Water distribution network modeling: Hydroinformatics approach

Abstract

This research suggests a scheme for design or rehabilitation, and for determination of the optimal parameter values of water distribution network. The traditional approaches are either not capable or computationally too cumbersome to tackle these problems. Heuristic algorithms such as Genetic Algorithm, Ant Colony Optimization Algorithms (ACOAs), Shuffled Frog Leaping Algorithm (SFLA), etc. are also computationally not very efficient in many instances. In the present study, an efficient global optimization algorithm, Shuffled Complex Evolution (SCE), is considered and applied to the above problems. SCE is coupled with EPANET, a water distribution network solver, and applied to three network design problems and one rehabilitation problem. The coupled models are also applied to determine the optimal network parameter values for optimal water delivery systems. Comparison between performances of various techniques is conducted. Results in all case studies show that SCE reaches to the optimal solutions very efficiently.