Stopping Outages Before They Start

Power grids everywhere are being challenged by extreme weather events, cyberattacks and increases in demand.

“It is imperative that we develop tools that modernize the system and make it more resilient,” said Souma Chowdhury, associate professor of mechanical and aerospace engineering at UB.

Chowdhury and a team of UB researchers have developed an artificial intelligence model that can remediate power outages by automatically rerouting electricity in milliseconds. The approach is an early example of “self-healing grid” technology, which uses AI to detect and repair problems all on its own as soon as issues occur.

In most cases, electrical outages are caused by failures of the distribution network. This means that the power is there—it’s just not getting to where it needs to go.

The U.S. grid is an extensive, complex network of transmission and distribution lines, generation facilities and transformers that distributes electricity from power sources to consumers. When issues arise, that intricacy makes determining alternate paths a painstaking process that can take hours.

The new AI-based system has the advantage of speed. Once trained, it’s able to find the optimal path to send power to the majority of users within microseconds, before service disruptions even occur.

To map the complexity of a power distribution network, the research team used algorithms that apply machine learning to graphs. The team also relied on reinforcement learning—a method in which a virtual agent is deployed in a simulation of the real problem—to systematically play out scenarios and progressively learn from the experience.

So, for example, if electricity became blocked due to line faults, the system could reconfigure using switches, drawing power from available sources in close proximity such as large-scale solar panels or batteries on a university campus or in a business park.

The UB researchers, who collaborated with engineers at the University of Texas at Dallas, said that while more research is needed before the system can be scaled to existing grids, the model, with its potential to minimize or even eliminate power outages, marks an exciting development for the nation’s beleaguered power grid.