Water distribution system modeling: An essential component of total system security

One of the major concerns for water utilities in the United States is how to protect the consumers from chemical/biological contamination resulting from surreptitious malevolent attacks within the water distribution system or from accidental contamination from any means. This threat is especially problematic because the ability to detect and then meaningfully describe the movement (migration) and dilution of contaminants is one part of an overall systematic management tool for sensing, predicting, controlling, and treating contaminants within the distribution system. Over the course of the last several years, the need to resolve these important issues has taken on greater urgency. Following September 11, 2001, the United States government directed that vulnerability assessments be completed for the medium and large utilities in the country. While these assessments were an important step forward, the outcomes indicated that distribution systems weren't adequately understood from a security standpoint, nor were the actual threats. Furthermore, given the potential diversity of contaminants, sensor technologies were (and generally still are!) by and large unavailable to detect and identify even a small percentage of the possible contaminants. To make matters even worse, water utilities generally don't have the research budgets or available financial resources to create an early warning system for the sole purpose of security; early warning systems will need to also serve the dual purpose of assessing general water quality. This paper will discuss early warning systems in the context of the need for water distribution system modeling as the tool that can be used to a priori define contamination risks and optimize sensor and response locations, but also will serve to transform data from early warning systems into actionable knowledge. Ongoing activities at Sandia National Laboratories (SNL) and collaborations with the United States Environmental Protection Agency/National Homeland Security Research Center (EPA/NHSRC) have resulted in development of a number of numerical models and schemes for determining the consequences from a contamination event (in terms of human health effects), tools to efficiently optimize sensor and sampling locations for large data sets, methods to identify the contamination source location in near real time, the ability to include uncertainty in the analyses, and a model to evaluate the likelihood of attacks to water systems. This paper provides an overview of these activities.