基于PSD-BPA的电力系统监测大数据属性约简方法

Power system monitoring big data attribute reduction method based on PSD-BPA

传统的大数据属性约简方法,在处理大数据时的约简速度较慢,导致在规定时间内不能获取有效监测数据,为此根据PSD-BPA软件的功能特点,提出对电力系统监测大数据属性约简方法。该方法利用决策数据分片,获取规则边界数值,从而设置数据属性提取规则;建立独立存在的状态空间,并将其划分成3层集合空间,结合特征矩阵建立偏序约简算法;按照直流联络线,将PSD-BPA与电力系统相连接,通过PSD-BPA分析数据属性的约简损耗量,确定大数据属性的约简程度,以此得到更加完整的约简结果。实验结果表明:与2种传统约简方法相比,所提出的约简方法,对数据属性的约简占比达到了97.2%,比传统方法分别高出17.23%、25.41%。由此可见,基于PSD-BPA的约简方法,更适用于电力监测大数据的属性约简。

Traditional big data attribute reduction methods have a slow reduction speed when processing big data,which leads to the inability to obtain effective monitoring data within the specified time.Therefore,according to the functional characteristics of psd-bpa software,a reduction method for big data attribute of power system monitoring is proposed.This method USES decision data sharding to obtain the boundary value of the rule and set the data attribute extraction rule.The independent state space is established and divided into three levels of set space.According to the dc tie line,psd-bpa is connected to the power system,and the reduction level of big data attributes is determined by analyzing the reduction loss of data attributes through psd-bpa,so as to obtain more complete reduction results.The experimental results show that compared with the two traditional reduction methods,the proposed reduction method accounts for 97.2% of data attributes,17.23% and 25.41% higher than the traditional method respectively.Therefore,the reduction method based on psd-bpa is more suitable for the reduction of attributes of big data of power monitoring.