Lithium Battery SOC Estimation Based on Type-2 Fuzzy Cerebellar Model Neural Network

Accurate estimation of the state of charge (SOC) of lithium batteries is critical for the safe and optimal operation of battery management systems (BMSs). Traditional SOC estimation methods are often limited by model inaccuracy and noise interference. In this study, a novel type-2 fuzzy cerebellar model neural network (Type-2 FCMNN) is proposed for accurately estimating the state of charge of lithium batteries. Based on the traditional fuzzy cerebellar model neural network (FCMNN), type-2 fuzzy rules are innovatively introduced to enhance the robustness of the model against uncertainties and noise disturbances. This enables the model to better cope with nonlinear complexity and external disturbances when dealing with SOC estimation of lithium batteries and significantly improves the accuracy of prediction. On this basis, by analyzing the working principle of lithium batteries, a battery equivalent circuit model is successfully established and simulated and tested by Simulink R2022b, which provides a theoretical basis for the selection of the size of the input parameters of the subsequent neural network. Then, this paper designs and implements the SOC estimation model based on Type-2 FCMNN and tests it in Matlab. Finally, this paper carries out simulation comparison experiments between Type-2 FCMNN and various classical neural network algorithms in Matlab R2022b including traditional FCMNN, a backpropagation neural network, radial basis function neural network, and Kalman filtering algorithm. The simulation results show that Type-2 FCMNN exhibits a significant advantage in SOC estimation accuracy, with mean absolute error and root mean square error values of only 43.1% and 36.0% of FCMNN’s, respectively, while FCMNN achieves the best results among the compared methods.