CRCHI Develops AI Model for Tunneling, Enhances Construction Intelligence

June 15,2026

Recently, a key breakthrough has been achieved in the development of the AI knowledge model for full-face tunnel boring machines, led by CRCHI, in collaboration with the Fourth Survey and Design Institute, China Railway 18th Bureau Group, and Hunan University. The shield tunneling operation and maintenance intelligent body (intelligent equipment intelligent body) has been put into engineering application, promoting the transformation of tunnel boring construction from traditional experience-driven to data-empowered and intelligent decision-making. Artificial intelligence is empowering intelligent construction of underground projects, helping to improve the quality and upgrade the new quality of productive forces in the infrastructure field.

CRCHI Edge AI Large Model

This project addresses the issue of data silos and lack of collaboration in tunnel design, construction, and equipment segments. The research team has built a tunnel AI large model that integrates three intelligent bodies: tunnel design, construction management, and intelligent equipment, spanning the entire chain of "design optimization - construction control - equipment operation and maintenance decision-making" in tunnel engineering.

During the development process, the three intelligent bodies work together efficiently. The tunnel design intelligent body focuses on tunnel scheme design and support structure optimization; the construction management intelligent body achieves integrated control of construction safety, quality, and progress; the intelligent equipment intelligent body integrates practical functions such as intelligent operation and maintenance, automatic correction of tunneling deviation, auxiliary decision-making for tunneling parameters, prediction of surface settlement, and TBM jamming early warning.

Tunnel Boring Machine Intelligent Tunneling System Control Room

As the core achievement of the project, the intelligent equipment intelligent body has built a "performance perception - risk early warning - intelligent decision-making" closed-loop technical system. With the help of intelligent algorithms dynamically matching the geological conditions, it achieves autonomous correction of deviation and automatic gear shifting of shields in complex geological environments.

In the development of the intelligent equipment intelligent body, the R&D team has gone through seven rounds of iteration and refinement, compiling nearly ten thousand engineering documents into a professional knowledge base, upgrading the intelligent operation and maintenance Q&A module, adding supplier information retrieval capabilities, and continuously enhancing the model's adaptability and stability under complex working conditions.

Currently, this intelligent system has been equipped with full-face tunnel boring machines and has been successively applied to a number of national key projects such as the plateau railway, the water diversion project from the Yangtze River to the Han River, and the expansion of the Hang-Jia-Hu South drainage follow-up western channel project, with the cumulative application on site exceeding 5000 rings.

Intelligent Body Application Interface in the Yangtze River to Han River Water Diversion Project

Under the traditional construction mode, surface settlement relies on manual measurement, with sparse data and strong lag, and tunneling parameters are entirely adjusted based on manual judgment. The tunneling attitude is generally controlled within a 50mm deviation range, and equipment failure diagnosis heavily relies on manual experience, with low diagnostic efficiency and high operation and maintenance costs.

After applying the intelligent equipment intelligent body, from the actual measurement effects of the project, the prediction accuracy of surface settlement has reached the millimeter level, the accuracy rate of auxiliary decision-making for tunneling parameters has reached 90%, and the control deviation of tunneling attitude is less than 30mm; the efficiency of equipment failure diagnosis has increased by 20%, the overall construction efficiency of the project has increased by 10%, and it can achieve full-process closed-loop control of risks such as adverse geology, tunneling over-limit, and excessive stratum settlement, effectively reducing construction safety risks caused by human misjudgment.

Intelligent Equipment Intelligent Body Operation Interface

At this stage, focusing on the direction of the integration of physical artificial intelligence and high-end equipment, the R&D team is steadily promoting the landing transformation of cutting-edge technologies such as multi-modal reasoning alignment and end-to-end model training in the field of tunneling equipment.

Next, the R&D team will continue to optimize the performance of the tunnel AI large model and the three intelligent bodies, accelerate the industrialization and promotion of scientific research achievements, rely on independently controllable intelligent construction technology, continue to ensure the construction of major national projects, and consolidate China's leading position in the domestication of shield machine technology.

Source : www.51cm.org

Views : 6486

Tags : CRCHI tunnel AI model shield tunneling