4 December 2024 in IEEE Sensors Journal
With advancements in industrial automation, the accurate prediction of The Remaining Useful Life (RUL) in bearings is crucial for the proactive maintenance and reliability of industrial machinery. Traditional machine learning approaches often rely heavily on manual feature engineering and struggle to capture complex, nonlinear interdependencies between features that are vital for understanding machinery behavior under varying operational conditions. Addressing these limitations, our research introduces an innovative deep learning framework that integrates Multi-Adaptive Graph Neural Networks (MAGNN) with Temporal Convolutional Networks (TCN), thereby harnessing the power of graph-based learning to model complex interdependencies directly from raw sensor data. Our MAGNN framework employs a dynamic adjacency matrix that adapts to reflect the changing operational states of bearings, enabling the model to maintain high predictive accuracy even under fluctuating conditions. This adaptability is enhanced through a multi-scale feature extraction strategy that captures temporal patterns across different resolutions, providing a comprehensive feature set that is robust against environmental noise and operational variability. Experimental validation on the PHM2012 and XJTU datasets indicates the advanced performance of our MAGNN framework, significantly outperforming established AI benchmarks such as GNN-TCN, GNN-GRU, and Transformer models. In particular, the MAGNN-TCN configuration achieves a substantial improvement, reducing RMSE by up to 36.04% and reducing MAE by up to 34.19% when compared to the best of these conventional graph-based and sequence models. This performance boost highlights the effectiveness of our approach in leveraging dynamic, adaptive graph structures and multi-scale feature extraction, which are crucial for capturing the complex non-linear interrelationships inherent in sensor data from mechanical systems.
20 August 2025 in IEEE Sensors Journal
In the context of industrial bearing prognostics, conventional deep learning methods have demonstrated remarkable prediction capabilities but typically provide only deterministic Remaining Useful Life (RUL) estimates without quantifying prediction uncertainty. This significant limitation hinders reliable decision-making in critical maintenance scenarios where understanding prediction confidence is essential. While statistical approaches attempt to model degradation uncertainty, they often struggle with the complex degradation patterns exhibited by modern industrial bearings. This paper presents UABP-BTNSCL, a novel framework integrating supervised contrastive learning with Bayesian deep learning to address these challenges.A key innovation is the unique integration of a Deep Residual Shrinkage Network (DRSN) enhanced Temporal Convolutional Network backbone with a dual uncertainty quantification mechanism that simultaneously captures aleatoric uncertainty through Negative Log-Likelihood optimization and epistemic uncertainty via Monte Carlo Dropout. A key innovation lies in our SupCon-RUL strategy, which discretizes the degradation progression into meaningful stages to structure the feature space, significantly enhancing uncertainty estimation reliability. Extensive validation on PHM2012 and XJTU-SY benchmarks demonstrates the framework’s effectiveness, achieving superior prediction accuracy with average RMSE values of 0.086 and 0.077 respectively, while providing well-calibrated uncertainty bounds (PICP: 0.97, MPIW: 0.40). Ablation studies confirm that each component—DRSN enhancement, SupCon-RUL feature structuring, and Bayesian uncertainty estimation—contributes synergistically to the framework’s overall performance in both point prediction accuracy and uncertainty quantification quality.