Model Study Of Sediment Flushing In Rantambe Reservoir

3rd International Symposium on Advances in Civil and Environmental EngineeringPractices for Sustainable Development (ACEPS – 2015)Model Study of Sediment Flushing in Rantambe ReservoirC.R.Ratnayesuraj1, K.M.Rifas1, S.Risly1 and S.B.Weerakoon11Department of Civil Engineering, Faculty of Engineering, University of PeradeniyaPeradeniya, SRI LANKAE-mail: ratnayesu@yahoo.comAbstract:About 1% of the total storage capacity in the world’s reservoirs is lost annually due tosedimentation. Sediments can also block intakes in reservoirs and damage tunnels or turbines. One ofthe most effective techniques to remove these sediments is flushing, whereby water level is loweredsufficiently to re-erode deposits and flush them through the intakes. Outflow sediment discharge maywell be related to the parameters such as the sediment characteristics in the reservoir, during flushing.The purpose of this research is to investigate sediment removal efficiency during flushing of sedimentthrough low level sluice in Rantambe Reservoir and increasing the flushing efficiency by changingdifferent discharges and water level combination. In this project laboratory experiments wereperformed on a physical model. Similitude and scaling principles were used in order to create themodel. Based on the physical properties of the media and data collected the coal is used to modelsand media. The combination of 0.1 & 0.2 l/s flow releases through low level sluice is the mosteffective to flush the sediment. The time ratio between model and its prototype is equal to 1:16.Keywords:Rantambe reservoir, Sediment flushing, Flushing efficiency, Physical modelling.1. INTRODUCTIONReservoir sedimentation not only decreases the storage capacity, but also increases the probability offlood inundation in the upstream reaches due to heightening of the bed elevations at the upstream endof the reservoir. In order to reduce reservoir sedimentation and remove sediment, many approachessuch as flushing, sluicing, dredging of the reservoir, and water & soil conservation in the catchmentare used. Among these approaches, flushing is considered an economical approach to swiftly restorethe storage capacity of the reservoir with severe deposition.The purpose of this research is to investigate sediment removal efficiency during flushing of sedimentthrough the low level sluices in Rantambe Reservoirunder different combination of discharges andreservoir water surface level. Laboratory experiments were performed on a physical model of thereservoir using coal as the sediment.2. LITERATURE REVIEWSubstantial amount of sediment could be flushed out and reservoir capacity could be restored tomaximum if periodical flushing is conducted (NeenaIsaacab, Eldho, T & Gupta, I 2013).For a given reservoir, the sediment flushing efficiency changes widely by various factors such asconfiguration of reservoir, volume and grain size of deposited sediment, discharge rate duringsediment flushing, duration time from the start of draw down flushing etc.(Shen, H1999).In a physical model study, the sediment diameter should not be less than 0.7 mm lest the bed of themodel forms ripples (Jean & Rene, T 2011).300

3rd International Symposium on Advances in Civil and Environmental EngineeringPractices for Sustainable Development (ACEPS – 2015)3. METHODOLOGY3.1. Experimental Set-upScaled down model of Rantambe reservoir was constructed with use of timber frames and concrete.The geometric scale ratio between the model and the prototype is equal to 1:250.Table 1 Dimensions of the prototype and the modelItemPrototype length (m)Model length (m)(1:250)Dam length (m)4151.66Reservoir length (Project area) (m)7002.8Width (m)2.50.010Height (m)3.80.016300.12Sluice gateWater depth (m)Figure 1 Constructed modelCoal was used as sediment in the model testing based on dimensional analysis for similarity condition.Sieve analysis was used to take out coal of the required sediment size of 0.8 mm. Sediment passingthrough 1 mm and retained on 0.6 mm were separated by sieves.Initially a scale down model of the reservoir was constructed and the periphery of the reservoir wasproduced to the scale. Sediment was laid on the bed for a thickness of 5 cm and allowed toconsolidate to simulate the reservoir bed. The time ratio between model and its prototype is equal to1:16.Sediment (coal) was placed in the model bed to form reservoir bed topography using grid lines markedto cover the model bed. After that sediments were allowed to consolidate and thereafter initial bedlevels were measured.301

3rd International Symposium on Advances in Civil and Environmental EngineeringPractices for Sustainable Development (ACEPS – 2015)3.2. Experiment ProgrammeReservoir impounding was carefully done by sending water at a very small flow rate as high flowerodes sediment particles. Low level sluice gates were opened and loss of sediments in the reservoirbed was measured for different inflow and gate opening combinations (Table 2).Table 2 Schedule of trialsCASE No.Gate openingDischargesCASE 1-i16mm10 mmCASE 1-ii0.2 l/sOne gate openCASE 1-iii0.1 & 0.2 l/s combinationCASE 2-i0.1 l/sCASE 2-ii2X10 mm0.1 l/sCASE 2-iii0.2 l/sTwo gates open0.1 & 0.2 l/s combinationAll experiments were conducted for 2 ½ hours. In the cases with combination flows, i.e. Case-iii, flowrate of 0.1 l/s was fed first for 30 minutes. Thereafter, flow rate of 0.2 l/s was fed for 5 minutes and flowrate of 0.1 l/s was again fed for 30 minutes.4. RESULTS AND DISCUSSIONCase 1- i & Case 1-ii: No sediment was flushed out due to low flow rate and only ponding occurred.Case 1-iii (0.1 l/s & 0.2 l/s combination): A little amount (108 g) of sediment was flushed through thelow level sluice gate. There wasn’t channel formation in the reservoir bed.Figure 2 Reservoir bed after Case 1-iii302

3rd International Symposium on Advances in Civil and Environmental EngineeringPractices for Sustainable Development (ACEPS – 2015)Case 2- i (0.1 l/s flow rate):In this case 0.1 l/s flow was continuously fed from upstream for 2 ½ hours. Sediment was continuouslyflushed through the gate. A curved channel was formed in the reservoir. At a certain point inflow wasequal to outflow, reservoir water level remained constant. In this case a substantial amount ofsediment was flushed and measured as 6300 g in weight.Figure 3 Reservoir bed after Case 2-iCASE 2- ii (0.2 l/s flow rate):In this case, 0.2 l/s of flow was continuously fed from the upstream. Initially upstream sedimentsmigrated towards the gates and start to pass through the gates. Ponding occurred after sometime andthere were not sediment flushed out through gates.Initially sediment movement occurred in the direction indicated in fig4. After that erosion path havechanged to the direction indicating in red colour. In this case Erosion rate is less than Case 2-i.Figure 4 Reservoir bed after Case 2-ii303

3rd International Symposium on Advances in Civil and Environmental EngineeringPractices for Sustainable Development (ACEPS – 2015)Case 2-iii (combination of 0.1 l/s & 0.2 l/s):The quantity of flushed sediment is high. During 0.1 l/s flow rate, sediments eroded at a small rate inthe vicinity of the gate. When flow rate was changed to 0.2 l/s at the upstream, sediment started tomigrate towards the gate. This substantially increased the amount of sediment flushing. The flushedsediment in this case was measured as 9500 g.Figure 5 Reservoir bed after Case 2-iiiGate size (mm)16mm10 mmTable 3 Summary of result in modelFlow rate (l/s)CASE 1-i0.1 l/sNo sediment collectedCASE 1-ii0.2 l/sNo sediment collectedCASE 1-iii0.1 & 0.2 l/s combination108CASE 2-i0.1 l/sCASE 2-ii2X10 mmWeight of sedimentcollected /(g)CASE 2-iii0.2 l/s0.1 & 0.2 l/s combination63006000340039008100980095008900304

3rd International Symposium on Advances in Civil and Environmental EngineeringPractices for Sustainable Development (ACEPS – 2015)Gate size (mm)16mm10 mmTable 4 Summary of result in prototypeFlow rate (m³/s)CASE 1-i0.1 l/s-CASE 1-ii0.2 l/s-CASE 1-iii0.1 & 0.2 l/s combination0.001CASE 2-i0.1 l/sCASE 2-ii2X10 mmIncreased volume (MCM)CASE 2-iii0.2 l/s0.1 & 0.2 l/s 3Table 5 Feeding time duration of 0.2 l/sFlow Rate (m³/s)Case No.Case 2-iiiFlushed sediment weight(g)0.1 l/s0.2 l/s1351598001401095001450589004.1. Comparison of Experimental Results with Flushing of Rantambe ReservoirRantambe reservoir was flushed during June 2014 and he reservoir bed after flushing was comparedwith that of the model. There were some small sand bars formed in the model though there wasn’tany bar formation in the prototype. This can be attributed to the fact that sediments in the prototypehad been consolidated for a very long period compared to the sediments in the model. The channelpaths observed in both prototype and model were similar. Sediment deposited in the sides of thechannel was not eroded.Duration of feeding of water at a rate of 0.2 l/s has a prominent effect on theflushing sediment volumes.305

3rd International Symposium on Advances in Civil and Environmental EngineeringPractices for Sustainable Development (ACEPS – 2015)Figure 6Comparison of channel formation between the model and the prototype5. CONCLUSIONSA scale model experiments were conducted to investigate the sediment flushing in Ranambereservoir. Based on series of model test results, the flushing efficiency is highest at the combination offlow rates of 0.1 l/s & 0.2 l/s. Use of constant flow rates of 0.1 l/s or 0.2 l/s is less effective than usingcombinations of flow rates. When combination of flows is used, the time duration of inflow of 0.2 l/s33has a prominent effect on flushing. Thus the combination flows of 100 m /s and 200 m /s in theprototype is recommended, and this is feasible by supplying discharge from the Randenigala releasestogether with flow of Uma Oya to improve the flushing efficiency. Flushing will recover about 0.124MCM of reservoir storage.6. REFERENCESAmar A. (January 2012). Erosion rate of reservoir deposit as revealed by laboratory experiment.Aruppola S. (December 2002). Desilting of Rantambe Reservoir.Bashar K. and Eltayeb A. (2010). Sediment Accumulation in Roseires Reservoir.Čedomil J., Stefan S., Georg S. and Hans, P. (September 2009). Hydraulic Flushing of AlpineReservoirs – Model Study. Water Management and Hydraulic Engineering.Francisco J. and Chih Ted Yang. (2001). A Numerical Model for Reservoir Sedimentation.Sedimentation and River Hydraulics.Gregory L. and Jiahua f. (1998). Reservoir sedimentation handbook.Jean and Rene, T. (2011). The use of plastic media in a movable bed model to study sedimentaryprocesses in rivers.Jian L. and Akihiro, T. (2003). New Development of Sediment Flushing Technique. Water &Environmental Resources.Neena Isaacab, Eldho, T. and Gupta, I. (01 Aug 2013). Numerical and physical model studies forhydraulic flushing of sediment from Chamera-II reservoir. Hydraulic Engineering.Shen H. (1999). Flushing sediment through reservoirs. Hydraulic Research.306

Initially a scale down model of the reservoir was constructed and the periphery of the reservoir was produced to the scale. Sediment was laid on the bed for thickness of 5 cm and allowed to a consolidate to simulate the reservoir bed. The tim