Employing machine learning to improve energy distribution

Electricity increasingly comes from renewable, carbon-free sources like solar panels and wind turbines, presenting new challenges for power generation and distribution. The intermittent nature of these energy sources, coupled with increasing demand from industry and electric vehicles, is driving the development of new tools and strategies to make grids more efficient.

Marković et al. outlined approaches to employ machine learning (ML) strategies in power distribution. They discussed the current state of ML technologies for distribution systems, highlighted areas where these technologies can have the most impact, and offered guidance regarding directions for future research.

Machine learning is a versatile tool, and can be used to forecast power generation, identify power consumption patterns, and replace traditional techniques with more efficient alternatives. For instance, ML algorithms can solve optimization problems and provide dynamic modeling of distribution systems faster and with fewer resources than conventional software.

“Our perspective delves into these topics and recommends reading materials for further exploration,” said author Marija Marković. “Moreover, it highlights certain application areas where ML shows promise but requires further research to realize its full potential.”

One area where ML methods have immense untapped potential is in analyzing incomplete or partial data sets, such as charting distributed solar arrays and improving the usefulness of poor-quality meter data, and missing measurements from last-mile sensors.

“With dedicated research efforts and the immense potential of ML, we are genuinely excited to witness the progress made in these research directions and the significant impact they will have on shaping the future of the field,” said Marković.

Source: “Machine learning for modern power distribution systems: Progress and perspectives,” by Marija Markovic, Matthew Bossart, and Bri-Mathias Hodge, Journal of Renewable and Sustainable Energy (2023). The article can be accessed at https://doi.org/10.1063/5.0147592 .