Lithium-Ion Batteries’ Energy Efficiency Prediction Using Physics-Based and State-of-the-Art Artificial Neural Network-Based Models

Abstract

Abstract The new generation of lithium-ion batteries (LIBs) possesses considerable energy density that arise the safety concern much more than before. One of the main issues associated with LIB safety is the heat generation and thermal runaway in LIBs. The importance of characterizing the heat generation in LIBs is reflected in numerous studies. The heat generation in LIBs can be related to energy efficiency as well. In this work, the heat generation in LIB is predicted using two different approaches (physics-based and machine learning-based approaches). A validated multiphysics-based and neural network-based models for commercial LIBs with lithium iron phosphate/graphite (LFP/G), lithium manganese oxide/graphite (LMO/G), and lithium cobalt oxide/graphite (LCO/G) electrodes are used to predict the heat generation toward shaping the LIB energy efficiency contours, illustrating the effect of the nominal capacity as a key parameter in the manufacturing process of the LIBs. The developed contours can provide the energy systems designers a comprehensive view over the accurate efficiency of LIBs when they need to incorporate LIBs into their devices. In addition, the effect of temperature on charge/discharge energy efficiency of LFP/graphite LIBs is obtained, and the performance of three typical LIBs in the market at a very low temperature is compared, which have a wide range of applications from consumer applications such as electric vehicles (EVs) to industrial applications such as uninterruptible power sources (UPSes).