Benefit/cost model for noise loggers� distribution - City of Montreal

One billion individuals worldwide do not have access to clean drinking water. According to Canadian Infrastructure Report of 2016, Canada�s water network is in a declining state, which significantly increases the risk of sudden pipes� breaks. The average processed water loss due to leaks in the city of Montreal is one of the highest among other Canadian cities with 23% losses. Accordingly, early leak detection and repair will help reduce those huge water losses and keep the water infrastructure sustainable. Thus, this paper proposes a benefit-cost analysis to investigate the potential benefits of expanding the acoustic noise loggers� coverage over 50% network as opposed to the manual (crew-based) leak detection. Those benefits include (1) risk savings in terms of smaller number of leaks, water losses, and breaks; (2) monetary savings in both direct and indirect costs; and (3) temporal savings in terms of time spent to detect and repair the network leaks. To compute those benefits, a framework that functions through five integrated models was built: (1) central database that contains several datasets (i.e. asset inventory, leak detection types, leak repair temporal and financial, network data); (2) leak detection systems� financial models that compute the CAPEX and OPEX of the manual and noise loggers leak detection systems; (3) leaks simulation model that simulates the appearance of different leak categories among different street categories; (4) leaks temporal and financial models that estimates the temporal and financial implications of the leaks for both systems (i.e. repair time/cost, water losses, etc.); and (5) risk model that calculates the probability and consequences of breaks among different street categories. The framework was applied on Montreal's water network and the analysis was carried out on both leak detection systems across 20 years planning horizon. The results displayed huge savings for the noise loggers 50% expansion scenario.