Dynamic Coherence–Based Battery Health Assessment via Longitudinal Trajectory Analysis

Battery degradation is traditionally evaluated using capacity fade, internal resistance growth, or threshold-based fault indicators. This paper proposes and validates a falsifiable alternative paradigm: that degradation first manifests as a loss of dynamic coherence in the temporal relationship between voltage, current, and thermal response. Using THP Battery as a chemistry-agnostic diagnostic instrument, we analyze a longitudinal sequence of high-fidelity experimental datasets extracted from the CS2 (CALCE) battery repository. Results demonstrate that coherence erosion is detectable well before conventional State of Health (SOH) indicators deviate from nominal values. Battery health is therefore more accurately described as a trajectory relative to the system’s own historical peak coherence (“grace state”) than as a static capacity-based metric.