Extreme weather pushes power grids to the brink – Is AI the answer?


Hospitals forced to use snow water to flush their toilets. Families resorting to burning fence wood in their homes to survive. This is not distant history, but true stories from those recently affected by the power blackouts in Texas and the southeast. The country deserves better, and artificial (AI) intelligence can deliver it.

This tragic situation has forced a spotlight on the nation’s fragile energy grids. Cold fronts, snow storms, heat waves, fires, hurricanes, tornadoes, and flooding as well as power fluctuations and unpredictable availability of green energy sources are becoming all too common. These factors make reliable energy planning nearly impossible.

The solutions to this increasingly frequent extreme weather problem are complex and varied. But AI technology exists today that can optimise energy distribution across the grid. AI applied to electrical grids helps utilities instantaneously know how much of what type of energy to deliver where, ensuring grid resilience in the face of the unexpected.

Have you read?
Texans heating snow to get drinking water
Texas Attorney General investigates ERCOT utilities over winter storm outages
Five key challenges pushing utilities to seek new asset management solutions

Texas grid operators planned a worst-case winter peak usage of 67GW of electricity. The state rapidly consumed that as the cold front bore down. Frigid weather disabled multiple energy sources including windmills and thermal power plants, such as gas-powered generators. Spinning reserves were quickly depleted. As temperatures plummeted, demand for energy spiked, even more, causing rolling blackouts across the state. AI could have helped.

AI technology can help utilities effectively manage energy distribution and demand, and prevent rolling blackouts that are needlessly affecting millions of people. As a nation, we need to continue to invest in alternative fuels and grid optimisation and resilience to avoid entire cities being disrupted due to fragile legacy systems.

AI-powered energy systems can constantly learn the environment and all clean energy sources on the grid, and update the models used to make energy dispatch decisions. With real-time AI-based grid modeling, alternative power sources could be at-the-ready when needed, whether from neighboring states, from another autonomous microgrid, or from behind-the-meter battery storage.

AI can continuously collect millions of data points – predictive data on weather, supply, demand, energy prices – and use that data to build optimal energy models for every device on the grid, including solar inverters, batteries, wind turbines, and other clean energy sources. Utilities and independent power providers can use these models to tell them how to most effectively dispatch energy and control clean energy sources – individually and collectively. An AI-powered optimised and synchronised grid means no one is left out in the cold.

Cities can add resilience to their grids with autonomous microgrids that offer critical services such as hospitals, schools, police and fire. With AI these services always have power even if disconnected from the main grid. Every energy grid across this country should include AI in their grid resilience plans.

AI isn’t just for grid emergencies, though. AI can also be used for ongoing energy management, as energy providers use AI to process massive amounts of real-time data to route energy to residences and businesses efficiently, reliably and cost effectively. AI can fuse together real-time weather and load forecasting, economics, rules, and grid learning feedback to make energy decisions that would be far too difficult for people to make, and faster that any human could execute. With AI controlling the grid, everyone has what they need, and blackouts can be prevented.

Extreme weather events like those in Texas are increasing and this problem is not going away any time soon. With artificial intelligence to predict weather and grid state, continuously model an efficient grid, and intelligently control energy flow where it’s needed most, we can make unfortunate events like those in Texas distant history.

About the author

Sean McEvoy leads the Energy AI business unit for California-based Veritone, Inc., which specialises in AI-based grid optimization and resilience.