Probabilistic Identification of Water Pipe Network Failure

Water is an indispensable resource for the daily needs of our society. Water pipe systems assure its supply to numerous destinations. However, these systems are complex, spatially distributed and cannot be fully monitored. Water pipe systems have an inherent susceptibility to failures, which lead to water losses. The identification of a failures is a challenging task because the available methods are limited and there is no single tool that combines the relevant available information.

This project aims to develop a reliable tool for a probabilistic assessment of the state of pipelines with the ability to confidently localize failures and estimate their severity under incomplete knowledge. For this, Bayesian statistical principles will be used to account for the scarcity of information in the pipeline networks and to improve the accuracy of failure identification. Improving the decision support systems in water pipeline networks and reducing the loss of drinking water are expected in the scope of this project.