基于联合对角化法与数据筛选的谐波责任划分

针对背景谐波波动的随机性，为了更加准确地划分谐波责任，考虑到谐波阻抗的符号特性和电力系统中系统侧谐波阻抗远远小于用户侧谐波阻抗的实际情况，在联合对角化法的基础上，提出了一套数据筛选准则。利用联合对角化法对谐波电压和谐波电流进行分解，得到混合矩阵；分别通过符号判别法和比例系数法纠正次序不确定性和幅值不确定性；应用数据筛选准则得到谐波阻抗；根据谐波责任定义式估算谐波责任。对IEEE 13节点系统进行仿真测试和实测数据分析，结果表明：与传统方法相比，所提方法的估算精度和适应性更好。

Harmonic responsibility determination based on JADE and data filtering

In allusion to the randomness of the background harmonic fluctuations and considering the symbolic characteristic of harmonic impedance and the practical situation that the harmonic impedance of the system-side is far less than that of the customer-side, a set of data filtering criterion is proposed based on the JADE(Joint Approximate Diagonalization of Eigen-matrices) to determine the harmonic responsibility more accurately. The mixed matrix is obtained by using the JADE to decompose the harmonic voltage and the harmonic current, the order uncertainty and the amplitude uncertainty are corrected by the symbol discriminant method and the proportional coefficient method respectively, the harmonic impedance is obtained by using the data filtering criterion, and then the harmonic responsibility is calculated according to the definition equation. The simulation test and measured data analysis of IEEE 13-bus system show that the proposed method has better estimation accuracy and adaptability than traditional methods.