基于Hadoop框架的营配调数据处理模型的设计与实现

Design and Implementation of a Data Processing Model Based on Hadoop Framework

多年来,营配调专业各自建有信息系统,存在数据管理标准不一致、各专业工作职责划分不清等问题。因此,提出基于Hadoop框架的营配调数据处理模型,采用Person相关系数计算线损电量和供电量的相关性,结合BP神经网络和时间序列计算进行线损电量预测,定量地分析供电量、售电量与线损波动的相关关系,并采用基于Hadoop框架的TF-IDF(词频-逆文档频率)算法智能识别线损异常原因,以提升电网的运行效益和精细化管理水平。以山西省阳泉市下辖的6个供电公司为例,采集供电量、售电量、线损电量和线损率等数据,进行相关性分析、线损电量预测以及对电力设备异常挂载的筛查等试验,验证了该数据处理模型的有效性和实用性。

In view of the problems such as the establishment of information systems,inconsistency of information system data management standards,and unclear division of responsibilities of various professional jobs over the years. This topic based on Hadoop camp with adjustable data processing model,the Person correlation coefficient to calculate the correlation of line loss power and power supply,combined with BP neural network line loss calculation of power and time series prediction,a quantitative analysis of the relation between power supply,electricity,and line loss fluctuations,and based on the Hadoop framework of TF-IDF(word frequency,inverse document frequency) algorithm intelligently identify abnormal line loss reasons,to enhance the level of power grid operation efficiency and fine management. Taking six power supply companies under the jurisdiction of Yangquan as examples,the data of power supply,electricity sold,line loss and line loss rate were collected,and the correlation analysis,line loss and power capacity prediction and the screening of abnormal load of power equipment were carried out,which verified the validity and practicability of the data processing model.