基于互近似熵数据筛选的谐波源责任划分方法

Determination method of the harmonic source contribution based on cross-approximation entropy data filtering

针对背景谐波波动影响谐波责任划分准确性的问题,提出了基于互近似熵数据筛选的谐波源责任划分方法。计算谐波电压、电流数据段的互近似熵(Cross Approximation Entropy, CAE)值,利用阈值筛选的方法,保留波形模式相似的数据区段,以达到排除背景谐波干扰的目的。利用M估计稳健回归法对保留数据进行回归计算,规避异常值对系统谐波阻抗回归计算的干扰,实现准确的谐波责任划分。仿真分析结果验证了该方法的准确性和优越性。

Aiming at the problem that background harmonic fluctuation affects the accuracy of harmonic contribution determination, this paper proposes a determination method of harmonic source contribution based on cross-approximation entropy(CAE) data filtering. Firstly, the CAE values of the harmonic voltage and current data segments are calculated. The threshold filtering method is used to preserve the valid data segments with similar waveform patterns to eliminate the interference of background harmonic. Then, the M-estimated robust regression method is used to perform regression calculation on the retained data to avoid the interference of the outliers on the system harmonic impedance regression calculation, thereby achieving accurate the result of harmonic contribution determination. Finally, simulation results verify the accuracy and superiority of the proposed method.