An ICA-Based Adaboost GRU Ensemble Model for Estimating SOH of Secondlife-Battery

When the SOH of a lithium-ion battery (LIB) falls below 80%, the electric vehicle may no longer meet its energy and power requirements, resulting in its disposal at the end of life (EOL). The disposal of used batteries presents environmental challenges because they pose fire and explosion risks when exposed to the external environment. Hence, alternative approaches that involve reusing discarded batteries in energy storage systems (ESS) with low energy and power demand have been actively researched to create both environmental and economic value. To efficiently and safely operate discarded batteries, reliable SOH estimation techniques based on data-driven approaches have gained popularity. This paper proposes an AdaBoost-GRU ensemble model to estimate the SOH of post-EOL batteries using incremental capacity analysis (ICA) as the primary feature. The combination of ICA-based features and the AdaBoost ensemble approach results in more accurate SOH estimations for post-EOL batteries, paving the way for effective and environmentally responsible reuse of discarded batteries in energy storage systems.