Optimization of MW-Class Dual Three-Phase Permanent Magnet Synchronous Machines for Electric Propulsion Vessels Considering Mutual Effects Based on Neural Network

The mutual effect on a dual three-phase permanent magnet synchronous machine (DTP-PMSM) refers to the magnetic coupling between two electrical systems. It can cause significant changes in the characteristics of the machine, potentially affecting the stability and performance of the overall system. Therefore, when designing and modeling a DTP-PMSM, it is crucial to consider the mutual effect and accurately account for its impact on the system's behavior. However, considering the mutual effect in operating ranges and design variables is not feasible in a wide design space owing to the complexity and variability of the system. To address these considerations, we optimize the DTP-PMSM for propulsion motors in electric propulsion vessels (EPV) by employing a neural network (NN) to account for the mutual inductance across the entire operating range. The proposed method is effective as it involves performing finite element analysis (FEA) at the rated operation points, while the remaining points are optimized using computed data obtained from the NN, which has learned the mutual inductance. This approach takes advantage of the fact that EPV propulsion motors has low saturation characteristics, making nonlinearity negligible. The optimized model has been validated through co-simulation, demonstrating that the results can contribute to the practical design and operation of DTPPMSM.