Disaggregation of environmental factors affecting sewer pipe failures

This paper documents the development, calibration, and evaluation of a novel simulation model for understanding and predicting failures of wastewater pipe systems. The model is presented and subsequently calibrated and evaluated, within the generalized likelihood uncertainty estimation framework, against a historical record of pipe failures in the Newcastle region. It is shown that the model has significant utility in many aspects of the operational management of the pipe system by providing a framework for disaggregation of the multiple environmental factors that contribute to the overall failure rate. The disaggregation is achieved within an uncertainty framework, thus providing quantified uncertainty estimates. Such analysis will be of significant use in life cycle analyses and costing of networks, and hence in optimizing replacement strategies, as well as in the assessment of regulatory compliance. Additionally, it is argued that the developed model can provide a rigorous basis for real-time forecasting of major trends in system failures through coupling to ENSO phenomena forecasts, thus providing valuable managerial insight into periods of expected high repair/replacement costs, to which current practice is largely reactive.