Performance Modeling for Sewer Networks

In spite of the pressing need to preserve sewer networks, sewer pipelines and manholes are prone to deterioration and hence to collapse. According to the American Society of Civil Engineers (ASCE) (2017), the sewer network’s grade of the United States (US) is grade “D+”, making it one of the worst infrastructure assets in the US. In addition, the Canadian Infrastructure Report Card (CIRC) (2016) states that more than half of their linear wastewater assets’ physical condition were ranked between very poor to good states, with a total replacement value of $47-billion. Despite the enormous studies conducted in this field, many of the efforts lack a comprehensive assessment of sewer components, leading to misjudged rehabilitation decision plans and continued asset deterioration. Improved cost-effective models that optimize sewer rehabilitation plans, given the scarcity of resources, are clearly needed. Accordingly, the paramount objective of this research is to design a decision-support system that optimizes the maintenance, rehabilitation and replacement (MRR) decisions of sewer pipelines and manholes. The first phase of the research is to identify several defects that impact the condition of sewer components and to model the erosion void defect utilizing fuzzy expert system. The model provided accuracy, true positive rate and precision values of 83%, 76%, and 80%, respectfully. The identified defects were then grouped into several robust models to study their cause and effect relationship through the application of the Decision-Making Trial Evaluation Laboratory (DEMATEL). The overall condition of the sewer pipeline is then found by integrating the DEMATEL method with the Quality Function Deployment (QFD), while the manhole condition is calculated using the aforementioned two techniques along with the Analytic Network Process (ANP). After validating the two models with the Royal Gardens neighbourhood’s sewer network in Edmonton, the average validity percentage (AVP) for the pipeline and manhole assessment models were 58.68% and 76.24%, respectively. Subsequently, Weibull distribution analysis is adopted to predict the future calculated conditions of sewer manholes and pipelines by modelling the deterioration of each. The research establishes an approach to aggregate the condition indexes of all pipelines and manholes in the network through a criticality model to supply the overall network performance index. Accordingly, the economic factors are deemed the most important ones compared to environmental and public factors. An informative optimized model that integrates the outputs of the previously developed models is designed through the Particle Swarm Optimization (PSO) approach to maximize the sewer network performance and minimize the total costs. Different trade-off solutions are then established by varying the weights of the objective functions and considering the defined constraints. The best network performance improvement attained is 1.47 with a total cost of $1.39- million. The comprehensive sewer network assessment performed in this research will improve current practices in sewer networks management, thereby reducing sewer network failures and avoiding catastrophic sinkholes