Hello friends,

I changed the time marching scheme from runge kutta to adams bashforth and the code blows up (it is not the time step problem).I am running a case of flow over airfoil .The point where the code starts to get unstable is the front side wall boundary of the airfoil.

Can anybody guess what the problem could be.

Thanks in advance vasanth

Adam BAshforth, contrary to popular belief leads to spurious solutions... Check the following paper... PLZ remember Runga-Kutta is always better for time stepping

A comparative study of time advancement methods for solving Navier-Stokes equations -Journal of Scientific Computing Vol-21(2), pp 225-250. Oct. 2004

Thank you srinath , I undestand that. It might give wrong results but it should not blow up after a few steps. I looked at the solution and saw that the code blows up at the wall. I dont know why this is happenening. It is because of the beta values.

Thanks in advance

Are you sure your problem is because of the time marching scheme? If the solution blows up at the wall, then you should probably take a closer look at the boundary conditions and their implementations. I suggest that you run your problem for the steady state case and if the solution still blows up then it has nothing to do with the time stepping algorithm. Good luck

Thank you Srinath. The problem at the wall was fixed.The code is actually blowing up at the trailing edge of the airfoil.I have to investigate why ths was not happening for runge kutta time marching and why is it happeing for adams bashforth time marching.

any suggestions friends ?

The numerical stability depends on CFL. Different kinds of Runge-Kutta, and Bashforth have different bounds of CFL. The specific marching scheme determines its upper bound of CFL.