Google may have dropped the motto “don’t be evil” from its corporate code of conduct, but it seems that the search giant still wants to use its superpowers for good. With that mission in mind, Google and its A.I. subsidiary DeepMind have been working on a way to increase the usefulness of green energy produced by wind farms.
The problem the company has been trying to solve is that, while wind energy represents an important source of carbon-free electricity, it is fundamentally unpredictable. As a result, despite its positive points, wind power is less useful to the power grid than power sources that can reliably deliver it at set times.
By using machine learning artificial intelligence to predict wind output, Google and DeepMind have trained a neural network to accurately predict wind power output 36 hours ahead of the power being generated. Using these predictions, a computer model can then make recommendations for “optimal hourly delivery commitments” to the grid a whole day in advance. This makes it far more valuable since it means that it can be relied upon to deliver a set amount of electricity at a particular time.
“We can’t eliminate the variability of the wind, but our early results suggest that we can use machine learning to make wind power sufficiently more predictable and valuable,” Google and DeepMind researchers note in a blog post. “This approach also helps bring greater data rigor to wind farm operations, as machine learning can help wind farm operators make smarter, faster, and more data-driven assessments of how their power output can meet electricity demand.”
According to Google, this research could prove useful for strengthening the business case for the use of wind power. Google notes that it recently achieved 100 percent renewable energy purchasing and is working to source carbon-free energy on a 24/7 basis. This is an approach that fellow tech giant Apple has led the way in with its use of sustainable energy to power all of its operations.
It could also turn out to be a valuable use case for DeepMind. While the company — which Google purchased for $500 million in 2014 — has been a constant generator of fascinating research, it has not been a big moneymaker. Some reports have suggested that since it’s been a loss-making company ever since Google purchased it, the search giant is concerned about finding ways to monetize DeepMind. This might not be a solution able to make DeepMind profitable overnight, but depending on what this latest research means in monetary terms (Google doesn’t share that information), it may represent one promising direction forward.
- Data storage and dirty energy: What Big Tech’s carbon neutral pledges leave out
- Facebook’s new A.I. takes image recognition to a whole new level
- What’s in Joe Biden’s $2 trillion climate plan?
- Researchers have found a new way to spot the latest deepfakes
- A.I.’s next big challenge? Playing a quantum version of Go