Conjunctive management of large-scale pressurized water distribution and groundwater systems in semi-arid area with parallel genetic algorithm
Document Type
Article
Publication Date
1-1-2009
Abstract
This study develops a production well management model for the conjunctive management of water resources in semi-arid areas. The management model integrates a large-scale pressurized water distribution system and a three-dimensional groundwater model under an optimization framework. The well pump operations optimization problem is formulated as a Boolean integer nonlinear programming (BINLP) problem to optimize the periodic 24-h pump on/off operations over a 1-week operation horizon. The management model considers multiple objectives and is solved by a parallel genetic algorithm (PGA) to overcome the difficulty of solving the BINLP problem. The PGA significantly reduces computation time for a case study in Chandler, Arizona. The Chandler water distribution model is built based on EPANET, and the Chandler three-dimension groundwater model is developed using MODFLOW. The high performance computing (HPC) of the genetic algorithm makes it possible to obtain 24-h real-time operations in the 7-day forecast model. The tank reliability, resilience, and vulnerability (R-R-V) are evaluated to infer the system reliability. The Pareto curve provides compromise solutions between the two competing objectives of energy reduction and pressure violation reduction. © Springer Science+Business Media B.V. 2008.
Publication Source (Journal or Book title)
Water Resources Management
First Page
1497
Last Page
1517
Recommended Citation
Tsai, F., Katiyar, V., Toy, D., & Goff, R. (2009). Conjunctive management of large-scale pressurized water distribution and groundwater systems in semi-arid area with parallel genetic algorithm. Water Resources Management, 23 (8), 1497-1517. https://doi.org/10.1007/s11269-008-9338-5