Dartmouth Engineering Study Presents New Framework for Building More Resilient Power Grids

Co-authors (l to r) Professor Junbo Zhao, PhD student Yitong Liu, and postdoctoral researcher Alaa Selim at Zhao's Dartmouth lab, "CREATE: Cyber-Physical Resilience Empowering A Transformative Energy Future."

Rather than treating outages as isolated failures, the proposed framework organizes outage management into five connected functions: data analysis, long-term improvement, real-time monitoring, proactive mitigation, and outage restoration. It shows that combining outage data with infrastructure records, sensing data, and analytical tools can help quantify resilience, identify vulnerabilities, and evaluate investments.  

The study, "From Electricity Outage Analysis and Monitoring to Mitigation and Restoration," was authored by Junbo Zhao, the Todd M. Cook and Elizabeth Donohoe Cook Associate Professor of Engineering at Dartmouth, together with his PhD student Yitong Liu and postdoctoral researcher Alaa Selim, both members of Zhao's lab, CREATE: Cyber-Physical Resilience Empowering A Transformative Energy Future. 

"Power outages are not caused by only a single failed component," said Zhao. "They often emerge from the interaction of weather, aging infrastructure, limited visibility, operational uncertainty, and evolving grid dynamics. Our goal is to show how utility outage records can be analyzed and combined with existing technologies in a practical framework that helps utilities understand outages, detect risks earlier, limit cascading impacts, and restore service more effectively."

The study further highlights how real-time sensors, smart meters, weather and geospatial data, and equipment-health monitoring can be fused to provide a more complete picture of grid conditions before and during an outage. Building on this improved situational awareness, artificial intelligence can help translate fragmented data streams into operational decisions, such as identifying vulnerable feeders, forecasting outage duration, supporting topology-aware monitoring, and guiding network reconfiguration. These decisions can then inform mitigation and restoration actions. During restoration, self-healing grid capabilities can be enabled by coordinating microgrids, grid-forming inverters, and battery energy storage systems to maintain service to critical loads, stabilize local islands, support staged re-energization, and guide safe reconnection to the main grid.

"As power grids become more data-rich, more decentralized, and more uncertain, utilities will need stronger ways to integrate diverse data sources, monitor system conditions with incomplete models and sparse measurements, and place new sensors and meters where they provide the greatest operational value," said Zhao. 

The study also highlights the need to validate restoration strategies under field conditions and develop clearer rules for reconnecting inverter-based resources. Ultimately, the study argues, outage management must move toward an integrated, cyber-secure, and self-healing workflow that links data, decision-making, control, and restoration across the full outage lifecycle. 

"By securing both the information used to assess outages and the automated actions used to recover from them, utilities can build outage management systems that are not only smarter and faster, but also trustworthy under disruptive and adversarial conditions," said Zhao.

In addition to Zhao, Liu, and Selim, co-authors include Ian Dobson of Iowa State University, Giovanni Sansavini of ETH Zurich, Bálint Hartmann of Budapest University of Technology and Economics, and Fei Ding of the National Laboratory of the Rockies.