Keywords
water network partitioning; district metered areas; recursive spectral clustering
Start Date
25-6-2018 2:00 PM
End Date
25-6-2018 3:20 PM
Abstract
Water Network Partitioning (WNP) in District Metered Areas (DMAs) is an effective management strategy for the Water Distribution Networks (WDNs), simplifying pressure control, water loss detection and protection from contamination. Nevertheless, the definition of permanent DMAs constitutes an arduous task, since, the closure of some pipes may worsen significantly the energy and the topological redundancy of the system. In recent years, several optimization procedures, relied on heuristic optimization methods, were focused on the optimal design of DMAs; they generally are based on the simultaneous insertion of gate valves in all boundary pipes defined in the clustering phase. Since these procedures are performed on the original network layout, they neglect the consequences of each pipe closure on the topology and the hydraulic performance of the network. They fail to face the optimization problem recursively since they find the optimal pipe closure at one time. This paper proposes a novel optimization procedure based on the recursive spectral clustering after each pipe closure combining, step by step, the clustering and the physical dividing phase. In this way, the optimal positioning of each next single gate valve is achieved on a novel cluster layout that takes into account the previous pipe closures. The proposed methodology is tested on a real WDN and is compared, with a non-recursive procedure, through some energy and topological performance indices.
An innovative approach based on recursive clustering to design optimal districts in water distribution networks
Water Network Partitioning (WNP) in District Metered Areas (DMAs) is an effective management strategy for the Water Distribution Networks (WDNs), simplifying pressure control, water loss detection and protection from contamination. Nevertheless, the definition of permanent DMAs constitutes an arduous task, since, the closure of some pipes may worsen significantly the energy and the topological redundancy of the system. In recent years, several optimization procedures, relied on heuristic optimization methods, were focused on the optimal design of DMAs; they generally are based on the simultaneous insertion of gate valves in all boundary pipes defined in the clustering phase. Since these procedures are performed on the original network layout, they neglect the consequences of each pipe closure on the topology and the hydraulic performance of the network. They fail to face the optimization problem recursively since they find the optimal pipe closure at one time. This paper proposes a novel optimization procedure based on the recursive spectral clustering after each pipe closure combining, step by step, the clustering and the physical dividing phase. In this way, the optimal positioning of each next single gate valve is achieved on a novel cluster layout that takes into account the previous pipe closures. The proposed methodology is tested on a real WDN and is compared, with a non-recursive procedure, through some energy and topological performance indices.
Stream and Session
Stream B: (Big) Data Solutions for Planning, Management, and Operation and Environmental Systems
Session B1: Modelling and Managing Urban Water and Energy Demands in the Era of Big Data
Organizers: Andrea Cominola, Ashlynn Stillwell, Stefano Galelli, Andrea Castelletti