Online estimation of internal resistance and open-circuit voltage of lithium-ion batteries in electric vehicles

State-of-charge (SoC) and state-of-health (SoH) define the amount of charge and rated capacity loss of a battery, respectively. In order to determine these two measures, open-circuit voltage (OCV) and internal resistance of the battery are indispensable parameters that are obtained with difficulty through direct measurement. The motivation of this study is to develop an online, simple, training-free, and easily implementable scheme that is capable of estimating such parameters, particularly for the lithium-ion battery in battery-powered vehicles. Based on an equivalent circuit model (ECM), the electrical performance of a battery can be formulated into state-space representation. Also, underdetermined model parameters can be arranged to appear linearly so that an adaptive control approach can be applied. An adaptation algorithm is developed by exploiting the Lyapunov-stability criteria. The OCV and internal resistance can be extracted exactly without limitations of a system input signal, such as persistent excitation (PE), enhancing the method applicability for vehicular power systems. In this study, both simulations and experiments are established to verify the capability and effectiveness of the proposed estimation scheme.

Research highlights▶ This study is motivated to develop a unified method for estimating open-circuit voltage (OCV) and internal resistance of a lithium-ion battery via online voltage and current measurements. These two parameters can be used to determine battery state-of-charge (SoC) as well as state-of-health (SoH) via the built-in lookup tables that define the relationships between SoC/SoH and OCV/internal resistance. However, the lookup tables that shall take into account the thermal effect are not defined in this study. ▶ An adaptive control approach is employed to develop an estimated algorithm such that the developed method with the following merits: (1) online estimation; (2) simple mathematical operation; (3) specific load pattern or system excitation input, typically for an estimation system, is unnecessary; (4) no need prior time-intensive training; (5) easy implementation in comparison with existing methods. ▶ The proposed method has been verified through simulations and experiments.