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.
Low Mach number flows in COMSOL
Posted May 29, 2015, 5:16 a.m. EDT Fluid & Heat, Computational Fluid Dynamics (CFD), Studies & Solvers 1 Reply
Please login with a confirmed email address before reporting spam
Hello all,
I am looking to model flow in a combustion chamber where the flow is compressible because of the temperature variations and not due to pressure. The Mach number <0.3.
Numerical solution of mildly compressible flows are known to be problematic. In finite volume methods there are two main problems :
1)Slow convergence
2)Inaccurate representation of pressure
To account for this in finite volume methods, preconditioning methods are used.(use of artificial compressibility etc)
How does COMSOL tackle the problem of low mach number flows? Is there special pre-conditioning applied to overcome the numerical difficulties?
I am looking to model flow in a combustion chamber where the flow is compressible because of the temperature variations and not due to pressure. The Mach number <0.3.
Numerical solution of mildly compressible flows are known to be problematic. In finite volume methods there are two main problems :
1)Slow convergence
2)Inaccurate representation of pressure
To account for this in finite volume methods, preconditioning methods are used.(use of artificial compressibility etc)
How does COMSOL tackle the problem of low mach number flows? Is there special pre-conditioning applied to overcome the numerical difficulties?
1 Reply Last Post May 31, 2015, 10:03 p.m. EDT
Please login with a confirmed email address before reporting spam
Posted:
10 years ago
I routinely solve problems of this type.
The problems you describe with the finite volume methods are probably due to the fact that the solver is not set up to be fully implicit; i.e., the full Jacobian is not formed to solve the Newton iteration. COMSOL has the full Jacobian, and allows for fully coupled. You can simplify the solver from there to slow the convergence down if you like in order to fit it into the machine memory.
The problems you describe with the finite volume methods are probably due to the fact that the solver is not set up to be fully implicit; i.e., the full Jacobian is not formed to solve the Newton iteration. COMSOL has the full Jacobian, and allows for fully coupled. You can simplify the solver from there to slow the convergence down if you like in order to fit it into the machine memory.