Equivalent Circuit Modeling and Parameter Identification for Lithium-ion Batteries Based on Improved Barnacle Mating Optimizer

Abstract

To address the current problems of accuracy and reliability of the equivalent circuit models for lithium-ion batteries and sensors, in this study we propose an improved barnacle mating optimizer (IBMO) to accurately identify parameters. First, an improved second-order RC model (ISO RC) is derived by segmenting the battery hysteresis voltage and introducing it into the second-order RC model (SO RC model) as a controlled source, which better reflects the dynamic effects of the battery than the SO RC model. Second, the tent chaos mapping, cosine control factor, and Levy flight strategy are introduced to enhance the convergence performance of the IBMO. Its convergence performance is verified by comparison with other existing intelligent algorithms using five single-peak and four multipeak test functions. Finally, the IBMO-ISO RC model is established by using the IBMO as the identification strategy of the ISO RC model in the simulation of two real operating conditions of lithium-ion batteries. The root mean square error of the model under the two operating conditions is only 0.0431 and 0.0483, and the mean absolute percentage error is only 0.38 and 0.56%, outperforming other existing models. The IBMO proposed in this study enables higher model recognition accuracy, making it highly convenient for lithium-ion battery and sensor-related research.

An improved variable forgetting factor recursive least square-double extend Kalman filtering based on global mean particle swarm optimization algorithm for collaborative state of energy and state of health estimation of lithium-ion batteries

“…In the first place, Tian et al and Tian et al use the second-order RC equivalent circuit model to estimate state of charge (SOC) and SOH, respectively [6,7]. To get a better result, Li et al introduction of battery hysteresis voltage segmentation into a second-order RC model to form an improved second-order RC model, and the estimation results error is under 0.0431 [8]. Furthermore, for researching charging safety and SOC, Zhang et al and Xu et al use an improved back propagation (BP) neural network to get results [9,10].…”