High-precision torque control of inverter-fed induction machines with instantaneous phase voltage sensing

Aachen / Institut für Stromrichtertechnik und Elektrische Antriebe (ISEA) (2018, 2019) [Book, Dissertation / PhD Thesis]

Page(s): 1 Online-Ressource (x, 221 Seiten) : Illustrationen


Model inaccuracies, parameter variations and parasitic influences of the voltage source inverter can cause significant torque error and detuning in field-oriented induction machine drives. This thesis describes an analytical minimization approach for the sensitivity of flux-linkage observers to parameter variations, measurement feedback deviation and higher-order harmonics. An extended model of the induction machine, which takes into account magnetic saturation, iron losses, rotor deep-bar effect, and temperature drift of the rotor and stator resistances is introduced. Based on an analytical sensitivity analysis, the flux-linkage observer is designed with respect to minimal sensitivity to uncertain parameters. The low-speed accuracy is improved by an instantaneous phase voltage sensing circuit integrated with the bottom gate driver of each inverter leg. The method is based on oversampling and digital integration. A high-bandwidth low-pass filter is introduced and optimized to minimize the instantaneous measuring error. A novel sensor-offset compensation method is developed, which allows independent offset calibration for all phase voltage and current sensors during operation. The method is fully decoupled from the control algorithm. Another important aspect is the dynamics of efficiency-enhanced IM drives. A model-predictive approach for optimal dynamic current sharing of the flux-producing and torque-producing current is proposed. The results are transferred into a lookup table for implementation on a low-cost microcontroller.



Schubert, Michael


de Doncker, Rik W.
Mertens, Axel


  • REPORT NUMBER: RWTH-2018-231364