Ò»±¾µÀÎÞÂë

Seed grant spotlight: Improving electric-power resilience

Just as climate change prompts discussions about carbon emissions, it also necessitates a conversation about power system resilience in the face of increasing episodes of extreme weather and natural disasters. 

With support from a Scott Institute seed grant made possible by the Rankin Family Grand Challenge, College of Engineering researchers Maryam Hamidi, Ramteen Sioshansi, and Granger Morgan are developing a robust optimization model that is capable of analyzing threats to the grid and providing commensurate responses that improve its resilience and mitigate supply disruptions. 

“The types of threats that are being thrown at the electricity system are more extreme than they were in the past, both in terms of natural disasters and human-caused attacks, which are becoming more destructive,” said Sioshansi.

Sioshansi added that despite the best of efforts, no power grid can be completely safeguarded against factors like weather, aging infrastructure, increasing demand, or supply chain issues. The alternative is to help them “fail in a graceful way so that a system can recover either naturally or with human intervention, as opposed to brittle systems, which have a very difficult time recovering from failure.” 

The optimization model is intended to serve as a tool for decision support and assessment of how resilient or brittle a system’s infrastructure is. It could be used both proactively and reactively, and the many potential end users include utilities and emergency managers. 

“When we talk about resilience, it means the power grid can anticipate, withstand, and rapidly recover from disruptions of any kind. To get to this point, we need to have different technologies and strategies that can be implemented before, during, and after an event,” said Hamidi, a doctoral student in engineering and public policy. 

Hamidi explained that what makes a grid “resilient” largely depends on the system, including where it’s located and what the weather conditions are like. For example, underground distribution lines might be helpful in areas prone to ice storms and strong winds, but this same method could make power restoration more difficult in areas that experience flooding. 

To train the optimization model, Hamidi is using data from weather events that occurred over a span of decades, including roughly 400 recorded hurricanes on the east coast of the United States and 58 hurricanes that were concentrated in North Carolina between 1900 and 2010. Hamidi is also using fragility curves to extrapolate a relationship between hurricane strength and factors like pole damage and power outage duration. 

The definite goal of this early-stage research is to produce a mathematical structure that other researchers could use as a foundation for their own projects.