Adaptive urban drinking water supply model using the effect of node elevation and head loss formula: A case study

Abstract

Along with population growth and health improvement, water demand due to urbanization is increasing and creating a need to develop a strategy for handling water supply networks (WSNs). In the last decade, software modeling of WSNs has been developed to evaluate the state of networks in terms of pressure control, leakage analysis, and overall demand determination. In the case of very complex and extremely large networks, it is very difficult to manage the water supply. Metropolitan Waterworks Authority (MWA) in Thailand has to supply drinking water to the three densely populated cities; Bangkok, Nonthaburi, and Samut Prakan, that cover an area of 2944.05 km2. Hence, MWA has developed a main pipe model using EPANET software as a managing tool. This tool can offer a good solution for the water supply, but there is approximately a 14 percent error, mainly due to not having the elevation data of the pipe network. The current research is based on demand and pressure modeling analysis with utilizing two important parameters, node elevations, and head loss. The first trial model was an initial revision of the node elevation based on a road surface map. It was found that the model with elevation data could offer a better solution and was 3.95% more accurate than the existing model. The result was significantly improved, but another error, which may have been caused by using an inappropriate head loss model, was found. As the introduced model is based on the Hazen-William model, it cannot offer an accurate solution for all Reynolds number ranges. Even though Darcy-Weisbach is more complex to use, it could provide a better solution. The results indicate the Darcy-Weisbach model produces results that are 8.65% more accurate than the Hazen-William model.