Document Type
Presentation
Publication Date
8-1-2014
Abstract
Water storage tanks are key components of water distribution networks (WDNs) and are primarily designed and operated to meet demand variations and pressure needs. However, the common practice in the design of WDNs is to incorporate large storage tanks that may possibly create long residence time. Long residence time is a major contributing factor for loss of disinfectant, increased formation of disinfection by products and microbial regrowth. Also, poor choice in tank geometry, location and operation can play a role in deterioration of water quality. Most of the previous approaches on optimisation of WDNs design and operation do not take into account tank operation explicitly. In this work, optimal tank design, location and operation strategy has been implemented to assess network performance from water quality perspective. The most recently developed genetic algorithm based optimisation model “Penalty-Free Multi–Objective Evolutionary Algorithm (PF-MOEA)” has been employed. PF-MOEA uses a pressure dependent analysis simulator that handles the node pressure constraints and the conservation of mass and energy inherently. The algorithm considers tank operation strategy as one of the objectives in the optimisation process. The optimisation model incorporates pipe sizing, tank siting, tank sizing and pump operation. PF-MOEA has been applied on the benchmark “Anytown” network that comprises multiple loadings, storage tanks and pumps. The model provided many feasible solutions that are cheaper than the best previous solutions. The solutions satisfy both node pressure and operational constraints for the different loading conditions. A significant improvement in water quality has been achieved in terms of water age, disinfection residual and disinfection by-product concentration in the entire network. Results demonstrated that explicit consideration of the tank operation objective has substantially enhanced the network performance in reference to hydraulic as well as water quality.
Comments
Session R65, Parameter Estimation: Optimization Applications in Water Resources