285 / 2025-06-16 10:26:57

Next-Generation Condition Monitoring: Digital Twin in RUL Predictive Maintenance for Main Bearings of Semi-Direct Drive

life prediction,semi-direct-drive wind turbine main bearing,digital twin,transfer learning

     Against the backdrop of China's national strategic goals and supporting policies for carbon peak and carbon neutrality, reducing actual operation costs has become an urgent need for wind power enterprises. As a key component of the transmission chain in Semi-Direct Drive Wind Turbines (SDDWT), the main bearing's unexpected damage due to delayed maintenance can affect the overall machine stability, even causing prolonged shutdowns. Excessive maintenance, however, also leads to a significant increase in operation and maintenance costs. High-precision remaining useful life (RUL) prediction for main bearings can effectively reduce the operation costs of SDDWT, but RUL research on large rotating machinery like SDDWT main bearings still faces critical challenges: unclear performance degradation mechanisms, limited valuable data samples, and severe lack of data labeling.  

To address the above issues, this paper takes the main bearings of SDDWTs as the research object. By integrating model-based Digital Twin technology with data-driven domain adaptation methods, it conducts studies on: (1) the analysis of performance degradation mechanisms for main bearings; (2) the theoretical methodology of model-data fusion for RUL prediction.With the continuous development of Digital Twin technology, the availability of its simulation data will gradually improve and increasingly approach measured data. Correspondingly, data-driven life prediction models should also be continuously updated and developed to achieve further integrated development of model-based Digital Twin technology and data-driven life prediction models. This will fully leverage their respective advantages, improve the accuracy of life prediction research results, and create greater value for society.