A Fast and Energy-Saving Neural Network Inference Method for Fault Diagnosis of Industrial Equipment Based on Edge-End Collaboration

Data-driven fault diagnosis algorithms represented by deep learning have been widely used in industrial equipment fault diagnosis. However, the lack of real-time performance has always restricted the development of such methods. With the development of edge computing, many edge and end computing devices are deployed in industrial environments. For this distributed computing environment, we propose a distributed neural network inference method with edge-end collaboration. This method uses an edge server to cooperate with multiple end devices for network inference. In the diagnosis of industrial equipment, it can increase the speed of inference, reduce the traffic of the edge network, and help the application of deep neural networks in industrial environments.