CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   FLUENT (https://www.cfd-online.com/Forums/fluent/)
-   -   what can cause divergence? (https://www.cfd-online.com/Forums/fluent/34523-what-can-cause-divergence.html)

 Jeffery August 31, 2004 10:51

what can cause divergence?

My model is simple,(a cylinder, two opening(top and bottom).Gas come from a a opening(velocity inlet) and out from another(pressure outlet).I take gas as incompressible idea gas. And this is transient model. My question is that in my model, continue equation does't converge(5e-2), even iterating over 1000 time. So I want to know what may be wrong in my model. Appreciate !!!

 SAM September 1, 2004 11:44

Re: what can cause divergence?

it is also convergence. if ur residuale line contiue to be on same value then it is convergene by defintion.

 Jeffery September 1, 2004 11:59

Re: what can cause divergence?

Thanks, but I am still not confident about this result.Is there any way to make this value down(such as less than 1e-3)? Meshing the grid fine will help?

 Jason September 1, 2004 12:28

Re: what can cause divergence?

The residuals you are looking at aren't the actual residuals. They are the average residuals normalized based on the initial residuals. Which means if your initial guess is close, and there isn't much change from the initialized solution, then you're residuals won't drop 3 orders of magnitude.

You shouldn't judge convergence just on the residuals though. You should monitor any important characteristics, such as lift coefficient, drag coefficient, pressure at a point, maximum velocity magnitude over the domain... Think about what you are trying to get out of the solution and monitor those values to make sure they are converging.

The residuals help you understand what your model is having problems with, and help you to adjust your model to make it solve better, but convergence shouldn't be judged by residuals alone. If the residuals are flat-lined, and your monitors are too, then you have a converged solution you can be confident in (assuming you didn't screw up the model :) )

Good luck, Jason

 Jeffery September 1, 2004 14:40

Re: what can cause divergence?

Jason,thanks a lot, you suggestion is very helpful for me.

 ppaitt September 16, 2004 12:54

Re: what can cause divergence?

Try to mesh the model again. And make more denser grid as you can. The mesh quality plays an important role on convergent.

 All times are GMT -4. The time now is 18:22.