Cloud–Edge Collaborative Intelligent Fault Diagnosis of Rotor-Bearing System: Methodology and Experiment

作者：Hongwei Fan, Haowen Xu, Wei Fan, Buran Chen, Qingshan Li, Teng Zhang, Xiangang Cao, Xuhui Zhang · 发表于：IEEE Sensors Journal · 年份：2024 · DOI：10.1109/jsen.2024.3513644 · 被引用次数：11 · 研究领域：Advanced Decision-Making Techniques、Fault Detection and Control Systems、Advanced Computational Techniques and Applications

The condition monitoring and fault diagnosis of rotor-bearing systems are crucial for rotating machinery. The traditional methods rely on either a single edge computing of limited data processing capabilities or a single cloud computing of large data transmission volumes. In this article, a novel monitoring and diagnosis framework for rotor-bearing systems is proposed by combining cloud and edge computing. First, a cloud-edge collaborative diagnosis architecture is designed, where data acquisition, signal processing, and feature extraction are assigned to the edge end, and a deep learning-based diagnosis task is assigned to the cloud end. The edge ends collect acceleration and displacement signals in real time through piezoelectric and eddy current sensors and uses empirical mode decomposition (EMD) and time-frequency feature extraction to realize preliminary feature expression of the obtained signals. Second, the feature data are sent to the cloud end from the edge end, where an improved 1-D residual network (1D-ResNet) is used to classify the feature data and realize data-driven fault diagnosis. To verify the effectiveness of the methodology, a cloud-edge collaborative fault diagnosis system is developed, including a rotor-bearing unit simulation platform, an edge-end condition monitoring device, and a cloud-end diagnosis and visualization platform. The results show that the cloud-edge collaborative fault diagnosis method and system excel in terms of accuracy and real-time ...