Position estimation and performance prediction for permanent-magnet motor drives

Abstract

This thesis presents a theoretical and experimental development of a novel position estimator, a simulation model, and an analytical solution for brushless PM motor drive. The operation of the drive, the position estimation model of the test motor, development of hardware, and basic operation of inverter are discussed. Starting with the well-known continuous-time model of brushless PM motor, a sampled-data model is developed that is suitable for th6, application of real-time position estimator. An analytical methodo f calculating the steady-stateb ehaviouro f the brushlessP M motor for 1200in verter operation is presentedT. he analysisa ssumesth at the machinea ir gap is free of saliency effects, and has sinusoidal back EMF. The analytical solution is derived for 60" electrical of the whole period. By experimental results, it is shown that the method of analysis is adequate to predict Ihe motor's performance for typical operating points including phase advance and phase delay operation. C) I A computer simulation model for prediction of the performance of brushless PM moto rs is presented. The model is formulated entirely in the natural abc frame of reference, which allows direct comparison of the simulation and corresponding experimental results. The equations and diagrams are put into a convenient form for the simulation and future developments and library modules. The simulation model and corresponding experimental data of the brushless PM motor drive is given. The thesis describes a modem solution to real-time rotor position estimation, which has been subject to intense research activity for the last 15 years. The implemented new algorithm for shaft position sensorless operation of PM motors is based on the flux linkage and line current estimation. The position estimation algorithm has also been verified by both off-line and on-line experiments (accomplished by a DSP, TMS320C30), and a wide range of steady-statea nd transient results have been 0gi0v en including starting from rest. The position estimation method effectively moves the position measurement point in the drive from the mechanical side to the motor's terminals. As well as eliminating the mechanical shaft position sensor, the investigated method can be used for high performance torque control of brushless PM motors. The thesis demonstrates that, in contrast to many other "sensorless" schemes, the new position estimation method is able to work effectively over the full operating range of the drive, and is applicable to a wide range of motor/converter types. Since the hardware is straightforward, only the new position estimation algorithm differentiates a system. Therefore, if a DSP control system is already implemented in the drive, the position estimator can be implemented at low cost.