CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   WHICH SOLUTION TO BELIEVE ? (http://www.cfd-online.com/Forums/cfx/20545-solution-believe.html)

Bob August 10, 2004 20:05

WHICH SOLUTION TO BELIEVE ?
 
Greetings to everyone!

My task is modelling a spiral subsonic flow inside of a cylinder that rotates around its main axis, in CFX-5.6. I am using the SSG turbulence model. The solution converges within 100 iterations gradually to the RMS residuals: U-Mom = 1.5E-4 uu-RS = 6.6E-4 mass = 9.5E-7 In CFX-Post, the flow demonstrates several swirls. Running the Solver further is resulted in steady growing of all residuals and at the iteration 620, I have:

| Equation | Rate | RMS Res | Max Res | +----------------------+------+---------+---------+ | U-Mom | 0.97 | 6.9E-04 | 1.9E-02 | | V-Mom | 1.06 | 4.6E-04 | 1.0E-02 | | W-Mom | 0.89 | 5.7E-05 | 1.6E-03 | | P-Mass | 1.05 | 9.3E-06 | 2.1E-04 | +----------------------+------+---------+---------+ | uu-RS | 1.05 | 7.0E-03 | 6.5E-02 | +----------------------+------+---------+---------+ | vv-RS | 1.07 | 7.7E-03 | 7.1E-02 | +----------------------+------+---------+---------+ | ww-RS | 1.00 | 6.1E-04 | 7.9E-03 | +----------------------+------+---------+---------+ | uv-RS | 0.98 | 9.1E-03 | 8.2E-02 | +----------------------+------+---------+---------+ | uw-RS | 0.90 | 2.1E-04 | 3.6E-03 | +----------------------+------+---------+---------+ | vw-RS | 1.06 | 3.7E-04 | 9.5E-03 | +----------------------+------+---------+---------+ | E-Diss.K | 0.98 | 4.6E-04 | 1.6E-02 |

There is NO SWIRLS in this solution at all !

Anyone would know which solution to believe and if I should continue iterations hoping it will finally converge?

Glenn Horrocks August 11, 2004 19:10

Re: WHICH SOLUTION TO BELIEVE ?
 
Hi Bob,

I would believe neither at this stage. If it does not converge to a credible solution then something is probably wrong with the simulation. I suspect your simulation is not fully converged, and if it is not converging nicely then either your mesh quality is poor, your timesteps are wrong, you are using an incorrect physical model or you need to try a transient simulation.

Glenn


All times are GMT -4. The time now is 10:49.