Modeling of Reservoir Inflow for Hydropower Dams Using Artificial Neural Network

The stream flow at the three hydropower reservoirs in Nigeria were modeled using hydro-meteorological parameters and Artificial Neural Network (ANN). The model revealed positive relationship between the observed and the modeled reservoir inflow with values of correlation coefficient of 0.57, 0.84 and 0.92 for Kainji, Jebba and Shiroro hydropower reservoir respectively. The established model was used to predict 20 years stream-flow for each of the hydropower reservoirs which were found to have similar statistics with the observed values. The predicted reservoir inflow were subjected to trend analysis which revealed an upward trend with percentage increase of 4.58%, 6.34% and 5.42% for Kainji, Jebba and Shiroro hydropower reservoirs respectively. The upward trend is an indication of increase in water availability for hydropower generation at the three stations given other constraints are brought under control.http://dx.doi.org/10.4314/njt.v34i1.4