Optimizing the operation of the Valencia water-distribution network

Abstract

The second of the two case studies in the POWADIMA research project, the Valencia waterdistribution network, serves a population of approximately 1.2 million and is supplied by surface water via two treatment plants which have significantly different production costs. The only storage available is located at the treatment plants, each of which has its own pumping station.The management of the network is a complex operation involving 4 pressure zones and 49 operating valves, 10 of which are routinely adjusted. The electricity tariff structure varies with the hour of the day and month of the year. The EPANET hydraulic simulation model of the network has 725 nodes, 10 operating valves, 2 storage tanks and 17 pumps grouped at the two pumping stations. The control system that has been implemented comprises an artificial neural network predictor in place of the EPANET model and a dynamic genetic algorithm to optimize the control settings of pumps and valves up to a 24 h rolling operating horizon, in response to a highly variable demand. The results indicate a potential operational-cost saving of 17.6% over a complete (simulated) year relative to current practice, which easily justifies the cost of implementing the control system developed.

Development of a real-time, near-optimal control process for water-distribution networks

“…While both were anticipated at the model formulation stage, the complexities they induced were not known until later. However, all of these issues have, of necessity, been addressed when applying the control system developed to the two case studies (see Salomons et al 2007;Martinez et al 2007). …”

PEPSO: Reducing Electricity Usage and Associated Pollution Emissions of Water Pumps

“…A small portion of real systems are similar to small test networks of these researches, but most of the time we face large networks with a couple of hundred pipes, junctions, and a considerable number of pumps, valves, tanks, etc. There are a few studies that tried to optimize a real and large-size WDS [4,29,30]. Most of the systems that were used in WDS optimization studies do not have variable speed pumps.…”

Use of an artificial neural network to capture the domain knowledge of a conventional hydraulic simulation model

“…In addition to the hydraulic variables, an attempt was made to include water quality in the decision-making process by distribution network (see Martinez et al 2007;Salomons et al 2007), thereby enhancing the prospect of making real-time, near-optimal control a practical reality.…”