摘要
电池荷电状态(SOC)准确预测是电池管理系统的关键任务。针对过去电池SOC预测精度低等问题,提出了一种采用极限学习机神经网络(ELM)的预测模型,以电池电压和电流作为模型的输入量,SOC作为输出量。在建模过程中,采用粒子群优化算法(PSO)对ELM随机给定的输入权值矩阵和隐层阈值进行寻优,降低了随机性给模型造成的影响,提高了模型预测精度。利用实验采集的数据进行模型训练和预测,结果表明,用粒子群算法优化后的极限学习机模型(PSOELM)与单纯的ELM以及传统的BP和SVM相比,具有更高的预测精度和泛化性能。为磷酸铁锂电池的SOC预测提供了一种新的方法。
Estimation of SOC of LiFePO4Li-ion battery is one of the key missions for battery management system. For the low accuracy of SOC prediction in the past,an estimation model based on extreme learning machine（ELM） neural network was proposed. In the model,voltage and current were used as input vector and the value of SOC was used as output vector. After the input weight matrix and hidden layer threshold of ELM are optimized by PSO,the effects of its randomness to model were reduced and the prediction accuracy of the model was improved. Using the experimental data collected to train the model and then predict the output,the results show that compared with ELM and BP and SVM neural network,the PSOELM can get higher prediction precision and has more advantages in the generalization performance. Therefore,the new method can be provided for the SOC prediction of LiFePO4Li-ion battery.
出处
《计算机仿真》
CSCD
北大核心
2015年第3期163-168,共6页
Computer Simulation
基金
广西自然科学基金资助项目(桂科自0832075)
关键词
磷酸铁锂电池
荷电状态
极限学习机
粒子群优化算法
预测
LiFePO4 Li-ion battery
State of charge（SOC）
Extreme learning machine
Particle swarm optimization（PSO）
Estimation
