Research On Data-driven Battery State Estimation Methods

The "14th Five-Year Plan" proposes to vigorously develop strategic emerging enterprises and accelerate the expansion of new energy,new materials,new energy vehicles,green environmental protection and other industries.Lithium-ion batteries are widely used in various projects such as electric vehicles and energy storage.Due to the unobservability of the internal chemical reactions of Li-ion batteries and the complexity of their operating conditions,the estimation accuracy of the online state of charge（SOC）and state of health（SOH）of Li-ion batteries is difficult to guarantee.Errors in estimating the SOC and SOH of lithium-ion batteries may cause accidents in engineering applications,resulting in damage to the user’s property and personal safety.In order to improve the use effect and safety of lithium-ion batteries,it is necessary to accurately estimate the SOC and SOH of lithium-ion batteries online.In this paper,a combination of least squares support vector machine（LS-SVM）and particle swarm-genetic optimization（PSO-GA）algorithm is applied to the SOC estimation of lithium-ion batteries.In order to optimize the effect of LS-SVM algorithm for estimating SOC of lithium-ion battery,PSO-GA algorithm is used to determine the penalty coefficient and kernel function in LS-SVM algorithm to improve the performance of the algorithm.A large number of lithium-ion battery charging and discharging experiments were carried out under standard conditions and UDDS conditions,and comparative experiments were designed using the experimental data to verify that the algorithm in this paper estimates the SOC of lithiumion batteries with higher accuracy and better robustness.This paper proposes an algorithm for SOH estimation of lithium-ion batteries based on fuzzy theory and long short-term memory neural network（LSTM）.In order to adapt to the nonlinear attenuation during the SOH decline of lithium-ion batteries,the LSTM model is combined with fuzzy theory,which improves the correlation of the original LSTM network model and improves the estimation accuracy.The cycle life experiments of lithium-ion batteries are carried out under standard conditions,and comparative experiments are designed using the experimental data,which verifies that the model established in this paper has better adaptability and estimation accuracy in the field of SOH estimation of lithium-ion batteries.The method helps to accurately grasp the use state and remaining life of the battery,and maintain or replace the lithium-ion battery in time,which can effectively improve the safety of battery use,and can be popularized and applied to practical projects.