Abstract: Real-time monitoring of force conditions on a shield cutterhead during tunneling is challenging, leading to delayed analysis, decision-making, and untimely feedback on the operational status. To address this challenge, a novel real-time stress analysis method for the cutterhead, which leverages digital twin technology, is proposed. First, a digital twin framework for real-time stress analysis of the shield cutterhead is established. Finite element analysis is then performed on the basis of actual working conditions to determine the stress distribution of the cutterhead under various working conditions. Second, to improve computational efficiency and real-time performance, open-source software is employed to extract and process all node data. Based on the extracted stress data, a stress data model for the cutterhead is developed, and its accuracy and reliability are validated through comparison with the measured data. Subsequently, proper orthogonal decomposition method is applied to investigate the model order reduction techniques for the cutterhead finite element model based on the developed data model. This order reduction technique significantly decreases the computational load while maintaining acceptable accuracy, enabling rapid acquisition and real-time analysis of the cutterhead′s stress state. The proposed method facilitates the efficient extraction of cutterhead stress data under different construction conditions, enabling dynamic monitoring. Finally, the research findings are integrated into the shield′s digital twin system, achieving real-time stress analysis and visualization of the cutterhead. The system can monitor the cutterhead′s stress conditions in real-time based on the actual working states and provide immediate feedback. The results demonstrate that the proposed method exhibits high virtual-physical consistency and reliability in practical applications, providing effective support for the digital monitoring of shields.

Key words: shield cutterhead, digital twins, real-time analysis, model order reduction, proper orthogonal decomposition method