Study on the fault diagnosis of rolling bearings based on the CNN-BiLSTM method

Study on the fault diagnosis of motor rolling bearings is particularly important to the use and safety management of bearings and the motor system. Therefore, a fault diagnosis method for motor rolling bearings based on the CNN-BiLSTM model is proposed in this work. The model uses the convolutional neural network (CNN) to extract features and reduce data dimensions of the input vector to the model, and employs the bi-directional long-short memory network (BiLSTM) to learn the time correlation information of the input time series in the local features, so as to implement the operation state monitoring and fault diagnosis of rolling bearings. The experimental results prove the effectiveness of the model, and its fault diagnosis accuracy of motor rolling bearings reached 98.62%, which can meet the requirements for practical use.