Genetic-Algorithm-Based Control Parameter Optimization of PMSLM Servo System

Based on the conventional field-oriented control (FOC) framework of permanent magnet synchronous linear motor (PMSLM), a hybrid strategy improved genetic algorithm (HGA) is proposed. The aim is to achieve multi-objective optimization of PI controller parameter. Firstly, Tent chaotic map is introduced ensure the diversity of initial population, and expand the search range. Secondly, the adaptive crossover and mutation probability is used, probability threshold adjust dynamically in the search process, so that can avoid the algorithm falling into local optimum and enrich the search mechanism. Thirdly, opposition-based learning of unstable parameters is introduced to deal with the unstable individuals of the system, optimization performance and convergence speed are improved. Finally, the fitness function is designed to make the system speed error, position error and speed fluctuation achieve better indicators. The simulation results show that compared with the three classical optimization algorithms, the performance of the proposed algorithm is improved in terms of convergence speed and convergence accuracy, the optimized parameters are effective and practical in PMSLM servo system.