Abstract: To enhance the precision and efficiency of mechanized tunneling in TBM projects, this research investigated a boreability classification method based on tunneling parameters and muck characteristics. Based on a TBM construction section of a tunnel in western Sichuan, tunneling parameters and muck samples were employed as the primary research objects. The time-series data of TBM tunneling parameters were preprocessed using techniques including redundant data elimination, change-point detection and conditional filtering, outlier removal based on boxplot analysis, and smoothing filters. The average total thrust, average cutterhead torque, and average penetration rate per excavation cycle were selected as clustering variables. By employing the silhouette coefficient, BIC, and the elbow method, the optimal number of clusters was determined to be four. Subsequently, a Gaussian Mixture Model (GMM) unsupervised clustering algorithm was applied to classify the tunneling cycles. A boreability classification method was then established by integrating the clustering results with muck morphology characteristics and field geological surveys. The clustering results were validated through subsampling validation and external validation, both of which confirmed a high level of consistency. These findings demonstrated the robustness of the clustering outcomes and the effectiveness of the proposed boreability classification method. This research provides a data-driven method for TBM boreability assessment, offering enhanced guidance for design and improved adaptability in construction.
刘德卫. 基于掘进参数与岩渣特征的TBM隧道围岩可掘性分析[J]. 隧道与地下工程灾害防治, .
LIU Dewei. TBM tunnel surrounding rock excavability analysis based on excavation parameters and muck characteristics. Hazard Control in Tunnelling and Underground Engineering, 0, (): 1-9.
