N
Since the modes of rock deformation in tunnels are governed by the interplay of numerous ill-defined variables, rock support requirements cannot be completely predetermined, whether by general formulas, empirical methods, or specific analytical studies. Yet much of value can be learned about the behavior and relative efficacy of supports for tunnels in jointed hard rocks from appropriate model studies.
The model methods of interest here are finite element analysis and scaled physical models. The intent of this paper is to discuss present capabilities of these methods in the context of two specific underground problems. The first is a "back-calculation" of rock deformation in circular tunnels subjected to a nuclear directly transmitted ground shock (the "Piledriver" experiment). The second is a physical model study to compare the response of tunnels, under increasing load, with different rock bolt configurations. The methodology and results will be elaborated in detail.
T
1. t
Limitations in our ability to predict deformations in structures in jointed rock stem partly from the newness of our models but even to a greater degree from our imperfect means of evaluating in situ rock and joint properties. Back calculation of joint deformability from a full scale test allows us to improve not only our grasp of the analytical tools but our knowledge of the rock properties as well. This was the case in the work discussed here - the Piledriver test.
The "Piledriver Experiment" consisted of a series of underground drifts arranged at varying distance from a blast as shown in figure 1. The drift complex is in granitic rocks with moderately spaced tight joints. During the test many tunnels were destroyed while others suffered varying amounts of distress. A pseudo static analysis was made using a finite element program for jointed rock as described by Goodman and Dubois, (1972). The program uses a stress transfer iterative approach for non linear joint behavior (Zienkiewicz et al, 1970); non linearities in rock behavior are superimposed by hand manipulations between runs. The program offers a procedure for analysis of rock excavations loaded to an extreme degree. Since the Piledriver experiment provides measured deformations in tunnels subjected to non-elastic deformation over appreciable rock volumes, it furnishes an excellent opportunity to check the program.
Not all drifts shown on figure 1 could be studied owing to time limitations. The sections chosen satisfied the following criteria:
a range of tunnel sizes should be offered
the lining type and strength should not vary among the selected sections.
The "addresses" chosen were CK1, CR1A, and BR12. These are steel lined circular tunnels with thick cellular concrete "back packing." The design concept presumed the lining and back packing to yield under the external blast pressure (Pe) thereby supplying a uniform radial pressure (Pi) to the rock. CK1, CR1A and BR12 are described in the table inset in figure 1. Their deformations are shown in figure 2, which compares the pre and post shock configurations of lining and rock(*)
