Internal short circuit diagnosis of lithium-ion battery packs considering incomplete charging under cell inconsistencies

Internal short circuit (ISC) diagnosis is a major means to prevent thermal runaway in lithium-ion battery packs. During practical operation, the initial state of charge of a lithium-ion battery pack during charging may not be 0 %, and inherent inconsistencies exist among individual cells. Existing ISC diagnosis methods heavily rely on complete charging curves while neglecting the impact of cell inconsistencies, which can lead to misdiagnosing inconsistent cells as ISC cells and vice versa. This work presents an ISC diagnosis method that accounts for cell inconsistencies and is applicable under incomplete charging conditions. Partial incremental capacity curves are extracted from incomplete charging voltage curves to characterize the abnormal state of the battery. During the offline training phase, kernel density estimation is used to obtain the probability density functions of partial incremental capacity curves under different states. During online monitoring, Jensen–Shannon divergence is employed to quantify the similarity between different probability density functions, enabling the detection and differentiation of normal, inconsistent, and ISC cells within the lithium-ion battery pack. For the detected ISC cells, a state-space model is established with short-circuit current as one of the state variables. The short-circuit resistance is estimated using a dual adaptive extended Kalman filter. Experimental results show a diagnosis accuracy of 96.88 % with a false alarm rate of 0. The maximum root mean square error in short-circuit resistance estimation is only $2.28\Omega$, which empirically validates the proposal.