Rock mechanics in the investigation and construction of Tumut 1 Underground Power Station, Snowy Mountains, Australia
lt is recognized that the particular assumption of isotropy of the rock mass was far from correct; nevertheless, the photoelastic analysis was considered valuable for indicating semiquantitatively the position and magnitude of the stress concentrations produced by various shapes of openings, and various arrangements of the multiple openings.
As already mentioned, the granite and gneiss even though distinctly different petrographically, exhibit quite similar elastic properties measured on specimens in the laboratory. These specimens show an almost perfect linear relationship between stress and strain within the range of stresses applied in the photoelastic tests. The site is not intersected by any wide structural discontinuities such as wide fault zones, although there are several persistent narrow fracture zones. A major deviation from the isotropic state is caused by the joint pattern. The rock mass is jointed differently in different directions and the gneiss is usually much more closely jointed than the granite. During construction, the walls and roofs of the excavations were mostly systematically rock-bolted; this treatment tends to maintain the structural continuity of the rock in the zone immediately surrounding the excavations.
The results of certain photoelastic investigations involving features of geological interest will be described in outline.
The effects of assuming different initial systems of stress were investigated by a series of tests on a model having the same cross-sectional shape as adopted for the machine hall, The model was loaded first vertically and then horizontally; the effect of the combined stresses was obtained by addition of the stresses. The stress-concentration factors around the boundary of the excavation for two cases are given in Figure 9.
Figure 9 – T. 1 Machine Hall, Stress-Concentration Factors
The first case assumed the ratio of the horizontal principal stress to the vertical principal stress to be 0. 25. This is the condition where he horizontal stress is due solely to the weight of overlying rock having a Poisson's Ratio of 0.2 For this case the analysis predicts, for the full excavation, conditions of near tension in the crown of the roof and in the floor, and compression elsewhere.
The second case assumed the ratio of the horizontal principal stress to the vertical principal stress to be 1. This is close to the same ratio of horizontal to vertical compression that was determined from the results of flat-jack tests at the site. For this case the analysis predicts high compressive stresses in the roof and a considerable decrease of the original natural compression in the walls.
The wide and important divergence between the stress concentrations in these two cases demonstrates the necessity of determining the actual natural state of stress. This is further illustrated by experience during construction.
At T. l power station the determinations of the state of stress of the rock mass were not made until the excavation was well advanced.
