Physically based electrochemical models can provide internal state information of batteries, exhibiting significant potential for application in battery management systems across various lithium-ion battery products. However, the complex structure and limited observability of electrochemical models make their integration into battery management systems challenging. In this paper, a novel nonlinearity-aware adaptive (i.e., NAA) observer based on a simplified electrochemical model is proposed for estimating the electrode surface concentration and state of charge. The proposed NAA observer can identify the nonlinearity of the electrochemical model and switches to the most suitable observer structure automatically. Experimental tests on a cylindrical battery were conducted to observe the performance of the observer. The results demonstrate rapid convergence and high accuracy of all estimated states under an initial state error of 20%. Furthermore, to thoroughly validate the performance of the proposed observer, a comparison was made with an interconnected sliding mode observer based on electrochemical models, considering initial state uncertainty, sensor noise, and computation time.

中文翻译：

---

一种新颖的非线性感知自适应观测器，用于估计锂离子电池的表面浓度和充电状态

基于物理的电化学模型可以提供电池的内部状态信息，在各种锂离子电池产品的电池管理系统中展现出巨大的应用潜力。然而，电化学模型的复杂结构和有限的可观测性使其集成到电池管理系统中具有挑战性。在本文中，提出了一种基于简化电化学模型的新型非线性感知自适应（即NAA）观测器，用于估计电极表面浓度和充电状态。所提出的 NAA 观测器可以识别电化学模型的非线性并自动切换到最合适的观测器结构。对圆柱形电池进行了实验测试，以观察观察者的性能。结果表明，在 20% 的初始状态误差下，所有估计状态的快速收敛和高精度。此外，为了彻底验证所提出的观测器的性能，与基于电化学模型的互连滑模观测器进行了比较，考虑了初始状态不确定性、传感器噪声和计算时间。