基于Root-MUSIC和支持向量机的间谐波参数估计

Interharmonics Estimation Based on Root-MUSIC and SVM Algorithm

为准确地检测电力系统中间谐波信号的参数,提出基于求根多重信号分类法(Root-MUSIC)和支持向量机(SVM)的间谐波参数估计方法。首先对采样数据构成的自相关矩阵进行特征分解,利用信号子空间和噪声子空间的正交性求得谐波和间谐波的个数及频率;然后通过支持向量机算法对间谐波信号的幅值和相位进行回归估计。Matlab仿真结果表明:该算法在低信噪比下频率估计准确,利用支持向量机在处理小样本数据上的优势,有效的提高了幅值和相位估计的精度。

In order to accurately detect the parameters of the interharmonics in power system, an algorithm based on root-multiple signal classification （Root-MUSIC） and support vector machine （SVM） is proposed. First, eigenvalue decomposition is used for the autocorrelation of matrix constructed by sampling data, and the numbers and frequencies of the interharmonics are obtained using the orthogonality between signal subspace and noise subspace; then SVM is used to estimate the amplitudes and phases of the interharmonics. Matlab simulation results show the accuracy of this algorithm for frequency estimation in the condition of low SNR. Benefiting from SVM in processing small sample data, the accuracy of amplitudes and phases estimating is improved effectively.