The granite in the penstocks near roof level was intersected by fault A and by the zone of close fracturing D, but otherwise had quite normal jointing. Apart from the immediate vicinity of these fractured zones, the rock mass appeared quite solid and stable, and only a few scattered rock bolts were used for support. In addition, four light-steel sets, packed between with concrete, were installed at the future portals. When the machine-hall excavation was deepened to the level of these tunnels, it was found that the granite in the roof and walls was generally loose, mainly along joint blocks. In less jointed areas spalling had occurred. There were a few small falls both of joint blocks and freshly fractured slabs. At the portals of three of the tunnels the concrete had developed one or more tension cracks dipping steeply toward the machine-hall and open to a total width of about half an inch. Additional supports in the form of steel rings and rock bolts were installed in these tunnels until the permanent steel liners were placed.

In the second case it appeared that, when the penstock tunnels were first driven, the surface layer of rock was strong enough to carry the stress concentrations. At this time the machine-hall roof excavation was sufficiently distant to have little effect. Later, the deepening by stages of the main body of the machine hall evidently created a progressively changing zone of stress concentrations extensive enough to envelop the tunnels and be superimposed on those caused by the tunnels themselves. The surface layer of rock around the tunnels was not strong enough to withstand these increased stresses and thus failed, The failure occurred mainly by sliding along joints, with spalling of local areas of rock that possessed interlocked joints where sliding could not take place.

On the other side of the machine hall the four draft-tube tunnels, corresponding to the four pen- stock tunnels, were driven after the main body of the machine hall was excavated. They were sup- ported by systematic rock bolting as excavation proceeded. No signs of unusual rock movements or spalling were seen in these tunnels. The reason for this is not clear - it may have been due partly to the effects of the bolting of the tunnels, partly to the somewhat better quality rock, and partly to the fact that the rock in the machine-hall wall was already partially decompressed but stabilized by bolts at the time the tunnels were driven into it.

These examples of failure by spalling were so localized that they were of only relatively slight significance in the construction operations. They are, however, of general interest, since rupturing of sound strong granite is independent evidence of the existence of very high stresses in those areas, created by excavation procedures or by compound shapes of openings. These failures seem analogous to the longitudinal splitting and surface spalling which sometimes are observed in cylindrical specimens of brittle rocks undergoing tests in simple compression in the laboratory, immediately preceding or accompanying ultimate failure.

Instrumental Measurements

The behavior of the rock mass around the machine-hall excavation was observed during construction by the following quantitative instrumental measurements:

(a) The strain in many of the reinforced concrete arch ribs was measured by means of electric resistance-type strain meters embedded in the concrete, and by observation of Huggenberger deformeter points fixed on the surface of some ribs.