Hello all,

I am trying to simulate the cold flow through a swirl stabilised combustion chamber with CFX 11. Besides of convergence problems due to the high swirl I am facing an unexplainable problem with increasing temperature. The case setup is without fuel and without combustion model. However the temperature is increasing at some wall adjacent cells. After approx. 50 iterations the temp. starts to increase and goes up to values of approx. 5e+03. This occurs regardless of the percision I use (single precision / double precision). Has anybody experienced that?

Thanx for your support.

Mic

P.S. The RMS residuals for momentum and energy are in the range of 8e-05 (using a blend factor of 0.5). The grid quality seems to be ok.

Hi,

Can't you open the results in CFX-Post and see what is causing the high temperature?

Glenn Horrocks

ps: blend factor of 0.5 is lower than is normally used, this will probably not increase accuracy significantly over a pure first order solution. Have a read of the documentation regarding tips for obtaining convergence.

Hello Glenn,

thanx for your answer. I started using the high resolution scheme, then shifted to upwind and finally tried to use a specified blend factor. The problem of diverging temperatures occurred irrespectively of the used advection scheme. I used different time steps and found the same result with regard to the temperature problem. CFX-Post revealed that the high temperature occurred at some sporadically distributed cells near the wall. The Mach number in the vicinity of those cells is quite low (M < 0.1) and the expansion factor is within the recommended range. I have no idea what is causing this.

Mic