Research On Identification Method Of Moment Of Inertia Of Permanent Magnet Synchronous Motor

In recent years,with the rapid development of industrial technology and control theory,the application of permanent magnet synchronous motor(PMSM)has become more and more frequent.In the field of servo control,PMSM plays a more and more important role.PMSM has the advantages of small size,high efficiency and strong overload capacity,which greatly expand the application scope of PMSM.Among the many parameters of permanent magnet synchronous motor,the identification of moment of inertia is particularly important,because the change of moment of inertia will change the control performance of the system,and the response characteristics of the system will decrease with the increase of moment of inertia.For high performance servo control system,identification of moment of inertia is more critical.The identification of moment of inertia can be divided into off-line identification and on-line identification.Off-line identification method is simple and easy to implement,but the recognition accuracy is not high and the identification time is too long.The common methods of on-line identification of moment of inertia include least squares method,model reference adaptive method,etc.On-line identification method can improve the identification accuracy and shorten the identification time,but each method has its shortcomings and needs to be improved.In this thesis,the basic knowledge of permanent magnet synchronous motor(PMSM)is introduced,including its mathematical model and control strategy.The least squares method,gradient correction algorithm and model reference adaptive algorithm are analyzed.The principles of these three methods are introduced,and the three methods are used to identify the moment of inertia.Then the system simulation model is built,the identification results are analyzed,and the characteristics of each identification method are summarized.A test platform is built to identify the moment of inertia during the actual operation of the motor.Finally,the application of moment of inertia in PI parameter self-tuning is studied.The recursive least squares method is improved by introducing forgetting factor to solve the problem of data saturation,so that the identification effect is not affected by the range of moment of inertia,at the same time,the convergence time of the identification results can be reduced and the convergence speed can be accelerated.The model reference adaptive algorithm is improved by adding variable adaptive gain coefficient,and then the identification of moment of inertia is carried out by combiningrecursive average filtering algorithm.The problem of the conflict between convergence speed and fluctuation amplitude of the identification results is solved,and the identification results with fast convergence speed,small fluctuation and strong anti-interference are obtained.By using the moment of inertia to change the parameters of PI regulator in real time,the problem that PI regulator can not be self-tuned is solved,the overshoot produced in the starting stage of motor is reduced,and the dynamic performance of the system is improved.