Predictive modeling and risk zoning for safety of large-section high-speed railway tunnels adjacent to karst caves: a case study of the Chongqing-Kunming high-speed railway tunnel

• Developed a predictive model for the critical safety distance between tunnels and adjacent karst caves. • Identified surrounding rock grade and lateral pressure coefficient as the most significant factors influencing critical safety distance. • Established a novel influence zoning standard based on plastic zone connectivity and energy mutation. • Quantified the influence zone extents for Grade IV and V rock under varying cave sizes. Focusing on the core issue of safety impacts during the construction of large-section high-speed railway tunnels adjacent to karst caves, this study comprehensively applied theoretical analysis, orthogonal numerical experiments, and multiple regression methods to systematically investigate the stability of the cave-tunnel system. A predictive model for the critical safety distance was established, and a zoning standard for construction safety influence was developed. The research shows that the surrounding rock grade and the lateral pressure coefficient have a highly significant influence on the critical safety distance, with their impact exceeding that of the geometric parameters of the karst cave and the tunnel burial depth. Through range analysis and analysis of variance, the primary and secondary order of influencing factors under different karst cave locations was clarified, and multiple regression prediction expressions for the critical safety distance were established for five typical karst cave locations: above the vault, outside the spandrel, outside the sidewall, outside the wall foot, and below the invert. The most critical conditions occur when the karst cave is located outside the tunnel wall foot or spandrel. Based on the criteria of plastic zone connectivity and energy mutation, a comprehensive discriminant standard centered on the critical safety distance was constructed, classifying the impact of karst caves on tunnel construction into strong, moderate, and weak influence zones. Combining the conditions of the supporting project, the influence zoning ranges for typical tunnel sections were determined, resulting in the zoning for the Changshui Airport Tunnel under Grade IV and Grade V surrounding rock conditions with cave sizes of 0.2D, 0.4D, 0.6D, and 0.8D. As the cave size increases from 0.2D to 0.8D, the extent of the strong influence zone expands from 0.33D–0.94D to 0.80D–2.23D in Grade IV surrounding rock, and from 0.97D–1.44D to 1.34D–3.47D in Grade V surrounding rock. This demonstrates a significant amplification effect of cave size on the disturbance range imposed on the tunnel. Compared to Grade IV surrounding rock, the influence zone induced by karst caves in Grade V rock is substantially larger and more sensitive to changes in cave size. The validity and engineering applicability of the proposed model and zoning criteria were verified using the case study of the Changshui Airport Tunnel. This research provides a theoretical basis and practical reference for the safe design and construction of similar tunnel projects in karst areas.