Leveraging 9-1-1 Calls for Early Detection of Water Contamination Events

Abstract Submission: The contamination of water distribution systems (WDSs) by hazardous chemicals—whether intentional or accidental—can lead to severe health risks and significant socio-economic impacts. Certain chemicals, often referred to as “super-poisons,” are so potent that they can cause fatalities within an hour of exposure. Given their extreme toxicity and the rapid onset of symptoms, 9-1-1 call centers are likely to be the first point of contact for affected individuals or their families. In the wake of such incidents, information from emergency calls can be crucial for detecting the contamination event and identifying potential intrusion points, thereby aiding in prompt containment and protection of public health. Thus, these emergency calls can serve as an early warning system for such intrusions.
This paper will cover the development of an AI algorithm designed to detect contamination events from 9-1-1 call data and will address the practical challenges encountered during the training and testing phases of the algorithm. The AI was tested on a case study involving a water distribution system with roughly 4000 km of piping and 26,000 demand nodes. The results demonstrated that the AI algorithm successfully identified contamination patterns and accurately pinpointed the source of the intrusion with over 93% accuracy.

Learning Objectives/Expected Outcome (Optional) :