面向变电站工程设计信息的数据挖掘技术研究

发布时间：2018-06-27 19:34

本文选题：变电站 + 工程设计　；参考：《华北电力大学(北京)》2017年硕士论文

【摘要】：随着电力体制改革和国民经济的迅猛发展,电网建设步伐在逐年加快。变电站工程设计的评价过程从人工评价向系统智能评价的过度是必然趋势,也是输变电系统工程可靠性的必然要求,虽然近年来在部分环节上采用了智能评价手段,为变电站工程智能评价提供了一定的参考,但是对于变电站工程设计整体的评价还是十分有限的,这就需要通过对变电站工程设计的基础信息,设计方案,参数数据等工程设计信息进行数据挖掘分析,对其特征参量进行提取、收集,作为变电站工程设计的评价基础,进而选择有效的工程设计评价方法,确定科学合理的评价策略,提高对变电站工程设计评价的可靠性和准确性。针对以上问题,本论文开展了面向变电站工程设计信息的数据挖掘技术研究。首先,分析了智能电网背景下变电站工程设计信息的特点并对其进行数据预处理工作,利用有限覆盖原理和聚类分析原理对变电站设计信息进行基于峰值聚类的数据清洗;其次,采用社会网络相关原理对清洗过的工程设计数据进行分析处理,利用属性空间投影、指标关联因子、空间关联因子以及子属性空间关联矩阵等概念构建变电站工程设计评价指标体系以及基于属性模式识别的评价模型;最后,以220k V变电站工程电气一次部分设计信息数据为例进行实例分析,利用本文所涉及的数据挖掘技术,对全国100多个220kV变电站工程电气一次设计信息数据进行数据清洗,并且对其中8项新建变电站工程进行基于社会属性网络的变电站设计电气一次评价。本论文实现了基于峰值聚类理论对多数据源变电站工程电气一次设计信息数据的数据清洗工作,并且,基于社会属性网络的评价结果也与国家电网公司评出的优质工程相符合,从而验证了本论文中所提出的社会属性网络SANs在变电站工程初步设计电气一次评价应用的合理性,进一步验证了面向变电站工程设计信息的数据挖掘技术的有效性。论文中所涉及的数据挖掘技术在输变电工程甚至其他领域的工程项目设计与评价环节中有着极大的参考价值。
[Abstract]:With the reform of power system and the rapid development of national economy, the pace of power grid construction is accelerating year by year. The excessive evaluation process of substation engineering design from manual evaluation to system intelligent evaluation is an inevitable trend and an inevitable requirement for the reliability of transmission and transformation systems, although in recent years intelligent evaluation methods have been adopted in some aspects. It provides a certain reference for intelligent evaluation of substation engineering, but the overall evaluation of substation engineering design is still very limited, which requires the basic information of substation engineering design, design scheme, The engineering design information, such as parameter data, is mined and analyzed, and its characteristic parameters are extracted and collected as the evaluation basis of substation engineering design. Then the effective engineering design evaluation method is selected, and the scientific and reasonable evaluation strategy is determined. Improve the reliability and accuracy of substation engineering design evaluation. In order to solve the above problems, this paper studies the data mining technology for substation engineering design information. Firstly, the characteristics of substation engineering design information under the background of smart grid are analyzed and the data preprocessing work is carried out. The data cleaning of substation design information based on peak clustering is carried out by using the principle of limited coverage and clustering analysis. Secondly, using the social network correlation principle to analyze and process the cleaned engineering design data, using attribute space projection, index correlation factor, The concepts of spatial correlation factor and subattribute spatial correlation matrix are used to construct the evaluation index system of substation engineering design and the evaluation model based on attribute pattern recognition. With the example of 220kV substation electrical once design information data, the data of one time design information of more than 100 220kV substations in China are cleaned by using the data mining technology involved in this paper. And 8 new substation projects are evaluated based on social attribute network. In this paper, the data cleaning of the primary design information data of substation engineering based on peak clustering theory is realized, and the evaluation results based on social attribute network are in accordance with the high quality project evaluated by State Grid Company. Thus, the rationality of the application of the social attribute network SANs proposed in this paper in the primary design of substation engineering electrical evaluation is verified, and the validity of the data mining technology for substation engineering design information is further verified. The data mining technology involved in this paper has great reference value in the design and evaluation of transmission and transformation engineering and even other engineering projects.
【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】：TP311.13;TM63

【参考文献】