依托关联规则挖掘的电力生产安全事故致因攫取

Cause grabbing of electric power production safety accidents based on association rules mining

电力大数据的蓬勃发展给电力生产安全态势的可靠感知提供了新的机遇和平台。首先,建立多层级、全方位的电力生产安全事故致因体系;其次,依托关联规则挖掘,建立事故致因对事故诱发程度的计算模型,包括关联规则的相关定义、不同类型事故的布尔离散化、关联规则的置信度、支持度、相关度的计算等;再次,借助Apriori算法进行事故关键性致因的攫取;最后,佐以实例证明。实证表明,利用本文所建模型,可对历史故障报告的全景信息予以科学提取,并抽象为各类事故事件的关键致因展示,以便形成非常有针对性的安全管控措施和策略。

The vigorous development of large power data provides new opportunities and platforms for the reliable perception of power production security situation.Firstly,a multi-level and all-round power production safety accident causation system is established.Secondly,relying on the mining of association rules,the calculation model of accident-induced degree is established,including the definition of association rules,Boolean discretization of different types of accidents,the calculation of confidence,support and correlation of association rules.Again,Apriori algorithm is used to capture the critical causes of accidents.Finally,an example is given to prove it.Empirical results show that the panoramic information of historical fault reports can be extracted scientifically by using the model in this paper,and can be abstracted to show the key causes of various accident events,so as to form a very targeted safety control measures and strategies.