Discrete Interval Binary Sequence for Stable and Stationary Impedance Spectroscopy of Li-Ion Batteries

To ensure stability, causality, and linearity in impedance spectrum measurements, excitation signals must be carefully considered. These signals should cover all necessary frequencies down to the mHz range, have sufficient signal energy, and not change the battery state. However, designing a binary excitation signal for dc/dc converters that meets these requirements can be challenging due to the presence of harmonics in the excitation signal and the influence of frequent battery cycling. To address this problem, we propose a design methodology for a discrete interval binary sequence (DIBS) based on binarized multisine signals that are optimized by evolutionary role-playing game theory. Several techniques, including sign,
$\Delta \Sigma $
, and pulsewidth modulation (PWM), have been studied to implement an appropriate binarization procedure. DIBS-
$\Delta \Sigma $
has shown a high signal-to-noise and distortion (SINAD) of 83 dB, but it has a high switching rate (~76%), which leads to more power consumption. The switching rate of the DIBS signal is only ~9%, but it has higher charge changes and a low SINAD of 23 dB. The proposed signals have been experimentally validated using a Samsung INR-18650-25R Li-ion battery cell. With a frequency density of ten frequencies per decade and a measurement time of 200 s, DIBS-
$\Delta \Sigma $
has an rms deviation of 0.05%. For DIBS-Sign, the accuracy is 0.11%. These deviations can be reduced to 0.02% and 0.03%, respectively, with sufficient stabilization time.