基于支持向量机的锂离子电池寿命预测方法分析-车辆工程专业论文.docxVIP

华中科技大学硕士学位论文 华 中 科 技 大 学 硕 士 学 位 论 文 II II ABSTRACT The li-ion battery is a research hot spot at home and abroad as a performance outstanding and widely used energy storage power. Establish life evaluation methods and life model, evaluation and predict battery life scientifically is important. Support Vector Machine (SVM) has many advantages and is a hot spot in field of artificial intelligence. It has been used in the battery capacity forecasting successfully and achieve good effect. This paper sets a life prediction model based on least square SVM, predict residual capacity of li-ion batteries effectively and realize the life prediction. The parameters selection is important for the model,and the genetic simulated annealing algorithm (GSA) in this paper can effectively improve the prediction accuracy and generalization ability. Finally, the simulation experiment results show that LS-SVM regression model based on RBF kernel function is superior to the Artificial Neural Networks (ANN) and the SVM. Chapter 1 it mainly introduced the aim, significance, research situation of lithium-ion battery life prediction and the application in life prediction of SVM. Chapter 2 it introduced the basic working principle, life influence factors, failure mechanism of lithium-ion battery. This analyze the lithium-ion battery performance characteristics and masters the change rule. Chapter 3 it introduced the basic principle of SVM based on statistical learning theory, explanation SVM classifier model and SVM regression model detailedly. Chapter 4 it establishes LS-SVM regression model based on RBF function and uses GSA for parameters selection. At last, it does a simulation experiment with the 30 sets of data of lithium-ion battery and compares the method in this paper, ANN and SVM. Key words: Lithium-ion battery Support vector machine Least square Life prediction PAGE IV PAGE IV 目 录 摘 要 I ABSTRACT II 1 绪 言 1.1 锂离子电池寿命预测的目的及意义(1) 1.2 国内外对锂离子电池寿命的研究现状(1) 1.3 支持向量机在电池寿命预