It has been confirmed that single-channel impellers can not be simulated with rotating region, if the center of mass is not inline with the rotational axis. Also if working with two different material densities for the impeller, the result is an elliptic rotation, so the case cannot be solved. Only when building a "balance mass" onto the impeller, the case can be simulated; but the torque results aren't usable (higher because of additional faces). So it would be nice to change the axis of rotation in the simulation program, to get usable results.