A Simplified Fractional-Order Model Adapted to Temperature and Aging for Fast Estimation of State of Power of Lithium-Titanate Batteries

The state of power (SOP) of lithium-titanate batteries with Li4Ti5O12 (LTO) anodes is a critical index that quantifies their capability to continuously supply or absorb energy over a specified period. For battery-powered locomotives, accurate SOP estimation is essential for developing control strategies for acceleration and regenerative braking. Nevertheless, the complexity of conventional fractional-order models (FOMs), combined with the nonadaptive estimation algorithms, poses challenges to achieving both efficient and accurate SOP estimation. In this work, a simplified fractional-order model (SFOM)-based multiconstraint algorithm (MCA) combined with an unscented Kalman filter (UKF) is proposed to fast online estimate SOP. First, the SFOM is developed by analyzing timescales of dynamic processes based on electrochemical impedance spectroscopy (EIS) across varying states of charge (SOCs), temperature, and aging conditions. Second, the MCA is designed based on the SFOM to calculate the peak current sequence required for SOP estimation, where SOP is quantified as an average of the power sequence over the peak current. Furthermore, the SOC, an essential parameter for online SOP estimation, is obtained using the SFOM-based UKF. Finally, the proposed method is validated through experiments. The results demonstrate that the proposed method not only reduces the average root-mean-square error in SOP estimation by up to 37.1% but also significantly saves calculation costs compared with traditional estimation methods.