On-line parameter estimation, current regulation, and self-sensing for IPM synchronous machine drives

This thesis presents several new contributions for improved control of interior permanent magnet synchronous machines (IPMSMs). A new on-line adaptive parameter estimation method is introduced first. The estimated parameters from on-line self-tuning are utilized for decoupling of the unwanted, reference frame-dependent, cross-coupling terms and for improving efficiency. A virtual translation technique is introduced to further improve current regulator performance for salient-pole ac machines. These current regulators with a virtually translated pole are analyzed and experimentally verified. Novel additions to self-sensing techniques are presented for operation from zero speed through high speed including initial rotor position with magnet polarity. Accurate IPM machine models including saturation for rotating and pulsating voltage carrier signals are introduced to detect the magnet polarity for initial position estimation at start-up. Effective signal processing techniques for both types of carrier signal injection methods are developed to identify the magnet polarity. The saturation-induced saliencies for both carrier injections are measured and analyzed in terms of their effect on motion state estimate at zero-low speeds. A saliency-based EMF estimate from a state filter is combined with a motion state observer including feedforward for zero-phase lag position estimation at high speeds.