A voltage reconstruction model based on partial charging curve for state-of-health estimation of lithium-ion batteries

Battery in-situ state-of-health (SOH) estimation has attracted considerable attention, but essentially, most studies focus on the normalized capacity or resistance, while the underlying aging mechanism is vacant. Therefore, we propose a voltage reconstruction model, which not only accurately estimates the SOH but also quantitatively identifies the aging modes. The model takes into consideration the limited battery operation range in practice and the over-potential caused by large current rate. Based on the matching relationship from the half-cell electrode equilibrium potentials to the full-cell terminal voltage, the complete terminal voltage is reconstructed via the partial charging data, and the SOH is estimated by a specific cut-off voltage range. We introduce matching differential voltage curves in the optimization objective of the model to further reduce the required data while achieving high accuracy. The model accuracy is investigated from the perspective of state-of-charge ranges and input data quality. The applicability is verified on various aging states and different cell designs. It is concluded that based on the limited data that contain one phase transition of the positive electrode and the negative electrode, the proposed model can estimate SOH with relative errors less than 2.5%, and quantify the potential aging modes, which holds promise for practical applications from a cell level to a pack level.

Automated summary

The study presents a voltage reconstruction model that accurately estimates battery state-of-health and quantitatively identifies aging modes. By optimizing the model with matching differential voltage curves, data requirements are reduced while achieving high accuracy. Experimental results show that based on limited data, SOH can be estimated and aging modes quantified, holding promise for practical battery applications.