MODELLING OF WATER LOSS IN HYDRAULIC DISTRIBUTION SYSTEM OF MINNA, NIGER STATE, NIGERIA

Abstract

The main objective of this study is to model water loss in the distribution network of Shiroro District Metered Area, Minna, Niger State capital which has been facing severe water supply scarcity in recent time. This scarcity has been further complicated with high water loses via leakages. Hence, there is need for new and modern approaches which will involve increased automation for estimation of water leakages. The hydraulic machine, EPANET was used for the hydraulic modelling of the water loss in the networks of Minna Water Treatment Plant and Shiroro District Metered Area, (DMA). Active Leakage Control mechanism was adopted by deploying the use of modern leak detecting equipment. Leaks were collected and measured from 37 nodes of Shiroro distribution network. 24 hours Extended Period Simulation, (EPS), was carried out from leak prone nodes by varying emitter coefficient from 0.1, 0.15, 0.2 and 0.3 respectively. The leakage, Q leaks generated from the model using discharge coefficient and the values of observed leaks Q measured from site were compared statistically using Nash-Sutcliffe efficiency model to check the performance of the model. Data set collated were further utilised in neural networks for the estimation of leakages. The leak generated using modern leak detecting equipment was 1.3m 3 for eight (8) hours which was large because it was unnoticed. Pearson Product Correlation of pressure and leak was found to be very high positive and statistically significant. The NSE values obtained at optimum leakage coefficient of 0.2 for 8th hour, 9th hour, 10th hour and 11th hour are 0.73, 0.68, 0.58 and 0.52 respectively. These values are good indictors of good performance. Model validation for 20 pm, 21 pm and 22 pm hours in M. I. Wushishi Distribution Network gave NASH values of 0.767, 0.668 and 0.601 respectively. For validation procedure using ANN, for estimation of leakage at M. I. Wushishi Network, R 2 was 0. 65 and relative errors for training and testing were 0.367 and 0.215 respectively. R 2 of 0.65 is an indication of better performance of the model.in terms of leakage estimation. The performance of the model has suggested that using the emitter coefficient of 0.2 can model the water leakage in Shiroro and M. I. Wushsishi Water Distribution systems because of high correlation between the simulated and the observed discharges. Variation of leakage exponent on the performance of emitter equation in estimating leakages in water distribution system is recommended.