CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   Siemens (https://www.cfd-online.com/Forums/siemens/)
-   -   Parameters of discretisation schemes (https://www.cfd-online.com/Forums/siemens/55830-parameters-discretisation-schemes.html)

george February 1, 2007 09:34

Parameters of discretisation schemes
 
Hello to everybody,

I have a question about the discretisation schemes. Firstly I used UD scheme for velocities and turbulence variables (standard high Reynolds k-e turbulence model) and 10000 iterations with maximum residual error 10E-05. My run stopped at 314 iterations (steady state). After I changed the scheme in LUD and number of iterations in 50000 without change anything else (e.g. Blending factors or sweeps or relaxation factors). My run went then in 50000 iterations and it did not converged. I do not understand which parameters usually we prefer to change (increase or decrease) in order to achieve the desired convergence (e.g. blending factors or sweeps or relaxation factors). Now I run the same case with the QUICK scheme, but I believe that it will not converge again due to the above reason (I do not change some parameters). Do someone can help me?

Thank you in advance

George

sidd February 1, 2007 11:03

Re: Parameters of discretisation schemes
 
George,

There is no thumb rule for differencing schemes. It depends a lot on the kind of problem you are handling.

UD scheme being first order, usually converges without much problem, giving less accurate result. LUD being second order, take time for convergence. And QUICK is 3rd order, it may take more time.

As per my knowledge, you should try a lower blending factor, as this factor decides the mixing of higher order and lower order differencing.

But i can't say that this blending only may be the problem in your case.

Good Luck


george February 2, 2007 01:45

Re: Parameters of discretisation schemes
 
Dear Sidd

Thank you very much, I will begin firstly by decreasing the blending factor to investigate the character of the run.

George

kathir February 2, 2007 03:07

how to control the yplus value
 
hi expert my problem is finding the heat transfer coefficient around the cylinder in a uniform flow. but i got the ypus value nearly .5 to 3. according to my theoretical knowledge for this yplus value k-e model overpredict the heat transfer. how to get the accurate heat transfer coefficient. but i cant provide the fine mesh in near walls because i want to find the boundary layer effects how to use wall functions for this kind of problem

by kathir

siva March 30, 2007 02:29

Re: Parameters of discretisation schemes
 
What is CFD? Whee it is used?, What are the types of Algorithm and Advantages?


All times are GMT -4. The time now is 18:59.