Research on piston aeroengine fault diagnosis method with acoustic emission signal and rough set theory synergy

Piston aeroengine is a general power device for small aircraft. However, the engine system is complex and is subjected to strong noise interference in addition to changing operating conditions, which make it difficult to diagnose the faults of the engine. This article proposes a novel intelligent diagnostic model for piston aeroengine in collaboration with the acoustic emission (AE) signal and rough set theory. First, the wavelet packet transform is used for the time-frequency analysis of AE signals; then based on the rough set theory, to select the sensitive components and features with a high contribution to the classification; and finally, the RGB-3-channel color feature maps are constructed based on the selected components and features for the recognition of convolutional neural network. We compare the proposed method with several conventional methods, and the results show that it has higher diagnostic accuracy as well as better generalization ability and universality.