Statistically Designed Power Processing Architecture for Second-Life Batteries

#Battery#SecondLifeBattery#BatteryManagementSystem#BMS#EnergyStorage#HardwareInLoop

--Technical Webinar Series by IEEE Benelux Section Transportation Electrification Council Chapter--

Topic: Statistically Designed Power Processing Architecture for Second-Life Batteries

Speaker: Dr. Xiaofan Cui, Assistant Professor, University of California, Los Angeles (UCLA)

Bio: Xiaofan Cui is an Assistant Professor in the Department of Electrical and Computer Engineering at the University of California, Los Angeles (UCLA). He was a postdoctoral researcher at Stanford University from 2022 to 2023. Dr. Cui received his bachelor's degree in Electrical Engineering and a dual bachelor's degree in Economics from Tsinghua University in 2016. He earned his Master's degree in Mathematics and his Ph.D. in Electrical and Computer Engineering from the University of Michigan, Ann Arbor, in 2022. His research interests include the modeling, control, and design of energy systems and high-performance power electronics. He has received the IEEE Power Electronics Society Best Paper Award, Hellman Fellow Award, UCLA Innovation Faculty Fellowship, the Michigan Translational Research and Commercialization Award, and the Towner Prize for Distinguished Academic Achievement.

SUMMARY: Second-life battery energy storage systems (SL-BESS) have emerged as a cost-effective and sustainable solution to the rapidly growing supply of retired electric-vehicle batteries. Unlike new batteries, however, second-life batteries exhibit highly stochastic availability and largely heterogeneous cell characteristics. Without proper management, these variations can lead to severe performance degradation. Moreover, conventional power-electronics architectures for managing SL batteries become inefficient and costly under such conditions, undermining the economic value of SL-BESS.

In this talk, I will begin by summarizing key insights gained from real experiments on commercially retired EV batteries, tested in a laboratory environment that emulates grid usage. I will then introduce a statistically designed partial-power-processing architecture to mitigate the stochasticity, heterogeneity, and thereby substantially reduce balancing costs. The underlying design principles generalize across a wide range of applications, including DC microgrids, behind-the-meter storage, and utility-scale energy storage. Validated through Monte Carlo analysis, statistical hardware-in-the-loop simulations, and hardware demonstrations, the proposed architecture consistently outperforms conventional designs across diverse battery configurations.