Clean Energy and AI: The Ultimate 'Power Couple'?

The International Energy Agency has labelled AI and energy as the ‘new power couple’ (pun definitely intended), with the technology being hailed as critical to the renewable transition. Indeed, from optimising scheduling, predicting maintenance, and forecasting energy flows, AI has been estimated to serve over 50 different purposes in the energy system already. This month’s instalment of The Digital Download will dig further into this claim, examining the ways that AI can help the adoption of future energy sources.

‍

‍

Smarter Grids

The Labour government’s manifesto labelled the national grid as the ‘single biggest obstacle to the deployment of cheap, clean power generation’. The UK’s power system is creaking as it is, and with electricity demand growing and energy flows becoming less predictable, integrating renewable sources whilst maintaining a reliable network is proving to be extremely challenging. There are several key ways in which AI can overcome such difficulties and aid the transition to net zero, even beyond its widely recognised ability to predict maintenance needs.  

For the average Brit, unpredictable weather patterns are usually no more than a minor frustration, or a useful nugget of small talk. However, they are a critical stumbling block for renewable energy sources – the sun doesn’t always shine, and the wind doesn’t always blow, so our grid needs to be able to handle this variability in energy supply. This is where AI comes in: neural networks can accurately predict future energy output up to 36 hours in advance using historical data. This not only allows for more reliable management of supply, storage and demand, but also allows power to be sold in advance rather than in real time, increasing the financial value of renewable sources and driving further investment.  

This is all well and good but hooking those now-predictable renewable energy sources up to the grid in the first place is easier said than done. Applications for connection to the UK’s energy grid currently face waits of over 5 years, significantly hindering the uptake of renewable sources. Reducing this queue to 6 months could not only provide enough energy to power the UK four times over for the next 25 years but would also save around £75 billion. AI matching algorithms can be leveraged to replace the current ‘first-come, first-served’ model with a more efficient model, allowing us to harness the renewable projects currently trapped in the queue.  

‍

Brighter Ideas

Whilst grid optimisation is essential for the adoption of future energy sources, AI’s potential is not limited to this sphere of efficiency. Innovation is crucial to the achievement of secure, affordable and sustainable energy, and AI has enormous potential to help us in these endeavours (and no, I’m not just talking about asking ChatGPT for ideas).

Researchers at the University of Saskatchewan in Canada recently made a significant breakthrough in the production of green hydrogen through the discovery of a more efficient and cost-effective catalyst. AI was a key driver of this advancement, analysing over 36,000 metal oxide combinations in just days – a process that would have taken years for humans to perform. The scientists were then able to focus their efforts on validating these findings in further laboratory tests, saving time, money and producing more reliable results. The clock is ticking when it comes to the energy transition, and AI can not only speed up these research processes but also improve efficiency in other areas of the commercialisation pipeline.  

Renewable energy research is a prime candidate for AI usage, with highly complex design spaces, large datasets, and performance trade-offs being exactly the kinds of problems it excels at solving. For example, less than 0.01% of possible perovskite materials for solar panels have been experimentally produced, despite the critical need for a more stable, cheap, and less space-intensive material – AI could significantly speed up this area of research. However, start-up data suggests that this is not yet widespread practice, with only 1% of energy-related patents referencing the use of AI in their innovation.

‍

‍

Early Days for the ‘Power Couple’

AI holds clear potential to help accelerate the energy transition, speeding up the innovation process for future sources of clean energy and optimising the grid for integration of that power. However, this power couple needs a little more quality time – there is an employment gap, hindering its widespread adoption in both areas, and more specific AI packages are still largely in development. Furthermore, the honeymoon period is definitely over: experts are raising concerns over the high energy usage of AI itself, as well as issues such as privacy and cybersecurity. AI could be a great match for renewable energy, but here’s hoping the stars will align.