基于RS并行约简算法的配电网故障诊断方法

摘要

随着电力系统对系统故障诊断智能化的要求越来越高，为了解决传统专家系统对不完整知识处理的局限性以及减少误判、漏判的情况，本文提出一种基于RS并行约简算法，将约简算法中计算相对正域的过程和计算核值的过程实现了并行化处理。实验结果显示，本方法提高了粗糙集中决策表属性约简的准确性，同时降低了属性约简的时间。本方法对配电网故障诊断的决策规则的自动化生成、实时故障信息进行分类判断和识别具有重要意义。%With the increasing demand of intelligent fault diagnosis of distribution network, an RS parallel reduction based approach is proposed in this paper in order to address the limitations of incomplete information processing in traditional expert systems. This approach computes relevant positive areas and core values of rough set in parallel. The experiment results show that our approach improves the accuracy of attributes reduction and reduces the performing time of executing reduction. Our approach has the significance to automatic generation of decision rules and classification and recognition of fault information.