改进汉明Burg和稳健支持向量机的间谐波检测  

Modified Hamming Weighted Burg and Robust Support Vector Machine Algorithm for Inter-harmonics Detection

作　　者：崔晓荣[1] 王军[1] 林莉[1] 孙章[1] 田攀 

机构地区：[1]西华大学电气信息学院,成都610039 [2]四川中电启明星信息技术有限公司,成都610000

出　　处：《电力系统及其自动化学报》2014年第6期30-36,共7页Proceedings of the CSU-EPSA

基　　金：四川省重点学科基金资助项目(SZD0503-09-0);西华大学创新基金资助项目(ycjj201290);四川省教育厅成果培训基金资助项目(12ZZ007);2011年科技部中小企业国际合作资助项目(SQ2011ZOF000004)

摘　　要：间谐波具有非线性、随机性及非平稳性等特点,在实时检测中具有一定困难。为精确检测出所含间谐波分量,文中先利用改进汉明Burg算法检测间谐波的频率,该算法按二阶预测误差输出功率最小原则求解一、二阶系数,求高阶系数时进行汉明加权,选取阶次时引入控制参数;再用稳健支持向量机法校正检测出的频率分量,并获得其幅值和相位。仿真结果表明,该算法检测频率时,减弱了谱峰偏移度且不易出现谱线分裂,低信噪比时不会出现谱峰消失现象,提高了频率检测精度;用稳健支持向量机法校正频率可精确得出间谐波的频率、幅值和相位,且计算量小。Due to the non-linearity,randomness,instability,etc.,it is highly difficult to carry out the real-time inter-harmonics detection.In order to detect the inter-harmonic component accurately,the modified Hamming weighted algorithm is initially utilized for the frequent components detection.The first and second coefficients are obtained by minimizing the totla output power of second order prediction error,and the other higher coefficients are calculated via hamming weighted,and the control parameter is introduced in the model order selectiom.Then each frequency of detected was corrected by robust support vector machine algorithm which could get accurate frequency,meanwhile amplitude and phase of inter-harmonic are obtained.Simulated results demonstrate that when detecting the frequency component,the peak shifting is reduced and spectral splitting is not manifest; the phenomenon of spectral disappear will vanish with low signal to noise ration,and the accuracy of frequency detection is improved; the frequency components are calibrated by robust support vector machine,meanwhile the amplitude and phase can be obtained,moreover the computation is curtailed.

关 键 词：间谐波 自回归模型 伯格算法 汉明加权 稳健支持向量机 

分 类 号：TM711[电气工程—电力系统及自动化]