Fault diagnosis in rotating machines using machine learning

Bearing fault detection plays an important role in machine maintenance and machine health control. In this study, data obtained from the Case Western University (CWRU) data center website is used. The purpose of using this dataset is to classify faulty and non-faulty bearing signals. In this study, a deep learning- based approach is proposed to classify faulty and non-faulty bearing signals. A dataset is obtained by creating spectrograms from Drive End and Normal Baseline Dataset taken from Bearing Data Center. Three different Convolutional Neural Network (CNN) models with low, medium and large number of parameters are created. These three models are compared to classify faults. Although satisfactory results are obtained from all three models, the accuracy rate of the high parameter model is almost error-free. This study emphasizes increasing accuracy in proportion to the complexity of the models on fault classification. The results suggest that deep learning models can distinguish between faulty and non-faulty bearings, indicating that spectrogram analysis is useful in machine fault detection systems.