Investigation of cavern distortion in the Gordon underground power station, Tasmania

A. J. Bowling

The Gordon River Power Development of the Tasmanian Hydro-Electric Corporation is located in the south west of Tasmania. The development uses the stored water in Lake Gordon and Lake Pedder to generate electricity in the 480MW Gordon underground power station. The development was constructed in the 1970s with the underground power station cavern being excavated between early 1972 and the end of 1974. The cavern is 96 m long in an east-west direction, 31 m high and 22 m wide. The cavern and associated water tunnel system was excavated to accommodate five 160MW generators. The intake shaft from Lake Gordon and the penstock manifold is on the northern side of the cavern (called the upstream side in the following) while the tail bay and tailrace tunnel are on the southern side of the cavern (the downstream side). Initially only two generators were installed and commissioning of the station was completed in 1978. In 1986 a third generator was installed. The western end of the cavern, which is in schist, see Section 2, is thus currently empty for its entire height and width over a length of about 30 m.

The generating machines consist of a vertical axis Francis turbine directly connected to a generator, the overall height of the two units being about 12 m. The equipment in the power station is serviced by two overhead cranes which run on rails supported by concrete beams cast against the top of the opposite long rock walls of the cavern. The reinforced concrete crane beams are anchored to the rock with stressing cables at 3 m spacing. There are thus 32 cables over the 96 m length of each crane beam. The cables are inclined upwards at 30 degrees into the rock walls, are 15 m long into the rock and are stressed to a load of 880 kN. The crane rails are connected to the crane beams at 0.6 m intervals with rail clips attached to a steel base plate bolted to the beams.

The first two generating machines were installed, accurately vertical to about 1 in 25,000 (40 micro-radians), within a few months of the completion of cavern excavation and the crane rails were also accurately aligned (nominally to ±⅛ of an inch) near the end of cavern excavation just prior to the installation of the two overhead cranes.

Accurate measurements on the numbers one and two machines in recent years have indicated slight upstream tilting of their vertical axes although number three machine has not shown any measurable tilt. In addition, and more significantly, recent surveys of the upstream and downstream crane rails have indicated that since their installation they have bowed slightly towards each other in a horizontal plane and with respect to their ends (Figure 1). The same recent surveys have also indicated that, in a vertical plane, the upstream (northern) rail has bowed upwards and the downstream (southern) rail has bowed downwards, again with respect to their ends (Figure 2). Since the crane rails are rigidly connected to the concrete crane beams, the crane rail distortions can be considered to reflect the distortions of the crane beams themselves and thus of the rock to which they are attached.

There is anecdotal evidence that slight misalignment of the crane rails was causing some problems with the operation of the overhead cranes within a few years of the completion of the power station. The problems included wear on the edges of the rails on the cavern opening side and corresponding wear on the crane bogie wheels and a tendency for the bogie wheels to ride up on the rails. However it was not until 2001 that the above described accurate surveys of the crane rail alignments were carried out.

This paper describes the analysis that suggests that these distortions of the crane rail beams and the tilting of the machine axes are most likely due to the development of creep strains in the rock surrounding the power station cavern due to an inclined principal stress field in the rock.