Research on Gearbox Fault Diagnosis Method Based on Multisensor Fusion and Maximum Classifier Discrepancy

Addressing the challenges posed by domain distribution differences and the complexities of features across various domains in gearbox fault diagnosis under variable operating conditions, this paper presents a fault diagnosis method based on multisensor fusion (MSF) and maximum classifier discrepancy (MCD). Initially, the continuous wavelet transform (CWT) is employed to convert the gearbox multisensor signals into grayscale images. The resulting grayscale maps are then fused to serve as input for the fault diagnosis model. An improved DenseNet, incorporating the Convolutional Block Attention Module (CBAM), is utilized as the feature extractor for the fault diagnosis model. To tackle the issue of feature distribution discrepancies under different operating conditions and the challenges associated with aligning features from distinct domains, MCD is applied to align the target domain features more closely with those from the source domain. Experimental results demonstrate that the model proposed in this study effectively diagnoses gearbox faults under variable operating conditions.