一种基于数学形态学的扰动信号分形检测方法  

A Fractal Detection Method Based on Mathematical Morphology for Disturbance Signal

作　　者：石佳[1] 黄纯[1] 李杨[1] 

机构地区：[1]湖南大学电气信息工程学院,长沙410082

出　　处：《电力系统及其自动化学报》2008年第5期86-90,共5页Proceedings of the CSU-EPSA

摘　　要：针对电能质量扰动信号检测和定位问题,提出了一种基于自适应形态滤波和网格分形的方法。该方法将信号通过改进的形态滤波器进行预处理。该滤波器综合了不同的结构元素且在开-闭、闭-开滤波器权系数的确定上采用了基于最小均方误差算法的自适应技术,具有更好的滤波性能和细节信息保留能力;根据网格分形理论,通过对信号网格变化规律的分析实现对电能质量扰动信号的精确定位。仿真结果表明了该方法的可行性和正确性。A novel approach based on adaptive morphology filter and grille fractal is presented to detect and locate the power quality （PQ） disturbance signals. At first, the signal is disposed by an improved morphology filter, which is composed of different structuring elements and adopts the least mean square （LMS） arithmetic the weight coefficients of open-closing （OC） filter and close-opening （CO） filter. So the improved filter has good performance for suppressing noise and preseruing derail information. And then, following the theory of grille fraetal, PQ disturbance signals can be located exactly by analyzing the grille change regularity. Simulation results show that the proposed approach is practicable and valid.

关 键 词：电能质量 扰动检测 数学形态学 最小均方误差算法 网格分形 

分 类 号：TM761[电气工程—电力系统及自动化]