الفهرس 
INTRODUCTIONOnly 14 pages are availabe for public view
 |  | from | 176 |  |  |
| | | from | 176 | | |
Abstract
This study uses Artificial Neural Networks to determine Non-Revenue Water (NRW) in Water Distribution Networks (WDNs). The study is divided into two parts. The first part is concerned with the prediction of Non-Revenue Water (NRW) via Artificial Neural Networks (ANN) and Multiple Linear Regression Modelling (MLRM). The second part concern a leakage assessment in water distribution networks using hydraulic model calibration. The study predicts the impact of the hydraulic parameters on NRW using artificial neural network models. The predictive models were developed using the historical data of the holding company for water and wastewater of various measured district-metered areas (DMAs) in Egypt.
Similarly, a hydraulic analysis of Egypt’s district metered areas (DMAs) was conducted to create a permanent monitoring system and eliminate leaks despite its high cost and time consumption. Results showed high accuracy in NRW final estimation. Using Artificial Neural Networks (ANN), two scenarios were assumed, one with complete data of the DMA investigated while the other assumed a total lack of knowledge. It led to the development of four models to forecast multiple parameters affecting the percentage of non-revenue water. The performance indicators (i.e., RMSE, MAE, and correlation) show that the machine-learning algorithm is better at identifying complex relationships between different parameters. The models developed in this research can be applied to other DMAs in Egypt. Using the Darwin Calibrator, the proximate location of water leakage in a DMA was also predicted. The study also found that more than 40% of water networks in Egypt are over 40 years old, leading to over 30% of non-revenue water losses with an average cost of 25.5M Egyptian pounds. In all networks where the node’s pressure and input discharge were introduced to the program as observational data, leakage location was determined at the exact location. These results were found to be similar to the leakage resulting from the field inspection. Water-GEMS has therefore been deemed a good tool for engineers to estimate the location and amount of leakage in water networks. This study’s results can provide the basis for decision-makers to reduce costs after applying the rehabilitation process to any new DMA. Overall, these findings indicate that the machine-learning model is adequate for water companies seeking immediate, cost-effective and long-term improvement of their water distribution systems. Thus, it will help decision-makers strategically enhance the overall NRW ratio across Egypt and elsewhere worldwide.