CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   Main CFD Forum (https://www.cfd-online.com/Forums/main/)
-   -   Best turbulance model for analysis (https://www.cfd-online.com/Forums/main/114489-best-turbulance-model-analysis.html)

Best turbulance model for analysis

I am doing a undergraduate research for calculating drag force and co-efficient of drag on heavy vehicle. please suggest me best turbulence model which one i should select for more realistic and accurate result. And please tell me should i scale down my model or i should perform analysis on full length model?? nd Wht should be OPTIMUM no of iteration for convergence?? I am using SOLIDWORKS FLOW & AUTODESK CFD.

 atmcfd March 15, 2013 16:18

1) There is no such thing as "best" turbulence model... but usually in these cases k-epsilon / Spalart - Allmaras models are preferred ( and well documented for external flows, so you can refer)

2) Scaling is really your choice, provided a good scaling factor exists for your problem. Else, its safer to do a full length sim. But if you want a good solution and want to observe the trailing vortices etc. from the vehicle, you might have to make the near wall mesh quite fine, and this will increase your computation time.

3) There is no such thing as the optimum no. iterations for convergence in turbulent flow simulations. you can give a reasonably low tolerance criteria for the residuals. The residuals going down is a good indication of convergence,but not the only one. You can also observe the flow statistics , like averaging of the velocity fields etc. and when you get something close to a statistically stationary/uniform field, you can say that the simulation has converged (almost)

Thank you sir!!!

One more thing i want to ask if i scale down my model then i have to increase the velocity in same proportion. Sir can u tell me what is the maximum speed till that we can neglect the effect of compressibility?? neglect the Effect of mach number on Coefficient of drag?? Please Reply..

 Aeronautics El. K. March 16, 2013 11:47

You have to adjust the velocity with respect to the characteristic length of your model in order to have the same Reynolds number.

If the flow you're studying is not dominated by viscous effects (which in your case I believe it's not) then you can approximate the flow as incompressible if the Mach number is less than 0.3 (Ma<0.3).