The model is trained using weather data and historical wind turbine performance data.
“Based on these predictions, our model recommends how to make optimal hourly delivery commitments to the power grid a full day in advance. This is important, because energy sources that can be scheduled (i.e. can deliver a set amount of electricity at a set time) are often more valuable to the grid,” the post reads.
Little data was provided on the overall accuracy of the system’s predictions thus far, but Google and DeepMind want to make wind power more reliable with machine learning in order to make it a more attractive form of energy in the future.
“Our hope is that this kind of machine learning approach can strengthen the business case for wind power and drive further adoption of carbon-free energy on electric grids worldwide,” the blog post reads.
Google set a target to reach 100 percent renewable energy by 2017, and last year signed a 10-year deal with a Finnish energy company to fuel its Nordic data centers. Google also uses AI to make cooling servers in its data centers more efficient.
This is the most recent form of artificial intelligence to be applied to make wind power cheaper and more efficient. Earlier this month, a group of researchers introduced WaveletFCNN, a classification model to predict the buildup of ice on wind turbines. The stop of energy production due to ice damage to wind turbine blades can reduce energy production up to 20 percent, according to wind consultancy firm TechnoCentre Éolien (TCE).