Evaluating water main failure risk using a hierarchical fuzzy expert system

Distribution networks often account for up to 80% of the total expenditure involved in water supply systems. In Canada and United States, there are 700 water main breaks every day and more than 2 million breaks since the beginning of this century costing more than 6 billion Canadian dollars as repair costs. Dealing with the risk of water pipelines failure has been undergoing through a great change in concept from being active to failure events to being proactive by taking corrective actions that maintain the water main in good working conditions. Risk of failure is defined as the combination of the probability and impact severity of a particular circumstance that negatively impacts the ability of infrastructure assets to meet the objectives of the municipality. This paper designs a framework to evaluate the risk of water main failure using hierarchal fuzzy expert system This system considers four main risk-of-failure factors which have sixteen sub-factors that represent both the probability of failure and the negative consequences of failure event. Some of the findings are that the pipe age gives a strong indication about the condition of the pipelines then the pipe material and breakage rate come into play, and that the damage to surroundings/business disruption has the most negative impact of a failure event. A set of water network real data is used as a case study tc apply and examine the developed model. According to a proposed risk scale of failure, about 13 km of the network pipelines are risky and require mitigation actions in the short-term.