Multi-sensor fault diagnosis, especially when using heterogeneous sensors, substantially enhances the accuracy of fault detection in asynchronous motors operating under high-interference conditions. A critical challenge in multi-sensor fault diagnosis lies in effectively fusing data from different sensors. Deep learning offers a promising solution by transforming multi-sensor data into a unified representation, thereby facilitating robust data fusion. However, existing approaches often fail to fully exploit inter-sensor correlations and inherent prior physical knowledge. To address this limitation, we propose a novel graph neural network-based model that emphasizes graph structure construction for heterogeneous multi-sensor information fusion. Our framework includes (1) a multi-task enhanced autoencoder for node feature extraction, enabling discriminative representation learning, particularly with heterogeneous sensor data; (2) an adjacency matrix builder integrated with physical prior constraints to improve the generalization and robustness of the model; and (3) a graph isomorphism network to derive graph-level representations for fault classification. Our experimental results demonstrate the model’s effectiveness in diagnosing faults, as it achieves superior performance compared to conventional methods on two heterogeneous asynchronous motor datasets.

Accurate Remaining Useful Life (RUL) prediction for bearings is essential for machinery safety and widely applied across industries. Data-driven RUL prediction became the mainstream for RUL prediction, owing to their effectiveness and broad applicability. However, Data-driven RUL prediction confronts the difficulties of modeling complex vibration signal patterns and the scarcity of bearing full lifecycle data. To address these issues, this paper proposes a novel framework that integrates a dual-path GNN with a progressive fault-data-driven transfer learning strategy. First, a dual-path GNN architecture is designed to process separately constructed temporal-causal and spectral-interaction graphs, followed by a late fusion mechanism to preserve domain-specific patterns. The temporal graph is built using Transfer Entropy to infer directed causal relationships. Second, a twostage progressive transfer learning framework is introduced to bridge the domain gap between abundantly available fault data (source) and limited run-to-failure data (target), enhancing RUL prediction accuracy in data-scarce scenario. Comprehensive experiments demonstrate that the proposed method achieves superior accuracy and robustness in RUL prediction compared to state-of-the-art approaches, particularly under data-scarce conditions.