RESERVOIR INFLOW FORECASTING AND OPERATIONAL PLANNING (THE CASE OF GENALE DAWA GD .. 3 HYDROPOWER SCHEME)

Abstract

To determine the optimal operation policy and management of reservoir systems, accurate forecast of the inflows to the reservoir are very crucial. Therefore, the aim of this research is to propose a best forecasting model that can efficiently forecast the random inflow to GD-3 reservoir and integrate it to its operational planning. Resulting forecast, simulation techniques and planning vary with the system purpose, physical characteristics, and availability of data. This paper presented an Artificial Neural Network (ANN) approach for reservoir inflow forecasting and Water Evaluation and Planning System (WEAP) model for its power simulation; using daily stream flow and rainfall data. Radial Basis Function (RBF) of ANN and the Multiple Regression models were used to compare the model output and the correlations between estimators were also examined. In developing the ANN models, different networks with different numbers of neuron hidden layers were evaluated. A total of 26 years of historical data were used to train and test the networks. The optimum ANN network with 5 inputs, 5 neurons in two hidden layers and one output was selected. To evaluate the accuracy of the proposed model, the Mean Squared Error (MSE) and the Nash-Sutcliffe Correlation Coefficient (NSE) were employed. The network was trained and converged at MSE = 0. 064 by using training data subjected to early stopping approach. The network could forecast the cross validation data set with the accuracy of MSE = 0.067 . Training and cross validation process of the best model developed during testing showed the correlation coefficient of 0.936 and 0 . 939 respectively . The Water Evaluation and Planning System (WEAP) was to set the operational planning for GD-3 Reservoir as a reference scenario and evaluate the effects on its downstream water resource development (two run of river hydroplants GD-5 and GD-6) following the construction of GD-3 Dam. In the simulations using WEAP, a monthly averaged flow of about 266Mm 3 is continuously released to produce a continuous power of 1632GWh/year for 95% of the time. The construction of GD-3 dam stabilizes the flow of middle Genale River by improving the flow pattern; adds about 65m 3 Is to the dry months flow (Dec to March) and reduces the peak flow (Aug to Oct) by more than 75m 3 /s and generated a total energy of 3389GWh/year for 96% of the time including the two run of rivers. Therefore this study justified that coupling inflow forecasting and operational planning of reservoirs is very crucial in deriving the benefits derived from GD-3 Hydropower scheme. Keywords: Genale Dawa, ANN, Forecasting, Reservoir Inflow, Simulation, WEAP