Continuous-time optical-electrothermal SOC model for lithium-ion batteries under random dynamic loads

Abstract

This study proposes a continuous-time optical-electrothermal State-of-Charge (SOC) and terminal-voltage prediction framework for lithium-ion batteries under complex mobile-device loads. Major hardware modules, including the processor, display, GPS, peripherals, and communication units, are modeled as external loads, and the total system power demand is obtained through linear superposition. A power-balance equation is introduced to couple external circuit demand with battery internal dynamics, enabling the discharge current to be determined as the excitation input. On the battery side, an electrochemistry-enhanced second-order RC equivalent-circuit model is developed. SOC evolution is described by continuous-time Coulomb counting, while the open-circuit voltage is dynamically updated using electrode equilibrium potentials. Solid-phase diffusion is simplified through a Padé approximation to provide an online-compatible expression for surface lithium-ion concentration. To capture multiphysics coupling, Bernardi-based heat-generation decomposition and a lumped thermal model are incorporated. Optical sensing is further introduced as an auxiliary observation channel to enhance thermal-field perception and improve SOC estimation reliability under dynamic loading conditions. Temperature-dependent electrochemical parameters, including electrolyte conductivity, exchange current density, and diffusion coefficient, are mapped to equivalent-circuit parameters, forming a self-consistent optical-electrothermal feedback loop. Validation using the NASA PCoE Randomized Battery Usage Dataset shows that the proposed framework accurately reconstructs OCV-SOC characteristics under quasi-equilibrium and intermittent constant-current conditions, particularly in the low-SOC region. Overall, the model balances physical interpretability and computational efficiency, providing a unified basis for mobile-device energy evaluation, power management, optical-assisted thermal monitoring, and battery state estimation.