Hallo

I am having troubles getting verified my CFD simulations. I have LDA data of the velociuty profile in a ball shaped valve. But when I make CFD-simulations I have quiet a difference between the measured and simulated velocity profile?? How accurate could one expect the CFD-simulaitons to be when a complex 3D flow is investigated?? Does anyone have any experience with this?? i.e. some literature.

Regards

Bo

I think you may check your results of CFD as following: 1,the differencing Schemes 2,turbulence model 3,Solution Algorithm I think the first is very important .Many CFD examples have verfied . If it is possible.You can support more details messages-

(1). If you change the mesh or refine the mesh and re-run the case, it is likely that they are not going to be the same. (2). And if you also chenge the numerical methods or the turbulence models, I am sure that the solutions will be quite different from one another. (3). It can match the test data only if: you modify and validate the numerical scheme and turbulence model specifically for your application! (4). If you are using a general CFD code, validated for other flow problems which do not include your applications, then the result is not going to match the test data. (5). There are "General codes", but they are not "universal codes or models". This is the most important concept in using CFD. CFD is trying to give you a vehicle, but unless someone also give you a road map, the vehicle alone is useless for you to reach your destination. Most poeple simple do not understand this principle. They normally think that their solution is preprogrammed inside the code. And they can find the solution by running the code. That's a big mistake. And it is hopeless.

IMHO in your case the main difference will be from the turbulence model. Do you use an isotropic model? May be it would be better to use some anisotropic model like LES, for example.