Addressing sediment problems

In most countries, sediment can be a major problem when developing hydro power projects. The regions contributing most sediment are western and southern US, southeast Europe, Asia, eastern Australia and New Zealand [Alam 2002]. About 20B tonnes of earth material are carried to the sea every year worldwide, of which nearly 6B tonnes are from the Indian sub-continent alone [Naidu 2000]. A maximum of 10,000 to 50,000 t/km²/year have been reported in China.

In the Himalayan regions, concentrations as high as 80,000 parts per million (ppm) lead to the accumulation of millions of tonnes of silt in reservoirs. Despite elaborate desilting arrangements, silt passes through generating units at the rate of thousands of tonnes per day [Naidu 2000]. The quartz content of the sediment can exceed 90% and causes heavy erosion leading to the removal of several tonnes of steel during every monsoon. Similarly, referring to Carson's 1985 article, Stole [1993] stated that sediment loads as high as 25,000 ppm are regularly recorded on major rivers such as Narayani in Nepal. Even higher sediment loads may be expected on smaller rivers with higher stream gradients and more potential for landslides blocking the whole valley. Moreover, Bishwakarma [1999] reported that the sediment concentration as high as 57,000 ppm was observed in the Jimruk river in Nepal during the monsoon of 1996.

A combination of factors leads to sedimentation. Soil erosion in tropical climates can be caused by high summer temperatures that cause soil to crack and grass to burn, leaving web-shaped spaces in the ground. The flooding monsoon waters then erode and transport soil. From an Asian perspective, additional factors responsible for soil erosion include: the immature geology of the Himalayas, glacial silt unleashed from melting snow, climatological anomalies-droughts followed by flood, landslides, uncontrolled tree felling, heavy pressures of grazing, and methods of cultivations [Naidu 2000].

As water resources are derived from precipitations, the development of water resources is bound to be affected by the pattern of precipitation. The Asian countries, particularly those in Southeast Asia, share a common pattern of precipitation, characterised by the occurrence of yearly monsoon or wet season. During a wet season of about four months, 60 to 80% of the annual precipitation can occur. Concentration of precipitation occurrences within a year while large variations in precipitation from year to year can lead to severe floods and droughts [Bingnan et al. 2000].

The main sources of sediment in Himalayan rivers are glacial deposits, land slides and intensively cultivated hill slopes. However, little qualitative or quantitative information is available on the sediment released from these sources. To combat further degradation of water resource by the river sediment, detailed knowledge is needed of the rate of supply, the characteristic size and shape of the sediment particles, hill slope and channel storage and downstream transport and attrition particles [Johanson, et al 1998].

Effects of sediment

Run-of-river projects are constructed to utilise the available water throughout the year without having any storage. These projects usually consist of a small diversion weir or dam across a river to diver the river flow into the water conveyance system for power production. Therefore, these projects do not have room to store sediments but should be able to bypass the incoming bed loads to the river downstream. The suspended sediments will follow the diverted water to the conveyance system. Settling basins are constructed close to the intake to trap certain fractions of the suspended sediment.

Many run-of-river hydro power plants built in sediment-loaded rivers are affected by sediment, both by a reduction in daily peaking poundage capacity and by rapid wear rates of the turbines and other mechanical equipment, such as gates and valves. The economic losses can be considerable because of the reduction in generation and the increase in the cost of repair and maintenance. The problem is acute in power plants built in Himalayan rivers. These rivers carry high sediment concentrations with high content of hard minerals like quartz and feldspar, which are highly abrasive, causing turbine wear. The hardness of quartz is 7 on Moh's hardness scale. Particles with Moh's hardness of more than 5 are harmful to hydraulic machinery. The intensity of erosion is directly proportional to the hardness of the particles irrespective of their size [Naidu 1997].

The Marsyangdi run-of-river hydro power project in Nepal was commissioned in 1990. Stole [1993] reported that Kayastha and Regmi studied this project and found that sedimentation in the small peaking reservoir upstream of the diversion dam reached its highest level after one year of operation and not 4-5 years as estimated during planning and design. The inflow of sediments to the settling basin was also several times higher than previously estimated. Similarly, Naidu [2000] reported that the Bira Siul project (3366MW) in Himanchal Pradesh handles nearly 10,000 tonnes of silt per day per machine during critical monsoon days. Material loss in guide vanes alone is 10% by weight. The trailing edge thickness of runner blades thins down to 2mm from 12mm. Approximately 15,000 welding electrodes weighing 1.6 tonnes are consumed per unit, per repair season. More than 90% of the silt passing through these machines is quartz.

Turbine wear

Flow containing suspended sediments of various sizes causes erosion in the turbine and its accessories. The places where the most severe sediment erosion occurs are in the guide vane systems and runner blade outlets of the high head Francis turbines and in the nozzles and buckets of the Pelton turbines. The erosion rate is directly proportional to the flow velocity. Brekke et al [1994] emphasise that the sediment-induced turbine wear problems cannot be overcome by the hydraulic design alone; however, the extent of damage can be reduced by careful design avoiding the strong accelerations and reducing the maximum velocity as much as possible. The extent of erosion depends on the turbine design, type of surface coating, velocity of flow, net head, and sediment characteristics.

According to Brekke [1991], when heavy sand erosion is expected, the Pelton turbine is normally chosen instead of the Francis turbine for high head operation (> 400m - 450m). This is because the time needed for exchanging runner and nozzles in a Pelton turbine is much shorter than changing guide vanes, seal rings and runner in a Francis turbine. Changing runner and nozzles in a vertical, multiple jets Pelton unit may be done in about four days while the work of exchanging and repairing a Francis unit will require approximately four weeks. Brekke [1991] further adds that the best weapon against sand erosion, however, may be the development of the ceramic coating, which will increase the time between necessary repairs by one to two decades compared with the surfaces of hardened stainless steel.

It has been observed that quartz particles contained in the sediment, even as small as 0.1mm, can cause significant damage. More attention therefore has to be devoted to the development of sediment resistant equipment. A choice has to be made between low initial investment, with subsequent higher maintenance costs as a result of sediment damage, or a higher initial investment to provide counter measure, with subsequent lower maintenance costs [Naidu 1997].

Laboratory test

Takagi et al performed a laboratory test to investigate the effect of suspended solids in water flow on the turbine performance. For this purpose, a model Francis turbine with specific speed of 100rpm was used. Alumina was used as a solid content in the water and the turbine performance and cavitations characteristics were measured by varying the solid concentration. The experiment revealed the following;

• The turbine's best efficiency decreased in a straight diagonal line by increasing the concentration of Alumina.

• The turbine's best efficiency unit rotating speed remained essentially the same for clear water.

• An empirical expression could be derived for the turbine's best efficiency.

• Clarification of the correlation between the critical Thoma cavitation coefficient and the Alumina concentration could be made.

Peihao et al [1996] also carried out a similar test for investigating the effects of water carrying suspended solid particles on the performance and operation of a Francis turbine. They concluded that the water containing suspended solids causes effects on the turbine performances such as efficiency, discharge, output, cavitations, etc. These effects will be more pronounced with the increase in solid concentration, characteristics of the suspended particles, net head, type of turbine and operating points.

Loss of turbine efficiency

The sediment induced wear not only damages the hydraulic machinery but also to a great extent reduces the efficiency of the turbine. Based on the efficiency tests carried out in the field, Brekke [1991] presents a comparison of the efficiencies of a Francis turbine during the period of sediment wear and after its maintenance. The study was conducted for Driva hydro power power plant in Norway, with an installed capacity of 71.5MW and net head of 540m. An increase of about 4% of relative efficiency was observed after the turbine repair. Pradhan et al [2004] reported that the sediment and turbine efficiency measurement carried out during the late monsoon (from 1 September to 11 November 2003) in this power plant revealed that about 4% of the hydraulic efficiency of the turbine was lost during the measurement period. The sediment load that passed through the turbine during the same period was about 6900 tonnes.

Planning and design considerations

The sediment process is a natural phenomenon. It may be possible to stop rapid increases in the sediment yield from the catchments by taking some precautionary measures, however full control of the sediments is not possible. Alam [2002] stresses the fact that the traditional approach of designing the reservoirs based on 50 or 100 years' lifetime, and then ignoring what will happen subsequently, should be abandoned. The dams can have significant impacts on river ecology and morphology, and if the reservoir fills with sediments it creates more serious problems. It is therefore essential to assess as accurately as possible the average annual sediment load entering the reservoirs.

To properly allow for the possible effect of sediment on the project also requires the planner/designer of the project to have a vision about the possible changes in future in soil conservation program and in land use, including the reclamation of land for agriculture, the construction of roads, the development of industries and the construction of upstream dams and reservoirs, among others [Bingnan et al 2000]. For instance, as a result of change in land use, the average suspended loads have increased 5 to 14 times after the completion of construction of Cameron Highland Scheme in 1963.

Bingnan et al [2000] suggest that the first step to be taken is to have more planning and design engineers informed of the grave consequences of underrating the importance of sediment problems. More training courses and technical functions, including symposia, seminars and workshops, should be organised to propagate knowledge of erosion and sedimentation.

Site selection

The sediment yield at the headworks site of a hydro power project depends on the location, catchment characteristics and the rainfall pattern. The catchment characteristics in this connection means the ground slope, land use pattern, soil type, vegetation cover, etc. The amount of sediment yield becomes significantly different for sites having different catchment characteristics even if the catchments are equivalent in terms of the covered area. Therefore, site selection plays an important role in estimating the amount of sediment that has to be handled during the operation of a project. A suitable site can help save a lot of money during operation.

Data collection

The primary task in handling the sediment related problems is to collect reliable sediment data from the proposed river during the early stage of planning. Without reliable sediment data, it is not possible to assess the impact of sediment-induced problems on the economy of energy generation and environmental consequences. Reliability of the sediment data depends on the sampling location, adopted sampling technique, frequency, skill of the involved personnel, etc.

Simanton et al [1993] emphasised the horizontal and vertical spatial variability of sediment concentration with the river flow and problems in getting the representative sediment samples. As practically and economically as possible, complete measurement of sediment data is necessary. Furthermore, Alam [2002] mentioned that the precise knowledge and information about the suspended concentration and the characteristics of the total sediment load that is being transported would enable the design engineer to foresee and provide adequate structural dimensions, structural layout and operation procedures which will ensure the sediment management.

With regards to the skills of stream gauging, the development is uneven among the regional countries. Only a handful of countries have extensive experience and necessary instrumentation to measure the bed load and to evaluate that part of the suspended load that passes through the 'unsampled' area near the bed [Bingnan et al 2000]. Measurement of bed load is tricky, but it is important for countries possessing mountainous rivers. Furthermore, as stream gauging is ordinarily time consuming, it is desirable if investigations are carried out on the possibility of extending the current remote sensing technique to determine depth, velocity and concentration of sediment-laden flows.

Appropriate design

Many attempts have been made to forecast sediment and water yield from a watershed, however it takes both insight and vision along with technical and socio-economic knowledge on the part of the planner/designer to arrive at a proper estimate or forecast of sediment influx and to plan or design the project accordingly [Bingnan et al 2000].

The design of different components of a hydro project is site dependent. A standard design cannot be applied in all projects to achieve the same level of performance. Therefore, the design of a project demands a wide range of experience so that pragmatic planning and design is possible for the sustainable operation of the project during its designed lifetime. In order to prevent the sediment entering the generating system, effective settling basin arrangements are important. In the civil design side, Alam [2001] suggests the following design considerations for a run-of-river hydro project:

• Simple structural shapes with adequate transition.

• Efficient sediment flushing arrangements.

• Trapping efficiency of the various particle sizes for a given sediment grain size distribution.

• Total construction cost of the installation, including flushing arrangements.

• Total costs caused by the loss of average annual power production during flushing plus the loss of excess water required for flushing.

• Total estimated costs of downtime resulting from abrasion damage or its repair.

• Total costs of plant shut down during excessive sediment concentrations in the water.

Furthermore, the existing modelling techniques, both physical and numerical, should be used extensively during the project design in order to obtain as much information as possible for the smooth operation of the plant after its completion.

Turbine design

In addition to a sound civil engineering design of the sediment control systems, the design of mechanical equipment should also be given equal importance. As mentioned earlier the hydraulic design of the turbine alone cannot fully solve the problem of sediment wear, however the extent of turbine wear can be reduced by careful design considerations. Naidu [2000] suggested guidelines should be adopted in the hydro-mechanical design of the turbine and its accessories. The main parameters that have to be taken into consideration during the turbine design are: turbine type, number of turbine units, turbine speed and specific speed, turbine design facilitating efficient maintenance, selection of base material, and application of appropriate surface coating.

Optimising the sediment exclusion

Sediment transport in the Himalayas is a natural phenomenon. Reliable and efficient systems for sediment control and removal of sediments from withdrawn water will therefore always be one of many preconditions for the successful utilisation of water resources in the Himalayas and other similar regions [Stole 1993]. In run-of-river hydro power projects, it is neither economically feasible to design settling basins which can trap all sediments that are coming along with the withdrawn water for power production nor it is possible to design a turbine with presently available technology which can take all the sediment load during its lifetime without any irreparable damage. Thus, there has to be a compromise between the investment on the turbine design capable of resisting sediment wear and the size of settling basin for preventing the undesired sediments reaching the power plant. Comprehensive research and development in this area has not been done so far, however some laboratory studies have been carried out in order to establish the relationship between the intensity of erosion and the sediment characteristics.

Naidu [1997] and Bishwakarma [1999] cite the following cost components to consider when carrying out an optimisation study with respect to sediment exclusion for a run-of-river hydro power project. The major cost components include; cost of construction and operation of the sediment exclusion system, cost of turbine repair and maintenance due to sediment-induced wear, cost of turbine replacement due to sediment induced wear, cost of loss of efficiency due to sediment induced wear, cost of generation loss due to repair and maintenance or replacement, and cost of power delivery cut-off or curtailment due to sediment problem.

Moreover, Bishwakarma [1999] emphasises that the value of water must be analysed before being utilised for power production. An economic analysis based soley on availability of water is not satisfactory during the operation, because during periods of high sediment concentration the run-of-river power plant has to be closed, which causes generation loss. If the plant is not closed, then the excessive sediment exposure will lead to damage on the machinery resulting in a high operation and maintenance cost. Naidu [2000] emphasised that the annual operation and maintenance costs of sediment-affected power stations can be as high as 5% of the capital costs, against 1.5% in normal cases. Therefore, for optimal operation of a power plant, an optimisation study is necessary to find out the level of concentration limit in the water. Instrumentation for measuring sediment concentration in real time plays an important role for adopting an optimal plant operation regime.

Real time sediment monitoring

The importance of a real time sediment monitoring system is two fold. Firstly, it provides an early warning system showing the level of sediment concentration in the water entering the hydraulic machinery. Secondly, it facilitates recording the sediment exposure of the turbine over time. Sediment concentration analysis in the laboratory takes time, and often it is too late to have information on the sediment concentration after the high sediment load has passed through the turbine.

Dr. Haakon Støle of Sediment System AS conceptualised a tool for sediment monitoring in hydro power plants. This tool is labelled as the Sediment Monitoring and Operation Tool for Hydro plants (SMOOTH) in cooperation with NTNU. A pilot test of SMOOTH was conducted at the 12MW Jhimruk hydro power plant, Nepal, and in the laboratory at NTNU. The results revealed that the tool was able to observe pipe flow as well as sediment concentration in the water flowing through the turbines with an acceptable level of accuracy. Støle and Karki [1999] reported that the tool could be used to produce real time data on sediment content in the water and can therefore be used to adopt an operation strategy to optimise the power production. However, further research and development was required on the hardware and software side.

Bishwakarma [2005] further developed SMOOTH by testing comprehensively in the laboratory as well as in hydro power plants in Nepal and India. He also developed a computer program for instrument reading, data processing, logging, and real-time display in the computer screen. Now the SMOOTH package is able to record the sediment concentration as well as the parameters required to compute the turbine relative hydraulic efficiency in real-time.

Sediment removal techniques

One of the challenges in sediment handling is to have a proper technique for removing the deposited sediments from the peaking reservoirs and settling basins. While designing the sediment exclusion systems, the removal of accumulated sediments as well as its environmentally friendly way of disposal must also be considered. Flushing of the big reservoirs is beyond the scope of this paper, so the main emphasis has been given to the flushing of settling basins and small reservoirs. The following new sediment removal techniques for the small reservoirs, head ponds and the settling basins were developed during the early 1990s.

Serpent Sediment Sluicing System (S4)

Støle [1993] invented the 'Serpent Sediment Sluicing System' (S4) concept in 1988. S4 facilitates the continuous operation of settling basins and head ponds of small reservoirs. The flushing process is intermittent. An S4 installation is flexible, and the flushing frequency may be adjusted to comply with the sediment load of the river [Støle 1993]. The flushing process is carried out using gravity forces alone, and therefore no power input is required. The two modes of operation, named 'opening' and 'closing' modes, has made the system robust, eliminating the risk of a 'point of no return' situation during operation. Monitoring and operation of the headworks with S4 is simple and it can be operated manually or automatically [Lysne, et al 1995]. S4 has been installed in a few power plants in Nepal and performing satisfactorily.

Sediment removal by pipeline

Jacobsen [1998] developed a concept of pipeline removal of deposited sediments. On the basis of this concept he developed two types of sediment sluicing techniques namely 'Slotted Pipe Sediment Sluicer' (SPSS) and 'Saxophone Sediment Sluicer' (SSS). The pilot tests were carried out in Nepal during 1994 and the tests revealed that the operation of both of the techniques was satisfactory in real conditions. Manufacturing of these sediment sluicing systems requires limited resources and low investment. Operation is also easy, and minimal training is required [Jacobsen 1998]. Therefore, Jacobsen claims that these methods should be suitable for sediment sluicing in any country. Both of the techniques have been tested at the laboratory as well as in the field.

Conclusion

In planning a run-of-river hydro power project on a river carrying relatively high sediment loads, appropriate attention should be given to the sediment management aspect. Plans for future watershed development should be collected from relevant agencies for analysis to ascertain possible effects on the sediment yield of the watershed. It is deplorable that all developing countries are not equipped adequately for the collection of relatively complete hydrological and sediment data.

A common knowledge or general awareness of sediment problems is not enough to tackle this issue. Hydro power engineers must be 'sediment conscious' during investigations, design, operation and maintenance, and even upgrading and refurbishment. More research and development is needed into the causes and mitigation of sediment erosion impacts. Experiments and field observations are needed to understand the true relationship between the sediment characteristics, wear resistance of base material, relative velocity of water, angle of attack of sediment particles, and chemical properties of water.

It is very important that more planning and design engineers be informed of the grave consequences of underrating the importance of sediment problems. More engineers should be trained in sediment engineering in order to raise a proper consciousness of the importance of sediment problems in the benefit of better planning, design and operation of run-of-river plants.

Sediment exclusion optimisation so far has been a challenge. Therefore, it is recommended to install efficient monitoring systems so that a comprehensive database can be generated. The use of a developed real-time sediment monitoring system for guiding the power plant operation together with an efficient flushing system may contribute immensely in mitigating the sediment-induced problems in run-of-river plants.

Meg B. Bishwakarma, PhD Fellow, NTNU, Norway, General Manager, Hydro Lab, Kathmandu, Nepal. Email: mbb@hydrolab.org or meg.bishwakarma@ntnu.no