ABSTRACT
Buried metal pipes are often deteriorated by time and spatial-dependent corrosion. In this paper, a methodology is developed to accurately predict the failure probability of a buried pipeline network considering the random nature of corrosion occurrence and their position. The random field theory is employed to re-map the distributions of soil properties along the pipeline construction field, which are used to predict the corrosion rates at different positions of pipe. A dynamic segmentation method, based on the correlation distance of adjacent peak values of corrosion rate, is proposed to facilitate the assessment of system reliability. A worked example is presented to illustrate the application of the proposed methodology on a steel pipeline buried in various types of soils. The paper concludes that the developed method can equip pipe engineers and asset managers with a tool in accurately predicting the failure of pipeline network with a view to cost-effective management of the pipeline network.
