With the green energy transition fully underway, artificial intelligence is a welcome helping hand – or head – in an energy industry challenged by a shortage of engineers and the need to find and deliver the much-needed sustainable solutions.
COWI’s Digital Transformation Director Stefan Skanderup Brandt sees a number of areas where the industry can benefit from AI, from predictive maintenance and data-driven decision-making to sharing knowledge from earlier projects and removing routine tasks.
“The industry is facing a huge challenge recruiting an adequate number of engineers and specialists capable of designing all the projects out there at the required pace,” says Stefan. “If AI can take some of the routine tasks off our hands, then engineers can use their skills to design what really matters. Engineers are a scarce resource, and AI can help the ones we already have to focus on the right, value-adding tasks.”
Some routine tasks have already been identified as obvious candidates for AI-driven automation, and to protect the proprietary data accumulated through numerous products, the data is kept in a closed environment.
“We’ve already established an AI playground based on GPT, and we’re looking to optimise some very simple tasks,” Stefan says. “For example, report generation.”
“The next step will be training internal chatbots in our expertise on specific domains, paving the way for a more active utilisation of AI in the actual design process.”
Data-driven decision-making
Offloading documentation is another way AI can play a role, and according to Stefan, having AI do some of the work will also help make the documentation more consistent.
“With the huge data sets we’re gathering in an organisation like COWI, and in these major design tasks with thousands of datasets in use, we can use AI to ensure uniform formatting. This enables us to learn more effectively from the data we have.”
This more structured approach to data can allow for better decision-making.
“Having a more structured approach to how we handle data could lead us to a more data-driven decision-making process,” notes Stefan. “With AI, we have the potential to identify valuable patterns that can lead us towards making the right decisions earlier in the project.”
Stefan highlights that data collected and processed by artificial intelligence could optimise renewable energy production relative to weather conditions, grid demand and more, leading to higher cost efficiency and reduced environmental impact.
“In order to deliver the complicated designs we do in complex projects, we sometimes have to be a bit conservative in our design assumptions. Sometimes, we deliver a good design, but in a broader perspective, I’m asking myself, ‘Could it have been a better design? A more sustainable design?’”
By using artificial intelligence to, say, run more design iterations of a future wind farm, it becomes possible to produce a more optimised design, leading to a reduction in the amount of steel needed in the construction, which will reduce the carbon footprint.
Predictive maintenance
Stefan highlights predictive maintenance for energy installations, which can extend their lifetimes, as another area in which AI can play a key role.
For example, AI can comb through sensor data on Power-to-X plants and solar and wind farms and identify incident patterns leading to downtime. Predicting and scheduling maintenance before a breakdown occurs save money and ensure that solar and wind farms have uptime whenever the sun is shining or the wind is blowing, eliminating the need to be supplemented by fossil fuels.
50 per cent of the way from the offset
In the quick-moving world that is global energy, change is constant. Staff and corporate structures change, yet retaining the proprietary knowledge gained from previous projects is crucial.
“We often have young engineers asking the experienced engineers, ‘Tell me about this project you did ten years ago. What was your design approach? What was it like to get the approvals?’”
This is another area where artificial intelligence can help. “What we’d love to do is train a chatbot in, for instance, all the wind farms we have delivered, so the next time we’re asked to design something like a boat landing, we could have the AI output the first draft of a design based on our previous projects.
“It might only be 50 per cent of the way there,” Stefan says, “but then we could use our time more efficiently to deliver a design with a lower carbon footprint. There’s no need to reinvent the wheel – or wind turbine – every time.”