The main challenge with lithium-ion batteries in vehicular applications is aging. It is known that the battery aging is sensitive to various factors such as current, temperature and depth of discharge. These elements have a considerable impact on the loss of the battery’s capacity, as well as on the increase of the internal resistance. So, these two parameters can be considered as the aging indicators. However, these indicators are not measurable and they must be estimated. In the first part of this study, we present an analysis of the aging of lithium-ion batteries in order to predict their failures. The comprehension of aging can retroact on the operating conditions in order to improve reliability. Thus, the work carried out involves the experimental analysis of a LiFePO₄ battery’s calendar aging under different discharge currents and temperatures.

In order to estimate the aging indicators and evaluate the aging evolution, the second part of this study proposes a hybrid State-of-Charge (SOC) and State-of-Health (SOH) estimation technique for lithium-ion batteries according to surface temperature variation. The hybrid approach uses an adaptive observer to estimate the SOH while an extended Kalman filter is used to predict the SOC. Unlike other estimation methods, the closed-loop estimation strategy takes into account the surface temperature variation and its stability is guaranteed by Lyapunov direct method. In order to validate the proposed method, experiments have been carried-out under different operating temperature conditions and various discharge currents. Results highlight the effectiveness of the approach in estimating SOC and SOH for different aging conditions.