simulation of a stirred tank
 

Hi, everybody,

I intend to simulate a stirred tank with Rushton-impellers using a snapshot method proposed by prof.Ranade. Has anyone tried this method?

He thought the mass sink Sm=-Rho * Abc * Wbc ( Abc and Wbc are the area of the computation cell adjacent to the back side of the blade and rotational velocity of blade respetively) and mass source Sm1= - Sm for the compuatation cell adjacent to the front side of the blade, which means Wbc = Omega * r ( r is the radius of the computation cell). However when I add the mass source and sink into the pressure equation, (by su=su+sm) and I add the corrosponding source terms and sink terms into the momentum equations ( by su=su+sm*phi, phi is U, V, W velocities of the last iteration). However, I found I can not proceed because there is a fatal error about floating-point mistakes.

Who can give me some suggestion?