Abstract
The 4,000km Thames Water trunk main network is approximately 62% cast iron, of which
over half of it has been in operation for more than 80 years, with some pipes being as old as the
Victorian era. When the modern design life of a pipe of fifty years is taken into consideration, a
signifi cant proportion of the network has been exposed to graphitic corrosion for a long period of
time. With inadequate methods to measure the current condition of these pipes and insufficient
models to predict the future condition, it is not known in sufficient detail what the condition of
the network is in. This research develops a corrosion model for cast iron trunk mains that can
predict their current and future conditions.
The corrosion model was developed using a modified cellular automaton method in three
dimensions. Using a database of the condition of past trunk main bursts and soil environment
conditions the model was calibrated to the Thames Water service region. The model shows a
good success at predicting the current condition of trunk mains, with results that are similar to
scans made of exhumed pipes. The future condition of a pipe can also be predicted with success
using NDE scans as condition inputs.
This research contributes to the asset management of water networks as it can be used to
help inform pipe replacement programs on the condition of their pipes around the network. The
result of this is that they are able to make more informed prioritisations on which pipes need to
be replaced. Ultimately, this research and the corrosion model will lead to fewer bursts occurring
across the Thames Water network.