Research On Control Method For Direct-drive Permanent Magnet Traction Motor System At Low-speed Operation

Direct drive permanent magnet traction synchronous motor has been widely used in modern elevator system with the advantages of high efficiency,highdynamic performance and high-power density.In the direct drive permanent magnet traction motor drive control system,the common resolution incremental encoder with high-cost performance is suitable for measuring position and speed.The weighing sensor installed at the bottom of the car may cause the transmission line disturbance when measuring the load gravity,which leads to the error compensation of starting torque.Thus,the starting control mode without weighing sensor is favored.In order to adapt to the development trend and practical application requirements of the elevator system,it is necessary to study the low-speed operation control strategy of direct drive permanent magnet traction motor drive system which adopts the incremental encoder and weightsensorless starting control mode.At present,the main research directions include: low speed position and speed detection method of traction motor using incremental encoder,zero-servo starting torque control strategy of traction motor without weight sensor,speed fluctuation strategy of traction motor.According to the above application requirements,this dissertation carries out the exploration of the core problems of low-speed operation,and provides key theoretical support and related technical support for the direct drive permanent magnet traction motor drive system.The sampling accuracy of the rotor position signal of the permanent magnet traction motor with incremental encoder is low when running at low speed,which leads to the pulse fluctuation of the rotating speed,thus causing the unbalanced operation of the traction motor.The traditional observer method needs the precise mechanical parameters of the traction motor system,and the mechanical parameters of the traction motor may change greatly during operation.Thus,the observer method is not suitable for the speed measurement of the traction motor.To solve this problem,a position and speed detection method based on adaptive extended window linear regression is proposed in this dissertation,which has the advantage of independent mechanical parameters of the system.Firstly,the moving-window linear regression method is analyzed theoretically,and the main influencing factors of position and speed detection errors are revealed.On this basis,the extended window sampling adjustment method is studied,which adaptively expands the sampling window length according to the speed change,and reduces the error of position and speed detection.Aiming at the problem of speed detection delay during dynamic operation of traction motor,a dynamic adjustment mechanism of sampling window is proposed to improve the speed detection accuracy during dynamic operation.The experimental results show that the position signal sampling accuracy can be significantly improved by using the proposed adaptive extended window linear regression position and speed detection method,which can effectively reduce the speed fluctuation and improve the operating stability at low speed.The traditional structure of the traction motor system needs to install a weighing sensor at the bottom of the car to balance the disturbance torque caused by the car load in the zero-servo stage during the starting process.However,the installation and maintenance of weighing sensor is more complex,and once the sensor measurement error or external signal interference,easy to produce torque error compensation phenomenon.To solve this problem,a zeroservo starting torque control method of the traction motor without car gravity feedback was studied,which reduced the backsliding distance of the traction motor.Firstly,the starting process of direct drive traction motor was modeled to reveal its mechanical characteristics and clarify the load characteristics of uncertainty,nonlinear and strong disturbance.On this basis,active disturbance rejection control methods are analyzed,and the limitations of traditional linear and nonlinear active disturbance rejection control strategies are revealed.Then,a starting torque control strategy based on high gain active disturbance rejection is proposed.By constructing a high gain error convergence function,the estimated error can converge more quickly under different disturbance states,and more accurate compensation can be achieved for the more complex starting torque during the electromagnetic brake release.The high gain nonlinear error feedback control law is adopted to optimize the configuration of parameters and improve the speed of error convergence.The experimental results show that the traction motor drive system based on high gain active disturbance rejection control can achieve smaller backslip distance and backslip time in the zero-servo state,and effectively improve the starting torque control effect.In the operation process of direct drive permanent magnet traction motor,the non-ideal characteristics of the driver will affect the running state of the motor,including the measurement error of the current sensor and the nonlinear effect of the inverter.The non-ideal characteristics will lead to current harmonics and speed fluctuations of the motor,which will affect the operating quality of the traction motor.To solve this problem,a speed fluctuation suppression method based on adaptive maximum speed error gradient descent is proposed.Firstly,a mathematical model of the current vector including measurement errors is established to reveal the harmonic current and speed characteristics caused by the driver’s non-ideal characteristics.Then,taking the frequency division position as the transformation matrix Angle,the current vector is equivalent transformed in the new coordinate system,and the fundamental wave component and harmonic component in the current vector are decoupled.Based on the reconstruction of the coordinate system,the adaptive maximum velocity error gradient descent harmonic suppression strategy is constructed,and the harmonic is eliminated by adaptive adjustment of filter weight and frequency to converge the current vector error.According to the change of running speed,variable step size convergence is used to obtain the ideal filtering effect at different speeds,so as to reduce the negative effects caused by the driver’s non-ideal state.The experimental results show that the speed fluctuation suppression method based on the adaptive maximum speed error gradient descent can effectively reduce the speed fluctuation and ensure the smooth operation of the traction motor.Finally,on the basis of the above research methods,the experimental platform of direct drive permanent magnet traction motor drive system is built.The proposed method was systematically verified by experiments on the experimental platform,and the starting performance,steady state performance and dynamic performance of the traction motor were tested.Compared with the traditional method,the effectiveness of the proposed low speed control method is verified,which lays a foundation for further improving the operation quality of the direct drive permanent magnet motor drive system.