Degradation Date Modeling And State Of Health Estimation Of Lithium-ion Battery

Lithium-ion batteries are widely developed and used as an emerging green energy source in various fields such as electric vehicles.State of health estimation and Remaining useful life prediction based on lithium battery degradation data are key issues in lithium battery health management.In this paper,two methods for predicting the state of health and remaining useful life of lithium-ion batteries are proposed by analyzing and modeling the degradation data of lithium-ion batteries.The specific work is as follows:First,combined with the working principle and degradation mechanism of lithium-ion batteries,the data of lithium-ion batteries under two different operating conditions of NASA PCo E laboratory,as well as the change of capacity with cycle are analyzed,indicating the feasibility of battery state of health estimation driven by degradation data.Secondly,a method for state of health and remaining useful life prediction of Li-ion batteries is proposed based on gaussian process regression.Five health features were extracted from the cyclic charging current of the Li-ion battery and grey relation analysis showed that these five features were highly correlated with the battery state of health.A new state of health estimation model for Li-ion batteries was developed by improving the basic gaussian process regression model and basing it on these eigenvalues.A new state of health estimation model for Li-ion batteries was developed by improving the basic gaussian process regression model and basing it on these features.Meanwhile,a loop-based polynomial regression model was developed to update future features.Further,the remaining useful life prediction framework for Li-ion batteries was designed by combining the state of health prediction model with a polynomial feature update model.The experimental results show that the proposed model has better prediction accuracy than the other base models,and the robustness of the proposed model is verified by random walk battery data.Finally,a lithium-ion battery state of health estimation method with a fusion data-driven model is proposed.In this method,the genetic algorithm genetic algorithm-back propagation neural network is used to express the relationship between health features and state of health of the battery,and a multi-feature degradation model of lithium-ion battery state of health is established.Furthermore,a measurable indicator based on battery surface temperature information is extracted and a method for estimating the degradation rate of lithium-ion batteries is designed using particle filtering.Finally,by fusing the genetic algorithm-back propagation neural network degradation model with particle filtering,a new method of state of health estimation is proposed that can quantify the uncertainty of the prediction results.The experimental analysis of data from two different types of lithium batteries shows that the proposed state of health estimation method not only has better accuracy compared to other methods,but also gives an assessment of the uncertainty of the prediction results,effectively improving the reliability of the state of health prediction.