Adaptive state of charge estimation for lithium-ion batteries based on implementable fractional-order technology

Abstract

Abstract State of charge (SOC) estimation is of vital importance in battery management systems to ensure both safety and reliability. However, the polarization effect, the stochastic disturbance and the highly nonlinear and dynamic natures throughout the whole lifetime of batteries bring many challenges for accurate online estimation. On account of these difficulties, this paper proposes a novel SOC estimation strategy based on open circuit voltage by means of some implementable fractional-order techniques. Firstly, the fractional-order equivalent circuit model that can reveal more intrinsic electrochemical characteristics of batteries is introduced and parameterized by particle swarm optimization algorithm, then it is combined with particle filter that based on Monte Carlo method for SOC estimation. To achieve rapid convergence and strong robustness, an adaptive noise variance updating algorithm is adopted to update the SOC estimations in particle filter. Moreover, considering the computational burden of fractional-order model, the infinite impulse response filtering technique that adapts to data-driven modeling is introduced to simplify the discrete state space model in estimation. Lastly, the proposed algorithms are implemented in static and dynamic experiments, and the results indicate that the aforementioned strategies can realize fast convergence and precise estimation with applicable calculations.