Comparative Study of Rock Support System Design Practice for Large-Scale Underground Excavations

Rock support system design practices for large-scale underground excavations in North America, Europe, Japan and some other countries are reviewed in a systematic manner for about 110 case histories. Different rock support system design methods are summarized and data related to support quantities, such as arch concrete thickness, shotcrete thickness, rockbolt and anchor length, support pressure for the surveyed large-scale caverns are presented. Special attention is being paid to the comparison of engineering practices in Japan and other countries. Data re-analysis using Finite Element Method was performed for 37 cases to establish the anticipated depth of rock mass fracturing and to justify the adopted support pressure. It is found that support pressures applied to most caverns are reasonable and comparable based on the Q-index. It is also found that while reinforcement lengths for cavern sidewalls are generally acceptable, unnecessary long rockbolts or anchors have been applied to cavern roofs in many cases. compiled information about 85 large-scale underground excavations, which includes size, overburden, rock types and conditions, and some description of rock support system used. More recently, a compilation of data on large-scale underground caverns, which focused on 450 hydropower caverns, was published in the WaterPower Yearbook 1997 (Hönish, 1997), and about 70 power caverns were analyzed based on the RMR rock mass classification system by Hönish and Nagel (1998). Unfortunately, these databases do not provide sufficiently detailed data for re-analysis purpose. In the present study, an independent data collection process was undertaken and data, which contains detailed information related to rock support system design for approximately 110 large-scale caverns, was complied. Figure 1. Evolution of the apparent cross section area Figure 2. Overburdens of large-scale caverns Figure 1 presents the evolution of the apparent cross-section area (width × height) for the data set collected for this study. These caverns are larger than the largest traffic tunnels which have apparent cross-section areas around 100 m (e.g., the Seikan Tunnel in Japan is 11 m in width and 9.12 m in height). Underground powerhouses have been located at depth up to 700 m (e.g., La Tasajera powerhouse in Colombia). There is a trend that underground excavations go deeper and deeper in the rock mass and this trend can be seen from Figure 2. The deepest large-scale underground cavern, which is not plotted in Figure 2, is the Sudbury Neutrino Observatory (SNO) in Canada. The SNO was built in a cavern as large as a 10-storey apartment building (22.9 m wide and 30.5 m high), in the deepest section (2072 m beneath the ground surface) of INCO Limited's Creighton Mine (Oliver, 1992). At such depth, the ground stress is extremely high and special attention had to be paid for rock support in burst-prone ground. 2 SUPPORT SYSTEM DESIGN METHODS 2.1 Design principles Very large span natural caverns are found in France (230 m), Spain (245 m) and Indonesia (400 m), which serve as good examples of cavern stability without any support (Duffaut et al., 1986). In most cases, economy can be achieved by taking advantage of the capability of the rock mass to support itself. If a rock support system is properly chosen, and properly installed at the right stage of excavation, the loads on the support system can be minimized. Thus the support system can often be regarded primarily as a reinforcement that helps the rock mass to support itself. It is necessary to distinguish between a structuralcontrolled stability environment and a stresscontrolled environment when designing the rock support systems for underground excavations, because the design methods used for different situations are not the same. Structural-controlled instability can be prevented by correctly installed rock support and reinforcement such as rock bolts, shotcrete and steel ribs. Stress-controlled instability, which is due to the overstressing of the rock mass around the cavern, can to some extent be controlled by optimizing the cavern shape and the alignment but stress-driven rock failure can often not be prevented (Stacey & Page, 1986). The geological condition around a cavern is generally complex so that it is difficult and often uneconomical to work out sufficient rock support quantities for all anticipating failure modes. It is suggested that in the design of the rock support system, a basic pattern of support or reinforcement can be nominated to cope with the majority of specified potential failures. This pattern is to be checked against observations during excavation and supplemented or amended where necessary (Douglas et al., 1979). 2.2 Design methods In structural-controlled environments, wedges falling from the roof or sliding out of the sidewalls are Cross section areas of underground caverns 0 50