Aiming at the problem that the loss and temperature rise of the pole-suspended rotor low-speed high-torque permanent magnet synchronous motor (LHPMSM) increase in the pursuit of high torque density, and the design cycle is prolonged due to the dependence on thermal post-verification. In this paper, a multi-physics trade-off design method based on weighted heating rate combined with a surrogate model and a multi-objective evolutionary algorithm is proposed. Firstly, the rationality of introducing a weighted heating rate is proved by mathematical proof and thermal network calculation. Secondly, the two-dimensional sensitivity analysis of the key structural parameters of the motor is carried out to identify the most influential structural variables, which are then used to construct a high-precision surrogate model based on gradient boosting regression tree (GBRT). Then, in order to effectively obtain the Pareto solution set of balanced torque performance and heat dissipation performance, the non-dominated sorting genetic algorithm (NSGA-II) is used for multi-objective optimization. Finally, the multi-physical field finite-element simulation verification and a 356kW prototype experimental analysis show that the optimized design significantly improves the torque performance while effectively controlling the temperature rise and realizes the fast compromise design of the multi-physical field of the motor. The effectiveness and advancement of the proposed method to achieve coordinated improvement of high power density and high steady-state thermal margin in motor design are verified.
Wang et al. (Fri,) studied this question.