Fault Degradation State Recognition for Planetary Gear Set Based on LVQ Neural Network

In order to ensure the safety and reliable operation of equipment, reduce accidents and economic loss caused by the mechanical fault or failure, prediction and health management (PHM) technology has attracted more and more attention. As the basis and starting point of fault prediction, degradation state recognition is one of the key steps of PHM, which directly affect the reliability of the equipment failure prediction and the selection of corresponding maintenance strategy. As to the degradation state recognition problem of planetary gear set, firstly, select the proper prognosis feature by evaluating various time-frequency features. Secondly, utilize the learning vector quantization neural network to recognize degradation state of planetary gear set. Finally, validate the effectively of presented method with pre-planted chipped fault experiment of planetary gear set. The results show that the proposed algorithm recognizes the multi-level degradation state effectively, and provide a useful reference for subsequent fault prediction.