In the configuration you are attempting to test, are your blades rotating? If so, you need to use the MRF approach. You need to create a small region around the turbine blades. This volume would be rotating relative to the absolute frame. The rest of the fluid region would be stationary.
If this is not the case and you are attempting to model a fully stationary case, then you do not need the MRF approach. You simply would be modeling everything as stationary.