The solver is encountering convergence difficulties when running a torque converter calculation in CFX, even for good quality meshes, why?
|
One method of diagnosing a problem of this sort is to save the results right before the solver crashes and examine the solution for clues on what might be causing the convergence issues. One common issue is the eddy viscosity at the stage interface. If the Eddy viscosity seems to be unusually large at the interface, that is usually due to large velocity gradients at the interfaces. These large values cause the Turbulence equations to diverge. This problem occurs when using k-epsilon and stage models. The large values of the eddy viscosity are caused by large production rates, which are not limited for the epsilon-based turbulence models with default settings. However, they are limited for omega-based turbulence models with default setting. So running SST model can help, or the user can turn on the Production Limiter under Advanced Model Controls. Set a value up to 3 for k-epsilon model. |
||
|