摘要
结合统计学习理论的方法,将最小二乘支撑向量机(LS-SVM)用于通信电源中蓄电池的荷电状态(SOC)检测。选定径向基核函数为支撑向量机算法的核函数,并选取矩阵分块求逆的方法改进最小二乘支撑向量机的算法,在此基础上建立了蓄电池荷电状态估计的模型。通过仿真实验验证了该算法具有较好的SOC估计效果,对于实际应用,可以选择合理的剪辑标准,从而得到更好的估计效果。
Combining statistical learning theory approach,least squares support vector machines（LS-SVM） is used to estimate the VRLA battery’s state of charge of the telecommunication power systems.The paper builds model of state of charge through choosing radial basis function kernel for support vector machines and selecting the better algorithm of least squares support vector machines which is based on block matrix inversion method,then the model of value-regulated lead-acid（VRLA） battery’s state of charge prediction is established.Simulation results show that the algorithm has good effects on the SOC estimate.For practical application,you can choose a reasonable clip standard to get a better estimate of effect.
出处
《电测与仪表》
北大核心
2010年第10期78-80,共3页
Electrical Measurement & Instrumentation
基金
广东省学研示范基地项目(H04010701CXY2009006)
关键词
通信电源
铅酸蓄电池
荷电状态
支撑向量机
分块矩阵
telecommunication power systems
lead-acid battery
SOC
support vector machine
block matrix
