Data-Driven Design of a Cascaded Observer for Battery State of Health Estimation

The design of a combined observer for battery state of charge (SoC) and state of health (SoH) using data-driven models is addressed. Knowledge of both SoC and SoH is an important topic in battery management system design for (hybrid) electrical vehicles. The proposed cascaded observer is designed for estimating SoC and SoH during normal vehicle operation based on measured signals, which are directly available in the car. For an SoC estimation, that is accurate in the entire battery lifespan, the underlying battery model has to be adapted based on the estimated SoH. For that purpose, ageing data analysis using data-driven models and the associated model adaptation is discussed. A cascaded observer structure for combined estimation of SoC and SoH is presented and the performance of the proposed concepts is demonstrated by means of battery ageing data covering the entire lifetime of a lithium-ion cell.