Phase 1A of the Lesotho Highlands Water project (LHWP) is practically complete, with the commercial transfer of water having begun in January 1998. The principal elements of Phase 1A include the majestic 185m high, concrete double curvature arched Katse Dam; the 45km long Transfer Tunnel; the 72MW Muela hydropower station and Muela Dam; and the 37km long Delivery Tunnel. Phase 1A boasts a firm yield of 18.1 cumec, with a maximum capacity, when Katse Dam is full, of 32.6 cumec. (Fig 1 shows elements of phases 1A and 1B.)
Currently under construction, Phase 1B comprises the 145m high, concrete faced rockfill Mohale Dam on the Senqunyane River, its reservoir being connected to Katse Reservoir via the uncontrolled 31.8km long Mohale Tunnel. A small but important part of Phase 1B is the 19m high, mass concrete Matsoku Weir and associated 5.7km drill+blast Transfer Tunnel, designed to divert floodwaters from the Matsoku River to Katse Reservoir. Approximate increases in LHWP yields for the Mohale and Matsoku elements are 9.7 and 2.2 cumec respectively. The Matsoku Diversion is cost effective and will come on stream about a year earlier than the Mohale Transfer.
Phase 1B planning
An in-house team from LHDA (in conjunction with Acres International (Canada) and the engineering division of South Africa’s Department of Water Affairs and Forestry) undertook the planning study of Phase 1B major works. This study concentrated on the Mohale area in respect of: preferred dam site location; optimum full supply; Mohale-Katse Transfer Tunnel layout; cost optimisation studies; and environmental considerations. In an effort to minimise costs attributable to the complexity of constructing the Mohale Tunnel after Katse Reservoir is impounded, the outlet works were built as part of the Phase 1A construction of the Katse Dam.
The planning study called for a feasibility study which included such elements as: geological conditions; long-term maintenance; operations; environmental considerations; and programme and cost.
There were to be three contracts: the Mohale Tunnel; the Mohale Dam; and the Matsoku Tunnel and Weir.
Mohale Tunnel
This tunnel has an intake in the Mohale Reservoir and an outlet in the Katse Reservoir. There is a shaft at either end where the intake and outlet tunnels respectively can be isolated from the waterway tunnel. The intake at Mohale consists of upper and lower bellmouth intake structures some 27m apart in elevation, connected by tunnels to a 92m deep, 9m i.d. intake shaft.
The subterranean gate operating gallery at the shaft top is 17m long x 12m wide x 9m high and will house a 400 kN capacity overhead gantry crane for operating the gates and stop logs. The gallery is accessed via a 75m long drill+blast adit. The TBM will be launched from the lower intake portal excavation (Fig 2)
At the outlet, permanent arrangements at Katse, constructed as part of the Katse Dam contract include: the outlet bellmouth structure; the 6m diameter, 130m deep outlet shaft; and a 220m long, 3.4m diameter concrete lined section of the waterway tunnel between the outlet and shaft. The outlet bellmouth in the reservoir is sealed at present by a steel bulkhead.
The outlet drive TBM will be assembled and launched from an enlarged underground chamber along the waterway, which will be accessed by a 700m long drill+blast adit to be excavated down a 1:8 grade (Fig 3). The waterway tunnel is about 31.8km long and will be constructed on a continuous downward grade of 1:3261 between the intake and outlet shafts. A short 942m length of waterway at the outlet between the TBM launch chamber and the outlet shaft will be excavated by drill+blast, the rest of the waterway being driven by two TBMs. Under normal operating conditions, the flow in the tunnel varies from zero to 50m3/s, depending on the differential head in the two reservoirs.
There are no controls for flow in the tunnel but gates will be provided in each shaft and will be either fully open to transfer water from one reservoir to the other or fully closed to stop transfer of water. The gate at the intake will be partially opened to fill the tunnel, the size of the opening limiting the filling rate to 7m3/s. The full supply level (FSL) of the Mohale Reservoir is 2075m asl and, for Katse Reservoir, 2053m asl. The respective minimum operating levels (MOL) are 2005m and 1989m asl.
The Mohale Tunnel will be constructed in the basaltic lavas of the Lesotho/Drakensberg Formation along its entire length. Dolerite intrusive bodies are generally observed to form near vertical dykes or sub-horizontal sills. The classification system for the tunnel excavation consists of four face classes based on varying proportions of amygdaloidal, non-amygdaloidal and intrusive rocks occurring in the face.
The amygdaloidal basalts are sometimes associated with low durability/rapid degradation and the results of the durability investigation programme undertaken for the Mohale Tunnel led to the recommendation that the tunnel be constructed as a fully lined tunnel.
All access adits have been excavated by drill+blast. Hand-held methods were used at the three shorter and smaller adits at the intake and a 2-boom drilling jumbo for the longer and larger cross section access adit at the outlet.
The 9m diameter intake shaft was sunk by first raiseboring a 1.8m diameter central bore and then reaming out using drill+blast methods with a single platform stage. The waterway tunnel is to be driven by a 5.4m diameter Wirth TBM and a 4.9m diameter TBM manufactured by NFM, Boretec and Mitsubishi (the NFM TBM). Both TBMs are telescopic double shielded machines and the cutterheads are fitted with 423mm cutters. The Wirth TBM is hydraulically driven, while the NFM is electrically driven, and both have variable speed drives. Both of them are secondhand and have been refurbished.
The philosophy of refurbishing the two TBMs has been quite different in each case. The Wirth machine had been mothballed since 1995, when it completed the Delivery Tunnel North for Phase 1A. The contractor therefore decided to carry out a full refurbishment, including some upgrading to modernise it. The NFM machine was brought directly from Ecuador immediately after it had finished a 12km drive there, so it was decided to carry out basic refurbishment of obvious problem sections and to get the ‘operational’ machine into the new tunnel as soon as possible. In both cases, a new main bearing has been fitted and the used bearing kept as a spare, as both of them were in good condition.
The waterway and intake drill+blast excavations for the permanent works will be lined with concrete cast in-situ. The intake shaft will be slid. The access adit and gate operating gallery will be lined with mesh reinforced shotcrete.
The precast concrete segmental lining forms a major part of the tunnel construction and a large segment factory has been set up at each end of the tunnel to manufacture segments. At the intake, the factory has been erected adjacent to the portal, but at the outlet, this has not been possible because of restricted space at the portal site.
The outlet factory has therefore been erected on the quarry site and the segments will be transported to the portal, where they will be loaded on to flat cars, shunted on to a shuttle, which is then lowered by winding hoist down the 1:8 access adit to the TBM handling chamber.
Each factory has 48 steel segment moulds and the contractor is aiming for a peak production rate of three casts in 24h (144 segments/day). A long-term average of 21¼2 casts/day is anticipated. Each tunnel ring consists of four segments. The intake segments are 1.4m wide and the outlet segments 1.3m wide.
The contract started on February 2 1998 and the stipulated completion date is January 3 2003 (end of month 59). The contractual start date for filling and testing the tunnel (when water is expected to be available in Mohale Reservoir) is October 31 2002. However, the stipulated completion date at the intake for all work below FSL is October 2 2001 (end of month 44), when the Mohale Dam will begin impoundment.
March 9 2002 was the date in the contractor’s tender contract programme to be ready for testing. After contract award, this date was moved forward to December 15 2001 in its proposed contract programme, which is some 11 months early, thus providing considerable float. However, should the intake TBM run late, this drive will have to be curtailed and the outlet TBM will drive the corresponding additional length.
Current position
The establishment at both sites is virtually complete. The segment factories are operational and aggregate is being supplied from two quarries. All drill+blast excavation work at the intake and all critical drill+blast work at the outlet has been completed. The intake TBM drive started on March 11 1999, about three weeks later than the programme start date. At the end of July 1999, this drive was 16 weeks behind programme, which was largely due to the fact that a longer time was taken for commissioning and the learning curve for the tunnelling crews was longer than anticipated. The contractor does not consider this delay to be serious at this stage, bearing in mind the 11 month float. A total length of 1276m was finished by August 11, of which the first 326m is unlined and on a 600m radius. This section of tunnel will be lined with an in-situ concrete arch cast against the installed precast invert segments once the TBM drive is finished.
The outlet TBM drive started on July 26, some eight weeks behind the contract programme. Thirty-three metres had been driven by August 11. The intake and outlet segment factories had produced 7900 and 4500 segments respectively by this date. Some 18 months after commencement, the contractor is now in a position to begin a sustained effort to complete the TBM tunnels.
Matsoku Tunnel
Because the transfer of water in the Matsoku Tunnel will range from zero under low river flow conditions to a maximum of around 55 cumec when a 200 year return period flood occurs in the Matsoku River, it was decided early in the design process that the tunnel would be designed as a free water surface conduit to prevent scenarios where the flow could ‘make and break’ during the transition from open channel flow to a pressure conduit, albeit at low pressure.
Another requirement was that the Froude number should remain below 0.7 so as to avoid an unstable flow regime. It was also necessary to ensure a velocity of around 2m/s under all but the lowest flow conditions to reduce silt accumulating in the tunnel. It was imperative that the outlet should be designed to discharge above Katse full supply level because there was no possibility of deliberately lowering the Katse Reservoir level during construction, and a cofferdam was considered an unnecessary expense. Of course, the tunnel had to be at an elevation that would prevent any reverse flow from Katse to the Matsoku Valley.
The LHDA decided that the tunnel would be excavated by drill+blast to make the maximum use of the experience many of the Basotho workers had accumulated by working in South Africa’s underground mines. The shape finally selected was an excavated 4.5m span modified horse-shoe having an excavated area of 18m2 (Fig 4).
The tunnel was designed with a concrete floor, which was considered to contribute more than sidewalls would when the hydraulic roughness of the conduit was assessed. It was also decided to design and pay for a temporary concrete invert (blinding concrete), which was to be at least 75mm thick and placed within 100m of the advancing face at all times. Benefits included protection of possibly degradable basalt in the invert where water was always present during construction. Furthermore, it would provide the contractor with an even floor, encouraging installation of high-quality rails as well as a surface which would be easy to clean during the final lining and concrete floor construction stages, which are on the tunnel critical path.
Design studies indicated that shotcrete lined walls would be more economical than concrete, particularly for the non-vertical sidewalls of the selected cross section. Shotcrete lining was extended to 1.5m above springline, above the highest expected water level. The crown would be left unlined except where rock conditions dictated otherwise. Circular or circular segment crowns were considered, based on the experience of achieving good profile control in basalt in drill+blast excavation of the adits for the Transfer Tunnel in Phase 1A and the resultant lower rock support levels needed because of the structural stability of the arches.
The tunnel gradient selected to suit the chosen weir location and the minimum velocity and Froude number constraints was 1:230.
The tunnel intake design was incorporated into the weir construction. A channel with a capacity of 650 litres/s was provided below the sill of the intake and discharging through the weir low level compensation outlet works via a 600mm diaphragm valve. No water is therefore transferred when the river carries less than this flow. When flow in the river increases, the diaphragm valve automatically controls the throughflow to the required 650 litres/s, and the flow over the tunnel intake sill passes through a coarse trash screen to exclude large objects from entering the tunnel.
In order to prevent excessive flow in the tunnel, a rectangular orifice 2m wide and 2.5m high has been designed to allow only 55 cumec through, even under conditions of safety evaluation flood of 2710 cumec in the Matsoku River. The jet from the orifice impinges on a baffle wall in a stilling basin, where the energy is dissipated and the water surface stabilised before water flows in to the tunnel.
The ‘spongey’ nature of the highly amygdaloidal basalt layers effectively creates mixed face conditions for blasting, requiring higher powder factors than might otherwise be expected. This, together with the advantage gained by matching cycle times to the shift pattern, results in an optimum drilled round length of 3.4m.
Excavation and mucking is performed from both ends using rail mounted equipment. The rounds are drilled using a 2-boom Atlas Copco Raildrill 282 jumbo. Mucking is carried out by a Haggloader type 8HR-2, with a 750mm conveyor loading a Rock Machines type-C self-discharging shuttletrain.
Initial rock support is governed mainly by the lava flows, that is, by the presence or absence of amygdales and flow contacts in the upper tunnel profile. The presence of large sub-vertical planar joints striking sub-parallel to the tunnel sidewalls often also requires early support. Initial support generally takes the form of pattern rockbolting over the crown and upper sidewalls, with unreinforced shotcrete where highly amygdaloidal basalt occurs in the upper profile.
At the end of July 1999, 18 months into the 37 month contract, 1288m had been excavated from the outlet portal, and 746m from in the intake or Matsoku portal. While it is true that this critical path activity is behind programme, a concerted effort by the contractor, the consulting engineer and the owner has been partially successful. Since the end of March 1999, the projected breakthrough date, based on projecting actual performance over the proceeding two weeks, has not changed from mid- September 2000, some 32.5 months late. However, the quality of workmanship is good and the project team is currently working on a plan to recover the time lost to date.
Related Files
Layout of Intake
Layout of Outlet
Location Map
Cross Section
