Research On Mechanical Resonance Suppression And Control Methods For Two-mass Systems

Servo system is widely used in industrial production because of its excellent characteristics.In the servo driving device,the load is driven by the motor through the flexible link,which is generally called two-inertia system.The flexible link of two-inertia system is easy to produce mechanical resonance,destroying the stability of the system.In addition,friction,external disturbance,parameter uncertainty and unknown dynamics also cause great difficulties to the control of two-inertia system.Therefore,it is of great significance and research value about how to overcome the influence of mechanical resonance and various disturbances and improve the control accuracy of the servo system.In this thesis,the resonance suppression,disturbance compensation and control of a two-inertia system affected by mechanical resonance and various nonlinear disturbances are discussed.The problems of nonsingular terminal sliding mode control,fixed time integral sliding mode control and improved neural network dynamic surface control are studied.The main research contents are as follows:(1)Different methods are proposed to suppress the resonance in different frequency bands for the mechanical resonance in the two-inertia system.The resonance model of twoinertia system is established;In the low frequency band,the load torque is estimated by linear extended state observer,and a new IP regulator with anti proportional and anti differential impact is designed to suppress resonance based on load torque feedback;In the high frequency band,a vibration suppression strategy based on enhanced notch filter is proposed.The identification method of resonance frequency and the parameter tuning method of filter are given.The simulation results show the mechanical resonance is effectively suppressed by the two methods.(2)A fast terminal sliding mode control strategy based on finite time extended state observer is designed for the two-inertia system whose position information of load side is difficult to measure.Firstly,a finite time extended state observer is used to estimate the motor speed and total disturbance,and the chattering of the nonsingular fast terminal sliding mode is reduced by feedforward compensation of the total disturbance,and the control performance is guaranteed.Then,a nonsingular fast terminal sliding mode controller is designed to improve the response speed and tracking accuracy.Finally,the finite time stability of the system is proved based on Lyapunov theory.The effectiveness of proposed control scheme is verified by simulation results.(3)A fixed time integral sliding mode control method based on disturbance observer is studied for the two-inertia system with external disturbances on the both sides.The fixed time convergence of the load speed tracking error is realized.Firstly,the fixed time disturbance observer is used to estimate the disturbance and its derivative,and then the model is transformed into standard integral series form.Then,a fixed time integral sliding surface is designed by using a fixed time high-order regulator.Finally,based on the fixed time reaching rate,a new integral sliding mode controller with global fixed time convergence is designed.Based on Lyapunov theory,the fixed time stability of the system is proved.The simulation results show that,compared with the fixed time convergence method only considering sliding mode variables and observer estimation error,the fixed time convergence of tracking error is achieved by the fixed time integral sliding mode surface.(4)For a two-inertia system with unknown nonlinear friction on both sides,an adaptive dynamic surface control method based on radial basis function neural network is proposed.Firstly,a continuously differentiable friction model is introduced to establish the mathematical model.Then,in order to improve the transient performance of the filter,the sliding mode differentiator is used to solve the problem of insufficient filtering accuracy of the first-order filter.At the same time,radial basis function neural network is designed to approximate unknown nonlinear function.Finally,an improved neural network dynamic surface controller is designed based on the approximation value of the function.Lyapunov method is used to judge the stability of the proposed algorithm.The simulation results show that the "flat top" problem of position tracking and the "dead zone" problem of speed tracking was solved by the method,caused by nonlinear friction when the load is tracking at low speed,and improves the dynamic performance and robustness of the system.