Condition monitoring analysis of synchronous generator based on an adaptive technique

This paper presents an adaptive technique for detecting internal faults in synchronous generators. The adaptive technique is works based on the Bat Algorithm (BA) and Recurrent Neural Network (RNN). The BA is a nature-inspired metaheuristic optimization algorithm, which is introduced for solving engineering optimization tasks. The RNN is a name for neural networks containing recurrent structures which is exhibited by machines. In the paper, the BA and RNN technique is utilized for classifying the internal faults of the synchronous generator. But the training process of RNN, the BA algorithm is utilized. Initially, the proposed technique analyzes the voltage and current of synchronous generator. Based on these variations of inputs, the internal faults are diagnosed and classified. The proposed powerful technique for fault detecting and classification is implemented in MATLAB/Simulink working platform and the performances are evaluated. Finally the statistical measures of proposed technique is determined and compared with DWT-ANN, DWT-ANN with GA, DWT-ANN with GSA and DWT-ANN with ALO technique.