A probability model for investigating the trend of structural deterioration of wastewater pipelines

Abstract

Canada’s infrastructure is aging and deteriorating. New legislation requires the municipalities to estimate operating and capital expenditures for running the systems into the future and to develop financial sustainability plans. Wastewater pipelines deterioration is currently not well understood and realistic deterioration models need to be developed. This paper demonstrates how the condition assessment data from trenchless visual inspections of wastewater pipelines can be used to understand the performance of wastewater pipelines. A new ordinal regression model for the deterioration of wastewater pipelines based on cumulative logits is elaborated. The model is presented using the Generalized Linear model formulation and takes into account the interaction effect between the explanatory variables. The new model is demonstrated and validated using the City of Niagara Falls high quality wastewater collection network condition assessment data for reinforced concrete (RC) and vitrified clay (VC) pipes. This new model is found to represent the City of Niagara Falls RC and VC pipes’ deterioration behavior for pipes in service for up to 110 years. RC pipes deterioration is found to be age dependent while VC pipes deterioration is not age dependent. This finding is contrary to other deterioration model studies that indicate that VC deterioration is age dependent. The service life for RC pipes is estimated to be approximately 75 years while VC pipes are found to have an indefinite service life if installed without structural damage. The cumulative logit model can be used to determine wastewater pipelines’ service life, predict future condition states, and estimate networks’ maintenance and rehabilitation expenditures. The latter is critical if realistic wastewater networks’ future maintenance and operation budgets are to be developed for the life of assets and to meet new regulatory reporting requirements. Further research is required to validate this new methodology for other networks and the deterioration modeling of pipe materials other than RC and VC.