There isn’t often a day that goes by now where the letters A and I, said consecutively, don’t cross the human consciousness. Once a concept only known to tech-focused minds has now been catapulted into the mainstream, artificial intelligence (AI) has well and truly taken over, with its benefits being harnessed left, right and centre. And that is no different when it comes to the energy sector.
Historically being regarded as an industry slow to take on and utilise emerging technologies despite literally powering modern life,the energy sector has come on leaps and bounds in recent times in not only the energy sources used, but when it comes to harnessing the benefits that come with digitisation and modern technology.
Here’s where predictive AI comes in. Not to be confused with generative AI (Gen AI), which has a primary focus on creating new and original content, predictive AI is a form of intelligence that uses patterns in historical data to forecast or classify, providing actionable insights and support when it comes to decision-making and strategy formulation.
AI’s role in changing the energy sector
AI, according to a 2023 GlobalData report, is set to revolutionise the potential of predictive maintenance, improving system efficiency and even international energy security, thanks to its ability to identify and resolve problems before they have a chance to disrupt operations.
The Thematic Intelligence: Artificial Intelligence in Energy report shines a light on how AI is changing the energy sector, suggesting that AI will increasingly be used to analyse real-time data. Key findings show that it can be leveraged to detect and repair faults, using monitoring, thermographic and analytical technology.
Predictive AI plays a vital role in increasing energy’s reliability
Colin Gault is the Head of Product at Scottish energy software firm POWWR. A chartered engineer with a energy-centric career, he has had a front-row seat witnessing the emergence of applications for predictive AI when it comes to maintaining and optimising energy assets. “The advances in the technology has been rapid,” he proclaims. “But the challenge has been in supplying the ‘right’ data. This is being overcome due to the wider digital transformation of the industry.”
In the words of Gault, predictive AI is helping to increase the reliability of the entire energy system, as it not only informs when an asset is at a higher risk of sustaining damage and in need of preventative maintenance but combines this with other data like weather and traffic to support dispatching engineers to the site optimally.
This is echoed by Bill Wilson, Chief Environmental Sustainability Officer and Head of Data and Intelligence Solutions at NTT DATA for the UK and Ireland. Wilson — who has 25 years of experience in consulting and a extensive experience across the energy landscape, including working for energy giants like bp, EDF and Trafigura — has also observed the power of predictive maintenance in other industries, such as transport, defence and telecoms.
“I’d like to see similar optimisations and problem-solving methods applied in the energy sector,” he shares, acknowledging its power and use in the industry thus far, while also noting it has a fair way to go in being used to its full potential.
“In manufacturing, audio information and audio learning models can partly replace the trained engineer who knows when a machine does not sound healthy,” he further explains. “Of course, the same predictive techniques can be used to gather data from several sources to inform condition maintenance. This requires less predictive power because it’s more concerned with accurately understanding the overall condition of an asset based on multiple inputs. Sensors can give false readings, but techniques such as cohort analysis allow these anomalies to be identified by comparing each sensor with others doing a similar role.”
Part of this, and what can be argued as a minor teething problem, is how the training data is sourced and subsequently implemented. As Wilson explains, traditional predictive AI relies on having enough of the right telemetry data, clean information about fault occurrence and relevant context. If the incidence of faults is rare, this makes it harder to train AI models, but techniques such as random forest can be used to mitigate this issue.
Offering a solution, Wilson details: “More recently, drone surveys of visible assets have started to complement this approach, while generative AI can produce additional training data that enables computer vision models to be trained to spot faults.”
Predictive AI’s role in predicting energy forecast patterns
Gault is a firm believer in that the transition to net-zero disrupts both the supply side and the demand side of the energy system.
“For example, electric vehicles (EVs), residential solar and electric heating are continuously changing demand patterns,” he shares. “At the same time, an increase in renewables on the grid is causing fluctuations in supply capacity. Add to this more frequent extreme weather events and you can have challenging supply and demand patterns.”
By being able to better predict when the energy system will experience an imbalance in supply and demand means that the charging of EVs can be scheduled better to ensure the balancing of the grid, with a reward being cheaper electricity and, if the charging can coincide with when there is a renewable energy supply, then the CO₂ output can also be reduced.
“The changing dynamics of battery and EV storage, as well as the activities of prosumers, are making the electricity grid increasingly unpredictable,” Wilson adds. “In a paradigm like this where statistical models are unable to deal with the number of unknowns, we can use machine learning models and start to trail more novel, or ‘alternative’, data sources.”
In his role, Wilson oversees how data generated by other sectors can be shared with the energy supply industry, for example, a telco client experiencing unexpected extremes in power demand.
“AI and energy efficiency is a topic all by itself and there are many mature solutions,” he starts, giving another example. “We partner with a firm that uses AI to predict when file servers will be in demand and turn down the clock speed at other times, saving 15-30% of energy needs.”
“I took part in an energy industry round table where the strong consensus was that AI is more than a measure to improve efficiency. It was clear that executives viewed AI-driven optimisation as an essential stop-gap while physical infrastructure catches up – taking us towards a more flexible, higher capacity, and lower-carbon future.”
Could predictive AI mitigate risks to the energy sector?
A big part of working to said lower-carbon future can be attributed to efficiency in other areas such as supply chain. This, along with other unpredictable problems like extreme weather events and cybersecurity threats, has a circular impact on the slickness of the sector’s operations and its green credentials as a result.
But both Wilson and Gault firmly believe that technology, such as predictive AI, enhances the sector’s overall resilience and subsequent operations.
“A big risk to the energy sector is energy imbalance,” Gault critiques. “The ability to accurately forecast is imperative to being able to mitigate supply-demand imbalance. Extreme weather not only impacts supply and demand profiles but can damage power lines and prevent power plants from operating properly. There are certain innovative projects, such as one by Scottish Power, that are aiming to better predict when extreme weather events will lead to power outages and where these outages will occur to provide them with enhanced intelligence.”
Wilson concurs, signposting the industry to understand that, to properly be conscious of the application of predictive AI, there must be a process of subdividing the categories of risk - on top of the others previously mentioned, including physical risks of climate change which put infrastructure at risk.
He elaborates: “Combine this with geology information which shows us that some buried assets are more at risk than others.”
“Other AI systems will be responsible for predicting the greatest pinch points in the grid so that new developments do not create local resilience problems. In other words: the UK energy grid is highly resilient today, but AI can help maintain that resilience while dealing with increased renewable generation and its fluctuating output. We only want fossil fuel capacity online when it’s needed, but it takes time to bring this capacity online.”
This is where detailed AI predictions about supply and demand can help optimise this process, Wilson shares, adding that as this trend is more widely adopted the energy industry will be able to make use of alternative datasets from outside the sector to make it stronger, and more resilient, than ever.
Gault concludes: “The use of predictive AI is already prevalent in many new projects, however, it is still in the emerging technologies phase and needs to overcome the challenges of scaling up. That is where a lot of digitalisation of the energy sector will focus over the coming year. The industry has started to envision a digital twin of the energy system where predictive AI and open data combine to better plan and operate a much more distributed and flexible energy system.”