基于灰靶理论与云模型的电压暂降事件数据挖掘分析方法

An Analytical Method of Data Mining on Voltage Sag Based on Gray Target Theory and Cloud Model

对电压暂降事件记录中所蕴含的知识进行总结,挖掘不同故障场景与节点电压暂降影响程度之间的关系,可以为管理部门制定决策提供有价值的信息。采用自适应高斯云变换算法,对与电压暂降相关的各类连续量数据进行离散,通过扫描一遍原始事务库构建维度矩阵,以此代替传统Apriori Tid算法中的项目集,形成了空间和时间效率都有所提升的改进算法。随后结合灰靶理论,基于隶属度构建指标序列,建立实际场景与各强关联规则之间的匹配模型,在当前分析里筛选出最能反映实际场景中节点受电压暂降影响规律的知识。实例分析验证了所提方法的实用性和有效性。

This paper summarizes the knowledge contained in voltage sag event records and mines the relationship between different fault scenarios and site voltage sag severity, providing valuable information for management departments to make decisions. The adaptive Gaussian cloud transformation algorithm is used to discretize the continuous data related to voltage sag. A dimension matrix is constructed by once scanning original transaction database to replace the item sets in traditional AprioriTid algorithm. An algorithm improving both space and time efficiency is proposed. Then, combined with gray target theory, index sequences are built based on membership to construct the matching model between actual scenes and strong association rules. Finally, the knowledge best reflecting the law of the effect of voltage sag in actual analysis is selected. Practical example analysis validates practicability and effectiveness of the proposed method.