Research On Rotor Position Detection Method Of Sensorless Servo Synchronous Motor

The method and realization of position sensorless detection of motor rotor position have received extensive attention from the engineering community,the rotor position detection accuracy of the servo synchronous motor directly determines whether the motor can start smoothly and run stably.However,the rotor position detection error is more obvious when the motor is at a low speed or at a standstill.Therefore,this article has carried out research work on this problem.This paper takes permanent magnet synchronous motor as the main research object,First,in-depth study and understanding of the nonlinear characteristics of permanent magnet synchronous motors,Establish mathematical model and vector control system of permanent magnet synchronous motor.Secondly,it analyzes the BP(Back Propagation)neural network algorithm,using an improved BP neural network algorithm to detect the rotor position of a permanent magnet synchronous motor,to make up for the shortcomings of traditional methods and the large detection errors of low-speed rotors.Then build a permanent magnet synchronous motor rotor position detection model based on BP neural network,because the rotor position detection error is large during the start-up process and commutation,improve and optimize the model.Extract motor parameter data based on SGMJV-08AAA61 permanent magnet synchronous motor control platform,provide data support for the established neural network model,conduct network training.On the basis of,for the single training sample data,it leads to the problem that the dynamic response ability of the rotor position detection system based on neural network is reduced.This article adopts the idea of updating training data,analyze the influence of parameter changes of permanent magnet synchronous motors on the detection of motor rotor position during operation in the full speed range,According to the motor parameters of the permanent magnet synchronous motor operating under different conditions as the training sample data,train the network offline,until the error meets the requirements,in order to realize the error correction of the training sample data,further improve the accuracy of rotor position detection.Finally,perform simulation and experimental verification,the BP neural network used in this paper can realize the detection of the rotor position of the permanent magnet synchronous motor with high accuracy.Suitable for permanent magnet synchronous motors running in the full speed range,solve the problem of large rotor detection error of permanent magnet synchrono us motor at low speed,it provides an effective method for detecting the position of the motor rotor without a position sensor.