ABSTRACT Water utilities must assess the condition of buried pipelines without excavation or interruption of service. We review 20 pipeline risk models – specifically for the likelihood of failure – and 10 leak detection methods. We find that common risk factors include physical attributes (e.g., pipe age, material, diameter), operational attributes (e.g., pressure, flow, repair history), and environmental attributes (e.g., soil type, bury depth, fault proximity). The models contain 4–27 risk factors, differing in inclusion and weight. For an industry perspective, we interview water utilities in Utah, USA, on their condition assessment and leak detection practices. Physical attributes are most available, and environmental attributes are least available. Many models are unworkable because the water utilities lack the required data. Instead, they use acoustic methods, if any, for reactive leak detection. We recommend developing data-light risk models (using a few common attributes) to inform the best leak detection method for a given water system.
Johnson et al. (Fri,) studied this question.