Classification model for surrounding rock based on the PCA-ideal point method: an engineering application

Scientific classification of rock masses surrounding tunnels has great significance for construction cost and risk in subway systems. Quantifying the surrounding rock simply, quickly, and accurately is always a challenging issue as well as an urgent requirement in construction. Surrounding rock classification considers many complex and variable factors with uncertainty and nonlinear characteristics. Using principal component analysis (PCA) and the ideal point method, a new classification model is built consisting of five key factors, uniaxial compressive strength (UCS), rock mass integrity coefficient (K_v), softening coefficient (η), joint surface coefficient (J_c), and groundwater (ω). In the model, weights of key factors are determined by PCA, then the level of the surrounding rocks is analyzed using ideal point theory. The new model is applied successfully to classify surrounding rock in the Qingdao Metro system. Results provide a reference for classifying surrounding rock quickly and guide the tunnel design and construction.