By Duncan Bain, Senior Energy Advisor at SAS UK, a data and AI leader
Duncan Bain
As energy providers and network operators work to understand how they can meet their Clean Power 2030 goals set by the UK government, many are already taking advantage of AI and machine learning in various forms to help them get closer to their net zero targets.
But in embracing the technology, some of the challenges they have faced include progressing AI and machine learning projects from the ideas stage through to production. This includes ensuring sufficient processes are in place for operations teams to manage model performance safely and securely, and dealing with the increased costs of cloud computing.
Many of these issues were highlighted by senior leaders from TOs (Transmission Operators) and DNOs (Distribution Network Operators) at a recent Utility Week roundtable where they shared their first-hand insights on AI use cases in the industry, the investments they’re making and how the technology could be used to deliver better networks in the future.
The contributors looked at how AI was helping networks negotiate their workloads and changes to their sector, and discussed how the technology might be used in years to come for everything from asset management to stakeholder engagement.
How is AI being used?
Network operators shared a multitude of uses across their industry and how AI has changed their working practices and processes. They included the following examples.
• Saving time
One senior engineer shared that their utility company is already using AI technology to rapidly detect overhead line defection. AI can recognise chipped insulators in video footage from drones and helicopters and analyse point cloud data from light detection and ranging (LiDAR) equipment to identify conductors during assessments of dynamic line rating. The upshot is this is “saving hundreds of thousands of system planning hours.”
AI is also saving time when it comes to stakeholder engagement. One leader at a transmission operator said they’re able to process feedback much faster using AI, “saving more than 80% of time…it’s down to minutes and hours, rather than days and weeks.”
• Speed
One of the valuable benefits of AI is in helping DNOs and TOs to digitise and decarbonise rapidly and they agreed that this alone was a determining factor as to why they had decided to implement the technology. One public sector AI expert said: “Where can AI or automation of a process speed things up by a factor of 10? How would massively reducing delays change the way the system worked?”
• Managing huge projects
AI is being used by leaders to forecast project performance and delivery times. A project director from a TO said eight out of 10 of their projects didn’t currently deliver the expected benefits but AI could take the programmes and schedules, compare them to previous projects and assess their deliverability and more accurately predict when the project would end. “The software will tell you if there are different critical paths you need to be made aware of,” they said.
• Sorting data
There is no denying that AI is helping to process huge volumes of data, but it’s important to choose wisely when it comes to which data to sort and how it is managed. One roundtable attendee said it could potentially take years or even decades for their TO to get their data in order and avoid “ending up with an infinite bill for data management.”
Having thousands of new data centres could end up drawing on much of the available clean energy, which all the experts recognised the irony of, but each of them understood that they were needed.
One expert shared that firms still need to recognise the value of data, adding: “One of the biggest roles for AI is data quality – going in and helping us cleanse our data.” Adding significant numbers of new assets to the field as networks grow and change to meet evolving demand will add further complexity. Using AI to leverage data from previous projects will enable project managers to manage these developments more easily by automating low value tasks, identifying errors and freeing up time for proactive management.
Working through the challenges
Network operators shared many challenges with using AI, such as the skills shortage as the existing workforce ages, and the increasing complexity of the network edge and energy transition. They also said it had been difficult to engage industry subject matter experts in the process of designing models.
One of the concerns discussed by everyone was whether AI was a threat to traditional roles at networks and whether it was replacing jobs?
The overall understanding was that while AI was speeding up tasks, its use still needed to be managed. One person shared their worries that it could affect engagement with employees as they may feel less motivated to use, or learn to use, AI if they thought it would later replace them.
They said one way to overcome this, which would be welcomed by staff, was to give people time to focus more on the rewarding elements of their job than the more mundane. “We are looking to enrich people’s jobs, not take them away,” they said. “People want to be out in the community talking to stakeholders, not analysing data and creating reports.”
Independent studies by the Futurum Group looked at how business challenges can be solved with AI and machine learning – from managing and cleaning data to automating and deploying insights. They highlighted how a platform like SAS Viya can help companies with these challenges, leading to faster performance, and delivering productivity gains across the AI lifecycle. Leveraging these performance and productivity advantages drives down the cost of cloud operations and using a trustworthy AI framework ensures value is delivered faster and more securely.
The future of AI in utilities
SAS has worked with partners in the utilities sector for many years, supporting firms to deliver AI projects through machine learning, natural language processing and analytics to overcome some of the biggest challenges. We want to help them to unlock the potential of emerging technologies and the unprecedented value they can bring at a crucial time for their sector.
Moving forward AI can help TOs and DNOs gain insights into their biggest projects to ensure they are working productively and efficiently. One of the engineers we spoke to shared that in the coming years they would like to be able to interrogate design documents using natural language processing, while a stakeholder manager said they saw the value in personalising stakeholder engagement for specific communities without the need for boots on the ground research.
Another discussed combining AI with robotics and how it can handle workloads and simplify processes, adding: “The volume of work is going to be massive: tens of thousands of connections and upgrades. If there is any aspect of these processes we can automate, that’s going to require AI.”
This sentiment was echoed by many throughout the event and another expert added: “Resources for decarbonisation are going to be key. The more you can do to de-skill tasks, the bigger the pool you can access.”
While there were concerns and challenges shared about using AI across the utilities sector, the positivity and optimism for how it can help TOs and DNOs attain their energy targets in the next five years was much stronger.
It is vital for energy leaders to continually have these discussions to learn from one another, figure out how they could be doing things differently and also share how they are embracing emerging technologies to make positive developments and changes to their organisations. And we hope they turn to trustworthy and secure AI platforms to support their digital transformations and achieve their sustainability targets.
