A fault source localization method for aircraft engine casing with dual-sensors based on acoustic emission

Accurate estimation of the position of the fault source in the aircraft engine is the key to achieve engine structural health monitoring (SHM). In this paper, a convolutional neural network and graph convolutional network (CNN–GCN)-based dual-sensor acoustic emission (AE) localization method is proposed for locating the fault source in the engine casing with multi-part coupling features. Firstly, the time–frequency map data sets of AE signals at different locations are established by using continuous wavelet transform to analyze the effect of multi-part coupling features on AE signals. Secondly, combined with its reverberation mode, multi-modal and dispersion characteristics, the effectiveness of CNN–GCN model is trained, verified and tested. Finally, the sensitivity of the localization results to the sensor types is analyzed, and the sensor combination mode with high localization accuracy is obtained. These results show that the proposed method in this paper can be used as an effective means for locating the fault source of the engine casing with complex coupling interface features.