| Digitalisation Microsoft data centre, Quincy, WA, USA (photo: Microsoft)
and assist with grid planning, line routing, transformer placement, asset management, and day-to-day operations. AI models help generate synthetic data for grid planning, visualise grid designs, predict maintenance needs, and analyse sensor data to identify potential issues early. Power companies like PG&E and Duke Energy are already using AI for infrastructure inspections and anomaly detection, enhancing grid reliability and reducing downtime.
AI also improves real-time assessments of power transmission line capacity through dynamic line rating, which uses environmental and load data to provide more accurate – and often higher – capacity ratings. AI-based models of customer load profiles are proving useful in grid modeling, demand response, and overall grid operations. Competitions like Learning to Run a Power Network (L2RPN) highlight AI’s potential to optimise complex grid operations, hinting at future automation.
AI plays a crucial role in energy storage by optimising battery charge and discharge cycles, reducing energy losses, and extending the lifespan of battery systems. DNV’s Battery AI service predicts battery life based on usage patterns, while
AI-driven simulations help determine the ideal size and placement of energy storage systems. Machine learning is also accelerating the discovery of novel battery chemistries, further advancing storage technology.
Machine learning is key to optimising the balance between energy supply and demand, improving the financial viability of renewable energy, and integrating it into the grid. Companies like DNV’s Solcast provide solar irradiance data and forecasts using AI algorithms, while DNV’s Forecaster service employs advanced statistical methods and machine learning to deliver short-term forecasting for wind and solar sites worldwide.
AI applications in optimising the siting and design of renewable energy installations are widespread. For example, researchers in China have developed a machine learning model that uses satellite and sunshine data to optimise the placement of double-sided solar panels. In the USA, DNV partner HST employs AI-based decision engines and data analytics to optimise large-scale solar installations, connecting developers with energy buyers and other key project stakeholders. AI is transforming materials science, particularly
in such areas as discovering and selecting materials for power generation components and new battery chemistries. The US National Renewable Energy Laboratory (NREL), for instance, has used invertible neural networks to significantly accelerate the design process for wind turbine blades. Additionally, AI is being used to assess molecules for their stability and chemical properties to enhance carbon capture technologies and identify new approaches for carbon capture and storage (CCS) projects. AI can also streamline complex and time- consuming approval and permitting processes. For example, in the nuclear sector, AI could enhance reactor design, plant operations, and training. AI is also being applied in environmental planning, where it utilises sensor networks and drones for biodiversity monitoring, aiding data visualisation and report generation. This could significantly speed up environmental assessments and permitting processes.
A double-edged sword While AI promises to drive substantial advancements in the energy sector – for example, enhancing power grid management, optimising energy storage, accelerating materials discovery, and improving renewable energy siting – it also brings challenges.
Leaving the moral and ethical debate to one side, the technology’s own growing energy footprint must be monitored and accounted for, or else it could become a blocker rather than an enabler.
The key question remains: how can we balance AI’s transformative potential in the energy transition with its rising energy demands? Though AI is not the primary driver of the transition to net zero, its role is significant, and ongoing efforts to improve energy efficiency, particularly in data centres and hardware, are essential. Its long-term impact on both the energy sector, and broader society, will likely be underestimated, and its influence will continue to grow in tandem with advancements in responsible AI development.
The Eco Data Center in Falun, Sweden, employs 32 mtu emergency backup gensets, which have been converted to HVO. Photo: mtu/Rolls-Royce Power Systems
www.modernpowersystems.com | November/December 2024 | 39
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45
