The most impressive element of the Kariba power scheme is the high double- curvature arch dam impounding water in the Kariba Reservoir. Most of the geological investigation on the site was concentrated on the dam foundations and on the rock abutments on both sides (see section 11.4.3) and on the reinforcement work on the South Bank.
The geology of the Kariba area was recorded and interpreted by the late Dr Francis Jones and Dr Dubertret (Paris). Dr J. L. Knill and Dr K. S. Jones have published an extensive geological description of the area, including the areas where the South Bank and the North Bank underground power houses had to be located. They reported that:
The foundations of the dam are mainly composed of biotite, with occasional layers of amphibolite which are more abundant on the south bank. A thick band of quartzite, folded into a syncline, outcrops on the upper part of the south bank. The quartzite is inter-stratified with the gneissic succession. Frequent dykes of pegmatite are present and the gneiss is so ramified by granitic veining that it locally grades into migmatitic rocks. The main trend of the foliation and the individual lithological horizons is north-east to south-west, although there is both regional and small- scale evidence for a later folding on axes normal to the dominant trend. The structural phase which has affected the bedrock at the site most significantly, from an engineering viewpoint, is the minor thrusting which resulted in the formation of the fractured synclinal flexure on the upper part of the south bank. This fold, which is associated with deep weathering in and below the quartzite, has created special engineering problems. A fault trending parallel to the gorge, as well as several on echelon splinter faults, has been located near the river bed and on the north bank.
The geology of the north bank is straightforward in that the sound gneisses are overlain by a weathered layer of fairly uniform thickness (10 m approximately).
[435]
\ Lusaka,
>
It
\ằ \?
Kariba access roads
— — Principal roads
-t-t-w-i- Railways L ^ ^
International ~ ^ r ^ boundaries ^ f {
f—
fChilanga Z g g g ^ ^
) A ,i r u n u. ô
^^••^^it K . a r i b a \ ^
f \ ^ Lions'den J V^ ><^ SinoiaT
^ C ^ u l a w a y o |
1 ^
0
"I miles
r / Olinrvi
/ i i i ^ n a m
s^Jmvuma
1fFort Victoria
150
— ^
\
^ ^
—J J
N /
UmtaFT \
V.
\
/
\
.y r
Fig. 16.1 Map showing Kariba and Kafue power-stations and access facilities\
(after Olivier, 1961).
biotite
gneiss
Y
0
i
schist ^)
? Lake
m
migmatite '" -7 ' ' '/
Kariba
300
i
/
/
quartzite y • ^\ •
f
'/ * mica seam
IT /
r *''
!
Fig. 16.2 Geological map of Kariba Dam (after Knill & Jones, 1965).
Kariba South Bank 437 m
500
450
\100%
100%
adit A l
adit Bl
quartzite
micaceous quartzite gneiss
amphibolite
-weathered -sound
Fig. 16.3 Geological section through the south bank of Kariba, illustrating the correlation of information obtained from boreholes and adits (after Knill & Jones, 1965).
The exploratory boreholes and adits, carried out before excavation of the dam foundation, provided an accurate assessment of the geological conditions. The investigations, carried out in connection with the Second Stage power station, further confirmed this geological picture.
The geology of the south bank is more complex in that the gneisses are overlain by quartzites which frequently dip steeply towards the river in a downstream direction. The quartzites are highly fractured and sheared, the individual joint surfaces being grooved and slickensided. The joint blocks of quartzite are generally only moderately weathered owing to a loss of cementation, but the intervening fractures are infilled by clayey and silty material. Intercalated bands of gneiss and micaceous quartzite occur within the main quartzite mass. A zone of soft micaceous gneiss, composed largely of coarse-grained biotite and muscovite, referred to as the 'mica seam' forms a curved zone within the quartzite. The mica seam appears to be inter-bedded with the quartzites and the boundaries of the seam are approximately parallel to the quartzite-gneiss contact. The outcrop of the mica seam above the dam was strongly sheared to a narrow strip about 3 m in width, yet in depth the breadth increased considerably. The contact between the gneiss and the quartzite has a curved profile, with the axis of curvature lying parallel to the valley.
The original investigations on the south bank (1956) were as usual by boreholes, pits and adits. It was realized that the bank in the fractured quartzite was structurally suspect as an abutment and that an extensive pro- gramme of jetting and grouting was required (Lane, 1964). On the surface, there was no sign of the mica seam, which was detected during the excavation of an access road. A new series of adits, at two levels, and boreholes were used
to determine the real structure of the rock. The quartzite was found to be deeply fractured and permeable, the gneiss weathered.
The consolidation work of the south bank has been described in section 11.4.3. Problems of rock elasticity, compressibility of the mass, its bulk resistance to shearing forces and the general watertightness of the abutment on the right-hand side of the arch dam were of paramount importance to the dam designers.
The survey of the south bank confirmed that the underground halls were located almost entirely in biotite gneiss, injected in places by some dykes of granite pegmatite and would in general lie below the zone of superficial alteration where the gneiss has disintegrated and become friable (fig. 16.4).
gneiss, underlying the quartzite
•*•**•* *
14.
v+ +
===== joints,
===== slickensides
quartzite ^ gneiss, overlying
b^2i the quartzite ;o strike, dip in degreees
Fig. 16.4 Kariba South Bank underground power-station. Typical cross-sections showing geology of site and locating the excavations in gneiss (after Olivier, 1961).
16.1.2 The design of the underground works (fig. 16.5)
The 128-m-high double-curvature arch-dam at Kariba stores the flow of the Zambesi River in the immense Kariba Reservoir. According to the records available in 1961, the flow at Kariba varies from 8000 cusecs (227 m3/s) to possibly 570 000 cusecs (16 150 m3/s). The firm flow is estimated to be 42 000 cusecs (1190 m3/s). The maximum water level was designed to be at 1590 ft with the dam crest at 1616 ft. The six 100 MW generating sets will generate an estimated 8500 million kWh a year.
Figure 16.5 shows the underground works with six vertical intake shafts, four permanent intakes at level 1510 and 2 temporary intakes at level 1370, the great machine hall, and the associated transformer hall. The discharge from the six turbines is through three unlined, horseshoe-shaped tailrace tunnels, 34 ft (10-40 m) in diameter, each provided with a concrete-lined surge chamber, 63 ft (19-25 m) in diameter.
The machine hall has a length of 468 ft (142-75 m), a width of 75 ft (22-85 m) and a height of 132 ft (40-25 m), the top of the excavation being at level 1326.
The concrete arch of the underground cavern has a radius of 36 ft (10-98 m)
RL 1370 V,
R
screen-rake
gantry. stop-beam,x gantry /sam //
max. flood storage level 1605 v max. retention level 1590 vyy min. retention level 1560 y->
screens R
/
gate sections through intake tunnels shaft j p s
\ Nos. 1, 2, 3, temporary nos. 1,2 A 4, 5, 6 16'-0" x 16'-0"
li\ sector 20'-0" x 16'-0"
1 V
20 ft dia./
:oncrete V penstock^
at level 1560 at level 1420 sections through penstocks
temporary low-level intakes 1 and 2
diameter steel-lined
penstock machine hall
shaft completed and temporary intake plugged after sector gate transferred to permanent position
Fig. 16.5 Kariba South Bank. Underground works, typical cross-section (after Olivier, 1961).
and a height of 45 ft (13-73 m) and is thus nearly circular. The general shape of the machine hall is ovalized, a shape well in line with the theoretical think- ing at the time it was designed.
16.1.3 The excavations, linings and grouting
The sequence of operations for excavation and concreting is significant: (1) the pilot galleries, one at the crown of the vault at level 1326 and two others, 40 ft (12-2 m) apart, at level 1307 (see fig. 16.6) were constructed and a 68-ft
Fig. 16.6 Kariba South Bank. Sequence of operations for excavation and con- creting of the machine hall (after Olivier, 1961).
(20-7 m) wide first-stage vault excavated and concreted. (2) Work was carried out in bays from between 8 to 24 ft wide (2-4 to 7-3 m), depending on the quality of the rock, with concreting of the bays following immediately afterwards. (3) Thereafter bulk excavation continued in bays down to level 1281, together with (4) concreting of the second-stage vault, 75 ft (22-9 m) in width. Once the concreting of the vault had been completed, bulk excavation in the machine hall was taken down by steps to level 1194 (5 & 6).
Commenting on the underground works, Olivier (1961) remarks that arch forms were generally used to form vaults of permanent excavations, the most important being the vault of the machine hall. Near-circular forms were adopted wherever possible for shafts and adits. Initially, it was intended
Kariba South Bank 441 to use a complete eggshell lining for the machine hall, but this was rejected on account of the greater cost and the difficulty of construction.
For the much smaller transformer hall, a similar programme was adopted.
The volume of excavation for the machine and transformer halls, adits, shafts and tunnels was 680 000 yd3 (519 000 m3) against 460 000 yd3 (351 000 m3) for surface excavations.
The average head of water to the centre line of the power-house was taken as 200 ft (91-5 m) and the linings were designed to take full water pressure. In addition, the rock behind the linings was grouted to reduce any free flow from the reservoir to the structure. Most rock joints appeared to be too fine to be grouted but there is no doubt that pressure will build up behind the linings, hence a comprehensive system of drains was provided to assist in relieving pressures against the linings. It was argued by the designers that grouting also assisted in consolidating rock which had been disturbed by blasting. It is known that redistribution of stresses in the rock occurs mainly during excavation, prior to concreting. Grouting of the power house and transformer hall was therefore purposely delayed to allow internal movements to take place.
Further strains will of course take place during the filling of the reservoir.
The diversion tunnels were left unlined, but rock bolting was resorted to during construction as a safety measure.
Accoustic strain meters were embedded in the concrete during construction of the machine-hall vault and in the transformer-hall vault, in the end walls of the machine hall and in two draught-tube gate shafts. It was found that a general state of compression exists along the intrados and extrados of the machine and transformer-hall vaults (see Olivier). In some cases there was evidence of minor tensile stresses in the extrados. It may be that there has been a slight initial inward deflection of the springing of the vaults. The maximum stress in the concrete is of the order of 1000 lb/in2 (70 kg/cm2), and it is believed that the concrete is working monolithically with the rock.
Olivier reports that, in the machine hall, one of the upstream transverse buttresses opposite a faulted zone (concreted between the machines to act as temporary struts until incorporated in the mass concrete foundations) has been subjected to sufficient compression to cause shear cracks in the concrete.
The indications are that the other buttresses were also subjected to high compression.
Extensive use was made of grouting to consolidate the rock. Radial holes were drilled in a 15-ft grid in the two major vaults, and grouted in three stages:
(0 3 ft (0-92 m) into rock, grouted at 20 lb/in2 (1-4 kg/cm2) (contact grout- ing between concrete and rock).
(ô) 10 ft (3-05 m) into rock, grouted at 50 lb/in2 (3-5 kg/cm2).
(HI) 30 ft (915 m) into rock, grouted at 150 lb/in2 (10-5 kg/cm2).
Unlined rock faces and skin walls were equally grouted. A lake-side curtain was provided to protect the shafts until the linings of the shafts were sufficiently
high to counteract the effects of the rising reservoir. The main object was to consolidate the jointed quartzite in continuity with the dam curtain in 32 months. Approximately 131 614 ft (40 142 m) were drilled, in 7003 holes, taking 8389 tonnes of cement, and 2663 tonnes of sand.
16.1.4 New rock support techniques. Comments on Kafue underground machine hall
The tensile stresses measured along the extrados of the concrete arch vault linked to possible inward movements of the side walls of the excavation. A more detailed analysis of such a problem would have indicated that the wall movements depend not only on the shape of the vault, but also on the quality of the rock on each side of the cavern and on possible reactions from other nearby excavations, such as transformer hall or surge chambers. The prediction of the wall movements, on which the stresses in the concrete arch depend, is most difficult. In fact, the concreted vault acts more as a lining to the rock vault, and not as an independent loaded arch. The concrete vault deforms as rock deforms. This is also the general assumption most mathematicians make when using the finite element method for calculating strains and stresses in the concrete lining of the rock vaults.
Shortly after Kariba South Bank power system had been put into service L. von Rabcewicz outlined the new Austrian tunnelling method (NATM) which he used mainly in galleries and tunnels. Very keen designers used the same method for large excavations like underground power houses (Rescher (1968) on Veytaux, Mantovani (1970) on Lago Delio, etc.).
The Kafue underground power station on the Kafue River, upstream of Kariba, is a very good example of designs adopted at that time. Sten Rosenstrom describes it (1972), as shown on fig. 16.7.
Photoelastic tests were carried out to get the stress distribution around the machine hall and the transformer Hall. The tests were carried out for a horizontal principal stress ph = 200 kg/cm2 and a vertical stress pv = 1 3 5 kg/cm2. The tests showed that tensile stresses might have occurred in the walls which could be avoided by proper shaping of the cavern. During the final design of the profiles of the cavern it was specified that the walls were to be as flat as possible, and that transition radii were to be as large as possible.
(Similar results were obtained when testing other caverns like Waldeck II, Germany, and Ruacana on the Cunene River.)
Rosenstrom mentions that, in order to ensure the stability of the rock vault, the roof surfaces were rock-bolted and netted. The high walls of the machine hall were also bolted, the bolting and netting being carried out con- secutively during the blasting operation. Figure 16.7 shows the importance of the anchorages used. The longitudinal crane beams, and the wall slabs above them in the machinery hall, were concreted against the rock walls from the first bench. They are anchored into the rock. In both caverns, spaces of 30 cm
crane beam and wall slab, concreted from bottom of first bench *
, M !! i !
rock anchor bars in zig-zag
Kariba North Bank 443
smooth blasting, deformation — ^ -'^measurement poles presplitting
first tench
•1911
Fig. 16.7 Kafue underground power-station (after Rosenstrom, 1972). A vertical section through the underground power and transformer caverns showing the routine bolting of roof and wall surfaces with 1 in (2*5 cm) deformed bars, having a yield point of 40 kg/mm2. The 4*5 m long bolts placed in the roof were at 2 m centres in a zig-zag pattern, and the 6 m bolts in the walls were at 2 m centres. The roof surface was also secured by mesh (after Rosenstrom, 1972).
against the rock had been planned for roof arches, but these arches were not constructed.
The photoelastic model test had shown heavy rock stresses. Disturbances due to such stresses were not observed in situ, although in the transformer hall a fissure was observed on part of the roof in a direction approximately parallel to the cavern axis.
The example of Kafue is an interesting link in the sequence of events to be related in the next paragraphs of this chapter.