NUMERICAL SIMULATION FOR SEDIMENT FLUSHING IN RESERVOIRS

Abstract

Globally there are about 50,000 large dams and among them 25,500 are the storage reservoirs with storage volume of about 6,464 Bm3. World’s annual reservoir storage loss in different regions due to sedimentation varies between 0.08-2.3%, with an average of about 0.6%. It is estimated that in 2030 the demand of water would be 8500 Bm3, but the existing storage would be around 7000 Bm3. To meet 1500 Bm3 shortfall, about 8100 reservoirs are needed and construction of so many reservoirs in future seems to be difficult. The only solution is to conserve the existing reservoirs by enhancing their lives by adopting appropriate measures. Various methods employed globally to conserve storage capacities of reservoirs are watershed management, conventional dredging, dry excavation, hydrosuction, sediment routing/sluicing, sediment bypassing, density current venting, and sediment flushing through the reservoir, used independently or in combination. Present study focuses on the flushing operation to enhance the lives of reservoirs and to answer several questions related to flushing operation, like, is reservoir flushable?, if yes, then what would be the flushing efficiency?, how many times in a year it should be flushed?, when it should be flushed?, how much would be the flushing discharge required?, how much should be the duration of flushing?, how much water would be sacrificed for the flushing operation?, and what would be the recovery in capacity of the reservoir considering the flushing operation?, etc. Flushing is a method by which the flow velocities in a Reservoir are increased to such a level that deposited sediments are mobilized and transported through low level outlets in the dam. Flushing sediments through reservoirs has been practiced successfully and found to be inexpensive in many cases, however, a significant amount of water is required during flushing operation. Hence, there is need to numerically model the flushing scenarios to check the performances of reservoirs in restoring the reservoir capacities. Flushing probably has been implemented on many hundreds reservoirs of the world, but in literature only about 50 reservoirs are documented as flushed, and flushing data is available for only 25 reservoirs. Among them in literature about 6 reservoirs had been reported as successfully flushed i.e. Baira-India, Gebidem-Switzerland, Gmund-Austria, Hengshan-China, Palagnedra-Switzerland, and Santo-Domingo-Venezuela reservoirs. Various flushing indicators used to assess the feasibility of sediment flushing from reservoirs are Sediment Balance Ratio-SBR, Long Term Capacity Ratio-LTCR, Drawdown Ratio-DDR, Sediment Balance Ratio during full drawdown-SBRd, Flushing Width Ratio-FWR and Top Width Ratio-TWR. The usually adopted critical values of these indicators are: SBR >1, LTCR approaching to unity, DDR >0.7, SBRd >1, FWR > 1 and TWR = 1-2. In the present study, the values of these six flushing indicators were computed for the selected 14 flushed reservoirs of various regions of the world and were compared with their critical values. Out of 14 selected reservoirs, 6 were successfully flushed and 08 were partially flushed. From the analysis it was found that for successfully flushed reservoirs critical values of all six flushing indicators were satisfied, but for the partially flushed reservoirs critical values were also satisfied except for the Flushing Indicator LTCR. It shows the significance of LTCR over the other Flushing Indicators. So it was learnt that LTCR is the most important flushing indicator among the six indicators to assess the feasibility of sediment flushing through the reservoirs. Analysis results of the 14 reservoirs also show that among the successfully flushed reservoirs maximum value of LTCR is 1 for Santo-Domingo and Palagnedra Reservoirs, whereas, Hengshan Reservoir has the least value of LTCR, i.e., 0.77, which is a successfully flushed reservoir, hence it was concluded that the critical value of LTCR may be taken as 0.77 instead of 1 for the successfully flushed reservoirs. Equations were developed for SBR and LTCR by Multiple Non-linear Regression Analysis, using the data of six successfully flushed reservoirs. These equations were tested on foreign and Pakistani reservoirs and the comparison revealed that developed equations results match well with the results of Atkinson equations. To get confidence in numerically simulating the flushing scenarios, flushing operations were modeled for three successfully flushed reservoirs for which data of entire flushing activities were available to calibrate and validate the Flushing Models. To numerically simulate flushing operations, initially these three reservoirs were modeled for the sediment deposition processes. These reservoirs are Baira of India, Gebidem of Switzerland and Gmund of Austria. Flushing processes have been modeled using three Models, i.e. SHARC, HEC-RAS 4.1.0, and Tsinghua University Equation. Results of the study show that SHARC Model well simulates the sediment deposition processes, but it underestimates the flushing durations. Results of the HEC-RAS 4.1.0 Model show that it can well simulate sediment depositions and sediment flushing operations. Then Tsinghua University Equation was used for simulating the sediment flushing operations through these three reservoirs. Results of the Tsinghua University Equation reveal that Model well simulates sediment flushing operations through these reservoirs. All small reservoirs of Punjab Small Dams Organization (SDO) of Pakistan were investigated, and 20 reservoirs were selected based on detailed data availability to assess their feasibility for sediment flushing. The results reveal that based on the computed Flushing Indicators, 5 reservoirs can be ranked as likely to be successfully flushed, these are Jammargal, Phalina, Dharabi, Talikna, and Jabbi reservoirs. Among the five likely to be successfully flushed reservoirs, Jabbi reservoir having 3.8 Mm3 storage capacity was selected for modeling the sediment deposition and flushing processes. Jabbi Reservoir was created after the construction of Dam across Jabbi Nullah by the end of year 1990. Hydrographic survey of the reservoir was conducted after 10 years of operation in 2000, which was used for the validation of the sediment deposition process in the reservoir. The survey results show that sediment deposition in 10 years was about 0.418 Mm3. As results of the flushing modeling on the three foreign reservoirs proved that HEC-RAS 4.1.0 and Tsinghua University Equation well simulate the flushing processes, hence flushing operations of Jabbi Reservoir were modeled using two Models i.e. HEC-RAS 4.1.0 and Tsinghua University Equation, under two scenarios, i.e. flushing after one year and ten years of sediment deposition. Results of the both the Models and both the options for sediment deposition show good agreement with the observed deposited sediments. A complete flushing operation includes the emptying of reservoir, flushing the sediment through the reservoir and refilling of the reservoir. Considering the results of complete flushing operation it was estimated that refilling time required for the reservoir is about 64% of the year as inflows to the reservoir are intermittent, hence annual flushing of the reservoir looks infeasible, however, large quantity of water for the flushing operation of the reservoir may be sacrificed after 10 years. HEC-RAS 4.1.0 and Tsinghua University Model Results were used to formulate the complete strategy for flushing the reservoir. Model results revealed that for flushing the Jabbi Reservoir after 10 years deposition, appropriate flushing months are July and August; suitable flushing discharge is 3 cumecs; time required to empty the reservoir is 0.34 day; time required to refill the reservoir is 235 days; flushing duration required to flush 10 years deposited sediments is about 4 days; average flushable sediment diameter is 10 mm; and water required for flushing the reservoir would be about 4.4 Mm3. Following the knowledge earned from this research work, similar procedures can be applied to other reservoirs of the world to check the degree of success in flushing operation, moreover, flushing plans / strategies can be formulated and relevant recovery in the reservoir capacities can be assessed.