Digital Twin Framework for Remote Water Distribution Systems
End-to-end monitoring and decision-support platform deployed in Unalakleet, Alaska
First Author
Yeji Kim, Yifan Huang, Rebecca Cantrell, Matthew Bartos, Lina Sela
Python (WNTR, EPANET, epanet-js), Flask, AWS EC2, InfluxDB, Plotly, Leaflet
Jan 2024 – Jun 2026
Published: Kim, Y., Huang, Y., Cantrell, R., Bartos, M., & Sela, L. (2026). ACS ES&T Water, 6, 3543–3554. DOI →
Live Dashboard: Unalakleet WDS Digital Twin →
Overview
Water distribution systems (WDSs) in remote communities often operate with limited staffing, aging infrastructure, and few digital tools. This work develops an end-to-end digital twin framework for resource-constrained utilities and demonstrates its implementation in the WDS of Unalakleet, Alaska — a community of ~800 residents in the Norton Sound region.
Approach
- Designed and deployed a custom wireless sensor network (pressure + temperature) on four distribution loops, with solar-powered cellular nodes feeding AWS EC2 / InfluxDB
- Built a dynamic hydraulic model (EPANET / WNTR / epanet-js) integrating sensor data, ANTHC SCADA flows, and tank levels for continuous state estimation
- Developed a browser-based dashboard (Flask + Plotly + Leaflet) for operator-facing visualization and decision support
Key Results
The digital twin reconstructs the hydraulic state with high fidelity:
| Variable | CRPS |
|---|---|
| System pressure head | 0.065 m |
| Tank level | 0.022 m |
| Loop pressures | 0.198 – 0.393 m |
Validated use cases:
- Anomaly detection — diagnosed a pressure-drop event in June 2025
- Maintenance outage support — risk-free testing of planned shutdowns
- Fire-flow capacity assessment — system-wide stress testing for infrastructure planning
Reference
Kim, Y., Huang, Y., Cantrell, R., Bartos, M., & Sela, L. (2026). From Field Data to Smart Operations: A Digital Twin Framework for Water Distribution Systems in Remote Communities. ACS ES&T Water, 6, 3543–3554. https://doi.org/10.1021/acsestwater.5c01515