Virtual ZiyouAI

The work in brief

Ziyou's work addresses how lithium-ion batteries are managed, monitored and reused in electric vehicles and stationary energy systems. The earliest papers (2014–2019) focus on battery/supercapacitor hybrid energy storage systems for electric vehicles and buses, covering topology comparison, component sizing, and energy management via dynamic programming, model predictive control and sliding-mode control, often with battery degradation or low-temperature operation as objectives. In the same period the work develops estimation methods grounded in identifiability analysis, including Cramer–Rao bound–based current profile design and active current injection for combined state-of-charge and state-of-health estimation.

From around 2020 the focus broadens to cell-to-cell variation and fault detection in parallel-connected cells, bias-compensated Kalman filtering for flat-OCV LFP cells, and integrated power and thermal management for connected and automated EVs using multi-horizon MPC. Second-life batteries, recycling automation and EV fleet charging also appear. The most recent papers lean toward data-driven and physics-informed learning, often with an industrial battery partner: state-of-health estimation from EV field data and partial charging profiles, physics-informed neural networks for degradation diagnostics, swarm and meta-learning for data-limited settings, and 'Discovery Learning' for predicting cycle life and evaluating battery designs from minimal experiments.

AI-written from the paper titles and abstracts; Ziyou hasn't reviewed it.