Design and Analysis of Current Sensor Fault Detection Mechanisms for PMSM Drives Based on Neural Networks

摘要

This paper describes a vector-controlled Permanent Magnet Synchronous Motor (PMSM)drive system with the current sensor fault detection mechanism. In general, the control structure isbased on the well-known Field Oriented Control (FOC) algorithm. The structure is equipped with anadditional algorithm for current sensor fault detection based on a neural network. The presentedcontrol structure is able to detect typical current sensor faults, such as lack of signal, intermittentsignal, variable gain and signal noise. The application of the NN detector guarantees a faster detectionof the sensor fault than classical detectors based on algorithmic methods or logical systems. Thiswork focuses on presenting the methodology of designing detectors and their analysis, based solelyon simulation analysis. The simulation results, conducted in the Matlab/Simulink environment, arepresented for the above-mentioned faults in phase A and phase B for different speed conditions.