Enabling early detection of lithium-ion battery degradation by linking electrochemical properties to equivalent circuit model parameters

Abstract

Finding an accurate and simple method to early detect degradation phenomena in lithium-ion batteries (LIBs) is a major objective to optimise battery use. Various detailed degradation models have been developed, but they are too sophisticated to be used in Battery Management Systems (BMSs) for online LIB state estimation. This paper aims at filling the gap between advanced degradation simulations and state estimation in BMSs by coupling a low computational equivalent circuit model (ECM, made of a series resistance and parallel resistor/capacitor RC network) with a physical/chemical description of the LIB via a pseudo-2-dimensional (P2D) model. After validation, the P2D model is used as a virtual battery to simulate the main degradation phenomena, by varying the associated electrochemical properties, and the ECM parameters are identified. Results show that electrolyte degradation affect all the ECM parameters but can be isolated in the first RC circuit which encodes fast dynamic phenomena, from 4 s to 5 minutes. The intercalation kinetics degradation is retrieved from the increase in ohmic series resistance R0, which represents very fast dynamic processes with time scale < 4 s, upon subtraction of the electrolyte contribution. Finally, the solid-state diffusivities degradation appears at slow time scales, from 3 to 100 min, in the second RC circuit. These results suggest a strategy to infer the nature and extent of the degradation via online monitoring of the ECM parameters.