Research On Inertia Identification Of Permanent Magnet Synchronous Motor

In order to suppress the influence of inertia disturbance, high performance servo drive system should have the function of inertia identification and controller parameters self-tuning. Among them, identifying the real-time inertia value accurately and rapidly should be solved first as the key technology. Research on inertia identification is mainly expanded in this dissertation.First of all, the mathematical model is expound, on which the vector contrl theory of the system is illustrated with the current control method of id=0. The control block diagram of the system is obtained and the simulation model of PMSM AC servo system is established under the MATLAB/Simulink environment.Secondly, several off-line inertia identification methods are introduced briefly, and two on-line real-time estimations are focused on. They are parameter estimations based on gradient algorithm and Landau discrete-time recursive algorithm. Research on identification theory of the two algorithms has been studied, and inertia identification formulas of permanent magnet synchronous motor(PMSM) are given through derivation. An parameters adaptive control method based on the principle of pole assignment is proposed. After that, bring the two identification algorithms to the MATLAB model of servo system, and do some simulation research on system identification with constant inertia and variable inertia. Though the simulatino results, the correctness of the two algorithms are verified, and the influence of identification parameters's changes to identification performance is discussed.Then, software design and debugging are completed on the PMSM servo system hardware platform which takes the Freescalse's high performance DSP chip MC56F8346 as the core. And based on it, gradient and Landau algorithms estimation procedures are programed to carry out constant inertia, variable inertia and parameter self-tuning experiments respectively. The experiments show that, gradient and Landau algorithms are both effective to identify the PMSM inertia; the shorter the period and the bigger the given range of given velocity, the better the real time property of inertia identification is; gradient algorithm is higher than Landau for the identification precision, but longer than Landau for convergence time, and they are for different occasions; adaptive control based on gradient algorithm inertia identification improves the velocity's dynamic performance and ineria identification performance significantly, and enhances the system's robustness.Finally, compared experimental results with simulation results, we know the hardware factors which affect the inertia identification performance of PMSM. The improved design scheme is proposed and illustrated in detail.