Investigation of Frequency Domain Reflectometry as a Degradation Monitoring Technique for Lithium-Ion Batteries

Abstract

Frequency Domain Reflectometry (FDR) is an impedance-based diagnostic technique traditionally used in power systems to assess impedance changes in power cables. Recently, FDR has been applied to battery systems to measure high-frequency impedance, which is valuable for understanding battery performance in power line communication networks and assessing electromagnetic compatibility (EMC). Previous studies have highlighted FDR's ability to detect high-frequency processes like skin effect and ionic shunt effect, and its sensitivity to factors such as charging current, state of charge (SoC), and temperature. However, its potential for monitoring battery State of Health (SoH) has not been thoroughly explored. This thesis investigates the use of FDR as a non-invasive, tool for monitoring SoH of lithium ion batteries. The study involved two main stages of analysis. First, FDR was used to measure the impedance of 19 commercial coin cells (LIR 2032) across a frequency range of 300 kHz to 1 GHz. These cells were aged to varying SoH levels through controlled cyclic aging, and their SoH was benchmarked using Electrochemical Impedance Spectroscopy (EIS). The FDR impedance measurements were then compared to health indicators like battery capacity and internal resistance to evaluate FDR's sensitivity and accuracy in detecting aging-induced changes. The second stage involved a statistical evaluation of FDR's effectiveness in data driven detection and prediction models, using techniques such as principal component analysis, multivariate statistical process control, and partial least squares regression. The findings show that while FDR can detect changes in battery impedance related to aging, it has limitations in sensitivity to slower degradation processes and accuracy at lower impedance values. FDR demonstrated potential for single-cell SoH tracking but was less effective for multi-cell detection and capacity prediction compared to EIS. Despite these limitations, FDR could complement other health indicators in a multi-metric battery monitoring system, provided that careful setup design and calibration are employed.