Research On Sensorless Position Detection Method Of Permanent Magnet Synchronous Motor

In recent years,the research about controlling Permanent magnet synchronous motor without position sensor has became a hot button in the filed of electric control.Using this methods and making the motor work at medium and high speed,we can gain control accurately.But there are still some problems about rotor accurate detection in starting when the motor is static and low speed control.The article will carry out research aiming at above problems.Firstly,the article will analyze the permanent magnet synchronous motor in detail and establish mathematical models and it will be the basis of the follow-up research about Sensorless Detection Technology of Permanent Magnet Synchronous Motor.Secondly,aiming at the saturation convex class charactersrics of surface mount permanent synchronous motor,this article introduces an arithmetic about the position observer based on pole assignment.The arithmetic achieves the accurate detection about the position of rotor when the motor operates at low speed,improving the anti-interference capability of the system.To solve the problem about long detection time and complex stucture in the iniatial position detction of motor rotor,the article also designs a device to test the rotor initial position based on BP neural network.The method is insentive to electric machine parameter and can achieve the test about the initial position without decoupling calculation.Then,the artical also does simulation experiments about above methods on the platform of Matlab/Simulink.The experiments indicate that the methods of improved observer can improve the system performance effectively.When the motor oprates in low speed,it can make the maximum estimated rotor position error 0.53°.And using this method,we can detect the initial position of the rotor quickly and effectively,and it can make the maximum detection error 2.36° through simulation,and make the average error 1.17°.Last,design the software and hardware of the system and put this method to the test.The results indicate that the overall system can achieve position detection about the sensorless accurately.