Path-dependent degradation diagnosis of commercial NCA/graphite 18650 cells under coupled cycle-calendar aging via dual-tank model identification and differential voltage analysis

Accurate diagnosis of lithium-ion battery degradation under alternating cycling and storage is important for lifetime assessment and health management in practical energy-storage applications. However, degradation under coupled cycle-calendar aging is inherently path dependent, and the stage-wise evolution of physically interpretable aging indicators remains insufficiently resolved. In this work, a stage-resolved diagnostic framework covering the aging interval from the fresh condition toward an approximately 90% state-of-health (SOH) aged condition is developed for commercial NCR18650GA nickel cobalt aluminum oxide (NCA)/graphite cells by integrating multi-stage aging experiments, dual-tank model identification, and differential voltage analysis. A sequence-dependent protocol at 50 ° C is designed to compare continuous cycling with cycle–calendar–cycle paths under different storage state-of-charge (SOC) levels. Based on full-cell voltage-capacity data and half-cell open-circuit potential (OCP) curves, an electrochemically interpretable dual-tank model is established to identify electrode lithiation states, effective electrode capacities, and a lumped resistance-related term. The results suggest that the inserted calendar-aging stage alters the subsequent degradation trajectory rather than merely interrupting cycling. The cathode lithiation parameter y 0 decreases by 4.68%–5.36% from Stage 1 to Stage 3 in all groups, indicating a consistent shift in electrode balancing. Among the identified variables, the fitted resistance-related parameter R shows the clearest path dependence, with much stronger growth in Groups A and B than in Group C. For the fitted parameters R and C p , lowering the intermediate storage SOC from 75% to 25% progressively mitigates the final degradation response; the electrode-level differential-voltage metrics are path dependent and are not strictly monotonic with storage SOC. Differential-voltage analysis further suggests that uninterrupted cycling is associated with a stronger late-stage graphite anode signature, whereas intermediate-SOC storage is more consistent with electrode-balancing migration that becomes amplified after cycling resumes. Overall, the proposed framework provides a physically interpretable route for comparing degradation modes under hybrid aging histories and offers useful support for storage-strategy design and aging-aware battery management.