Hybrid Experimental and AI-Based Framework for Life Prediction of Lithium-Ion Batteries under Mechanical Stress

Abstract

The Lithium-ion batteries (LiBs) have enjoyed a lot of popularity in electric vehicles and portable electronics, as well as in energy storage systems, because of its high energy density and efficiency. Nevertheless, mechanical loads, including compression, vibration, impact, and nail penetration have a considerable effect on the electrochemical performance and safety of batteries. The mechanical loading conditions have the capability to modify the important electrical parameters such as voltage response, internal resistance, temperature, and capacity degradation, which in the last analysis influence the battery life and reliability. Throughout this literature review paper, the relationship between mechanical testing conditions and the resultant electrical performance of LIB of various form factors, such as cylindrical, pouch, and prismatic cells is examined in detail. Moreover, the review examines how machine learning methods have increasingly been involved in the analysis of complex data produced as a result of mechanical abuse and structural testing. SVM, neural networks and ensemble methods of machine learning have proven to be highly effective in forecasting battery health, failure modes and remaining useful life (RUL). The paper also informs about the recent development of data-driven modeling to predict safety and detect fault in advance. Combining the knowledge of mechanical testing with the forecast of machine learning, this review gathers research gaps and outlines the way forward to build more dependable, secure, and long-lasting lithium-ion battery systems in the future use of energy storage.