A Bearing Fault Diagnosis Method Using Multi-Branch Deep Neural Network

Feature extraction from a signal is the most important step in signal-based fault diagnosis. Deep learning or deep neural network (DNN) an effective method to extract features signals. In this paper, novel vibration bearing diagnosis using DNN proposed. First, measured signals are transformed into new data form called multiple-domain image-representation. By transformation, task of transferred image classification. After that, with multi-branch structure proposed handle representation data. The helps multiple domains simultaneously, and lead better feature extraction. Better leads performance effectiveness was verified via experiments conducted actual its comparisons well-established published methods.