Integrated SOC Estimation for Grid-Interactive EV Batteries using Temperature-Dependent Kalman Filtering

The integration of electric vehicles (EVs) into smart grids through Vehicle-to-Grid (V2G) operations transforms them from passive loads into active energy assets. This evolution necessitates precise and robust estimation of the State of Charge (SOC), especially under operating scenarios characterized by thermal variability and long-term battery aging, where many existing SOC estimation techniques demonstrate reduced reliability. This paper presents an enhanced Kalman Filter (EKF)-based SOC estimation framework specifically developed for lithium-ion batteries engaged in bidirectional V2G applications. The estimator integrates three key components: (i) a nonlinear Open Circuit Voltage (OCV)-SOC relationship derived from empirical battery profiles, (ii) a temperature-sensitive internal resistance model that captures the impact of aging, and (iii) a physicsinformed thermal model accounting for entropy-driven reversible heat generation. The approach is validated using comprehensive MATLAB simulations under synthesized V2G load profiles that replicate realistic grid-interactive conditions. Across three distinct scenarios-nominal operation, elevated thermal stress, and accelerated aging-the framework consistently achieves SOC estimation errors below 3% under standard conditions and maintains errors within 9 % under intensified stress, outperforming traditional EKF implementations. These results demonstrate the estimator's resilience and its applicability for integration within advanced Battery Management Systems (BMS) in smart grid-connected EV systems.