Abstract
Large diameter trunk mains are the life line of the water supply system. They convey large volumes of water between treatment works and local distribution networks, sometimes over quite significant distances (tens to hundreds of kilometres). Compared to smaller diameter distribution mains, trunk mains tend to have low failure rates, but when they do fail, the consequences are potentially much more significant, with direct, indirect and societal costs. Worldwide, a significant proportion of trunk mains are still made of aging cast iron material. Remarkably, these aging assets have in some cases outlived the pipes that replaced deteriorated parts of the network. Even so, many cast iron pipes are beginning to approach, or have already exceeded, their design life: consequently, out of a large population of pipes, some are failing whilst some still have considerable residual life. Asset management, in this context focussing on the targeted replacement of degrading main, requires tools and models for the prediction of the future performance of the network. Several mechanistic deterioration models have been developed in recent years, which attempt to predict the condition of cast iron pipes, but few methodologies have specifically targeted water trunk mains. Nevertheless, the requirement has remained for a robust deterioration and failure model for cast iron trunk mains, worldwide. This project, being part of a wider, collaborative project between Thames Water Utilities and the University of Surrey, has reviewed existing failure models for cast iron trunk mains and sought to modify these based on information arising from other areas of the project. This has included, new understanding of the corrosion of cast iron trunk main, the use of fracture mechanics to predict failure and non-destructive evaluation data gathering techniques has provided significant insight into improvements that can be made to failure models. In particular, the present research has shown how traditional loss-of-section approach to the residual strength of corroded pipes can be used alongside a fracture mechanics approach, in order to provide boundaries to the failure “envelope”. This novel methodology has been incorporated as part of an enhanced modelling framework, which has shown to improve the failure predictions across the network. The enhanced model also enables more detailed analysis of sections of the pipes that have been surveyed on site.