Intelligent Fault Diagnosis of Rolling Bearing Based on DGAC-SNN

In recent years, the field of fault diagnosis has developed rapidly, and spiking neural network (SNN) has become a research hotspot because of its biological interpretability and low energy consumption. However, the existing SNN structure still has a lot of room for improvement. In this paper, we propose a dictionary gated attention coding spiking neural network (DGAC-SNN) model, which promotes the GAC coding module into a sparse DGAC coding module, and embeds the DGAC module with an improved spike-element-wise (SEW) ResNet18 network. Due to the visualization of the coding layer and the combination of spiking neurons, this method has a good interpretability. The results of bearing dataset obtained from Case Western Reserve University show that the model has high diagnostic accuracy, which verifies the effectiveness of the proposed method.