Online Weld Breakage Diagnosis for the Battery of Electric Vehicle: A Data-Driven Approach

A data-driven diagnostic approach is developed and proved successful to identify the weld breakage of the connector in battery pack according to the data-upload by the remote monitoring platform. Firstly, two modes of the weld breakage: complete breakage and loose contact, as well as their statistical patterns are illustrated. To further assure the weld breakage fault, the capacity of battery pack was estimated using probability density function approach, and the statistical patterns are further analyzed to identify the characteristics of weld breakage. The approach proposed in this paper provides inspirations for the development of diagnostic methods from a data- driven view.