Reservoir drawdown and sluicing Mechanical excavation Pressure flushing
8.28 Mm³ (original)
Date of commissioning:
Bakaru hydropower plant makes a significant contribution to power supply in its region, making it difficult to implement drawdown and sluicing, because this would involve the temporary shutdown of the powerhouse. Pressure flushing and mechanical excavation are implemented as additional measures to clear sediment.
The Bakaru hydropower plant represents the major power supply source in the South Sulawesi region in Indonesia, where the total installed capacity was about 400 MW in 2008. The Bakaru dam, completed in 1991, is located on the Mamasa River, as shown in figure 1. The Bakaru plant’s current installed capacity is 126 MW (comprising two 63 MW Francis turbines), operating under a maximum head of 322.2 m, with a design discharge of 45 m3/s.
The project consists of a 16.5 m high concrete gravity dam with a crest length of 122.5 m, fitted with two steel roller flushing gates, two steel roller regulating gates and four radial spillway gates. The maximum operating elevation is 615 masl and the minimum operating elevation 612 masl. The original reservoir volume was 8.28 Mm3 with a live storage capacity of 2 Mm3. The flood discharge capacity of the spillway is 2,500 m3/s. The dam is located at the upper edge of a significant drop in the river, where the original river bed slope (prior to sedimentation) was 1/570 immediately upstream of the dam, increasing to 1/5.8, with large amounts of rock immediately downstream of the dam, as illustrated in figure 2.
Hydrology and sediment
The catchment area of 1,080 km2 upstream of Bakaru hydropower plant produces an annual average mean flow of 1,794 Mm3, with an annual coefficient of variation of 0.22. The small variability in annual flows, as well as the mean distribution of monthly flows, are shown in the hydrographs in figure 3.
The sediment load in the Mamasa River at Bakaru is currently estimated as 988,000 t/yr (760,000m3/yr), which is equivalent to a specific sediment yield of 915 t/km2/yr. The bed load largely consists of sand, and the suspended load contains silt and clay. The catchment conditions are poor and erodibility high. The hydrographs for water and sediment discharge for three monitoring events, cumulative sediment inflow and a sediment rating curve are shown in figures 4, 5 and 6.
The main sediment challenges experienced at the project are storage loss and abrasion of turbines. By 2000, storage loss amounted to about 70 per cent of the total original storage, and then stabilised, as shown in figure 7, at about 73 per cent. The live storage was affected, but to an unknown degree. Sediment management efforts were intended to keep the reservoir volume stable at about 27 per cent of the original total storage volume. Damage to the turbines, as shown in figure 8, was significant due to abrasion by sediment flowing through the system.
The principal reasons for the high rate of sedimentation at Bakaru is that the actual sediment yield is much greater than that estimated during the design phase. The sediment yield used in the design was 133,000 m3/yr (about 172,900 t/yr), which amounts to a specific sediment yield of 160 t/km2/yr. The actual sediment load was later estimated at 760,000 m3/yr (about 988,000 t/yr), amounting to a specific sediment yield of 915 t/km2/yr. The sediment yield was therefore underestimated by a factor of five to six.
The project was originally designed to manage sediment through drawdown sluicing. The standard operating procedure as regards drawdown sluicing has not been implemented consistently, due to the fact that Bakaru supplies a significant amount of power to South Sulawesi. Dispatching needs overruled the implementation of sediment management procedures, as indicated further on in this case study.
Sediment management strategies
The project was originally designed to execute sluicing of sediment whenever the discharge in the river exceeded 400 m3/s. However, the low frequency of these flood magnitudes resulted in the frequency of sluicing not being sufficient. The standard operating rule (SOP) was changed in 2002 to sluice the reservoir whenever discharges in the Mamasa River exceeded 200 m3/s. Figure 9 illustrates the operating rule for a precipitation event in November 2008. The SOP decision process is shown schematically in figure 10.
In many cases when the reservoir could have been sluiced, the dispatch centre chose not to, giving priority to power generation. Figure 11 illustrates that during the period 2000 to 2009, there were 24 events when discharge exceeded 200 m3/s, but sluicing was only implemented nine times, and the duration of the sluicing events were limited to about four hours, which is not sufficient. When sluicing, the flushing gates were sometimes only partially opened to maintain the water level for concurrent power generation. In such cases, pressure flushing rather than drawdown sluicing was implemented.
The difference in the amount of sediment that can be removed when the gates are fully opened and when they are not fully opened is illustrated by the following example. On 4 February 2000, sediment was sluiced and power plant operation was completely suspended. This amounted to 863,000 m3 of sediment being sluiced while the daily inflow of water was 584.7 m3/s. Conversely, on 5 April 2001, a pressure flushing operation was carried out while simultaneously generating power; ie the water surface elevation was kept high while the gates were only partially opened. This resulted in removing about 256,840 m3 of sediment, while the average water inflow to the reservoir was 100 m3/s. Sluicing was not performed when the flows exceeded 200 m3/s, but pressure flushing was executed at a lower flow. The events in 2001, 2005, and 2009 were likely to have been pressure flushing events that did not follow the standard operating procedure to sluice. The relationship between the number of events when the discharge exceeded 200 m3/s and the number of sluicing events performed can be observed in the graphs.
After sedimentation problems became significant, the sluicing of sediment was complemented by dredging. Comparison of achieved and contracted mechanical excavation amounts of sediment are shown in figure 12, and figure 13 illustrates the activity.
The sedimentation problems at Bakaru are partially due to the sediment yield being underestimated by about a factor of 5.7 during design. This is a common problem in many projects, due to a lack of concurrent historic sediment concentration and water discharge data. This emphasises the importance of collecting sediment concentration and water discharge data on a regular basis, and over the long term.
Another challenge is that Bakaru contributes a significant portion of the energy demands of the South Sulawesi region, making it difficult to implement the SOP for sediment management when the opportunity occurs. The difficulty that arises if drawdown sluicing is not regularly carried out during high flow events is that the sediment that discharges into the reservoir deposits and is not conveyed through the reservoir.
Experience also shows that pressure flushing only clears sediment immediately upstream of the gates, as shown in figures 12 and 13. To remove deposited sediment further upstream it is necessary to implement drawdown flushing (instead of drawdown sluicing), which will likely be unsuccessful in the case of Bakaru because its reservoir is relatively wide compared to achievable flushing flows of 200 m3/s. To successfully implement drawdown flushing, it will be necessary to open more gates and use higher flushing flows, which is not likely to be possible given the important role that Bakaru plays in supplying power to the South Sulawesi region.