To simulate the flow nearby the wastewater discharge, I densify the mesh near the discharge. Now the size of the biggest mesh is about 70 times that of the most small ones. As we know the timestep is close related the size of mesh, I have to use a very small timestep after the mesh densified. I want use different timestep in differnet size mesh. I will be grateful if anyone can give me some advice about my idea. Thanks

Liu,

It is OK to use individual time steps for various elements if you want to simulate the steady-state. It is known as a distorted transient approach.

However, if your aim is to simulate a real transient behavior, you need to use a consistent time step (a true transient approach). It can be the time step related to your smallest elements if you use an explicit method. Implicit methods allow bigger time steps.

regards

DML

DML, Thanks for your advise. Fortunately, I just need a steady simulation result. I search a concept of "local timestep" yesterday. Is it just the approach to solve my question? Can you give me some reference paper?

Liu,

You may consult Chapter 7 by Raithby and Schneider (Elliptic Systems: Finite-Difference Method II) in the book by Minkowycz et al. (eds.) Handbook of Numerical Heat Transfer, J.Wiley 1988.

Have a look at the section with the Eq.(7.27).

regards

DML