the “volume” from the 14 arteries shall drive the pulsation. I can’t define a mass flow because I don’t have the CFD license.
Materials: The brain and spinal cord shall be modeled as an linear viscoelastic material model, Generalized Maxwell. The elasticity of the spinal canal shall be modeled as an membrane at the bottom of the spinal canal. I also have been thinking of removing that mebrane and replacing it by an outlet with a time dependent pressure which is known, as I do have But then mass seems to be out- or inflowing into the system which is supposed to be closed.
My ideas what one should check:
- the way the reality is modelled (are further simplifications modeled)
- initial values
- boundary conditions (especially inlets/outlets,)
- fixed constraints
- solver settings
- do I have too many degrees of freedom
- material model (especially the viscoelasticity because by disabling it the DOF are decreased, from 400k to 200k)
- is it really necessary to model the movement of the structure as I receive really small movement, movements is seen with a scaling of the deformation by 1e10)?
- is the mesh I came up with (different element types) okay? Is it too coarse for the first try? Shall I add a boundary layer as I deal with CFD?
The .mph file, some photos and the geometry are attached to this post. I highly appreciate all your efforts and look forward to your comments and ideas. Feel free to ask in case of any questions or uncertainties!
Many regards
Artem
I am sorry but when clicking the "Attach File" button nothing happens so here are some links to the files in my Dropbox.
www.dropbox.com/s/b3fj2fo38gsc...rain_liquid_pulsating.mph?dl=0
www.dropbox.com/s/nud3zebchjm74hj/brain_geometry.mphbin?dl=0
www.dropbox.com/s/a4119buq3tie375/membrane.png?dl=0
www.dropbox.com/s/fynuw24pkrclyln/brain.png?dl=0
after reading the COMSOL documentation and this forum for a couple of weeks I am reaching out to you. I am using COMSOL V4.4.
I have been desperately trying to run my simulation model and either failing or receiving a very slow convergence (time step in the order of 1e-6). So I am hoping for some fresh ideas and help. Or just one or the other comments would be really helpful as most of you had a whole lot more experience and expertise as myself who is a master’s student of biomedical engineering.
It is a 2D FSI model and I am trying to model the pulsatile movement of the liquid that surrounds the brain and its. The pulsatile motion comes from the pulsation of the expansion of the arteries that are all over the human brain. So the model consists of the brain and the spinal cord. Both of them are surrounded by the liquid. Within the brain there is a wormlike structure which also contains some of that liquid. In the ongoing research a parameter dependency study shall be conducted.
Inlets: I do have 14 inlets , “holes” all around the brain which are supposed to model the arteries. Fluid comes at a given time dependent velocity and goes back, netto volume flow is equal to zero. Together with the area it is supposed to be a kind of Volume flow as Volume flow=velocity*area. The same idea is behind the the artery which is in the brain itself. “Volume” is added to the area and the expansion together with the “volume” from the 14 arteries shall drive the pulsation. I can’t define a mass flow because I don’t have the CFD license.
Materials: The brain and spinal cord shall be modeled as an linear viscoelastic material model, Generalized Maxwell. The elasticity of the spinal canal shall be modeled as an membrane at the bottom of the spinal canal. I also have been thinking of removing that mebrane and replacing it by an outlet with a time dependent pressure which is known, as I do have But then mass seems to be out- or inflowing into the system which is supposed to be closed.
My ideas what one should check:
- the way the reality is modelled (are further simplifications modeled)
- initial values
- boundary conditions (especially inlets/outlets,)
- fixed constraints
- solver settings
- do I have too many degrees of freedom
- material model (especially the viscoelasticity because by disabling it the DOF are decreased, from 400k to 200k)
- is it really necessary to model the movement of the structure as I receive really small movement, movements is seen with a scaling of the deformation by 1e10)?
- is the mesh I came up with (different element types) okay? Is it too coarse for the first try? Shall I add a boundary layer as I deal with CFD?
The .mph file, some photos and the geometry are attached to this post. I highly appreciate all your efforts and look forward to your comments and ideas. Feel free to ask in case of any questions or uncertainties!
Many regards
Artem
I am sorry but when clicking the "Attach File" button nothing happens so here are some links to the files in my Dropbox.
www.dropbox.com/s/b3fj2fo38gsc...rain_liquid_pulsating.mph?dl=0
www.dropbox.com/s/nud3zebchjm74hj/brain_geometry.mphbin?dl=0
www.dropbox.com/s/a4119buq3tie375/membrane.png?dl=0
www.dropbox.com/s/fynuw24pkrclyln/brain.png?dl=0