基于高性能磁链算法的永磁同步电动机无位置传感器控制

准确的磁链辨识是基于磁链观测法永磁同步电动机无位置传感器控制的关键,传统电压型磁链辨识算法消除了纯积分存在漂移问题,对圆形磁场的磁链可以准确辨识,但对实际中电机内普遍存在的椭圆形磁场的磁链辨识时存在稳态幅值和相位误差。分析了椭圆形磁场时磁链矢量与反电动势矢量的空间关系。对传统算法辨识椭圆形磁链的误差进行了分析,提出一种改进的新型磁链算法。改进后的新算法运算量与传统算法相当,但对圆形与椭圆形磁场的磁链均能准确辨识。将新算法应用于一台表贴式永磁同步电动机无位置传感器控制系统的转子位置估算,仿真证明,在稳态及转速突变时新算法均具有良好的位置和转速跟踪效果。利用数字信号处理器TMS320F2812DSP对其进行了数字化实现。实验结果表明,在中高速范围内,调速系统在稳态与电机转速突变时具有良好的动静态特性,但在低速区会出现系统不稳定。仿真与实验证明了新算法的有效性和实际可应用性。

Position Sensorless Control of Permanent Magnet Synchronous Motor Based on High Performance Flux Estimation Algorithm

Accurate flux estimation is the key to position sensorless control of permanent magnet synchronous motor based on flux observation. Traditional flux estimation algorithms,based on the voltage model,can effectively solve the problems associated with a pure integrator,and can estimate the flux exactly in the round rotating magnetic field situation. But in real motors,the elliptical rotating magnetic field will introduce magnitude and phase errors in the steady state. The relationship between the back electromotive force(EMF)vector and the flux vector in the elliptical rotating magnetic field is analyzed. After an error analysis of the traditional flux estimation algorithms in the elliptical rotating magnetic field,an improved flux estimation algorithm is proposed. The new algorithm can estimate the flux exactly either in the round rotating magnetic field or in the elliptical rotating magnetic field with similar computational complexity to traditional algorithms. The new algorithm has been used for rotor position estimation in position sensorless vector control of a surfacermounted permanent magnet synchronous motor. Simulation results show that,the new algorithm has good speed and position tracking effect either in the steady state or when the motor speed has step changes. The experiment is implemented using the digital signal processor TMS320F2812DSP. Experimental results show that,in the medium and high speed region,the speed control system has good steady state performance as well as dynamic performance when the motor speed has step changes. But it will be unstable in the lowrspeed region. Both simulation and experiment results show the validity and applicability of the new algorithm.