Multi-Source Data Fusion Diagnosis of Rotating Machinery Based on Cloud-Edge Collaborative Architecture
Keywords:
data fusion, rotating machinery, cloud-edge computing, fault diagnosisAbstract
Rotating machinery forms the operational backbone of industrial systems, where early and accurate fault diagnosis directly impacts safety, efficiency, and lifecycle management. Traditional diagnostic approaches often rely on single-source sensor data or centralized processing, leading to latency, bandwidth bottlenecks, and limited adaptability in dynamic operating environments. This research introduces a cloud-edge collaborative architecture specifically engineered for multi-source data fusion in rotating machinery diagnosis. The framework distributes computational responsibilities across hierarchical layers: edge devices perform real-time signal preprocessing, feature extraction, and lightweight anomaly detection using time-frequency domain representations; the cloud layer orchestrates heterogeneous data integration, model retraining, cross-machine knowledge transfer, and long-term health trend analysis. Data fusion occurs at two critical stages—sensor-level synchronization and semantic-level alignment—ensuring consistency across vibration, acoustic emission, temperature, and current signatures despite varying sampling rates and physical units. A unified normalization protocol and adaptive weighting mechanism enable robustness against missing modalities and sensor degradation. Validation was conducted across three industrial-grade testbeds simulating bearing faults, gear mesh defects, and rotor imbalance under variable load and speed conditions. Results demonstrate a 94.7% average classification accuracy with sub-second diagnostic latency at the edge and a 32% reduction in false alarm rate compared to monolithic cloud-only baselines. The architecture maintains diagnostic fidelity during network intermittency through local state retention and incremental model updates. This work establishes a scalable, modular, and operationally resilient paradigm for intelligent condition monitoring in distributed industrial ecosystems.References
1. W. Zhang, M. Li, and X. Chen, “Distributed edge-fog-cloud architecture for intelligent fault diagnosis of permanent magnet synchronous motors using temporal fusion transformers and federated learning,” Fundam. Sci. Rep. Multidiscip. Areas, vol. 2, no. 02, pp. 121–131, 2026.
2. G. Li, J. Wu, C. Deng, and Z. Chen, “Parallel multi-fusion convolutional neural networks based fault diagnosis of rotating machinery under noisy environments,” ISA Trans., vol. 128, pp. 545–555, 2022.
3. H. Wang, H. Zhu, and H. Li, “A rotating machinery fault diagnosis method based on multi-sensor fusion and ECA-CNN,” IEEE Access, vol. 11, pp. 106443–106455, 2023.
4. C. Li, C. K. Kwoh, X. Li, L. Mo, and R. Yan, “Rotating machinery fault diagnosis based on multi-sensor information fusion using graph attention network,” in Proc. 17th Int. Conf. Control, Autom., Robot. Vis. (ICARCV), Dec. 2022, pp. 678–683.
5. T. Gao, J. Yang, B. Zhang, Y. Li, and H. Zhang, “A fault diagnosis method based on feature-level fusion of multi-sensor information for rotating machinery,” Meas. Sci. Technol., vol. 35, no. 3, p. 036109, 2024.
6. T. Liao, X. Fang, Y. Feng, and S. Wang, “Reliable and Efficient Decision-Making for Large Language Model Agents through Uncertainty-Guided Adaptive Reasoning,” 2026. Available: SSRN 6844659.
7. S. Lu, B. Qin, Q. Zhang, Y. Qiu, and G. Sun, “A collaborative denoising and multi-sensor local-global fusion framework for intelligent fault diagnosis of rotating machinery under noisy environments,” IEEE Trans. Instrum. Meas., 2026.
8. W. Guo, W. Zhang, S. Wang, and R. Liu, “Efficient Test-Time Adaptation for Dynamic Distribution Shifts,” 2026.
9. X. Li, Y. Wang, J. Yao, M. Li, and Z. Gao, “Multi-sensor fusion fault diagnosis method of wind turbine bearing based on adaptive convergent viewable neural networks,” Reliab. Eng. Syst. Saf., vol. 245, p. 109980, 2024.
10. Z. Jiang, D. Liu, and L. Cui, “Adaptive weighted data fusion-driven multi-layer discriminative dictionary learning method for intelligent fault diagnosis of rotating machinery,” Mech. Syst. Signal Process., vol. 235, p. 112888, 2025.
11. L. Pan, H. Li, X. Li, D. Bi, L. Peng, and Y. Xie, “Parallel feature fusion networks for rotating machinery fault diagnosis with multi-sensor information,” Sci. Rep., 2026.
