Deriving Optimal Health Indicators Based on Fuzzy Logic using Lithium-ion Battery EIS Measurement Data and Developing SOH Estimation Based on LSTM Algorithm

Research on state estimation for lithium-ion battery, which is applied to various power conversion device fields, is receiving increased attention. Among various state estimation indicators, the estimation of the state-of-health (SOH) emerges as a critical factor, integral to the stability and efficiency of battery utilization. This paper explores an innovative approach to SOH estimation using electrochemical impedance spectroscopy (EIS). The sensitivity to impedance change due to battery degradation was quantified using fuzzy logic after analyzing the expected SOH estimation error for each battery state-of-charge (SOC) region. Therefore, optimal health indicators considering the two conditions were derived. Model performance verification was conducted by applying the long short-term memory algorithm for SOH estimation.