A Novel Fault Monitoring Method for Vertical Roller Mill Bearings Integrating Vibration Transmission Path Analysis and Knowledge Graph

As the critical load bearing component within vertical roller mills (VRMs), the grinding roller (GR) bearing system operates under extreme conditions, including low speeds, heavy loads, intense impacts, and high dust contamination. These conditions make it highly susceptible to failure, highlighting the importance of reliable real-time fault monitoring of rolling bearings for safe and stable VRM operation. However, the harsh environment and complex mechanical structure lead to attenuation and aliasing of the fault signals of GR bearings, resulting in extremely low signal-to-noise ratios (SNR) that pose a fundamental challenge to accurate measurement and monitoring. To address these limitations, this paper proposes a novel fault monitoring method that integrates vibration transmission path analysis with a knowledge graph (VTP-KG) for GR bearings. The proposed approach begins by quantifying the energy contribution of vibration transmission paths using power flow theory. Subsequently, a knowledge graph (KG) for GR fault monitoring is established by integrating multi-source information. Finally, a triple matching method is designed to accurately identify both the location and type of faults under multi-fault coupling scenarios. Compared to existing methods, the proposed VTP-KG approach achieves superior monitoring performance, including high accuracy and low false alarm rates, and demonstrates strong robustness against data imbalance.