Fault monitoring for rotating equipment under varying working conditions

Fault detection of rotating equipment is the key to ensure the reliability and safety of complex systems. However, rotating equipment always operate under varying working conditions due to changes in environmental temperature, humidity, and other factors during the operation process. Commonly used fault detection methods often have poor adaptability under varying working conditions. To cope with the problem mentioned above, this study puts forward an intelligent monitoring approach based on genetic algorithm optimization of long short-term memory network(GA-LSTM) and K-means clustering algorithm optimization to achieve fault detection, state recognition and fault prediction of rotating equipment. Firstly, the genetic algorithm globally explores to find the optimal solution for the number of neurons and iteration times with the long short - term memory network. Then, the adaptive ability of the model in the fault detection process is improved effectively. Secondly, an identification model for operating status is established by integrating the support vector machine and K-means clustering algorithm (K-means-BT-SVM). Then, data complexity is reduced, classification accuracy is improved, and different operating conditions of rotating equipment are identified. Thirdly, the extreme gradient boosting (XGboost) algorithm is employed to construct a prediction model, which is designed to predict possible faults in rotating equipment. Then, the reliability and safety during equipment operation is improved. Finally, experimental verification is conducted with a publicly available dataset, and the results indicate that the proposed monitoring method perform well in fault detection, condition recognition, and fault prediction.