Fault Diagnosis of Point Machines Based on Empirical Mode Decomposition and Transformer

To address the limitations of single-source data and the challenge of ensuring high diagnostic accuracy in the fault diagnosis of railway point machines, this paper proposes an intelligent fault diagnosis and decision-making model for the ZYJ7 point machine based on multi-source data fusion. The model first performs standardized processing on data collected from multiple sensors—including current, vibration, and hydraulic pressure—through data preprocessing and feature extraction techniques, thereby fully mining diagnostic information from each data source. Subsequently, a multi-algorithm diagnostic approach combining machine learning and deep learning is employed to construct individual fault recognition models. To integrate the diagnostic results from different models, a confidence-weighted intelligent voting strategy is introduced. This strategy utilizes an adaptive weight allocation mechanism to effectively combine the strengths of various algorithms and produce a final fault diagnosis decision. Experimental results demonstrate that the proposed model significantly outperforms traditional single-algorithm and simple ensemble methods in terms of diagnostic accuracy, robustness, and real-time performance. This study not only provides a theoretical basis for fault prediction and maintenance of ZYJ7 point machines, but also offers a novel approach for intelligent monitoring of railway signaling equipment, showcasing high engineering application value and promising prospects for widespread adoption.