New artificial intelligence (AI) tools could enhance the reliability of the power grid and electric vehicle (EV) charging infrastructure, according to a study by the University of Michigan Transportation Research Institute (UMTRI) and Utilidata, an AI-powered technology company.

Researchers are leveraging AI to analyse EV charging behaviour, aiming to improve the experience for EV users and help utilities prepare for the surge in electricity demand.

Preliminary findings from this initial study says EV charging can cause unexpected rapid swings in current draw and it lowers and causes variability in local voltage. Moreover, EV charging lowers power quality by introducing current harmonics generated from the conversion of AC to DC power. Lastly, EV charging in the presence of a grid fault showed potential to support voltage recovery, according to the findings.

These underlying issues waste energy and could lead to malfunctioning EV chargers. AI's ability to identify and predict these problems could revolutionise the industry. AI models could provide utilities with early warnings on how charging might impact the power grid. They could also advise EV users on optimal charging times and locations, and assist EV charging companies in maintaining their equipment more effectively.

For the study, researchers installed electric metre adapters equipped with Utilidata's AI platform Karman at six EV charging stations at the University of Michigan campus. Devices were also installed on the vehicles of ten drivers who frequently charge on campus to monitor their habits. Data for the study was collected between 3 March and 27 June of last year which included start and stop time for charging, location of charging, trips taken, and acceleration/deceleration. UMTRI collaborated with Utilidata to initiate this pilot study, which will inform a larger research project addressing the same concerns.