A High-Precision Aeroengine Bearing Fault Diagnosis Based on Spatial Enhancement Convolution and Vision Transformer

In aeronautical engineering, monitoring and diagnosing intershaft bearings of aeroengine is crucial for flight and life safety. Given that sensor signals are affected by strong noise environments, this study proposes a high-precision multisensor feature fusion fault diagnosis method. The method is based on the lightweight spatial enhancement convolutional module (SConv) with channel shuffling and combined with vision transformer (CSST-Net). First, the original 1-D time-series signals acquired from multiple sensors are converted into 2-D time-frequency images using wavelet transform. The data are then fed into the model by utilizing data layer feature fusion. Subsequently, the interaction of information can be better facilitated by introducing channel shuffling operation into the convolution module. In addition, the spatial enhancement algorithm (SEA) leverages human visual perception properties to extract deeper fault features from the samples. Finally, the global information are extracted by the vision transformer (VIT) to obtain the fault diagnosis results. By validating on aeroengine intershaft bearing and rolling bearing datasets and comparing with other classical and advanced fault diagnosis methods. The results show that the method proposed in this study is not only state-of-the-art but also extremely robust in strong noise environments. Code is available at: <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/DABINB/CSST-Net</uri>.