基于粗糙集的配电系统连续信号故障诊断方法

提出了一种基于断点重要性的配电网连续属性离散化方法,证明了该方法的有效性。综合考虑了线路的选择系数、灵敏系数和支持权值,应用属性离散指标作为离散化的评价标准,证明该指标可以作为离散化彻底的充分条件。与同类算法的比较可以发现算法在基本不损失分类信息的基础上有效降低了时间复杂度和空间复杂度。首次在配电系统故障诊断中将连续和离散信号统一的应用粗糙集进行约简, 使电压、电流等配电系统的重要模拟量可以参与到故障诊断系统中,提高了故障辨识的精度。同时,也使对于配电系统的临界稳态运行情况的判定提供了一个可能的解决办法。

Fault Diagnose Algorithm of Distribution System by Continuous Signals Based on Rough Sets

A continuous attribute discretization of the electric power distribution system is put forward based on the breakpoint importance, which is proved effectively. The distinguish breakpoint as evaluation standard of discretization is applied based on choice coefficient, sensitive coefficient and support value. And the degree is considered to be the sufficient condition. It can be seen that the algorithm is effective. In the fault diagnosis system of electric power distribution system, it is very important to make unified reduction of continuous and discretizing signals with the rough sets at first. In these way, we can make the important analogues of distribution system such as voltage, current take part in the fault diagnosis, and improve the precision of fault identification. At the same time, one possible solution can be found for determination of the critical stable status of distribution system.