永磁同步电机的转速自适应控制算法研究

发布时间：2018-04-01 00:09

本文选题：PMSM　切入点：粘滞摩擦系数　出处：《江苏大学》2017年硕士论文

【摘要】：永磁同步电机(PMSM)由于其高可靠运行性能、高功率密度、控制较简单等特点,广泛应用于高性能伺服控制领域,如工业机器人。而永磁伺服系统经常存在频繁加减速等复杂工况以及多变载荷的负载条件,伺服系统的转动惯量和负载转矩将发生显著变化,而控制器参数不变则无法满足控制精度的要求。因此,需要自适应辨识参数以提高控制的效果。传统的模型参考自适应算法在辨识机械参数时忽略了摩擦系数的影响,不影响摩擦较小的永磁伺服系统惯量和负载辨识结果的准确性,但对于大粘滞摩擦系统,会导致转动惯量和负载辨识结果波动较大、不易稳定。由此,本文提出了一种考虑粘滞摩擦系数的模型参考自适应辨识算法。利用该算法可以同时得到不同转速幅值和频率以及不同惯量初值条件下的转动惯量、负载转矩和摩擦系数的准确辨识值,提高了对存在较大摩擦系统机械参数辨识的准确性。基于反步法设计的控制器实质上是由kp调节器以及各补偿项构成,当各项参数准确时能够精确补偿,但对于永磁伺服系统来说,参数是未知的或者参数随运行环境改变,此时反步控制器无法精确补偿,使系统误差较大甚至失稳。因此,本文对转速环提出了自适应反步控制,将辨识结果用于实时修正控制器参数以实现精确补偿,提高系统鲁棒性。传统的未辨识摩擦系数的自适应反步控制器会导致大粘滞摩擦系统的辨识结果波动、无法稳定收敛于真实值,从而设计了实时辨识摩擦系数的自适应反步控制器,可以得到不同转速、不同负载和不同惯量初始值下的惯量、负载和摩擦系数的准确辨识值,从而实现转速环的自适应反步控制,使系统具有较好的速度跟踪性能。本文提出的转速自适应控制算法考虑了摩擦系数的影响,能得到机械参数的准确辨识值,并实时修正控制器参数,使得系统转速跟踪性能良好,并基于dSPACE公司DS1103系统实验平台对所提算法进行实验验证。控制方法结构简单、易于实现,适用范围更广,对于电流环参数的自适应辨识具有指导意义。
[Abstract]:PMSM (permanent Magnet synchronous Motor) is widely used in the field of high performance servo control because of its high reliability, high power density, simple control and so on. For example, for industrial robots, permanent magnet servo systems often have complex working conditions such as frequent acceleration and deceleration, as well as load conditions with variable loads. The moment of inertia and load torque of servo system will change significantly. Therefore, adaptive identification parameters are needed to improve the control effect. The traditional model reference adaptive algorithm neglects the influence of friction coefficient when identifying mechanical parameters. The inertia of permanent magnet servo system with small friction and the accuracy of load identification results are not affected. However, for large viscous friction system, the results of inertia and load identification will fluctuate greatly and are not easy to be stabilized. In this paper, a model reference adaptive identification algorithm considering viscous friction coefficient is proposed. By using this algorithm, the moment of inertia can be obtained under different rotational speed amplitude and frequency, as well as under different inertia initial values at the same time. The accurate identification value of load torque and friction coefficient improves the accuracy of mechanical parameter identification for friction system with large size. The controller based on backstepping method is essentially composed of kp regulator and each compensation item. When the parameters are accurate, the parameters can be compensated accurately, but for the permanent magnet servo system, the parameters are unknown or the parameters change with the running environment, so the backstepping controller can not compensate accurately, which makes the system error large and even unstable. In this paper, an adaptive backstepping control is proposed for the rotational speed loop. The identification results are used to modify the controller parameters in real time to achieve accurate compensation. The traditional adaptive backstepping controller with unidentified friction coefficient can cause the identification results of large viscous friction systems to fluctuate and can not converge to the real value stably. An adaptive backstepping controller is designed to identify the friction coefficient in real time. The accurate identification values of inertia, load and friction coefficient can be obtained under different rotational speed, different load and different inertia initial value. In order to realize the adaptive backstepping control of the rotational speed loop and make the system have better speed tracking performance, the speed adaptive control algorithm proposed in this paper takes into account the influence of friction coefficient, and the accurate identification value of the mechanical parameters can be obtained. The controller parameters are corrected in real time to make the system speed tracking performance better, and the proposed algorithm is verified experimentally based on the DS1103 system experiment platform of dSPACE Company. The control method is simple, easy to realize and has a wider range of application. It is of guiding significance for adaptive identification of current loop parameters.
【学位授予单位】：江苏大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP273;TM341

【参考文献】