摘要
锂离子电池因其长寿命而导致其设计和测试周期较长,因此,建立一个锂离子电池寿命预测模型对锂离子电池的开发和测试具有重要意义。基于数据驱动的锂离子电池寿命预测是一种有前景的预测方法,它可以在不需要专业知识的情况下进行预测。但是,基于数据驱动的预测方法对数据的要求较高,同时,大量的数据特征可能会降低学习模型的精度。为了解决这个问题,提出了一种基于进化算法的特征选择策略,该策略利用进化算法对锂离子电池训练数据的特征进行筛选,从而提高算法的精度。最终的实验结果在麻省理工学院提出的测试集上表现出了良好的效果。
Lithium-ion batteries are known for their long lifespan,which often results in longer design and testing cycles.Therefore,establishing a battery life prediction model is of great significance for the development and testing of lithium-ion batteries.Data-driven battery life prediction is a promis-ing prediction method that can be performed without expert knowledge.However,data-driven pre-diction methods have high requirements for data and a large number of data features may reduce the accuracy of the learning model.To solve this problem,this paper proposed an evolutionary algorithm-based feature selection strategy,which used evolutionary algorithms to screen the features of lithium-ion battery training data,thus improving the accuracy of the algorithm.The final experimental re-sults show good performance on the test set proposed by MIT.
作者
周继香
闫泽远
李淼林
ZHOU Jixiang;YAN Zeyuan;LI Miaolin(School of Automobile and Transportation Engineering,Guangzhou College of South China University of Technology,Guangzhou Guangdong 510800,China;School of Computer Science,Sun Yat-sen University,Guangzhou Guangdong 511400,China)
出处
《电源技术》
CAS
北大核心
2024年第4期679-684,共6页
Chinese Journal of Power Sources
基金
国家自然科学基金(62206313)
教育部产学研项目(51/Z100008)。
关键词
锂离子电池
进化算法
数据驱动
寿命预测
lithium-ion battery
evolutionary algorithm
data-driven approach
life prediction
