Abaqus, Ansys or COMSOL?
I need to model and simulate a coupled problem of solid mechanics, fluid mechanics and light propagation. I personally prefer to use Abaqus but not sure what could be advantages of using COMSOL as a famous multiphysics simulation over Abaqus or Ansys?
Is it possible to simulate solid and fluid in Abaqus and then couple with electromagnetic part in COMSOL?
How about open source coupling frameworks?
It is always preferred that to have confidence in the computations that you perform, you have control over the finite element code - something that is not possible through the commercial packages. You can very easily get a wrong answer from a commercial package as soon as the problem becomes tough/nonlinear. For example, see this paper http://dx.doi.org/10.1007/s00466-013-0871-6 where Abaqus gives wrong solutions for a nonlinear problem.
That being said, writing your own code is very tough, so the next best thing is to use open-source packages where you can view the code. One FEM package that many of my colleagues use to solve electro-mechanical and magneto-mechanical coupled problems is deal.II https://www.dealii.org/
A lot of people use COMSOL for electromagnetic coupling with solid mechanics, but I think it works well only as long as your problem is linear.
Three way coupling (Solid, fluid, electromagnetic) is a particularly hard problem and I'm not sure if you can get a straightforward solution from commercial packages.
http://dx.doi.org/10.1007/s00466-013-0871-6
For a large part I agree with Prashant Saxena. However, I would certainly not dismiss the commercial codes. The code-makers spend a lot of time in improving, verifying, and accelerating their code. But you have to remember that a FE code is a tool, like a hammer: if you don't know how to use it, you will have accidents. If you have a though problem, you probably can solve it using a commercial code, nowadays. But for sure, you cannot do that using the default settings of the code. The paper cited by Prashant is a good example of not being sufficient careful. Also commercial codes provide you with all kind of means to control the solution process, but they may be hidden (advanced menu's, or outside the graphical pre/postprocessor: via the input file), as you probably don't need it for many engineering assessments.
Still, if I had to choose my path in the problem you describe, I would probably go for COMSOL. Three way coupling is possible (I even did 4 way coupling) and you can also go non-linear. BUT, start simple, start linear, make sure you are confident you have physically sensible solutions, then make it more complex. NEVER expect that code will solve your problem automatically. Doing good FE simulations is a craft!
I think that you should use COMOSOL, but it is possible to use Abaqus with Matlab to resolve a coupled problem (multiphysics simulation).
- Badges
- Science topic