Proportional-Integral Parameter Setting of Permanent Magnet Synchronous Motor Based on Surrogate-Assisted Multimodal Optimization Algorithms

This paper primarily investigates the application of Surrogate-Assisted Multimodal Optimization Algorithms (SAMMOA) to solve the Proportional-Integral (PI) parameter optimization problem of the three-phase permanent magnet synchronous motor (PMSM) and find multimodal solutions more efficiently. In the PMSM vector control process based on the PI controller, a simulation model is constructed using Simulink, and the synchronous motor module is built in Ansys. The time-integral absolute error (ITAE) generated by the PI module is used as the optimization objective for the PMSM PI parameter optimization, which is time-consuming and has multimodal solutions. Therefore, it is necessary to design an algorithm that incorporates global and local search strategies, archive selection strategies, and multimodal solution selection strategies to address this problem. At the same time, considering the accuracy of model construction, a local surrogate model is built by selecting a certain number of individuals around the target individual for prediction. Simulation and experimental results show that using the expensive multimodal algorithm for PI tuning can quickly find multiple multimodal solutions.