Canadian municipalities have noted that 59% of their water systems needed repair and the status of 43% of these systems is unacceptable. In the United States, ASCE assigned a near failing grade of D- to the condition of water system infrastructure. Therefore, municipalities face a great challenge of managing the expected large replacement and new installation projects of water mains. This research aims at designing a robust model in order to assess the condition and predict the performance of water mains. Data are collected from three different Canadian municipalities: (1) Moncton (New Brunswick); (2) London (Ontario); and (3) Longueiul (Québec). An integrated model and framework, using an analytic hierarchy process (AHP) and artificial neural network (ANN), are developed. In addition, an automated, user-friendly, web-based infrastructure management tool (CR-Predictor) is developed based on the integrated AHP/ANN model to assess water main condition. The developed tool and models are validated in which they show robust results (98.51%)-the average validity percent. They are expected to benefit academics and practitioners (municipal engineers, consultants, and contractors) to prioritize inspection and rehabilitation planning for existing water mains.
|Number of pages||14|
|Journal||Journal of Infrastructure Systems|
|Publication status||Published - 27 Nov 2008|
- Neural networks
- Water distribution systems
ASJC Scopus subject areas
- Civil and Structural Engineering