关键词: 隧道掘进机(TBM), 小断面, 土石组合, 可掘性分级, PCA-RF模型

Abstract: The efficient construction and intelligent control of TBM heavily rely on the classification and real-time identification of surrounding rock excavability. To address this, we establish a classification standard for surrounding rock excavability in TBM construction under geological conditions characterized by small sections and soil-rock combinations based on actual data (penetration thrust and torque per unit penetration) from the Anyang Western Water Diversion Project. Moreover, we introduce the PCA-RF model for real-time identification and prediction of surrounding rock excavability, and then compared the results with those of BP, SVR, and RF models. Our research yields the following conclusions: 1) The classification standard for surrounding rock excavability in TBM construction under the geological conditions of small sections and soil-rock combinations proves to be applicable. This standard resolves the limitations of traditional methods for classifying surrounding rock in soil-rock composite environments. 2) The PCA-RF model demonstrates an identification and prediction accuracy of 98.3% for the surrounding rock excavability in TBM construction under the geological conditions of small sections and soil-rock combinations. This accuracy surpasses that of the BP, SVR, and RF models and fulfills the demands of engineering construction.

Key words: tunnel boring machine (TBM), small section, soil-rock combination, classification of excavability, PCA-RF model