I have a problem statement: “ A closed tube, water bubble into it, what happens to its droplet dynamics if we vary internal pressure conditions in a fixed temp.” how to approach and model this using ANSYS Fluent. Can anyone suggest tutorial video or help me with the CFD modelling approach?
Hi guys! I am very bad at COMSOL and so far have only simulated things that I had the instructions for so never my own project, so sorry if my question is very basic.
I am trying to model a very simple 2D disc with a cut/slit in it in the middle. I have the geometry (created the slit with very thin, long rectangle and difference boolean operator, since i could not just do it with a line for some reason). In real life the material is temperature responsive hydrogel and shrinks as we raise the temperature of the liquid it is in. In the model my goal is to simulate the disc shrinking but with constant diameter (boundary condition), so that instead of the disc getting smaller, the slit opens up from the stress. I want to see the shape of the opened cut and the stresses ont he disc (later want to check how much flow it would let through as it is opening) but so far I cannot even manage to compute the study and see proper stress distribution. It is acting very strange.
I used simple solid mechanics physics and added thermal expansion to the linear elastic material. I also tried having the heat transfer in solids physics (then disabled the thermal expansion). None of them seem to work.
I was wondering that maybe its my geometry somehow, but it is so simple I dont see how it would be the problem.
If someone is more experienced, could you walk me through how you would do this or what could be the problem? I attached a picture to what I see when I compute the study.
I have been using COMSOL Multiphysics 6.4 to do the meshing of complicated STLs of Porous materials.
Here is the pipeline I am using
I have got the tomography scans
So I segment them into 4 different phases using Avizo 3D
Then I generate the STL file using FIJI
Finally I load them into COMSOL
However, every time I am stuck with the same error that the faces are overlapping or something like this. I have tried numerous ways like reducing the volume of element size, cleaning using software called Meshlab, deleted the flagged faces, but nothing seems to work.
It will really be a great help if someone could suggest a way out of this mess ; )
Hello all!
I'm running an auxiliary sweep for the first time and everything looked to be going as planned, but on my fourth iteration my segregated solver is stuck at 92% for more than half that iteration. All the important residuals; velocity, pressure, turbulence and temperature; are 100% converged, so I don't quite understand what is happening here. Does anyone have any idea what is happening?
Some additional information about the simulation. It is a fluid dynamics and heat exchange simulations. It has two different fluid domains, one turbulent and another laminar. Separating the two fluid domains is a solid domain.
As I said, it's an auxiliary sweep. The parameters that are being varied are the inlet velocities at both fluid domains. Three different values for each parameter, sweep type set to "All combinations", so it should perform 9 iterations in total.
Hope I have provided enough information. I'll try to provide more if needed. Thanks for the help in advance!
I've been using comsol for an internship project involving photonic crystal fibres and recently I noticed something peculiar. While calculating the effective mode index at a single wavelength im getting a very different answer at the same wavelength but this time when its part of a parametric sweep while every other thing is the same. These discrepancies reduce as I approach a larger wavelength. Secondly for some even shorter wavelengths I am getting the guided mode profile in a single wavelength run but that completely disappears when im running the sweep. Can someone please help out and tell me what is going on and how to fix it. Im losing my mind out here.
Say you were to fine-tune a NN on experimental data, would it be possible to run a pipeline like
Python updates NN weights
→ rebuild/export model
→ run COMSOL
→ extract reaction forces
→ compute loss
→ update weights
→ repeat
In this comsol is called iteratively to solve for the constitutive quantities(here the overall response). and then compare it to your ground truth(experiments) --> update weights and repeat until a satisfcatory loss.
Or should one go with a custom FE code or a FE software like Abaqus that might offer better scripting