Transfer learning with inception ResNet-based model for rolling bearing fault diagnosis

With the development of information technology and sensor technology, people have paid more attention to data-driven fault diagnosis. As one commonly used methods in diagnosis, deep learning has achieved significant results. However, engineering practice, insufficient number labeled samples for diagnosis poor targeting extracted features lead a limited structural depth models inadequate model training, limiting diagnostic accuracy A novel method is proposed this paper by implementing model-based transfer Inception-ResNet-v2 model. Firstly, process applies signal-to-image transformation feature extraction stage merge frequency weighted energy operator (FWEO), kurtosis, raw vibration signals into RGB images as input dataset diagnosing type rolling bearing faults. Secondly, new combined CNN (TL-IRCNN) under minor sample conditions. Finally, The performance was validated using motor from Case Western Reserve University (CWRU) local laboratory. results show that TL-IRCNN achieves high classification conditions