Abstract: So far tunnel boring machine(TBM) boring still depends heavily on the pilot′s experience and it is difficult to respond in time to the abnormal situations during TBM boring, resulting in TBM boring delay or aggravated cutter wearing. Therefore, it is not only necessary to obtain the accurate information of the rock mass at the tunnel face in time without suspending the TBM boring, but also to realize intelligent auxiliary decisionmaking after understanding the information of the rock mass at the tunnel face. In this paper, with the Water Supply Phase Ⅱ Project in North Xinjiang of China as the study object, 〖HJ5.2mm〗TBM boring indices such as the rock mass classes, the TBM boring efficiency, and the TBM boring parameters are analyzed; Based on image perception, data mining and machine learning technologies, realtime perception of rock mass information is realized by means of muck image perception, cutterhead vibration monitoring and advance geological prediction; Database of geological information, TBM boring parameters, equipment and supporting parameters is built to carry out big data preprocessing and correlation analysis; The multiobjective intelligent optimization algorithm is adopted to optimize the TBM boring parameters, with the TBM boring speed and the cutters′ service life as the objectives; On this basis, the intelligent auxiliary decisionmaking for the TBM boring parameters, supporting patterns, and suggestions on countermeasures against TBM jamming are put forward. The application of the intelligent auxiliary TBM boring technology in the trial sections of XE tunnel has increased the TBM boring speed by 15.6% and the cutters′ service life by 4.5%, without suspension of TBM boring or abnormal damage to the TBMs due to improper selection of TBM boring parameters. The research results can be popularized and applied to similar projects to improve the TBM boring efficiency and reduce the costs.

Key words: tunnel engineering, tunnel boring machine(TBM), intelligent auxiliary TBM boring, rock mass perception, big data analysis, optimization of TBM boring parameters