Dan Wang, Sicheng Zhou, Zhibo Feng, Jun Wu, and Guanjie Wang
Lithium-ion battery, state of charge (SoC), intermediate observer, H∞ technology, SoC estimation
An intermediate observer that no longer requires the observer- matching condition is proposed for real-time state-of-charge (SoC) estimation of lithium-ion batteries under dynamic operation. Built upon a dual-polarization RC equivalent-circuit model, the algorithm introduces an intermediate variable to generate a residual that com- pensates for model errors, measurement noise and ageing distur- bances online. Lyapunov analysis proves that the estimation error is uniformly ultimately bounded, while the computational burden is noticeably lower than that of conventional sliding-mode or high- gain observers. Neware bench tests show a maximum error of 1.5 % under noise-free conditions, 53 % lower than the H∞ method, and only 2.8 % when 0.1 A sensor noise is injected, still outperforming the comparison methods. Without offline training, large datasets, or observer-matching conditions, the scheme is readily deployed in em- bedded battery management systems and offers a robust, lightweight SoC solution for diverse lithium-ion chemistries.
Important Links:
Go Back
