Real-time detection of rotor misalignment using a non-contact triboelectric sensor integrated with deep learning

The stability of the rotor system is crucial for the reliability of mechanical equipment, and rotor misalignment is one of the main issues that can severely affect the performance of machinery. This study introduces a non-contact triboelectric misalignment detection sensor (TMDS) designed for real-time, quantitative detection of rotor misalignment. The TMDS exhibits high sensitivity to rotational speed, gap size, and dielectric materials, offering a precise and dependable method for misalignment identification. By integrating the deep learning model ResNet18-1D, the system has achieved a recognition accuracy rate that exceeds 96 % for misalignment identification. Furthermore, the practical application on a rotor dynamics simulation test bench proves the effectiveness of the TMDS in complex mechanical systems, and it is equipped with a user-friendly interface for data visualization and analysis. These features make the TMDS an effective tool for enhancing the health monitoring and fault diagnosis capabilities of rotating machinery.