Resistance dissymmetry localization method based on vector space decomposition approach for six-phase induction machines

Fault detection is a necessary requirement to exploit one of the most important multiphase drives advantages: the enhanced fault tolerance without additional hardware. Among open-circuit faults, those in power switches of voltage source converters (VSCs) are the most common ones. Nevertheless, in other cases a poor connection or incipient faults may cause an imbalance in the value of the stator resistances. In this case the phase is not open but this anomalous situation requires detection and eventual isolation. This phenomenon is usually called resistance dissymmetry and it leads to current imbalance that in turn reduces efficiency and produces overheating. Moreover, the variation of the machine parameter is an important issue in those control methods based on the machine model, as it is the case in model predictive control (MPC). For this reason, this work presents a resistance dissymmetry detection method based on the vector space decomposition (VSD), implemented together with a MPC-based six-phase induction motor drive. Simulation results are employed to validate the goodness of the proposed method.