ARTIFICIAL INTELLIGENCE APPLIED TO

Electricity Generation Prediction

PROBLEM

A Renewable Energy company in Montana (USA), sells renewable electricity to the wholesale markets. Their ability to predict future energy generation impacts on their business profitability, because margins are directly related to supply guaranteed. Usually, trained personnel study weather patterns and make guesses about future energy generation.

The company wonder if Artificial Intelligence could help to perform a deeper analysis and better pattern recognition from weather data, in order to increase the business profitability.

CHALLENGE

Pattern recognition in complex weather data

SOLUTION

Terminus7 team solved limitations on towers data access in order to centralize training processes, and at the same time we installed the running model on each tower.

The information from every tower is used to predict final energy production. At the same time, Terminus7 uses combined information from different towers to enrich each single tower prediction.

Finally, Terminus7 returned substantial improvements on energy production prediction, with the following confidence levels: 95% in 1 hour prediction, 89% in 2 hours prediction, 78% in 4 hours prediction. Particularly, the 2 hours prediction has the strongest impact on business profitability.

HOURS PREDICTION

%

CONFIDENCE LEVEL

Tell us your problem and we find the best AI solutions for your business