基于大数据的电力系统信息质量评估

发布时间：2018-07-27 17:10

【摘要】：随着电力系统规模的增大、量测技术的发展与成本下降,电力系统的数据量呈现快速增长趋势,逐步具备了大数据特征。充分利用大数据来改善电力系统的规划、运行与控制已受到越来越广泛重视,这样如何评估大数据的质量就是一个值得研究的重要问题。在数据质量提高技术如数据清洗、数据整合、相似记录检测等方面,已有相当多的研究报道。然而,在数据质量的评估方面,研究工作还相当初步。在此背景下,针对电力系统特征和电力大数据质量特性,提出一种电力大数据质量综合评估方法。对大批量的评估总体,有时各类质量的实际分布差别很大,那么采用组合的评估方法就比较合理,因此将熵权法和灰类评估法结合起来,在Hadoop平台上构建电力大数据质量评估模型,该模型详细地列出电力大数据状态评估的步骤。本文将电力系统数据作为主要研究对象,阐述了电力系统数据质量面临的主要问题,参照国际标准ISO/IEC 25012,并从电力系统存在的质量问题及电力大数据的特点,首先提炼出电力系统数据质量评估的一般指标,构建电力大数据质量评估的指标体系,针对大数据处理的时效性问题,利用MapReduce并行化K-means聚类算法来实现大数据样本集的快速预处理。然后利用熵权法计算指标体系中的指标各类数据集的客观权重,最后采集某市电力公司所采集的用户用电负荷数据进行算例分析,采用灰色评估法判断数据质量所属等级,在此基础上实现对样本数据集的综合评价。计算结果表明,所提出的方法可以通过业务规则和需求对指标体系有一个定量的描述,灰色熵权综合评估法评价大数据质量在时间效率上具有较好的表现。
[Abstract]:With the increase of the scale of power system, the development of measurement technology and the decrease of cost, the amount of data in power system is increasing rapidly and gradually has the characteristics of big data. Making full use of big data to improve power system planning, operation and control has been paid more and more attention. Therefore, how to evaluate the quality of big data is an important issue worth studying. Many researches have been reported on data quality improvement techniques such as data cleaning, data integration, similar record detection and so on. However, in the evaluation of data quality, the research work is still quite preliminary. Under this background, a comprehensive evaluation method of power big data quality is proposed for the characteristics of power system and power big data quality. For mass evaluation, sometimes the actual distribution of various kinds of quality is very different, so it is more reasonable to adopt the combined evaluation method, so the entropy weight method and the grey class evaluation method are combined together. The power big data quality assessment model is constructed on Hadoop platform, which lists the steps of power big data state evaluation in detail. In this paper, power system data is taken as the main research object, the main problems of power system data quality are expounded, referring to the international standard ISO/IEC 25012, and the quality problems of power system and the characteristics of power big data are discussed. Firstly, the general index of power system data quality evaluation is abstracted, and the index system of power big data quality evaluation is constructed. Aiming at the timeliness of big data processing, MapReduce parallel K-means clustering algorithm is used to realize the fast preprocessing of big data sample set. Then the objective weights of all kinds of data sets in the index system are calculated by using entropy weight method. Finally, the data collected by a power company in a certain city are analyzed by an example, and the grade of data quality is judged by grey evaluation method. On this basis, the comprehensive evaluation of the sample data set is realized. The calculation results show that the proposed method can describe the index system quantitatively through business rules and requirements, and the grey entropy weight comprehensive evaluation method has a good performance in evaluating the quality of big data in terms of time efficiency.
【学位授予单位】：华北电力大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM73

【参考文献】