12.3 Foundation investigations for Morrow Point dam and power-plant and Oroville dam and power-plant
12.3.4 Rock tests at the Oroville dam and power-house
A detailed description of the Oroville dam and underground power-house rock tests was given in a paper by Stoppini & Kruse (1964) from which the following information has been taken.
An extensive exploration programme was conducted at the power-house site as a preliminary to the construction of the underground works. An exploration adit and cross drifts within the left abutment of the dam were used. Similar to the Karadj tests, the Oroville plate-bearing tests used a circular plate with four dial gauges placed at equal intervals around its edge. And, as with the Karadj tests, the deflection pattern of the loaded
Investigations for Oroville dam 371 plate failed to conform to any standard formula found in textbooks: it revealed a tendency for the plate to tip.
Three jacks, each of 100 tons capacity, were placed in line along a re- inforced WF beam which in turn delivered the thrust along a short length of 6-in bar stock to a lj-in bearing plate 6 | in in diameter. It was intended to obtain high loading intensities. The adit was of minimum size (5 ft by 7 ft), unsupported, and driven on a slight gradient into the power-house site.
Similarly, from the main adit several drifts and cross cuts were driven. The plate-bearing tests were performed in these and they provided access for extensive core drilling throughout the power-house site. They were also used to measure in situ stresses with flat jacks and boreholes. Flat jacks were also used to determine the rock modulus.
Rock relaxation tests were also conducted in the first of the two diversion tunnels (circular tunnels with a finished diameter of 35 ft), approximately 200 ft north-west of the power-house area. The rock mass here is homologous to that in the power-house area and the depth of overburden is virtually the same. Rock bolt extensometers were installed in a radial pattern in five rings of five bolts each, at intervals of approximately 110 to 190 ft. They were anchored at the base and grouted securely at the collar of the borehole.
Measurements were to an accuracy of 0-001 in.
Since most of the convergence occurs within the first one or two heading advances following installation, it was important to install these extenso- meters as near to the heading as drilling would permit. The diversion tunnel was excavated with top heading full width, at which time the extensometers were installed. When the lower bench was removed, further convergence occurred. (Convergence of the tunnel causes the extensometers to record an elongation of the borehole in which they are installed.)
The method of computing the modulus is based on the assumption that displacements within the rock surrounding a tunnel cross-section are analogous to those in an elastic and isotropic material surrounding a hole in an infinite plate when subjected to biaxial stress. The in situ stresses in the rock were measured by flat jacks and confirmed by the borehole defor- mation method. In situ normal stresses of approximately 500 lb/in2 were found to exist vertically and horizontally, both parallel and perpendicular to the river channel. With uniform stresses in all directions, the analysis was somewhat simplified. The modulus was obtained from the following equation:
in which p is the uniform primary stress in rock mass (existing prior to excavation of the tunnel), a is the tunnel radius, r is the tunnel radius plus length of rock-bolt extensometer and d is the elongation recorded by exten- someter over the length of the bolt. The value of the Poisson's ratio deter- mined from core samples was assumed to be a reasonable approximation
for the rock mass and the value of 0-25 was used in the formula. (It is suggested that in even moderately jointed rock with a ratio of vertical and horizontal in situ stresses equal to one, a higher Poisson ratio should have been assumed.)
Main results. Rock at the dam site is hard, dense, generally fine-grained with specific gravity about 2-94. Unconfined compression tests on core samples indicated a Young's modulus averaging 12 900 0001b/in2 (910 000 kg/cm2) and a compressive strength of 40 000 lb/in2 (2820 kg/cm2). Tensile and cohesive strength determined from the same cores were 3300 lb/in2
(233 kg/cm2) and 5400 lb/in2 (381 kg/cm2) respectively. The rock is meta- morphic with high percentages of amphibole and is best described as an amphibolite. It appears massive although a rough schistose structure is not uncommon. Narrow bands of schists and a random system of well-healed joints and occasional shear zones are found throughout the test area.
Results from tunnel convergence measurements show that the modulus for each array of extensometers averaged from 1 200 000 to 3 310 000 lb/in2. The five arrays together averaged 2 600 000 lb/in2. The flat jacks yielded much higher values, ranging from 5*15 X 106 to 12*5 X 106 lb/in2 for ten sites and an average of 7-5 x 106 lb/in2 (530 000 kg/cm2). Modulus determined from the plate-bearing tests had to be corrected because the plate tilted. Values are given as follows: minimum E= 995 000 lb/in2 corrected to 1 272 000 lb/in2, maximum E = 1 555 000 lb/in2 corrected to 1 834 000 lb/in2, average (5 sites) 1500 000 lb/in2.
13 Incidents, accidents, dam disasters
A paper entitled 'Dam disasters' was read before the Institution of Civil Engineers, London, in January 1963 by E. Gruner. He mentioned a Spanish publication (Revista de Obras Publicas, 1961) which listed 308 dams where serious accidents had happened. The causes of failure were attributed as follows:
foundation failure 40 % inadequate spillway 23 poor construction 12 uneven settlement 10 other causes 15
Most of the foundation failures concern earth-fill and rock-fill dams built on pervious alluvium. The number of failures in concrete dam foundations is relatively small. A few are listed as follows.
(a) The designers of the Bouzey dam (France) ignored uplift forces and the dam collapsed on 27 April 1895. A committee, headed by Maurice Levy, was set up in the same year to investigate the disaster.
(b) The Gleno buttress dam failed in the same way on 1 December 1923.
Poor design (excessive shear stresses) and bad construction were blamed but Mary (1968), attributed the failure to uplift forces under the foundations.
(c) On 13 August 1935, one of the two dams at the Molare development in the Italian Appenines was destroyed by floods. The maximum flood expected on the small catchment area (141 km2) was about 850 m3/s. But during an unexpectedly heavy storm the floods rose to 2500 m3/s, over- topping the dam and deeply eroding the dam toe, which collapsed.
(d) In April 1959, the catchment area of the Rio Negro (Uruguay) severely flooded for a period of four weeks. The inflow into the Rincon del Bonete reservoir reached a peak of 605 000 cusecs. Of this, a total of 340 000 cusecs were discharged over and around the dam which flooded the power- house to generator top level, but did not destroy the river barrage.
(e) On 13 March 1928, the St Francis dam belonging to the Los Angeles Water Board failed through defective foundations. The rupture of the 16-m- high Moyie thin-arch dam (Idaho, USA) and the 19-m Lanier thin-arch dam (North Carolina) were also caused by flooding and rupture of the dam lateral rock abutment.
An unsuccessful attempt was made during the Spanish Civil War by the troops of General Franco to blow up the Ordunte Dam near Bilbao. In
[379]
September 1941 the retreating Soviet Army destroyed the Dnjeprogues dam.
On the night of 16 March 1943 the Mohne dam and the Eder dam in the Ruhr (Germany) were successfully bombed by the Royal Air Force.
Three more recent cases of dam accidents or dam disasters will now be analysed in detail. All three were caused by rock weakness.