Predicting water main failures: A Bayesian model updating approach

Abstract

Water utilities often rely on water main failure prediction models to develop an effective maintenance, rehabilitation and replacement (M/R/R) action plan. However, the understanding of water main failure becomes difficult due to various uncertainties. In this study, a Bayesian updating based water main failure prediction framework is developed to update the performance of the Bayesian Weibull proportional hazard (BWPHM) model. Applicability of the proposed framework is illustrated with modeling failure prediction of cast iron and ductile iron pipes of the water distribution network of the City of Calgary, Alberta, Canada. The Bayesian updating models have effectively improved the water main failure prediction whenever new data or information becomes available. The proposed framework can assess the model performance in the light of uncertain and evolving information and will help the water utility authorities to attain an acceptable level of service considering financial constraints.