The government already recognises that artificial intelligence will play a key role in decarbonising energy.
That’s why, along with agriculture, the built environment, and manufacturing, energy is a sector targeted by the Artificial Intelligence for Decarbonisation’s Virtual Centre of Excellence (ADViCE) programme, which aims to promote collaboration between AI innovators and carbon-emitting industries.
The Digital Catapult is leading ADViCE, with support from Energy Systems Catapult and Alan Turing Institute, as part of a much wider initiative to decarbonise industry with the help of AI. But what are the true implications of the technology for energy, and where will it be of most benefit?
Utility Week caught up with Sam Young, practice manager for data science and AI at Energy Systems Catapult, to find out.
Sam, let’s start with the big question: what contribution do you think AI will make to hitting net zero?
SY: The way to think about the energy system is that there are lots of challenges to rapidly decarbonising. AI is not a silver bullet, but some of the factors in the system will require increasing automation like digitalisation and decentralisation.
On the one hand, AI can be an accelerator – for example, it could help accelerate the assessment of properties for installing heat pumps and automating some of the design process. We are also going to have so many devices to control and so much data we will need AI to deal with them.
AI has been hyped so much over the past year and a half or so. How do you view its current capabilities? Are some of the tech industry’s concerns justified?
I would draw a distinction between predictive AI and generative AI. A lot of the hype is around generative AI. With predictive AI, there aren’t huge new elements of risk that need consideration: you need people who understand the technology, good governance, and to make sure your data isn’t rubbish.
The shift from an engineering, physics-based approach to a data approach is a cultural shift we need to manage a sector, but that’s not unachievable. We can have good processes around that.
With generative AI, we are in the middle of the hype cycle. Someone was saying to me recently we are going to have an ‘AI winter’ where there is all this hype and then it collapses with nothing really having changed. Or maybe it will be like the dot com boom in the early noughties where the bubble burst but in the aftermath the world had changed and the whole structure of industry was different.
Should we worry about AI? My sense of it in terms of generative AI is that there are a whole load of challenges and limitations with generative AI. It’s easier to produce a cool demo than a robust production system.
There will be some really good, transformative applications, but many others that aren’t.
As an expert, have you been surprised by the development of AI over the past few months?
I think the jump in capability of generative AI and the level of awareness surprised everyone. As a data scientist, I think to some extent AI is to data science what data science was to statistics a few years ago. It’s the new label you apply.
A lot of what my team does is not building deep neural models. Instead, we have two billion measurements from heat pumps, and it’s about calculating efficiently how those heat pumps are performing. Do we need to remove anomalies or resample at half hourly intervals?
There is a lot of data cleaning that needs to be done with statistics and relatively simple machine learning algorithms can process the data and draw those conclusions. You hear a lot about AI focused on language and images and video, and not so much about time series and sensor data, which is where there is a lot of value for the energy sector.
It’s not you just putting your request into an online portal and it solving your problem.
Where can AI be of most benefit to decarbonisation of energy?
As part of ADViCE we’ve identified a number of challenges where AI can help and we are bringing people together to create the ecosystem to deal with them. One challenge is unlocking and accelerating domestic decarbonisation, including unlocking financing and getting buy-in from consumers. Another is around utilities getting to net zero with advanced construction timelines and all the new grid connections that are needed; the third is that AI will be crucial to the introduction of flexibility services.
We’re already seeing AI helping with speeding up planning overhead lines. And we are working at a domestic level on making the installation of heat pumps easier by automating assessment of heat loss and calculating savings from retrofit.
That is the type of application where there is lots of value to be had from AI.
Where are people arguably going wrong with the technology?
People are thinking ‘can I solve this problem with AI’, rather than ‘will solving this problem accelerate the journey to net zero’? Pick the problems that really need solving rather than thinking about whether AI can be applied for the sake of it.
In terms of innovation, people should think about whether new business models can be introduced with an AI tool – heat as a service, for example. The business model won’t work without the data-driven AI component. By pairing AI with a new way of doing things, you can completely change a process, rather than merely automating an existing one.
What are some of the areas that are going well?
One of the big success stories is open data in the energy sector. It is really good to see networks participating in sharing aggregated smart meter data. One of the areas we are looking at is how to establish an infrastructure for sharing data that can’t be publicly disclosed.
We need to build even more momentum behind sharing data so people can use it to build new products and business models and train AI.
How does it feel to be working for an organisation with a strong focus on decarbonisation?
I moved from financial services because I decided I didn’t want my life to be about getting people into more debt; I wanted to do something more productive. Tackling the massive challenge of how we decarbonise our whole society is vital for the future.
It’s not going to happen with inertia: we need to work really hard to accelerate and innovate on our journey. We won’t achieve what we need to achieve unless we do things differently. I’m excited by that challenge.
Want to hear more from Sam? He’s one of the speakers at this year’s Utility Week Live in May. Register for your place now here.