[steve20024]

While I'm doing simulation of a 5 blade VAWT, I've encountered a problem.
I'm using CFX, turbulence model is SST, and doing steady-state analysis with frozen rotor.
The problem is, there's a big difference between the torque I get from different angle I place the blade. Not the angle of the blade, but the position of blade.
Does that means I have to do like 6 simulation from 0 degree to 72 degree, and then get the average of 6 simulation?
I've checked some paper of VAWT simulation, but I didn't see description abouut this problem.

Attachment are the demo pic of the angle, first is 0 degree, second is 12 degree

Thanks

[bluebase]

Hi Steve
I haven't yet used the frozen rotor approach of cfx.
As far as i know, the flow for a VAWT has many intransient effects. So using a steady state approach might not cover those a effects well.
Also, i suggest you to read about MRF-Simulations to understand what the frozen rotor option of cfx does, the manual might have some answers.
Another point, i would check your y+ values. The SST model needs very fine meshed boundary layers, some papers say down to y+<1.
And regarding meshing, did you check coarser and finer meshes with your setup?
With regards,
Sebastian

[steve20024]

Thanks for answering, Sebastian
I'm sure about the y+ are below 1, and I didn't use structure mesh.
As the CFX manual, I've checked it but not pretty sure there are answers.
I know the intransient effects are muuch, but I got many cases to simulate, so I want to try if steady state approach works...
Thank you for replying, I'll keep trying.

[sreenivas]

Steve

the VAWT physics is transient in nature,torque generation is essentially sinusoidal in nature and each blade is going through the wake of other blades which is again unsteady behavior.By doing steady state MRF simulations you wont get values closer to the actual torque values.You need to simulate at least 3 revolution of the rotor to get Average torque experienced by the turbine.
You might be advised to use MRF to speed up the transient simulation convergence
There is a tricky way to calculate the time averaged torque generated by all the blades (which is an average of Sinusoidal torque variation)

Srinivas

[sreenivas]

Hello steve
it looks something wrong with your turbine blades orientation...is the flow passing from pressure side to suction side or Vice verse...if this is the case your turbine might act like pump