An Integrated Framework for Aging Analysis, Modeling and Impedance Diagnostics of Lithium-Ion Batteries

Lithium-ion batteries (LIBs) have been widely used as energy storage systems in both transportation and stationary applications due to their high energy density, long cycle life, and low self-discharge compared to other storage technologies. However, their performance and service life degrade under varying operating conditions and external stress factors. To investigate these degradation mechanisms and support the development of efficient battery management systems (BMS), this thesis proposes a practical, experimentally validated framework for diagnosing and optimizing the state of health (SOH) of LIBs. Compared with other external factors such as temperature, current rates, and mechanical stresses, current waveform modes have received comparatively less attention in literature. The influence of the current waveform on battery aging, particularly in balancing applications, was investigated through 100-cycle comparative aging tests using constant-current (CC) and pulsed-current (PC) protocols with the same mean current of 3C. Both protocols exhibited similar trends; however, the PC case showed approximately 2% higher capacity with a post-processing fitting slope of -0.27 mAh/cycle as opposed to -0.60 mAh/cycle (CC). To investigate degradation further, resistance was measured, showing that the mean internal resistance was approximately 4.3% lower in PC cycling than in CC cycling. Thermal analysis also showed a 2.5°C lower temperature rise for the PC cycling. Overall, these observations suggest that pulsed-current operation may promote partial relaxation effects and reduced polarization, whereas continuous operation may lead to increased thermal stress and faster degradation trends under the tested conditions. Similarly, the 1-RC Thevenin model was developed to enable single-step parameter identification and estimation without prior characterization. The proposed approach jointly estimates open-circuit voltage (OCV) and other dynamic parameters (R0, R1, C1) that remain physically stable and reasonable, yielding accurate terminal voltage predictions. Evaluation across laboratory and public datasets shows consistent agreement, suggesting that the proposed model provides a computationally efficient alternative to more complex approaches under moderate accuracy requirements for onboard BMS applications. Additionally, to provide a practically implementable alternative to conventional electrochemical impedance spectroscopy (EIS) systems, an integrated bench, load-driven galvanostatic EIS workflow was designed. Accurate impedance extraction was achieved through synchronized time-domain sampling, with excitation provided by a programmable electronic load using stepped-sine and square-wave methods. Validation against a reference HP4284A over an overlapping frequency range (20–3 kHz) demonstrated agreement in both Nyquist and Bode analysis. The implementation of an excitation quality metric, harmonics screening, and first order offset correction for fixtures and alignment compensation reduced the mean complex impedance deviation from 2.50 mΩ (stepped-sine) to 1.13 mΩ and from 2.42 mΩ (square-wave FFT-EIS) to 1.36 mΩ. Similarly, multi-SOC validation using an additional reference instrument (Squidstat) increased confidence in the method's reliability and robustness across different operating (SOC) conditions. Overall, the study provides measured aging health indicators and time-domain modeling of battery management system (BMS) functionality, along with a validated frequency-domain impedance measurement chain. These results provide a consistent framework for integrating state-of-health (SOH) monitoring, diagnosis, and modeling.