Optimal sensor network design for early intrusion detection and identification in water distribution networks

Water Distribution Networks (WDNs) are an important part of our society. Contaminant intrusion into WDNs could potentially affect a large cross-section of the population. Deliberate intrusion is the most serious threat to a WDN and can result in chemical and/or biological substances spreading throughout the WDN causing serious consequences. A contamination warning system needs to be installed in a WDN through sensors to detect the presence of contaminants as early as possible in order to protect the public from exposure. In this work, we develop algorithms for sensor network design for contamination detection and identification for a specific level of service using observability and identifiability concepts. We also estimate the total population exposed to the intrusion before detection and identification of contaminants is realized. The proposed methodologies are illustrated on a fairly large-scale network.