hi forum, i'm designing a new kind of wind turbine, I need to find out the rotor performance, i've done it in fluent, now which one is best for validation among the following, 1) Blade-momentum theory. 2) 3D Panel methods 3) integral boundary layer method, I cant afford for experiments because of limited facility here. thanks in advance, Sharuk

You cannot expect to "validate" a Navier Stokes solver, with an inviscid flow solver, or with an even-further simplified model. I would guess that your fluent solution would be way too diffusive to be correct, but at the same time using panel methods would be valid only in a narrow range, which excludes flow separation. That is, if you are 100% sure that the flow is attached, then I'd say go for the panel method as it probably is a better model of turbulent flow at very high Reynolds numbers than fluent could be. But if there is separation, then you cannot use standard panel methods (unless the code you have has been enhanced with the capability to somehow account for boundary layer separation)

Adrin Gharakhani

dear Adrin, it's good to 've a reply from u, my model is small wind turbine, n it's a concept of hubless type windturbine, i've submitted even one conference paper to WREC,NREL, is it good to model for one Blade as i did in fluent or i've to solve it on whole geometry in 3D panel methods, the rotor dimension of model is 2m dia n no. of blades are 4, i fabricated one model and when i made it run before a wind speed of 3.67m/s it rotated @ 240 rpm, the rotor was freely rotated with out any load on it. sharuk

Did you model the 4 blades using one blade and a periodic boundary condition? If not, I'd say you are solving the wrong problem altogether. If yes, then your answer is closer to reality. The problem with the assumption of symmetry is that in 3D unsteady turbulent flow you can't expect to see symmetric flow even for symmetric geometries (I'm sure you know that already). Depending on the design, there may be wake-blade interaction and/or wake-wake interaction (due to the 4 blades) that will affect the load distribution, and you want to be able to capture that.

Adrin Gharakhani