Artificial neural network-based fault diagnosis of gearbox using empirical mode decomposition from vibration response

This paper presents a gearbox defect diagnosis based on vibration behaviour. In order to record the vibration response under various circumstances, an industrial gearbox was used as the basis for an experimental setup. The signals resulting from gear wear were processed using an empirical mode decomposition for two operating time intervals (zero-hour running time and thirty-hour running time). The first three intrinsic mode functions and the corresponding frequency response were detected. The ten statistical parameters most sensitive to gear wear were selected using an evaluation method based on Euclidean distance. Using the identified features, an artificial neural network (ANN) was trained to track the gearbox for the selected future data set. The neural network received its input from the statistical parameters, and its output was the number of gearbox running hours. To achieve faster convergence, the radial basis function and the backpropagation neural network were compared. The superiority of the proposed strategy is demonstrated by comparing the performance of ANN. For monitoring the condition of industrial gears, the proposed strategy is found to be effective and trustworthy