基于EEMD和关联维数的小电流接地故障选线

Fault Selection for Small Current Grounding System Based on EEMD and the Correlation Dimension Algorithm

提出一种基于总体平均经验模态分解（EEMD）和关联维数相结合的小电流接地故障选线新方法。EEMD对非线性、非平稳信号的处理，不仅能达到与经验模态分解（EMD）相同的效果，同时又能有效地抑制模式混叠，非常适用于对小电流接地故障信号的处理。关联维数作为反映系统状态的特征量，能定量分析故障状态，提高故障诊断能力。在计算关联维数前，需要进行相空间重构，采用极大联合熵算法求取最佳延迟时间，以往用互信息求取延迟时间法，该方法简化了算法，缩短了计算关联维数的时间。最后采用G-P算法计算零序电流相关分量的关联维数，通过比较关联维数，实现故障选线。实验结果表明该方法能快速准确地选出故障线路，为小电流接地故障选线提供一种有效的新方法。

A new fault line selection method for small current grounding system based on ensemble empirical mode of decomposition （EEMD） and correlation dimension is proposed. EEMD is ideal for the fault signal processing of small current grounding system, which could not only achieve the same effect with the empirical mode decomposition （EMD）, but also suppress aliasing mode effectively when processing nonlinear and non-stationary signal. The correlation dimension could reflect the feature quantity of the system state, analyze the fault condition quantitatively, and improve the capabilities of fault diagnosis. Phase space reconstruction is necessary before calculating the correlation dimension, and the maximum joint entropy algorithm is introduced to get the optimal delay time, which simplifies the algorithm, and shortens the calculation time of the correlation dimension compared with the mutual information requirements delay time. Finally, G-P algorithm is adopted to calculate the correlation dimension, realizing fault line selection by comparing the numerical of correlation dimension. The experiment results show that the proposed method could select fault line rapidly and accurately, providing an effective method for the small current grounding fault line selection.