Real-time impedance spectroscopy of lithium-ion batteries via composite current pulses

Workflow of the proposed CP-EIS method. The composite current pulse sequence is transformed via FFT and subsequently subjected to truncated SVD reconstruction for effective outlier suppression. Conventional sinusoidal electrochemical impedance spectroscopy is often impractical for real-time control or on-board diagnostics because measurements at low frequencies require long dwell time, resulting in lengthy test duration. To address this issue, the composite current pulse excitation is implemented in this work for real-time impedance spectrum acquisition, using the discrete Fourier transform. Pulse sequences and sampling conditions are designed to balance bandwidth and accuracy of the impedance results while satisfying hardware constraints and system relaxation requirements. To improve repeatability under noise and dynamic operating conditions, outliers are mitigated by introducing truncated singular value decomposition reconstruction. Two pulse widths (1 and 100 ms) are applied to overcome the bandwidth limitation of a single-width excitation, enabling an accurate spectrum across 1 kHz to 1 Hz within ∼1 s. On a commercial 18650 lithium-ion battery, a mean relative impedance deviation of 2.1% compared with galvanostatic electrochemical impedance spectroscopy results is achieved across state of charge from 5% to 90% at 10 and 25 °C. Time-domain voltage simulations using pulse-calibrated parameters reproduce the measured dynamic responses, achieving accuracy comparable to simulations parameterized from galvanostatic electrochemical impedance spectroscopy.