The goals of this Phase I research project are to investigate the feasibility of using an artificial neural network (ANN) and an optimization-based water distribution security system ("smart security") for accurately detecting contamination events, and given the nature of the threat, identify optimal crisis management decisions that minimize the threat and maintain a safe water supply. Given the vulnerability of large water supply systems and their suitability for attack by terrorists seeking to inflict maximum physical harm in a dramatic and psychologically devastating manner, security of the Nation's drinking water systems is of paramount importance. As shown by Neural Optimization Applied Hydrology, LLC (NOAH), and others in many different applications (e.g., groundwater resources management, wastewater and drinking water treatment, etc.) ANNs are adept at accurately simulating and effectively managing complex hydraulic environments that are highly variable in both space and time. Research conducted by NOAH demonstrates that the ANN-derived state-transition equations may not only achieve superior prediction accuracy in complex hydraulic-driven systems, but also are ideal for simulating large numbers of possible scenarios in real time, as well as conducting formal optimization analyses that otherwise might be infeasible with traditional physical-based models. Real-time continuous monitoring of water distribution systems is becoming widespread, but these data are vastly underutilized in daily operational assessment and decisionmaking. The American Water Company (the Nation's largest water provider) will provide NOAH (a minority-owned business) with hydraulic and water quality data collected with automatic sensor technology from one of its water distribution systems. NOAH will use these data for developing and testing the ANN and optimization methodology, and identify critical components necessary for achieving high performance. In addition to testing the system for detecting anomalous water quality changes that otherwise might be missed by an operator, the ANN-derived state-transition equations combined with optimization algorithms will identify crisis response decisions. The ANN crisis management performance will be compared against a traditional water distribution system hydraulic model in terms of computational speed and accuracy. Increased time saving and prediction accuracy are critical for reducing the risk of terrorist attacks disrupting critical water distribution systems and exposing large populations to contaminated water. If successful, the ANN-based security and real-time crisis management system has widespread application, and could be installed in any water distribution system utilizing automatic data collection systems. Based on the large water market both nationally and internationally as well as the existing pervasive climate of global terrorism, the proposed water security system has significant importance and market potential. Supplemental
Keywords: small business, SBIR, artificial neural network, ANN, terrorist attack, homeland security, water security, water distribution systems, drinking water, monitoring, EPA. , Ecosystem Protection/Environmental Exposure & Risk, INTERNATIONAL COOPERATION, RFA, Scientific Discipline, Water, Analytical Chemistry, Chemical Engineering, Chemistry, Drinking Water, Engineering, Chemistry, & Physics, Environmental Chemistry, Environmental Engineering, Environmental Monitoring, Monitoring/Modeling, analytical methods, artificial neural network, biopollution, chemical detection techniques, community water system, crisis management, detection, drinking water system, drinking water contaminants, drinking water regulations, drinking water security, environmental contaminants, environmental measurement, field monitoring, field portable monitoring, homeland security, measurement, monitoring
