流形正则化框架下的极限学习机预测锂电池SOC方法

SOC Prediction Method of Lithium Battery Based on MRELM Framework

为提高锂电池荷电状态建模预测的精度及泛化能力,提出一种流行正则化框架下的极限学习机建模预测方法。首先,为了解决极限学习机建立预测模型存在过拟合泛化能力弱的问题,以流形假设为依据,在数据输入空间构建图拉普拉斯算子,在其框架内求解极限学习机隐层和输出层之间的权重,达到正则优化目的。其次,针对正则化参数难以选择的问题,提出将差分进化算法融入基于流形正则化框架的极限学习机中以优化其正则化参数。最后,利用采集到的锂电池数据进行了实验验证。结果表明:该方法建立的预测模型预测锂电池SOC精度高,泛化能力强,为锂电池SOC的预测建模提供一种新方法。

In order to improve the accuracy and generalization ability of modeling and prediction for the SOC of lithium batteries,an ELM modeling and prediction method based on MR framework was proposed. Firstly,in order to solve the problem of over-fitting and weak generalization ability in the prediction model of the ELM,based on the manifold hypothesis,a graph-Laplacian operator is constructed in the data input space,and the weights between the hidden layer and the output layer of the ELM are solved within the framework of the operator to achieve the goal of regular optimization. Secondly,in order to optimize the regularization parameters,a DE algorithm is introduced into the ELM based on MR framework. Finally,the data collected from lithium battery were used for experimental verification. The experimental results show that the prediction model established by this method has high accuracy and strong generalization ability for lithium battery SOC prediction,which provides a new method for lithium battery SOC prediction modeling.