Researchers from the Finnish Meteorological Institute (FMI) and Finland’s Aalto University are using machine learning

Thunderstorms can disrupt power grids, cause blackouts and cause significant damage to communities. To help electrical providers better predict which storms will likely cause outages, a team led by Roope Tervo, a software architect at FMI and Ph.D. researcher at Aalto University, used machine learning, a subset of artificial intelligence (AI) that relies on data analysis to build analytical models, to teach a computer model how to categorize storms based on data from associated power outages.

With historical data provided by Finnish energy companies Järvi-Suomen Energia, Loiste Sähkoverkko and Imatra Seudun Sähkönsiirto, the researchers determined the number of power outages that have disrupted all three networks. Those storms were then grouped into four different classes: Class 0 storms did not disrupt electricity to any power transformers; Class 1 storms cut off power to as much as 10% of the transformers; Class 2 storms cut off power to up to 50% of the transformers; and Class 3 storms cut power to more than 50% of the transformers.

Once categorized, the team referenced storm data from FMI and turned that data into information the computer model could easily understand.

“We used a new object-based approach to preparing the data, which was makes this work exciting," said Roope. "Storms are made up of many elements that can indicate how damaging they can be: surface area, wind speed, temperature and pressure, to name a few. By grouping 16 different features of each storm, we were able to train the computer to recognize when storms will be damaging."

According to the research team, the algorithm successfully predicted Class 0 storms and Class 3 storms. However, the team will continue to work with the algorithm to improve its predictions for distinguishing Class 1 and Class 2 storms by adding more data to the model.

"Our next step is to try and refine the model so it works for more weather than just summer storms," said Roope, "as we all know, there can be big storms in winter in Finland, but they work differently to summer storms so we need different methods to predict their potential damage."

The research appears in the article Short-Term Prediction of Electricity Outages Caused by Convective Storms, published in IEEE Transactions on Geoscience and Remote Sensing.