Abstract: With the trend of transformation and upgrading of the engineering construction industry, the key technologies and implementation method of intelligent operation and maintenance of railway tunnels based on digital twins are proposed to address the problems of low control efficiency, poor interpretability of health status evaluation, and complex process of operation and maintenance decision-making in the operation and maintenance of railway tunnel projects. The overall framework of railway tunnel intelligent operation and maintenance is established with digital twin as the driving core, which organizes the layers of tunnel operation and maintenance data flow and application, clarifies the main function and core value of digital twin in the system framework, and conducts research on the information mapping mechanism between physical entities and twin models and key technologies of intelligent operation and maintenance. First, the data classification system and information storage and expression methods in railway tunnel operation and maintenance are summarized, and the mapping mechanism between dynamic perception data and digital twin data in tunnel intelligent operation and maintenance is revealed. The key technologies that need to be solved in intelligent operation and maintenance, including intelligent perception of tunnel state, intelligent data analysis, and auxiliary decision-making in operation and maintenance, are analyzed. Based on the system framework, fusion mechanism and key technologies, the implementation method of intelligent operation and maintenance of tunnels is proposed. In addition, the reliability and practicability of the technical framework are verified through the intelligent operation and maintenance platform of railway tunnels. Furthermore, the feasibility of the digital twin technology in improving the efficiency of tunnel operation and maintenance management and the degree of intelligence is preliminarily verified.

Key words: railway tunnel, digital twins, intelligent operation and maintenance, data fusion