Classification of Single Current Sensor Failures in Fault-Tolerant Induction Motor Drive Using Neural Network Approach

In the modern induction motor (IM) drive system, fault-tolerant control (FTC) solution is becoming more and popular. This approach significantly increases security of system. To choose best strategy, fault detection (FD) classification (FC) methods are required. Current sensors (CS) one measuring devices that can be damaged, which in case system with IM precludes correct operation vector structures. Due to need ensure current feedback flux estimators, it necessary immediately compensate for detected damage classify its type. drives, there individual suggestions regarding classifying type CS during operation. article proposes use classical multilayer perceptron (MLP) neural network implement classifier. The online work this classifier was coordinated active FTC structure, contained an algorithm compensation failure two CSs used rotor field-oriented (DRFOC) structure. describes structure method designing (NN-FC). NN-FC verified by simulation tests integrated strategy. These showed high efficiency developed operating post-fault mode after compensating previously ensuring uninterrupted