Spur Bevel Gearbox Fault Diagnosis Based on Wavelet Packet Transform for Feature Extraction and Flow Graph Data Mining

摘要

Gearbox vibration signal contains a wealth of the gear status information, used wavelet packet transform (WPT) refinement of the partial lock ability to extract the fault signs attribute information in the vibration signal. Extracted signs attribute information as the input of the flow graph (FG), generated decision rules to achieve the purpose of fault diagnosis. FG was a knowledge representation and data mining method to mine the intrinsic link between the data and improve the clarity of the potential knowledge. The results confirmed that used of WPT feature extraction and FG data mining method can accurate detection the gear fault.