Palo Alto teen’s lab model estimates EV battery health with 2.36% error

At 18, Colin Jie Chu was already working on a question that follows every electric vehicle into old age: how much useful life remains in its battery? The Palo Alto student developed a laboratory model that estimates a lithium-ion battery’s state of health with a reported prediction error of 2.36%.

The problem is hidden inside the battery-management system, which cannot directly inspect a cell’s degradation. Engineers instead infer its condition from electrical behavior as the battery is charged, discharged and exposed to changing temperatures and operating conditions. Chu used data from 22 deliberately aged batteries, with electrical signals designed to mimic changing vehicle driving behavior.

His framework pairs two approaches. An equivalent circuit model represents a battery’s complex electrical behavior through mathematical components that are easier to analyze; machine learning then helps estimate health as operating conditions change. The work was conducted through Stanford University’s Young Investigators Program at Professor Simona Onori’s Stanford Energy Control Lab, alongside researchers and industry partners.

The practical payoff is less guesswork around one of an electric vehicle’s most expensive components. More accurate health estimates could help battery-management systems guide charging and maintenance, improve decisions about whether a battery should be reused or recycled, and clarify how much performance remains as usable capacity gradually declines. For owners, battery health could eventually matter alongside mileage when judging a vehicle.

The result is not yet a universal battery-life predictor. The model was tested on research data, and it still needs validation across different battery chemistries, battery ages, vehicle platforms and real-world driving conditions. Chu’s research was presented at the Modeling, Estimation, and Control Conference in Chicago and published in the Journal of The Electrochemical Society; his work now also extends to battery “kneepoints,” when degradation can begin accelerating.