A model-driven approach to ultrasonic detection of state of charge in lithium-ion batteries

Abstract

Ultrasonic inspection is one non-destructive method to monitor the internal state of lithium-ion battery (LIB) cells. During charging, lithium ions intercalate into the graphite anode, causing moderate volumetric expansion (approximately 10%) and as much as a three-fold increase in Young’s modulus of the anode. Many researchers have observed changes in the time-of-flight (TOF) of through-thickness ultrasonic waves that correlate with changes in the state of charge (SOC). We introduce a transfer matrix method and Bloch-wave formalism for periodically layered media to show that the observed changes in TOF are partially due to the associated changes in anode stiffness. However, more detailed models that consider battery heterogeneity in structure and properties are needed to predict changes in TOF and signal amplitude (SA) for sophisticated ultrasonic measurement configurations and different cell chemistries. We therefore employ a quasi-static homogenization scheme for layered media to estimate the effective anisotropic stiffness of LIB pouch cells as a function of SOC as an input to a time-domain finite element model to determine the influence of changes in the anode, cathode, and separator on the direction-dependent propagation modes. This model is then employed to interpret experimentally obtained TOF and SA data for more robust SOC detection.