摘要
准确预测超级电容器在不同工况下的容值,对超级电容器的安全使用、延长寿命及性能的充分发挥具有重要意义。为了通过有限次的试验数据预测超级电容器在多工况下的容值,提出基于支持向量机的超级电容器容值预测方法。首先设计超级电容器在多工况下的特性试验,并分析温度、电流倍率等工况因素对超级电容器容值的影响;结合试验数据,通过对比不同数量的样本预测效果和考虑实际应用,确定参与训练建模的样本数量,利用网格搜索与交叉验证的方法优化模型参数,进而建立超级电容器容值预测模型。结果表明,预测值与试验值基本吻合,且对全工况范围下容值的预测值在合理范围内,验证了支持向量机用于超级电容器容值预测建模的有效性。
Accurate prediction of the capacitance of supercapacitor under different operating conditions is of great significance for supercapacitor’s safe using,life extension and performance optimization. In order to predict the capacitance of supercapacitor under various operating conditions by using a small amount of experiments,this paper proposed a prediction method for capacitance of supercapacitor based on support vector machine（ SVM）. Firstly,the characteristic experiments of supercapacitor under various operating conditions were designed. And the effects of operating factors such as temperature and current on the capacitance were analyzed. Then,Combined with experimental data,the training sample size for modeling was determined by considering the practical application and comparing the prediction effect using different sample size. The model parameters were optimized by the grid search and cross validation,and the prediction model of supercapacitor’s capacitance was established. The results show that the prediction values are basically identical with the experiment values,and the predicted capacitance under whole working conditions range is reasonable. It validates that the prediction method for capacitance of supercapacitor based on SVM is effective.
出处
《电器与能效管理技术》
2016年第3期42-47,77,共7页
Electrical & Energy Management Technology
基金
国家青年自然科学基金项目(51207108)
国家863高技术项目(2011AA11A233)
关键词
超级电容器
不同工况
支持向量机
交叉验证
网格搜索
supercapacitor
different operating conditions
support vector machine
cross validation
grid search
