Revising Terzaghi's Tunnel Rock Load Coefficients Available to Purchase

In the USA the cost of steel ribs for tunnels approximates $100 million each year. European practice has long since abandoned heavy steel ribs for tunnel support, and tunnel costs in Europe (normalized for wages) are about half the costs incurred in the USA. A feeling has existed for some time that the USA practices of supporting tunnels with heavy steel ribs have been overly conservative. The USA practice is based on rock load coefficients by Karl Terzaghi. In 1946, Karl Terzaghi published a tunnel rock classification system (Terzaghi, 1946), with the rock loads on tunnels predicted for each class of rock. No data on real rock loads were presented by Terzaghi and his rock load coefficients may have been largely intuitive. His classification system and associated predicted rock loads have been widely used in the USA. Each rock class, taken from Terzaghi's classification system, is assumed to generate loads of loosened rock resting as dead load. The dimensions of the loosened rock are described by rock load coefficients multiplied by the tunnel diameter. The resulting rock loads are used by design engineers, contractors and agencies to size supporting steel ribs. Terzaghi clearly stated that he had conservatively assumed high rock loads for his classes 4, 5 and 6 to take a presumed effect of groundwater into account. If the tunnel was not affected by groundwater, Terzaghi stated those rock loads should be reduced by 50 percent. Twenty two years later Tot Brekke (1968) wrote a paper clearly showing that based on observation of more than 250 tunnels groundwater does not affect tunnel rock loads. In 1969, Deere, et al., published measurements made on a number of tunnels, and a well-known curve was drawn by Deere to illustrate Terzaghi's coefficients graphically. Deere ' s text stated that Terzaghi's predicted rock loads were too high and should be reduced. Cording and Deere (1972) more recently published data indicating that very large real rock caverns and chambers were being successfully built using rock bolts, which can take only a very limited rock load, thus indicating that real rock load is much smaller than Terzaghi or even Deere predicted. Real rock load is actually dependent on how discreet blocks of rock actually interlock and create an "arch action" which supports the roof of tunnels. Voegele (1978) published a two-dimensional computer program which showed this effect. More recently, Goodman and Hua Shi (1981) published what may be a definitive explanation of this arch action, showing in 3-D how certain "key blocks" serve to hold the rock mass intact in a tunnel roof preventing failure and acting to self-stabilize the tunnel. Provided that normal care is taken not to allow the rock to unravel, real rock loads are not large, and heavy support systems are not required. Rose, et a1.,(1981) parametrically showed how Terzaghi's 1946 rock classification system leads to given predicted rock loads, which then require that certain sizes of steel rib be provided.