A Review of Data-Driven SOH and RUL Estimation for Lithium-ion Batteries

Lithium-ion batteries have a promising future in the application of large-scale energy storage systems. However, there are certain issues with the consistency and safety of lithium-ion batteries, so it is important to estimate the state of health (SOH) and remaining useful life (RUL) of lithium-ion batteries. Due to the complex nonlinear degradation mechanisms of lithium-ion batteries, traditional SOH and RUL prediction methods still have many deficiencies. In recent years, with the development of big data and machine learning technology, data-driven estimation methods can effectively solve this problem. Data-driven models do not need to involve complex chemical reactions and various prior knowledge inside the battery. They only need to utilize machine learning models to train experimental data, and eventually achieve data prediction. This is currently a research hotspot in SOH and RUL prediction. The paper first introduces the concepts of SOH and RUL of lithium-ion batteries and the aging mechanism, and then details the current research status of data-driven lithium-ion battery SOH and RUL estimation, classifying and elaborating on several major data-driven algorithms, and comparing and analyzing the advantages and disadvantages of these intelligent algorithms. Finally, the paper presents prospects for the development direction of data-driven methods in this field.