Research on Inertia Identification and Resonant Reduction Methods for Servo System
|School||Harbin Institute of Technology|
|Course||Mechanical and Electronic Engineering|
|Keywords||servo system inertia identification mechanical resonance IIR filter acceleration feedback|
In servo motor drive applications, variation of load inertia can cause parameters mismatch of classical controller, and degrade drive’s performance. Compliantly-coupled load in servo system always brings up problem of mechanical resonance, which would decrease stability margin and bandwidth. So, it requires that high-performance servo system have ability of inertia identification and resonant depression.In this thesis, methods of inertia identification and resonant depression are researched deeply through theoretical analyses, combined with simulations and experiments.First, MRA and least-squares (including several derivative algorithms) inertia identification methods are researched theoretically. Then simulation and experimental study based on the SIMULINK model and test-bed of PMSM servo system is completed respectively.The results show that MRA method can not give consideration to convergence rate and precision and is suitable to the off-line identification of slowly varying load inertia for one-mass system .While least-squares method has characteristic of fast convergence, high precision and small vibration. It is suitable to the on-line identification of fast changing load inertia for one-mass servo system which has relatively small disturbance load torque.Also, FFT inertia identification method is put forward. This method is based on a simple principle and has high precision. Identification experiment shows that it has great practical meaning for inertia identification for two-mass servo system.Then, aiming at the resonant phenomenon brought by compliantly-coupled load, IIR filter (including low-pass filter, notch filter & bi-quad filter) resonant depression method is studied through theoretical analyses and experiments. The results show that in the aspect of reliability, low-pass filter is superior to notch filter and the latter is superior to bi-quad filter. While in the aspect of the ability to improve bandwidth, notch filter is superior to bi-quad filter and the latter is superior to low-pass filter. The conclusion of research can provide guidance for the selection of resonant depression filter in servo system. At last, SIMULINK model of PMSM servo system with compliantly-coupled load is founded and acceleration feedback method is researched for its resonant depression ability through theoretical analysis, combined with simulation. Then using this method, resonant depression of test-bed of PMSM servo system with compliantly-coupled load is elementarily realized.The related results in this thesis can be applied to high-performance servo system for the further improvement of its performance index.