RBFN based efficiency optimization method of induction motor utilized in electrically driven marine propellers

Thruster controllers of electric propulsion system with fixed pitch propellers are conventionally aimed to control only the shaft speed without utilizing the capabilities of the controllers to apply any other control strategies. In fact, the dynamic operating conditions lead to the fluctuation of motor load. For this reason, utilizing conventional controllers is hard enough due to the critical constraints and limitation of the ship power source. The paper presents study and analysis of efficiency optimization strategy in thruster shaft speed controllers driven by induction motor. The control strategy based on intelligent method called radial basis function neural network (RBFN) is implemented. A set of training data derived from a loss model controller of the induction motor working under indirect field-oriented-control (IFOC) drives is used for training process of RBFN. The loss model controller utilizes schematically the flux generating current as controlling variable. Estimation of the flux generating current through the RBFN process shows significant improvement in motor efficiency especially for low speed and ship transit system.