Development of a graph dynamic sectorization tool using QGIS/PostgreSQL to plan and operate water distribution networks

Abstract

[EN] This paper presents a definition of the different types of sectorization used by water distribution companies, as well as a methodology for their identification and application in the dynamic zoning of water distribution networks. The graph model of large water supply networks is made up of thousands or hundreds of thousands of nodes and pipelines, as well as valves, tanks, pumps and pressure groups. Its generation from the data contained in a GIS does not report difficulty with the currently existing tools, the GISWATER module of QGIS is an example of this. Problems arise when these models have to be operated from a high number of conditions and scenarios. To facilitate the operability of the system, one of the strategies is to simplify its operation through topological analysis. One of the proposals is clustering where the nodes belonging to a cluster have more and better connections with the internal nodes than with the external ones. Our work analyzes water distribution networks from five different points of view, which it represents five sector classes according to the parameter analyzed. As result, one of them, the District Metering Area (DMA) delimited by headers (flowmeters) and stoppers (other flowmeters or closed valves), is extensively covered in the literature. The other four, less common in the literature, have a definition equivalent to that of the DMA but based on other class of headers and stoppers. Therefore, if we are looking to operate the network from the point of view of pressure, we can use the dynamic sectorization algorithm by configuring Tanks, Head pumps, flow pumps, PSV, PBV or PRV as headers. Moreover, to operate and analyze the network from the point of view where water comes from, we can use again the dynamic algorithm by configuring the inlets of the system as headers (Tanks and Reservoirs). On the other hand, if the goal is to be focused on water quality, the algorithm can be configured using clorinathors, reclorinathors, water treatment equipment or water sources as headers. Finally, special attention deserves the minimum sector which represents the minimum structural unit of the network, defined by the region covered by network where there are shutoff-valves on their borders. In order to operate water networks as optimized as possible, the status of shut-off valves is dynamic. The opening and closing of valves allows a continuous reconfiguration of the network and the classes or types of sectors defined. As result, to analyze the network graph using this approach enables the water operator to analyze the best status as possible as well as it is amazing scenario to analyze resilience of network using stochastics methods.