Optimal sensor placement for leak detection and identification in pipeline networks


Detecting and identifying leaks in a pipeline network (citywide, industrial, domestic, etc.) is a challenging activity, one that has significant implication on safety and performance of the network. 

The performance of any leak detection methodology critically depends on the location of sensors in the network. Towards this end, in the current work, we will investigate optimal location of sensors in the network. In particular, we propose to use structural information and ideas from information theory to characterize performance of the placed sensors. The information theoretic measures will be used to judge leak detection and identification performance enabled by the placed sensors. The optimal design problem will be to then optimally place sensors which maximize the information theoretic measure. The ideas will be demonstrated using data generated from open-source simulators which are capable of simulating leaks across different sized networks. 

Scope of the work: Theoretical/computational

Pre-requisites: Keen interest/desire to explore and learn concepts in set theory, graph theory, information theory, optimization aspects, and also do heavy programming. 

Name of Faculty