Design And Bench Verification of PSO-Optimized Fuzzy PI Control Strategy for PMSM Based on UAV Flight Load Simulation

The development of the electrification technology of unmanned aerial vehicle (UAV) puts forward higher requirements for the control performance and verification methods of permanent magnet synchronous motor (PMSM). In this paper, a fuzzy PI control strategy based on a particle swarm optimization (PSO) algorithm is proposed to optimize the parameters of the PI controller and improve the dynamic response and control accuracy of PMSM. Firstly, Matlab/Simulink is used to build an online vector control model of PMSM, and the PSO algorithm is used to optimize the controller parameters online to adapt to the dynamic characteristics of the motor under different working conditions. Secondly, a current reconstruction scheme is designed, which reshapes the current waveform by sampling at the center point of the pulse width modulation (PWM) signal to make it closer to the ideal sine wave, so as to improve the motor operation efficiency. Finally, Speedgoat is used to build a HiL hardware in-loop simulation platform for UAV electrification applications. HiL simulated the load torque of the UAV propeller and loaded it to the test motor, and the drive motor realized closed-loop control according to the flight speed command. The response speed and anti-interference ability of fuzzy PI control strategy based on PI, fuzzy PI, and particle swarm optimization algorithm are compared and analyzed experimentally.The results show that, compared with the traditional PI control algorithm and fuzzy PI control algorithm, the speed amplitude of the PSO-optimized fuzzy PI control technology is reduced by about 50% at low speed and 30% at high speed under typical flight condition curve. The effectiveness of the algorithm is proved. This study effectively verifies the high responsiveness and robustness of the proposed control method, and also verifies the effectiveness of the designed semi-physical hardware in the loop simulation platform.