Sewer infrastructure assessment using logit statistical models

As the sewer infrastructure ages with time, it is imperative that sound rehabilitation strategies and alternatives be developed and implemented. The expected service life of a sewer system is variable and depends on several parameters that affect pipe condition including flow loading, wastewater quality, soil conditions, material of pipe, location, and construction methods. Traditionally, some sewer repair projects are executed on a reactionary basis. In other words, the discovery of damage in the sewer system initiates immediate action. Infrastructure projects executed on a reactionary basis more often than not result in high monetary costs due to their urgent nature. Therefore, it is important to develop a sound sewer rehabilitation strategy for proactive implementation scheduling. Currently, the City of Edmonton uses closed circuit television (CCTV) to inspect an average of 220 km/year of their local sewer infrastructure. The decision-maker is faced with the dilemma of trying to decide which portions of the network to schedule for inspection. This paper presents the analysis of the local sewer infrastructure at the City of Edmonton using Logit Statistical Models. Logit statistics are used to model the relationship between one or more dependent categorical variables and a set of independent variables in predicting the probability of a given sewer line being in a "deficient" condition based on historical data. This information can provide the decision-maker with viable alternatives for programming efficient inspection schedules.