| As one of the most important control and actuators in traditional industry and high-tech fields,AC servo system has been widely used in CNC machine tools,robots,aerospace and military industries.In recent years,with the rapid development of intelligent manufacturing technology,the control performance of the servo system has been put forward higher requirements.For example,when the transmission devices are elastically connected,mechanical resonance problems are prone to occur.In view of the above problems,in order to obtain higher dynamic performance and stability,this paper will launch research on the model identification and vibration suppression algorithms of AC servo systems.This paper starts from the vector control structure of Permanent Magnet Synchronous Motor,and establishes the mathematical model of the double inertia elastic servo system through the mechanism analysis method.Then combined with theoretical derivation and simulation results,the mechanical resonance phenomenon of the servo system is summarized.Finally,this paper makes classification based on the relationship between resonance characteristics and system bandwidth.If the anti-resonance peak is the main factor restricting the bandwidth improvement of sytem,it is defined as the mechanical resonance problem of Type I.On the contrary,if it is the resonance peak,it is defined as Type II resonance problem.Thus,this classification method lays a foundation for the design of vibration suppression strategy.In order to implement the resonance suppression strategy more effectively,this paper uses the Pseudo-Random Sequence as the excitation signal for servo system identification,and designs an identification algorithm based on the ARMA parameter model.Then the Least Square Method is used to calculate the high order fitting model of system.Furthermore,this paper also proposes a model reduction method based on the equilibrium theory,which extracts the dominant mode of the system by comparing the size of Hankel singular values.Then,the evaluation function is established according to the fitting residuals in the time domain and frequency domain,and a more accurate low-order models is obtained by using the Balanced Residuals Method.Aiming at the Type I mechanical resonance problem of the servo system,this paper proposes LQR control strategy based on the principle of Internal model theory.First,the dominant pole theory is used to achieve the second order reduction of the model.Then,the paper uses this model as the control object to design the LQ evaluation function,and introduces the Internal model control law.Finally,the optimal controller of the servo system is obtained by solving the Riccati equation,thereby improving the dynamic performance and stability of the system.For the mechanical resonance phenomenon of Type II,this paper proposes a vibration suppression strategy based on PI controller and Notch filter.In view of the special case that the open-loop and closed-loop resonant frequencies are consistent,an adapative notch filter is designed based on the online frequency identification algorithm of Second-Order Generalized Integrator Frequency-Locked Loop and the parameter setting rate.It not only guarantees the effectiveness of resonance suppression,but also improves the efficiency of vibration suppression strategies.Synthesizing the theoretical analysis and simulation results of the full text,this paper completely establishes suppression strategy for the mechanical resonance of the AC servo system.The software and hardware platform of the AC servo system based on DSP + FPGA is designed and built,and comparative experiments are carried out on three different types of mechanical resonance platforms.The experimental results prove the effectiveness of the model identification algorithm and resonance suppression strategy in this paper. |
