Please use this identifier to cite or link to this item:
Scopus Web of Science® Altmetric
Type: Conference paper
Title: A metamodeling approach to water distribution system optimization
Author: Broad, D.
Dandy, G.
Maier, H.
Citation: Critical transitions in water and environmental resources management [electronic resource] : proceedings of the World Water and Environmental Resources Congress : June 27-July 1, 2004, Salt Lake City, UT / sponsored by Environmental and Water Resources Institute (EWRI) of the American Society of Civil Engineers ; Gerald Sehlke, Donald F. Hayes, and David K. Stevens (eds.): CDROM
Publisher: American Society of Civil Engineers
Issue Date: 2004
ISBN: 0784407371
Conference Name: World Water and Environmental Resources Congress (2004 : Salt Lake City, Utah)
Abstract: Genetic Algorithms (GAs) have been shown to apply well to optimizing the design and operations of water distribution systems (WDS). The objective has usually been to minimize cost, subject to hydraulic constraints, such as satisfying minimum pressure. More recently, water quality considerations have also been incorporated into WDS optimization by requiring a minimum chlorine disinfection level throughout the system. Due to uncertainty in WDS data, such as the demands and pipe roughness factors, reliability has also been incorporated into optimization. However, to date, there have not been any attempts to simultaneously incorporate reliability and water quality into the optimization of WDS. This is due to the fact that the computational time requirements are extremely large. Considerable time savings can be achieved by using a technique known as metamodeling. A metamodel is a surrogate, or substitute for a complex simulation model. The type of metamodel used in this research was an Artificial Neural Network (ANN). ANNs are capable of approximating the non-linear functions that govern flow and chlorine decay in a WDS. A metamodeling approach has been applied here to optimize a water distribution design problem that includes water quality. The ANNs were calibrated so as to provide a good approximation to the simulation model. Large time savings occurred from training the ANNs to approximate chlorine concentrations (over 600 times faster than the simulation model) and reliability (expected to be 10 7 times faster). The solutions obtained by linking an ANN to a GA were shown to be similar to those when using a simulation model linked to a GA, with the added benefit that the solutions were found much faster.
Description: © 2004 ASCE Also titled: Water and enviromental resources management
DOI: 10.1061/40737(2004)453
Appears in Collections:Aurora harvest
Civil and Environmental Engineering publications
Environment Institute publications

Files in This Item:
There are no files associated with this item.

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.