在线粒子群优化PMSM 无传感器控制研究

Research on Sensorless PMSM Control Based on Online Particle Swarm Optimization

针对永磁同步电机(PMSM)无传感器控制存在速度值的估算精度容易受到多参数干扰而产生误差的情况,特别是转子在低速运转情况下更明显,提出一种有效的在线粒子群优化方法。设计了永磁同步电动机无传感器控制中的前向通道中PI控制器参数和反馈环中电流、速度PI控制器参数。转子速度和定子绕组磁链采用滑模观测器来进行估算,定子电阻值动态补偿量通过速度修正量计算获得,而速度修正量通过d轴电流估算的误差值计算获得。由于采用了粒子群优化方法,6个控制器参数以及定子电阻的变化量可以同时设计,避免了交叉耦合,减少了速度估算的误差。试验中,采用硬件DSPACE1103,指令信号采用一阶阶跃信号,并增加负载扰动信号。试验结果表明,采用的在线PSO优化方法可以在单采样周期内完成单次迭代任务,完全可以达到实时在线优化计算的目标,实现了无传感器控制。该方法较传统的PI控制具有更好的鲁棒性和动态特性。

In the sensorless control of permanent magnet synchronous motor(PMSM),there is error caused by multi-parameter interference in the estimation accuracy of the velocity value,especially in the case of low speed operation of the rotor.A novel online particle swarm optimization method is proposed to design PI controller parameters and feedback loop current in the forward channel and the speed PI controller parameters.Stator flux and rotor speed estimations are derived by sliding mode observer,and the stator resistance variation is compensated with a speed correction term which is derived from the estimation error of d-axis current.Due to the particle swarm optimization method has been used,the variation of the six controller parameters and the stator resistance can be designed at the same time,avoiding cross-coupling and reducing the error of speed estimation.In experiments,the hardware DSPACE1103 is used;the command signal adopts the first-order step signal,and the load disturbance signal is added.The results show that the online particle swarm optimization(PSO)method can complete a single iteration task in a single sampling period,which can achieve the goal of real-time online optimization calculation and sensorless control,with better robustness and dynamic characteristics than traditional PI control.