Facebook seeks to super-charge storage

Technology giants such as Facebook, Google and Microsoft have faced serious scrutiny in recent years.

And arguably it is Facebook that has come under most pressure, from the role of Cambridge Analytica in the UK’s Brexit vote to Netflix series ‘The Social Dilemma’ about the corrosive effects of social media.

The firm has inserted itself in our lives, our phones, and many of the biggest trends – good and bad – facing our world today.

On the positive side, Facebook has been one of the technology companies leading the corporate shift to renewable energy. It unveiled 1GW of solar and wind power purchase agreements in August in four US states and the Republic of Ireland. And the company now plans to wield its influence in storage too.

On 14^th October, Facebook AI – the company’s artificial intelligence research arm – unveiled plans to launch the ‘Open Catalyst Project’ to speed up the technology of storing renewable energy.

The company said it is using the project to help to accelerate the growth of energy storage to tackle the intermittency of renewables and combat climate change. It is the latest example of a tech heavyweight moving into storage.

In an analysis piece in Energy Storage Report last month, we wrote about how tech giants Google and Microsoft are shaping their battery storage plans in response to the falling price of battery cells. Specifically, Google and Microsoft are set to direct significant investment into energy storage to help meet their net zero targets.

This isn’t an altruistic move to help a new sector. It is cold hard business.

Unleashing renewables

Google and Microsoft run services that generate huge amounts of data and require huge amounts of electricity to manage it. They are also under pressure to go green, so signing renewable energy PPAs – as Facebook has also done – makes sense. And investing in storage is the next logical step to protect the flow of green electrons into their data centres.

But Facebook AI wants to take this a step further. It doesn’t just simply want to invest in renewables or commit to long-term sustainability targets. It wants to play a central role to unlock new innovations in the industry. The ‘Open Catalyst Project’ is its drive to work with scientists to find new elements that can make batteries more powerful.

Now here’s the science bit. Facebook AI and Carnegie Mellon University plan to use artificial intelligence to “accelerate quantum mechanical simulations by 1,000x in order to discover new electrocatalysts” in battery storage.

Essentially, electrocatalysts are combinations of elements that can produce chemical reactions under the right conditions.

For example, lithium-ion is a combination of two elements – lithium and oxygen – that produce a reaction that creates electricity that can be deployed when required. But while lithium-ion can be effective as a short-duration form of energy storage, it is not suitable as the long-term storage needed to offset the intermittency of wind and solar power. It isn’t the silver bullet that can answer all storage questions.

Therefore, the industry must test combinations of 40 known elements that can be used as electrocatalysts in billions of different ratios and configurations.

Currently, scientists can try three or four possible combinations each year, which is not nearly enough. That’s where Facebook AI’s technology comes in.

By using AI to predict the outcomes of configurations, instead of experimenting by hand, this project hopes to make the project cheaper and faster. It doesn’t yet have the information to accurately predict the results of the configurations, but it’s a start.

That’s why the partners have released a huge data set with information on 1.3million electrocatalyst structures so that researchers can use AI and machine learning to try billions of possible catalysts per year. Each could be tested in “a matter of seconds”.

Larry Zitnick, research scientist at Facebook AI, wrote that the work has "the potential to significantly accelerate the global shift toward renewable energy" as it would "remove the high costs associated with current electrocatalysts".

"As energy needs continue to climb and the fight against climate change grows more urgent, this problem offers a chance to advance AI in a way that will have a significant real-world impact," he said.

If this helps to unlock further evolution in energy storage and growth in renewables, then we think it should be good news. Perhaps not enough to mitigate some of the more corrosive impacts of social media, but potentially a big step forward. With great processing power comes great responsibility.