Please use this identifier to cite or link to this item: https://hdl.handle.net/2440/80953
Type: Conference item
Title: San Antonio water system's use of modeling and GA optimization to integrate new water sources
Author: Mann, E.
Duncker, A.
Andreu, J.
Eddy, A.
Haby, J.
Simpson, A.
Citation: Texas Water 08, American Water Works Association, San Antonio, Texas, USA: 18 p
Publisher: Publisher unknown
Issue Date: 2008
Conference Name: Texas Water 08 (2008 : San Antonio, Texas, USA)
Statement of
Responsibility: 
Elsinore Mann, Alana Duncker, Jamily Andreu, Adam Eddy, Jeff Haby, Angus R Simpson
Abstract: The San Antonio Water System (SAWS) has been experiencing substantial growth in its service area for several decades, and steady growth is expected to continue for the next 15-20 years. The SAWS network has historically been supplied primarily from wells in the Edwards Aquifer, and more recently supplemented by water from the Trinity Aquifer and Canyon Lake in the north. In recent years, withdrawal restrictions have been instituted for the Edwards Aquifer, prompting SAWS and neighboring water suppliers to develop additional water sources to meet future growth. SAWS recently constructed the Twin Oaks Water Treatment Plant in the southeast corner of the system to treat water from the Carrizo-Wilcox Aquifer and to store Edwards Aquifer water in an Aquifer Storage and Recovery facility. Large capacity pumps and a 60-inch diameter transmission main deliver Twin Oaks facility water to primary pump stations on the east side of the system. This arrangement has worked well to integrate up to 30 million gallons per day (MGD) of Twin Oaks water into the system, but given the Edwards Aquifer restrictions during drought periods, SAWS desires to integrate up to 60 MGD to help meet projected growth and to offset their use and dependence upon the Edwards Aquifer in times of drought. This paper describes how SAWS used hydraulic modeling and Optimatics Genetic Algorithm (OGA) optimization to determine a hydraulically viable low-cost set of new infrastructure improvements, as well as a flexible operating strategy, designed to integrate up to 60 MGD of Twin Oaks water into the existing distribution system. A 24-hour extended period simulation scenario of the 2013 projected minimum demand day was constructed with the existing hydraulic model. The OGA optimization routine was linked to the model and customized by inputting potential improvement alternatives, hydraulic constraints and unit cost data. The objective of the optimization was to identify minimum-cost improvements that met demands using the full 60 MGD from the Twin Oaks facility while satisfying constraints for minimum pressure, maximum velocity, and storage tank return levels. Millions of improvement combinations were evaluated against hydraulic constraints and cost by the genetic algorithm optimization process. Ultimately a hydraulically-viable, low-cost solution of capital improvements was chosen which successfully integrated 60 MGD of Twin Oaks water as well as 9 MGD from other sources. New pipes were recommended near and between 5 primary pump stations to overcome existing capacity constraints and convey Twin Oaks water into the existing system. New pumps at two primary stations and three new control valves were also part of the solution. With these improvements, most of the secondary wells and some of the primary wells that currently supply Edwards Aquifer water did not need to operate, allowing non-Edwards Aquifer water to supply as much of the system as possible. The total estimated capital cost of the solution was $32 million.
Rights: Copyright status unknown
Appears in Collections:Aurora harvest
Civil and Environmental Engineering 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.