Abstract:

 Currently, two classification methods exist for evaluating the stability and excavatability of tunnel surrounding rock. These methods correspond to the surrounding rock classification method by drilling and blasting method and that by TBM method, respectively. In this study, these surrounding rock classification methods are discussed. First, the development status and application conditions of the surrounding rock classification method by drilling and blasting method are reviewed, and suggestions for its future development are provided. These suggestions include paying attention to the quantitative characterization of construction factors on the grading deterioration characteristics of surrounding rock, actively exploring subgrade classification methods of surrounding rock, and establishing an integrated system of surrounding rock classification methods. Next, the common surrounding rock classification methods are summarized used with the TBM method and a comprehensive surrounding rock classification method based on geological factors, TBM tunneling parameters, and slag characteristics is proposed. For excavatability classification, it is suggested to use TBM net tunneling speed, rock mass drivability, and onsite penetration index as evaluation indexes. For adaptability classification, it is suggested to use machine jam risk, TBM utilization rate, or construction speed. Finally, the latest developments in "artificial intelligence+" classification methods for surrounding rock in tunneling are analyzed. This analysis focuses mainly on the application dynamics and development direction of new methods and technologies such as artificial intelligence algorithms, intelligent image processing, cloud computing, and intelligent decision making. The research aims to provide a reference for the construction and intelligent application of the surrounding rock classification system in tunnel engineering.

Key words: tunnel engineering, surrounding rock classification, drilling and blasting method, TBM method, artificial intelligence+