Model Structure Determination Of Motor Drive System Based On Two-mass Parameter Identification

According to the magnit ude of transmission stiffness,the motor drive system can be divided into one-mass rigid system and two-mass elastic system.Because of the existence of elastic connection device,the two-mass system will cause the phenomenon of mechanical resonance.Introducing the state feedback into the system is an effective and direct method to suppress the mechanical resonance.However,the setting of the state feedback parameter s depends on the system mechanical parameters.Therefore,accurate identification of syste m parameters is particularly important.At present,both domestic and overseas,parameter identification is based on a known system.For unknown systems,it is important to determine the structure type and then use the corresponding identification algorith m to improve the identification accuracy and the system control performance.In this paper,the recursive least squares algorithm with forgetting factor is used to identify the parameters of two-mass system.Taking the identification value of motor inertia as a reference index,a method for judging the model structure of motor drive system is put forward.Specific research contents are as follows :Firstly,the two-mass elastic system model is established and the mechanical resonance in the stage of speed op en-loop and closed-loop is analyzed theoretically.At the same time,the simulation model is built in MATLAB.By observing the motor speed,load speed,electromagnetic torque and other signals,the phenomenon of the mechanical resonance at different stages is verified.The simulation proves that the Anti-Resonance Frequency(ARF)resonance is the main factor in the stage of speed closed-loop,and the Natural Torsional Frequency(N TF)resonance is the main factor in the stage of speed open-loop.Secondly,based on the background of suppressing resonance,a method of two-mass parameter identification is proposed.The six mechanical parameters of the system can be identified by collecting the motor speed and electromagnetic torque signals and using the recursive least square algorithm with forgetting factor.Changing the given speed of the motor,the identification accuracy of the algorithm is improved.Thirdly,a one-mass identification algorithm is presented.Furthermore,this algorithm and two-mass identification algorithm are used to identify the parameters of different structures respectively.The simulation results showed that corresponding algorithm gets the best effect.Meanwhile,the accurate identification of parameters is very important for optimizing system control and avoiding mechanical resonance.Therefore,as for unknown systems,it is necessary to evaluate the system structure first,and then select the corresponding algorithm to identify the parameters.O n the basis of the two-mass identification,the error between the motor inertia moment identification and the actual value is taken as the index evaluation,then a decision of motor drive system model structure is proposed.Finally,a rigid system and an elastic system experiment platform are built.The feasibility of identification algorithm and model structure determination method can be validated on the platform.The motor speed and electromagnetic torque of the platform are collected by data acquisit ion card,and then imported to the parameter identification program to realize the off-line identification.The two parameter identification algorithms are validated on two experimental platforms,and the model structure determination method of the motor drive system is also proved to be effective.