永磁同步电机参数辨识及无传感器控制策略研究

发布时间：2018-04-23 11:31

本文选题：永磁同步电机 + 参数辨识　；参考：《江南大学》2017年硕士论文

【摘要】：永磁同步电机(permanent magnet synchronous motor,PMSM)具有结构简单、性能可靠、转速范围宽等优点。此外,随着永磁材料成本的持续降低,控制技术的不断改进和DSP等微处理器的飞速发展,永磁同步电机在智能制造、工业自动化等高性能伺服控制领域得到广泛应用。然而永磁同步电机是一个高阶耦合的非线性系统,电机定子电阻、电感等参数容易受到温升、磁饱和等因素影响,导致电机可靠性和控制性能的降低。另一方面,PMSM可靠的位置和速度信息是永磁同步电机能够实现高性能控制的关键。传统方法是通过增设硬件传感器来获取这些参数,但由于成本以及维护等因素制约其在永磁同步电机上的应用。因此,参数辨识和无传感器控制成为电机研究领域的热点。论文主要工作如下:1.给出了永磁同步电机的数学模型,并就参数辨识和无传感器控制的一些方法做了对比。接着简要阐述了矢量控制技术,选择了0di=的解耦方式,同时针对电气参数对系统解耦以及转速估算等电机控制性能的影响进行了分析,说明了进行PMSM参数辨识的重要性。2.针对教与学优化算法在求解复杂优化问题存在陷入局部解和收敛精度低等不足,提出了一种改进教与学优化算法对永磁同步电机进行参数辨识。该算法在教学阶段引入辅导教学机制加强教师的教学能力,提高算法收敛速度,在学习阶段,采用科目分步学习提高学员学习效率,并融入反向学习策略进行小概率变异来增加算法跳出局部最优的可能性。基于基准测试函数进行性能测试,验证了上述分析的正确性。通过MATLAB/Simulink仿真表明,此算法能够有效辨识出PMSM电气参数,具有较好的收敛性和可靠性。3.考虑滑模结构的抖振现象,从数学角度论证了滑模控制的抖振产生机制,在此基础上结合静止坐标系下PMSM数学模型,设计了一种改进滑模观测器,来对其转速和位置进行观测。针对高频抖振带来的系统稳定性问题,该观测器引入分段式双曲正切函数作为切换函数和可变滑模增益,以此减弱抖振。考虑到参数不确定性对观测精度影响,在观测过程中引入定子阻值估算环节,实时修正观测器参数,进一步提高观测精度。同时采用分数阶锁相环估算出转子位置和速度信息,从而有效避免了因传统相位补偿带来的计算复杂性。通过在MATLAB/Simulink环境中对不同工况下转速、位置进行仿真,验证了改进滑膜观测器的有效性。
[Abstract]:Permanent magnet synchronous motor (PMSM) has the advantages of simple structure, reliable performance and wide speed range. In addition, with the continuous reduction of the cost of permanent magnet materials, the continuous improvement of control technology and the rapid development of microprocessors such as DSP, PMSM has been widely used in the fields of intelligent manufacturing, industrial automation and other high-performance servo control fields. However, permanent magnet synchronous motor (PMSM) is a high order coupled nonlinear system. The stator resistance and inductance of PMSM are easily affected by temperature rise and magnetic saturation, which leads to the deterioration of reliability and control performance of PMSM. On the other hand, the reliable position and speed information of PMSM is the key for PMSM to achieve high performance control. The traditional method is to obtain these parameters by adding hardware sensors, but its application in PMSM is restricted by cost and maintenance. Therefore, parameter identification and sensorless control have become a hot spot in the field of motor research. The main work of this paper is as follows: 1. The mathematical model of permanent magnet synchronous motor (PMSM) is given, and some methods of parameter identification and sensorless control are compared. Then the vector control technology is briefly described, and the decoupling mode of 0di= is chosen. At the same time, the influence of electrical parameters on the control performance of system decoupling and speed estimation is analyzed, and the importance of PMSM parameter identification is explained. Aiming at the shortcomings of teaching and learning optimization algorithm in solving complex optimization problems, such as local solution and low convergence accuracy, an improved teaching and learning optimization algorithm is proposed to identify the parameters of permanent magnet synchronous motor (PMSM). The algorithm introduces guidance teaching mechanism to strengthen the teachers' teaching ability and improve the convergence speed of the algorithm. In the learning stage, the subjects are used step by step to improve the learning efficiency of the students. In order to increase the probability of the algorithm jumping out of the local optimum, the reverse learning strategy is combined to carry out small probability mutation. The performance test based on benchmark function verifies the correctness of the above analysis. The MATLAB/Simulink simulation shows that the algorithm can effectively identify the electrical parameters of PMSM, and has good convergence and reliability. Considering the chattering phenomenon of sliding mode structure, the chattering mechanism of sliding mode control is proved from the angle of mathematics. On the basis of this, an improved sliding mode observer is designed to observe its rotational speed and position combined with PMSM mathematical model in static coordinate system. Aiming at the stability of the system caused by high frequency buffeting, the observer uses the segmented hyperbolic tangent function as the switching function and the variable sliding mode gain to weaken the chattering. Considering the effect of parameter uncertainty on observation accuracy, the stator resistance estimation is introduced in the observation process, and the observer parameters are revised in real time to further improve the observation accuracy. At the same time, the information of rotor position and velocity is estimated by fractional order phase-locked loop (FPLL), which effectively avoids the computational complexity caused by traditional phase compensation. The effectiveness of the improved synovial observer is verified by simulating the rotational speed and position under different operating conditions in MATLAB/Simulink environment.
【学位授予单位】：江南大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM341;TP273

【参考文献】