Fuzzy Logic Applications for Water Pipeline Risk Analysis

Abstract

There is an urgent need to advance the processes that drive current water pipeline renewal decision support systems. The aging infrastructure needs more efficient decision support systems to help water utilities make renewal prioritization and renewal decisions. Fortunately, with the developments in computational techniques and data collection through advanced sensors, it is now possible to support prioritization of critical water pipelines with data and knowledge. Previous studies have found that around two trillion gallons of treated drinking water are wasted through 240 000 water main breaks per year in the United States. The costs associated with fixing these water pipeline failures are not only for replacing the pipe but also include many other indirect costs like loss of human-hours due to traffic disruptions, property damage due to flooding, and disruption of service to critical facilities like hospitals, medical centers, and primary schools. This underlines the need to further enhance the consequence of failure (CoF) assessment tools available to water utilities to identify critical pipelines and understand the risk for renewal prioritization. This study was performed to advance the state of the art in assessing the CoF based on a comprehensive list of parameters and utilizing a novel hierarchical fuzzy inference system (FIS) to support better renewal decisions for water pipelines based on criticality. This study will help water utilities to build data-driven decision support systems and to advance their asset management programs. The results in this study are based on analysis of water pipeline data collected from over 500 water utilities in the US and can provide an indication of overall criticality of water pipelines in the US. PIPE i D has been created to help in understanding the critical drinking water pipeline infrastructure. It is envisioned to be a national database platform for advanced asset management addressing the major management levels, including strategic, tactical, and operational, that will assist water utilities of all sizes in sustaining targeted levels of service with acceptable risk. It will also provide secure access to aggregated data, models, and tools, which will enable the synthesis, analysis, querying, and visualization of the data for decision support. With the help of PIPE i D, it becomes possible to incorporate the many previous models and develop a systematic and comprehensive approach in the statistical modeling of water pipeline performance prediction.