Core Temperature Prediction of Lithium-Ion Batteries Using Dual Adaptive Kalman Filtering With Monte Carlo-Enhanced Multiobjective Particle Swarm Optimization

Accurate internal temperature monitoring is critical for optimizing performance and ensuring safety in lithium-ion batteries. This study presents a novel multi-objective framework that combines a dual adaptive extended Kalman filter (DO-AEKF) with a Monte Carlo-enhanced Multi-Objective Particle Swarm Optimization (MOPSO-MC) for precise core temperature prediction in lithium-ion batteries. By coupling an equivalent circuit model system with a dual-mode thermal model, the framework enables simultaneous estimation of electrical and thermal states. The integrated MOPSO-MC optimizes the noise covariance matrix in the DO-AEKF, thereby enhancing the adaptability and robustness of the system. Simulative and experimental results under standard cycles and dynamic profiles at -10°C to 50°C show that the proposed framework reduces the maximum prediction errors by 31.8% and stabilizes temperature RMSE within 0.43°C. Notably, the framework achieves exceptional statistical fidelity with Pearson correlation coefficients exceeding 0.96, confirming its superior trend-tracking capability and robustness against nonlinear thermal inertia. These results highlight its potential for reliable battery thermal management system applications.