Abstract: In order to improve the fault-tolerant ability and output ability under fault conditions of permanent magnet synchronous propulsion motor, a new adaptive sliding mode fault-tolerant control method based on PSO-RBF optimization was proposed for the phase failure of propulsion motor in marine electric propulsion system. The switching gain of sliding mode controller was adjusted in real time by use of the learning ability of neural network, and the parameters of neural network were optimized by particle swarm optimization algorithm. Then the fault-tolerant control strategy was adopted to adjust the phase and amplitude of the other healthy phase currents after phase failure, so as to realize the rapid convergence and chattering weakening of the marine propulsion motor under normal operation and fault conditions, and then improve the fault-tolerant control performance of the propulsion motor. The simulation results show that compared with the traditional sliding mode control, the speed recovery time of the proposed method is shortened by 1.7 s, the steady-state error of speed is reduced by 0.17%, and the steady-state error of torque is reduced by 7% under phase failure conditions. The more stable operation of marine propulsion motor, the better the control performance after fault.

Key words: transportation equipment engineering; permanent magnet synchronous propulsion motor; RBF neural network; particle swarm optimization; adaptive sliding mode control; fault-tolerant control

摘要： 为提升船舶永磁同步推进电机的容错能力与故障条件下的输出能力，针对船舶电力推进系统中推进电机缺相故障，提出一种基于PSO-RBF优化的新型自适应滑模容错控制方法，利用神经网络的学习能力对滑模控制器中的切换增益进行实时调节，并通过粒子群算法对神经网络参数进行寻优，随后采用容错控制策略对缺相后其余健康相电流的相位和幅值进行调节，实现船舶推进电机在正常运行和故障状态下系统的快速收敛和抖振削弱，提高推进电机的容错控制性能。仿真结果表明：笔者方法较传统滑模控制在缺相故障下，转速恢复时间缩短了1.7 s，转速稳态误差降低了0.17%，削减了7%的转矩稳态误差，船舶推进电机运行更稳定，故障后控制性能更优良。

关键词: 交通装备工程；永磁同步推进电机；RBF神经网络；粒子群算法；自适应滑模控制；容错控制