Collaborative Research: AI-Enabled Digital Twins for Resilient Freshwater Infrastructure and Communities under Cascading Hazards

The continuity of freshwater service underpins public health, commerce, and emergency response across the United States, and improvements in its reliability benefit every community served by a public water system. The objective of this project is to support research on a new AI-enabled cyber-physical-social paradigm for predicting, managing, and mitigating cascading disruptions in freshwater infrastructure systems. Freshwater infrastructure, including intakes, treatment facilities, storage assets, distribution networks, sensing platforms, and utility operations, is becoming increasingly difficult to manage due to aging assets, deferred maintenance, climate-driven hazards, power outages, and water-quality variability. Disruptions that begin in one component, such as an intake, pump station, or treatment unit, can propagate through connected infrastructure and service networks, producing operational instability and uneven community impacts. Current utility practices often rely on monitoring systems, historical records, and reactive emergency protocols that provide limited ability to anticipate how local anomalies escalate into broader service disruptions. This research reframes freshwater infrastructure not only as a physical asset network, but as an adaptive cyber-physical-social system in which infrastructure dynamics, real-time sensing, AI, institutional decisions, and community experience interact under uncertainty. This project advances the field in three primary directions: (1) enabling AI-driven predictive digital twins for small and mid-sized freshwater infrastructure systems; (2) developing cyber-physical-social models that connect infrastructure performance with community exposure and human response; and (3) creating socio-technical decision-support methods for evaluating service-disruption mitigation strategies under uncertain and evolving hazard conditions. The research will leverage heterogeneous system-level data, including sensor measurements, water-quality observations, geographic asset information, weather and river forecasts, maintenance records, historical disruption events, and community-reported service experiences, to understand how disruptions emerge, propagate, and affect people. AI methods, including anomaly detection, graph-based modeling, digital twins, probabilistic forecasting, and uncertainty quantification, will be used to develop scalable and interpretable tools for infrastructure monitoring, prediction, and risk-informed decision support. By enabling earlier warning of system vulnerabilities and cascading disruptions, the project can support more effective operational decisions, reduce service interruption duration, improve public communication, protect households during hazard events, prioritize vulnerable service areas, and strengthen human well-being through safer, more reliable water service. Project outcomes will include transferable computational tools, digital twin workflows, socio-technical vulnerability maps, decision-support dashboards, and implementation recommendations to inform utilities and public agencies. Research findings will be disseminated through publications, open-source tools, educational materials, stakeholder workshops, and practitioner-oriented guidance, ensuring benefits for researchers, government agencies, utilities, community organizations, and the public. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.