数据清洗与电能质量评估技术的研究与实现
发布时间：2018-08-07 16:09
【摘要】：稳定而优质的电能是当今社会科学技术发展、国民经济飞跃的主要驱动力。为了保障电力系统可以安全、可靠地运行,必须进行电能质量的管理工作。电能质量管理的前提是可以准确地评估电能质量、查找电能质量存在的问题,它可以为管理电能质量提出科学依据与实际数据支持。电能质量的评估又依赖于电力系统所产生、采集和统计到的大量数据,这些数据中可能存在各种各样的问题与杂质,只有利用数据清洗技术,去除杂质与错误,才有可能提高电能质量评估的效率及电能质量管理的水平。本文首先对数据清洗技术的主要内容进行研究。采用“局部切比雪夫定理”与“邻近数据平均值法”针对属性值的错误进行清洗,利用“排序邻居算法”清洗重复数据,最终得到相对较为合理的统计数据。其次研究了电能质量的含义及其性质,并重点研究了构建评估体系的指标含义及计算公式。再次研究了熵权法、变异系数法、层次分析法这三种客观主观权重法,确定了评估指标的权重及评估结果。为了弥补和修正客观主观赋权法所自有的片面性与不合理性,又采用灰靶理论、基于粗糙集和信息熵的方法、最小二乘法得到综合权重及评估结果。之后又采用平均值法、Copeland法、模糊Borda法对上述评估结果进行组合评价,最终得到最全面的评估结果。最后利用编码等方式针对某110kV监测点实现了数据清洗与电能质量的评估。本文涉及的六种赋权方法,三种组合评价方法可以强有力地得到最为准确的电能质量评估结果,对提升评估效率与准确性做出了贡献,具有较强的推广价值。
[Abstract]:Stable and high-quality electric energy is the main driving force of the development of science and technology and the leap of national economy. In order to ensure the safe and reliable operation of power system, power quality management must be carried out. The premise of power quality management is to accurately evaluate power quality and find out the existing problems of power quality. It can provide scientific basis and practical data support for power quality management. The evaluation of power quality depends on the large amount of data generated, collected and counted by the power system. There may be a variety of problems and impurities in these data, only by using data cleaning technology, the impurities and errors can be removed. It is possible to improve the efficiency of power quality assessment and the level of power quality management. In this paper, the main content of data cleaning technology is studied. The method of "local Chebyshev theorem" and "the average value of adjacent data" is used to clean the error of attribute value, and the "ranking neighbor algorithm" is used to clean the repeated data, and a relatively reasonable statistical data is obtained. Secondly, the meaning and properties of power quality are studied, and the index meaning and calculation formula of constructing evaluation system are studied. Three objective subjective weight methods, entropy weight method, coefficient of variation method and analytic hierarchy process method, are studied again to determine the weight of the evaluation index and the evaluation results. In order to remedy and correct the one-sidedness and irrationality of the objective subjective weighting method, the grey target theory and the method of rough set and information entropy are used to obtain the comprehensive weight and evaluation results by the least square method. Then the method of average value and fuzzy Borda method are used to evaluate the above evaluation results, and finally the most comprehensive evaluation results are obtained. Finally, data cleaning and power quality evaluation are realized for a certain 110kV monitoring point by coding and so on. The six weighting methods and the three combined evaluation methods can obtain the most accurate results of power quality evaluation, and contribute to the improvement of evaluation efficiency and accuracy.
【学位授予单位】：华北电力大学(北京)
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM732;TP311.13
[Abstract]:Stable and high-quality electric energy is the main driving force of the development of science and technology and the leap of national economy. In order to ensure the safe and reliable operation of power system, power quality management must be carried out. The premise of power quality management is to accurately evaluate power quality and find out the existing problems of power quality. It can provide scientific basis and practical data support for power quality management. The evaluation of power quality depends on the large amount of data generated, collected and counted by the power system. There may be a variety of problems and impurities in these data, only by using data cleaning technology, the impurities and errors can be removed. It is possible to improve the efficiency of power quality assessment and the level of power quality management. In this paper, the main content of data cleaning technology is studied. The method of "local Chebyshev theorem" and "the average value of adjacent data" is used to clean the error of attribute value, and the "ranking neighbor algorithm" is used to clean the repeated data, and a relatively reasonable statistical data is obtained. Secondly, the meaning and properties of power quality are studied, and the index meaning and calculation formula of constructing evaluation system are studied. Three objective subjective weight methods, entropy weight method, coefficient of variation method and analytic hierarchy process method, are studied again to determine the weight of the evaluation index and the evaluation results. In order to remedy and correct the one-sidedness and irrationality of the objective subjective weighting method, the grey target theory and the method of rough set and information entropy are used to obtain the comprehensive weight and evaluation results by the least square method. Then the method of average value and fuzzy Borda method are used to evaluate the above evaluation results, and finally the most comprehensive evaluation results are obtained. Finally, data cleaning and power quality evaluation are realized for a certain 110kV monitoring point by coding and so on. The six weighting methods and the three combined evaluation methods can obtain the most accurate results of power quality evaluation, and contribute to the improvement of evaluation efficiency and accuracy.
【学位授予单位】：华北电力大学(北京)
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM732;TP311.13
【参考文献】
相关期刊论文 前10条
1 欧阳森;李奇;;供电企业的电能质量服务体系研究[J];华南理工大学学报(社会科学版);2015年04期
2 张逸;林焱;吴丹岳;;电能质量监测系统研究现状及发展趋势[J];电力系统保护与控制;2015年02期
3 付学谦;陈皓勇;刘国特;金小明;;分布式电源电能质量综合评估方法[J];中国电机工程学报;2014年25期
4 欧阳森;石怡理;;改进熵权法及其在电能质量评估中的应用[J];电力系统自动化;2013年21期
5 董伟杰;白晓民;朱宁辉;周子冠;李慧玲;;间歇式电源并网环境下电能质量问题研究[J];电网技术;2013年05期
6 刘俊华;罗隆福;张志文;许加柱;;一种考虑排序稳定分析的电能质量综合评估新方法[J];中国电机工程学报;2013年01期
7 叶鸥;张t，
本文编号：2170580
本文链接：https://www.wllwen.com/kejilunwen/dianlidianqilunwen/2170580.html
