A space-time probabilistic model for pipe network reconstruction planning

Abstract

A space-time probabilistic model can be combined with a dual objective management model to rank reconstruction strategies for a water supply pipe network. The network is characterised by similar age and environmental and traffic conditions. A homogeneous spatial Poisson process represents the location of pipe failures, while the number of failures is modelled by a non-homogeneous Poisson process. Thus, a space-time Poisson process describes failure probabilities in space and time. A simulation procedure is used to generate failure patterns in any time and estimate expected costs and expected failure consequences. A dual objective management model may be developed to aid ranking reconstruction strategies for a selected time horizon aiming at minimising cost and failure impact. A Failure Impact Index FI is formulated by calculating the areal extend of unacceptable pressure drops due to individual failures. The results indicate a trade-off exists between cost and FI.

Keywords:

• leakage
• decision support
• network rehabilitation
• planning
• costs