摘要
为了避免支持向量机预测结果易受惩罚因子和核函数参数参数选择的影响,提出一种DA算法优化SVM的电能质量扰动诊断和识别模型,实现电能质量扰动最优化诊断和识别。首先运用EMD将电能质量扰动信号进行分解,之后计算各尺度下的IMF分量的样本熵,并将其作为电能质量扰动信号的特征向量,建立SVM的电能质量扰动信号的识别模型。实验结果表明,与GA_SVM、PSO_SVM和DE_SVM相比,本文提出的算法DA_SVM可以有效提高电能质量扰动识别的准确率,收敛速度快,为电能质量扰动诊断和识别提供新的方法和途径。
For preventing the prediction result of the support vector machine (SVM for short) from the punishment factor and parameter selection of kernel function parameter optimal, a kind of algorithm is proposed to optimize power quality disturbance diagnosis and identification of SVM so to the most optimal diagnosis and identification of the power quality. Firstly, power quality disturbance signal is decomposed by EMD method, then the sample entropy of different scale signal of EMD decomposition is calculated and is considered as a signal of power quality disturbance characteristic vector and the identification model based on SVM for the power quality disturbance signal is established. It is shown by the experimental result that compared with GA_SVM, PSO_SVM and DE_SVM, the algorithm DA_SVM proposed in this paper can effectively improve the accuracy of the power quality disturbance identification, has fast convergence rate and provide new methods and means for the power quality disturbance diagnosis and identification.
作者
屈高强
王承民
齐彩娟
赵亮
刘涌
陈万喜
QU Gaoqiang;WANG Chengmin;QI Caijuan;ZHAO Liang;LIU Yong;CHEN Wanxi(Institute of Economy and Technology,Ningxia Electric Power Company,Yinchuan 750001,China;Shanghai Jiaotong University,Shanghai 200240,China;Shanghai Boying Information Technology Co.,Ltd.,Shanghai 200240,China)
出处
《电力电容器与无功补偿》
北大核心
2019年第1期115-122,共8页
Power Capacitor & Reactive Power Compensation
基金
"网源荷"互动的调峰调频规划关键技术研究(5229JY160003)
关键词
样本熵
支持向量机
蜻蜓算法
电能质量
经验模态分解
sample entropy
support vector machine
dragonfly algorithm
power quality
empirical mode decomposition
