Development of an Algorithm for Vulnerability Zoning of Water Distribution Network

摘要

Water Distribution Networks (WDNs) are highly vulnerable to failures of the system components especially pipes. Availability of a map showing different levels of WDN vulnerability will help to set the priority of rehabilitation and maintenance activities in different parts of the system. For developing vulnerability zones, it is needed to have an evaluation about the vulnerability of system components. Water mains are the main components of WDNs which are more exposed to failure. The extent and frequency of pipes' failure are related to the physical and hydraulic characteristics of pipe in a WDN. In this study, a set of these factors is used to predict the vulnerability of each pipe and they are then classified based on severity of probable failure. Due to relatively large number of factors affecting pipe failure, MRMR (Minimum Redundancy and Maximum Relevance) method is used to choose the main factors in pipe failure prediction with higher correlation with failure probability and less redundancy (transinformation) among themselves. This approach is an effective way in determining the optimal set of factors in system's vulnerability evaluation. Each of these factors affects a pipe condition in different levels therefore it is needed to prioritize them in evaluation of the general condition of a pipe. In this study, a combined vulnerability assessment of a pipe network is carried out that includes assessment of each pipe based on predefined criteria using Analytical Hierarchy Process (AHP). Then 5 classes of pipe vulnerability are defined based on their failure probability and the vulnerability zones of the system are developed. The proposed framework helps decision makers to have a better understanding of the system vulnerable components and needed rehabilitation practices of WDN.