Methodology for Optimized Location of Boosters in Water Distribution Networks

Abstract

The operation of pumping stations in a water distribution network (WDN) represents one of the largest expenses due to its energy consumption. The selection and location of pumps are a complex problem, since the dynamic behavior of the WDN imposes different operating conditions, altering the efficiency of the pumps, which may be far from their best efficiency point, thus increasing the overall consumption of energy. This work proposes a methodology to optimally place pumping stations, carried out through the analysis of the trunk network (TN), obtained with the graph theory. The TN pipes are crucial for the entire system, as those pipes conduct water from the sources to many consumers. Thus, these are considered the possible locations to install the pumping stations. To understand the functionality of the proposed model, a comparative study was conducted by applying the optimization model to a fictitious network widely studied in the literature, D-Town. Three scenarios are proposed to evaluate the methodology: the first two consider the original topology with single or multiple pumps in parallel; and the third searches for the best location for the pumping stations. In the scenarios, the selection of a pump’s operating point and operational speed is defined by an optimization procedure, using the particle swarm optimization algorithm to minimize energy costs and water losses. The research shows that allowing a more flexible operational routine for the pumps and optimizing the placement of boosters in the network can provide a more efficient operation.