Study On Improved Direct Torque Control Of Permanent Magnet Synchronous Motor Without Position Sensor

Permanent magnet synchronous motor（PMSM）has the advantages of high efficiency,simple structure and reliable operation,and has been widely used in civil electrical appliances,wind power generation,rail transportation and military industry.Direct torque control（DTC）technology is a traditional high performance control technology,which directly controls the electromagnetic torque and flux of the motor in the way of stator flux orientation,and has good speed regulation performance.However,the application of DTC is limited by the unsteady switching frequency and large torque pulsation.In the speed regulation process of PMSM,the actual speed is usually obtained by using mechanical sensors,but the mechanical sensors make the motor structure more complex,and it is difficult for the mechanical sensors to work in extreme environment,which increases the cost of the motor and reduces the reliability of the system.Therefore,the direct torque control technology and sensorless control algorithm of permanent magnet synchronous motor are studied as follows.Firstly,the traditional direct torque control technology is analyzed,and the direct torque control based on SVPWM is proposed for the torque and flux pulsation problems,and according to the analysis,the traditional PMSM direct torque control system and the improved direct torque system simulation model is built,simulation and analysis.The results show that SVPWM can effectively suppress the torque and flux pulsation of traditional DTC.Then,the model reference adaptive（MRAS）sensorless control algorithm is analyzed,and the MRAS algorithm is used to replace the mechanical sensor to estimate the motor speed and rotor position,and the direct torque control simulation model based on the traditional MRAS is built for simulation and analysis.The results show that the PMSM direct torque control based on MRAS can track the motor speed well,but the error convergence speed is slow and is greatly affected by the variation of motor parameters.Finally,the slow response speed and poor robustness of traditional MRAS,a fractional sliding mode reference adaptive observer with feedback correction was proposed,and build an improved MRAS simulation model.The results show the improved sensorless control of MRAS can accelerate the response speed and enhance the robustness of the system.