基于数据仓库的机场能源信息管理研究

发布时间：2018-01-11 17:05

本文关键词：基于数据仓库的机场能源信息管理研究　出处：《中国民航大学》2017年硕士论文　论文类型：学位论文

【摘要】：机场领域在日常运营中积累了大量的能源数据信息。随着机场规模的不断发展壮大,机场能源消耗的急剧上升,机场能源信息科学管理显得愈加重要,而基于传统数据库的统计查询分析方式越来越不能满足机场能源信息管理与分析的需求。为此本文提出将数据仓库技术运用到机场能源信息管理中,构建机场能源信息数据仓库,并在数据仓库的基础上进行聚类挖掘。首先分析机场能源管理数据仓库系统需求,设计了数据仓库模型;确立了机场能源管理数据仓库的两大主题结构,并为两个主题设计了逻辑模型结构和物理模型结构;根据机场能源数据特点设计两层ETL架构,对机场能源数据进行处理;实现远程异构数据库的访问和数据的抽取装载,对数据进行转换,包括数据清洗、数据集成。其次提出以机场能源数据仓库为基础进行聚类分析,实现对历史数据的有效信息挖掘;对K-Means算法在K值选取和初始化聚类中心上的不足进行改进。构建以逐月能耗值为数据点的多维能耗时间序列,确定用能单位的能耗模式。最后,基于改进K-Means算法对机场能源数据进行聚类实验分析。通过经典数据集测试改进K-means算法的有效性;以机场月度综合能耗数据进行聚类,建立不同类单位群体能耗基准值;最后通过对能耗水平序列以及能耗变化率序列进行聚类,建立群体细分模型。
[Abstract]:The airport has accumulated energy a lot of data in their daily operations. With the continuous development and expansion of the scale of the airport, the sharp rise in energy consumption of the airport, the airport energy information scientific management becomes more and more important, and the traditional database query based on statistical analysis method could not satisfy the airport's energy information management and analysis in this paper will demand. Data warehouse technology is applied to the airport energy information management, construction of the airport energy information data warehouse, and Clustering Mining Based on data warehouse. Firstly, the airport energy management data warehouse system requirements, design the data warehouse model; established two thematic structure of Airport energy management data warehouse, and the design of logic structure model the physical model and structure for the two theme; design of two layer ETL architecture according to the characteristics of airport energy data, airport source number According to the processing; remote access to heterogeneous databases and data loading, data conversion, data cleaning, data integration. Secondly, to the airport energy data warehouse based on clustering analysis, realize the effective information of historical data mining; K-Means algorithm for value selection and initialize clustering center on K improved. To construct the monthly energy consumption value of multidimensional time series data of energy consumption, energy consumption can be determined by model units. Finally, the improved K-Means algorithm of airport energy data clustering experiment analysis based on improved K-means algorithm. The classic data set test; clustering to airport monthly comprehensive energy consumption data, establish the same unit energy consumption benchmark value; finally, through clustering on energy consumption and energy consumption level sequence change rate sequence, establish group segmentation model.

【学位授予单位】：中国民航大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：V351.3;TP311.13

【参考文献】