Key Takeaways
- A multidisciplinary team led by the University of Maine is using AI to make electric grids more resilient to cyberattacks, equipment failures, and extreme weather.
- The project spans four states—Maine, Hawaii, North Dakota, and Puerto Rico—each contributing specialized expertise in hardware security, anomaly detection, self‑healing systems, and microgrid simulation.
- By integrating power‑systems engineering, AI, cybersecurity, communications, sensing, economics, visualization, and workforce training, the effort aims to produce solutions that work in real‑world utility environments.
- Funded by a $2.4 million two‑year grant from the U.S. Department of Energy’s EPSCoR program, the initiative will develop and test AI‑driven, federated‑learning self‑healing technologies and microgrid testbeds.
- Workforce development is embedded throughout, with plans to train students, create educational modules, and build long‑term capacity for maintaining smart, secure grids.
- Collaboration with local utilities ensures that innovations can transition from simulation to operational grids, improving reliability, security, and recovery from cyber and physical disruptions, especially in vulnerable communities such as those affected by Hurricane Maria in Puerto Rico.
Project Overview and Objectives
Researchers at the University of Maine, under the leadership of Prabuddha Chakraborty, are launching a two‑year, $2.4 million initiative funded by the U.S. Department of Energy’s Established Program to Stimulate Competitive Research (DOE EPSCoR). The goal is to harness artificial intelligence (AI) alongside advances in cybersecurity, hardware security, and workforce development to create electric grids that can withstand cyberattacks, equipment failures, and extreme weather events. By improving grid resilience, the project seeks to enhance the reliability, affordability, and security of electricity service, particularly for communities that are historically vulnerable to power outages.
Leadership and Multidisciplinary Collaboration
Prabuddha Chakraborty, the Waldo Libbey Assistant Professor in UMaine’s Department of Electrical and Computer Engineering and a faculty member of the Advanced Structures and Composites Center, serves as the principal investigator. He emphasizes that no single discipline can solve the complex challenges facing modern power systems; therefore, the project deliberately integrates expertise from power systems, AI, cybersecurity, hardware security, communications, sensing, economics, visualization, and workforce training. This holistic approach aims to produce solutions that are not only technically sound but also practicable in real‑world utility settings.
Jurisdictional Contributions and Research Themes
The research consortium includes partners from the University of Hawaii at Manoa, the University of North Dakota, and the University of Puerto Rico Mayagüez. Each jurisdiction brings a distinct focus that addresses a specific barrier to adopting AI in grid operations:
- Maine (UMaine): Acts as the central hub, connecting the various research themes into a unified framework and overseeing overall project coordination.
- Hawaii: Leads work on hardware security, ensuring that the physical components of smart grids resist tampering and unauthorized access.
- North Dakota: Concentrates on anomaly detection and forecasting, using AI to identify irregularities that could signal cyber intrusions or impending equipment failure.
- Puerto Rico: Focuses on self‑healing systems and microgrid simulations, aiming to enable grids to automatically reconfigure and restore service after disruptions.
By dividing labor in this way, the team can pursue deep, specialized investigations while maintaining a coherent vision for a resilient, secure grid.
Integration Framework and System‑Level Vision
The UMaine team’s leadership role involves synthesizing the individual contributions into a broader architectural framework rather than treating each technical effort in isolation. This integration ensures that advances in, for example, anomaly detection can be effectively communicated to self‑healing algorithms, and that hardware security measures are compatible with AI‑based control strategies. The resulting system‑level approach is designed to support real‑time operating decisions, maintain functionality during cyberattacks or poor‑quality data, and enable rapid recovery from equipment failures or communication breakdowns.
Workforce Development and Education
A core component of the project is the intentional cultivation of a skilled workforce capable of sustaining and advancing smart, secure grid technologies. Each partner institution will train graduate and undergraduate students, develop educational modules, and create outreach programs that embed cybersecurity, AI, and power‑systems concepts into curricula. Daisy Green, co‑principal investigator at the University of Hawaii Manoa, highlights the importance of inspiring a new generation of engineers who are passionate about integrating secure sensing, learning, and visualization techniques into electricity systems to improve grid resilience and uplift local communities.
Utility Partnerships and Real‑World Translation
To bridge the gap between laboratory research and operational deployment, the consortium collaborates directly with local utilities in each participating state. These partnerships facilitate testing of AI‑driven, federated‑learning‑based self‑healing technologies on simulated microgrid testbeds. Prakash Ranganathan, director of the Center for Cyber Security Research at the University of North Dakota, explains that working with utility partners allows the team to refine innovations for practical implementation, ultimately enhancing the reliability, security, resilience, and recovery speed of actual electric grids when faced with cyber or physical disturbances.
Impact on Vulnerable Communities
The project places special emphasis on regions that are prone to extreme weather and have historically unreliable power supplies, such as Puerto Rico, which continues to recover from the devastation of Hurricane Maria. Juan Patarroyo Montenegro, co‑principal investigator at the University of Puerto Rico Mayagüez, notes that the research aims to bolster the island’s ability to deliver reliable electricity to citizens, especially in areas that remain off‑grid after major storms. By modernizing the grid with AI‑enabled security and self‑healing capabilities, the initiative seeks to reduce outage durations, lower restoration costs, and increase community resilience to future natural disasters.
Funding, Timeline, and Expected Outcomes
Supported by the DOE EPSCoR award, the two‑year project will progress through phases of algorithm development, hardware security prototyping, simulation testing, and field trials with utility partners. Anticipated outcomes include:
- A validated AI‑based anomaly detection framework adaptable to diverse grid configurations.
- Secure hardware modules that resist cyber‑physical tampering.
- Self‑healing control strategies capable of autonomous reconfiguration during faults.
- Scalable microgrid testbeds demonstrating improved resilience under simulated cyber and weather stressors.
- Trained students and educational resources that expand the national workforce skilled in secure smart‑grid technologies.
Collectively, these results are expected to contribute to a more robust, affordable, and secure electric power infrastructure, serving as a model for other regions confronting similar challenges.