# Is COMSOL Multi Physics is suitable to solve complex flow problem?

 Register Blogs Members List Search Today's Posts Mark Forums Read

 October 25, 2018, 22:46 Is COMSOL Multi Physics is suitable to solve complex flow problem? #1 Member   Steve Join Date: Mar 2018 Location: South Korea Posts: 69 Rep Power: 5 Hello. I used about 6 years STAR-CCM+ in automotive company. But! I changed my job recently. Now I have to simulate some physics related semiconductor equipment. Simulating semiconductor equipment require to solve multi physics problem. Traditional CFD codes are not suitable to solve multi physics problem. (or have to buy other codes for coupling) So my company want to use COMSOL multi physics solver. I experienced COMSOL about 3 weeks(yes, it's very short). But I think COMSOL is not suitable to solve complex flow problem (I have no idea about other physics). Whether the problem is my ability to use COMSOL solver or the solver's limitation. there're the problem i think. 1. Too much spend time to solve CFD problem 2. Convergence problem when use pressure B.C 3. Result accuracy please give me a comment! Thanks. Last edited by steve lee; October 26, 2018 at 02:02.

 October 26, 2018, 00:18 #2 Senior Member   Arjun Join Date: Mar 2009 Location: Nurenberg, Germany Posts: 1,045 Rep Power: 28 Multiphysics is very broad. It is hard to comment without know what exactly you are trying to simulate. So what is the simulation that comsol is not able to do. A bit of detail would be help. Time4Tea likes this.

 October 26, 2018, 01:56 #3 Member   Steve Join Date: Mar 2018 Location: South Korea Posts: 69 Rep Power: 5 Dear arjun Thanks for reply I didn't try to solve multi physics problem yet.. This step is just CFD simulation. My question is.. this I think COMSOL is good for multi physics simulation(many people use COMSOL. I think they have enough reason using this code). But I wonder how good this code for just CFD simulation 1. I tried to solve simple cylinder(Radius 1 m, Height 10 m) with pressure B.C(velocity inlet and pressure outlet is OK. Almost all examples using velocity inlet..). but simulation was not converged.. I'm sure mesh is enough.. even though mesh is too coarse, it's just simple cylinder.. i used not only STAR-CCM+ but also FLUENT, CONVERGE, OpenFOAM.. any other solvers were converged.. 2. In gas mixing simulation, can i use more than 3 gases? i can not find more than 3 phases simulation module.. (without reaction.. just flows) 3. I get axisymmetry geometry(chamber with nozzle). and i tried to solve using symmetry function. and than i tried to solve full 3D model (not use symmetry function) but result was different.. 4. Solving time.. I made 200,000 meshs and tried to solve laminar steady flow with 4 core. but i felt solving time is too long compared other solvers i refered..

October 26, 2018, 02:23
#4
Senior Member

Arjun
Join Date: Mar 2009
Location: Nurenberg, Germany
Posts: 1,045
Rep Power: 28
Quote:
 Originally Posted by steve lee Dear arjun Thanks for reply I didn't try to solve multi physics problem yet.. This step is just CFD simulation. My question is.. this I think COMSOL is good for multi physics simulation(many people use COMSOL. I think they have enough reason using this code). But I wonder how good this code for just CFD simulation 1. I tried to solve simple cylinder(Radius 1 m, Height 10 m) with pressure B.C(velocity inlet and pressure outlet is OK. Almost all examples using velocity inlet..). but simulation was not converged.. I'm sure mesh is enough.. even though mesh is too coarse, it's just simple cylinder.. i used not only STAR-CCM+ but also FLUENT, CONVERGE, OpenFOAM.. any other solvers were converged.. 2. In gas mixing simulation, can i use more than 3 gases? i can not find more than 3 phases simulation module.. (without reaction.. just flows) 3. I get axisymmetry geometry(chamber with nozzle). and i tried to solve using symmetry function. and than i tried to solve full 3D model (not use symmetry function) but result was different.. 4. Solving time.. I made 200,000 meshs and tried to solve laminar steady flow with 4 core. but i felt solving time is too long compared other solvers i refered..

What you wrote is pretty basic stuff actually other than the number of gases thing.

The way i understand is that there are three situations
1. All the phases (gases) are well separated and then you need multiphase VOF type method (that i am not sure available there).

2. Gases or phases are well mixed and you need mixture multiphase type model (this again i am not sure available).

3. There is a multi-component mixture that is reacting (treated as single phase that is undergoing reactions) this i expect to be available.

What usually these softwares can't do is a situation where combination of above three conditions appear.

An example would be

https://youtu.be/9kkBMeEl2R8

Here the VOF method is doing the tracking while one of the phase is undergoing reactions (or some physics is applied).

PS: If you are just doing cylinder then a simple axis-symmetric model is good enough.

 October 26, 2018, 03:07 #5 Member   Steve Join Date: Mar 2018 Location: South Korea Posts: 69 Rep Power: 5 Thanks for your reply.. But, my question is not that.. Are you COMSOL user?.. Already I have experiences to solve many CFD problem using other code.. all i want to know is that how to set some physical models (or i want to know it's possible or not) in COMSOL.. My goal is that how to use this code effectively and want to know about the limitation. I felt this code is quite different than other CFD(STAR-CCM+, FLUENT, CONVERGE etc..) code..

 October 26, 2018, 14:12 #6 Senior Member   Lucky Tran Join Date: Apr 2011 Location: Orlando, FL USA Posts: 4,459 Rep Power: 53 COMSOL has some limitations in available turbulence models for CFD, but since you are only doing laminar... There shouldn't be any issues. You should also be able to converge a simple flow in a cylinder easily. I'm going to put this one on your inability to use COMSOL. Maybe doing some tutorials will help?

October 27, 2018, 04:53
#7
Senior Member

Filippo Maria Denaro
Join Date: Jul 2010
Posts: 5,905
Rep Power: 63
Quote:
 Originally Posted by steve lee 1. I tried to solve simple cylinder(Radius 1 m, Height 10 m) with pressure B.C(velocity inlet and pressure outlet is OK. Almost all examples using velocity inlet..). but simulation was not converged.. I'm sure mesh is enough.. even though mesh is too coarse, it's just simple cylinder.. i used not only STAR-CCM+ but also FLUENT, CONVERGE, OpenFOAM.. any other solvers were converged..

Laminar???

Are you using oil as fluid?

 October 28, 2018, 22:02 #8 Member   Steve Join Date: Mar 2018 Location: South Korea Posts: 69 Rep Power: 5 Actually i tried many case test with COMSOl.. I used laminar and tubulent model (different geometry..).. Also i tried to solve examples. Example is OK. well converged but when I changed B.C as velocity inlet to pressure inlet.. It's not converged..(even though the pressure drop is very small)