dear all. in my compressible and unstable les code,i use upwind scheme to convective item,and use LU approximate method.smagorinsky model is used.i tyied to apply cs=0.1-0.3 to the model,but the code did never to be convergent.

the flow chart of my program : 1,generate 3d mesh, 2,apply initial condition 3: main computational program (1)subroutine filter , (2)subroutine LU ,using finite difference method to solve N-S equations. (3)subroutine bounder,supply BC. (4)subroutine dynamics,integrate pressure field. 4,analyze flow field. the end thanks for your help,best wishes for you.

Dear baal:

My LU-SGS methold can go on now and the result is correct. It only fast 51s than the inital methold per time step.

It saved about 35% time.