基于多新息的永磁同步电机参数辨识研究

发布时间：2018-08-02 10:51

【摘要】：永磁同步电机结构简单,效率高,性能优越,并且具有良好的稳定性,也正因为如此,永磁同步电机广泛应用于风力发电、机器人、工业生产等各个领域。永磁同步电机因其内部转子通过嵌入一块高性能的永磁体产生固定的磁场,省略了励磁绕组,降低了电机的无功功率,从而提高了电机的功率因数,增加了电机的效率。同时简单的结构使得电机内部不易损坏,提高了电机的稳定性,降低了后期维修的成本。合理的控制电机的运行,充分发挥电机的性能,不仅可以增加产能,同时还可以节约能源。随着技术的进步,已经发展出多种控制方法,但是无论是哪一种控制方法,控制器的设计都离不开精确的电机参数。电机参数的获得方法大致分为离线参数测量与在线参数辨识两类。其中通过离线参数测量方法例如电机堵转实验,激励响应等方法所获得的电机参数都是电机处于静止状态下电机参数,无法体现电机运行过程中的真实电机参数。针对这一缺点,在线参数辨识则具有一定的优势,具体在于辨识结果可以跟踪电机参数的变化,进而控制方法中控制器参数可以随着电机参数的变化实时调整,提高电机控制方法的精确性,提高了电机的运行性能。众多的参数辨识算法由于其侧重点的不同,各自具有不同的优缺点,可以用收敛速度与收敛精度来表示算法的性能。多新息算法的优势在于可以通过改变新息长度的大小改变每次计算所需要的数据量的多少,从而增加算法对数据的利用效率。针对某些算法辨识过程中数据的利用率较低导致算法的性能不太理想的情况,将多新息算法对原算法进行改进,在保留算法原有优势的基础上增加数据的利用效率,提高算法的性能,并通过引入遗忘因子的概念,对改进后的算法进一步优化。本文首先介绍了永磁同步电机的数学模型以及辨识模型,在此基础上,运用已经非常成熟的矢量控制方法搭建永磁同步电机的矢量控制模型,通过采集电机运行时的参数,采用随机梯度类、最小一乘类、正交投影类等算法辨识电机参数,并根据算法的优缺点用多新息算法与遗忘因子进行优化,通过辨识结果的对比验证算法的有效性。
[Abstract]:Permanent magnet synchronous motor (PMSM) has a simple structure, high efficiency, superior performance and good stability. Because of this, PMSM is widely used in wind power generation, robot, industrial production and other fields. The permanent magnet synchronous motor (PMSM) produces a fixed magnetic field by embedding a high performance permanent magnet into its inner rotor, which omits the excitation winding, reduces the reactive power of the motor, improves the power factor of the motor and increases the efficiency of the motor. At the same time, the simple structure makes the motor not easy to damage, improve the stability of the motor, and reduce the cost of later maintenance. It can not only increase the production capacity but also save energy by controlling the motor operation reasonably and giving full play to the performance of the motor. With the development of technology, a variety of control methods have been developed, but no matter which control method, the controller design can not be separated from accurate motor parameters. The methods of obtaining motor parameters can be divided into two types: offline parameter measurement and on-line parameter identification. The parameters obtained by off-line parameter measurement methods such as motor shutoff experiment and excitation response are all motor parameters in static state, which can not reflect the real motor parameters in the process of motor operation. In order to overcome this shortcoming, on-line parameter identification has some advantages, which is that the identification result can track the change of motor parameters, and then the controller parameters can be adjusted in real time with the change of motor parameters. The accuracy of the motor control method is improved, and the performance of the motor is improved. Many parameter identification algorithms have different advantages and disadvantages due to their different emphases. The convergence rate and precision can be used to express the performance of the algorithm. The advantage of multi-innovation algorithm is that it can increase the efficiency of data utilization by changing the size of innovation length and changing the amount of data needed for each calculation. In view of the fact that the performance of some algorithms is not satisfactory due to the low utilization of data in the process of identification, this paper improves the original algorithm to increase the efficiency of data utilization on the basis of preserving the original advantages of the algorithm. Firstly, the mathematical model and identification model of PMSM are introduced in this paper. On this basis, the vector control model of PMSM is built by using the very mature vector control method, and the parameters of the PMSM are collected when the PMSM is running. The motor parameters are identified by random gradient class, least one multiplication class and orthogonal projection class. The algorithm is optimized by multi-innovation algorithm and forgetting factor according to the advantages and disadvantages of the algorithm. The validity of the algorithm is verified by comparing the identification results.
【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

【参考文献】