Because of their great energy density and efficiency, lithium-ion batteries (LIBs) are essential to renewable energy systems and electric vehicles. Effective battery management requires precise estimation of the state of health (SoH) and state of charge (SoC). In order to overcome the difficulties caused by parameter fluctuations and real-world disturbances, this work presents a novel μ-analysis-based methodology designed to improve the resilience and accuracy of online SoC and SoH estimations in LIBs. In contrast to conventional techniques, the suggested strategy successfully manages both structured and unstructured uncertainties in battery systems by combining μ-analysis with model-based estimation. The framework creates an estimator that is resistant to parameter drift and outside perturbations by combining model-based estimation approaches with μ-analysis tools. Simulations using UDDS, US06, and HWFET driving cycles are used to verify its performance. When evaluating battery health and condition in dynamic and uncertain operating scenarios, the μ-analysis-based estimator demonstrates superior accuracy compared to conventional H∞-pole placement filter methods. The proposed approach enhances system robustness, achieving an 8 dB improvement in disturbance attenuation, as verified through MATLAB/Simulink. Stability analysis reveals the μ-analysis controller maintains robust performance up to ‖∆‖∞ = 3.5 at 10 Hz, compared to only ‖∆‖∞ = 1.5 for the H∞-pole placement controller—demonstrating significantly greater tolerance to parameter variations and unmodeled dynamics. These capabilities make the μ-analysis approach particularly suitable for electric vehicle applications requiring next-generation battery management systems.
Nohra et al. (Mon,) studied this question.