Study On The Parameters Identification And Control Strategies Of Permanent Magnet Synchronous Motor

ABSTRACT: Due to its uncomplicated structure, excellent control performance, small size, low loss and high efficiency, AC PMSM (Alternating Current Permanent Magnet Synchronous Motor) is widely used in the fields of Numerically-controlled machine tool, electric automobile, and robot, etc. Based on the research for its principle of work, its mathematical model and SVPWM (Space Vector Pulse Width Modulation), the author does a lot of theoretical and practical research for the tests as well as identification of PMSM's parameters. Besides, control strategies of PMSM based on the theory of ADRC (active disturbance rejection control) and Gray Theory are also studied deeply. Simulation and experiment results demonstrate the effectiveness of new ways to test and identify parameters, as well as ADRC.A new method to test the several main parameters of PMSM is proposed in this paper. Since the magnet meter cannot be avoided in Voltage Integration method which is used to get Ld (direct-axis inductance) and Lq (quadrature-axis inductance) and considering the conditions of lab and worksite, this paper proposes an alternate method. Steps of this method are as follows. To begin with, data should be gathered by general devices of voltage collecting, such as oscilloscope. Then the discrete voltage signal can be integrated through Numerical-integration method to get PMSM's flux linkage which can be used to compute its inductance. This method is simple and easy to implement. Experiment results have proved the effectiveness of this new method.Since Rs (wingding resistance), Ld and Lq will vary a lot with temperature, identification of these parameters are studied though mathematical analysis and simulation. The author has built a simulation model of PMSM in MATLAB whose parameters can vary online, and identified the parameters using EKF (Extended Kalman) filter & Elman neural network and RLS (Recursive Least Square) method. Despite that both of the two methods can guarantee high identification accuracy, they are unsuitable to identify online and have big compute loads. Based on the above disadvantages, WLS (Windowed Least Square) method is proposed. In this method, the length of the window can be adjusted according to practical conditions. Besides, it has a small computational burden. That is why it can guarantee the identification accuracy and excellent timeliness. And at the same time, the issue of data saturation of RLS can be avoided. Compared with simulation results of the other two methods, WLS's results show that this method is the most effective.To improve the working stability and control performance of PMSM, ADRC is applied to reject disturbances from power supply and torque. Simulation results prove that the control performance of ADRC is better than PID control. Then, based on the ADRC theory, the author combines Optimal control, Neural networks and ADRC method together and proposes two algorithms named Self-optimizing ADRC method and ADRC based on neural networks, respectively. Compared with ADRC method, the Self-optimizing ADRC method has better control performance and least control energy. To eliminate the influence of varying disturbance like external load on the control performance of rotation speed, neural network is bedded in ADRC, which can approximately narrow the vary range of original system's parameters, and at the same time improve ADRC's ability of state estimation so that it is possible for ESO (Extended State Observer) to work with lighter load. What's more, by doing this, controller's computing and response speed will be improved. It is proved by study results that both of the methods can provide better control effect.In order to identify and control parameters of PMSM online, the author builds a control model of PMSM and simulates change process of Rs, Ld and Lq with temperature varying from 20℃to 80℃. Then the parameters are identified online using WLS method while ADRC is applied in speed loop. Simulation results demonstrate that the performance of ADRC system using online estimation strategy is better than that of the system using only PID method.In this paper, Gray Compensation PID algorithm and Gray Predict PID Control algorithm are applied in PMSM Servo System for the first time to control PMSM with less data, and results show that this control strategy has more satisfied robustness and adaptive capacity.A set of AC PMSM Servo System has been developed to realize the SVPWM control based on three closed loop. The load test and no-load test prove that ADRC has the advantages of fast response, no over modulation, and strong ability of rejecting disturbance, compared with classical PID control.In addition, since parameters of IGBT will change in work site, the author has developed a new kind of circuit which can modulate the turn-on and turn-off time of IGBT. Moreover, this driving circuit is also able to protect itself when power supply is under-voltage. And this kind of circuit is proved to be effective.