Font Size: a A A

Analysis About Heat-Metering Data Based On Data Warehouse And Cluster Technology

Posted on:2017-05-28Degree:MasterType:Thesis
Country:ChinaCandidate:Y WangFull Text:PDF
GTID:2308330485482190Subject:Control engineering
Abstract/Summary:PDF Full Text Request
With the development of City Centralized-Heating system, some problems about heating quality exposed. For example, the heating temperature is often on higher side or lower side, some users steal thermal energy by using some toolkits, and the heat meter has its own quality problems. Selecting strategic information from enormous records is a troublesome thing for the traditional database system. Data warehouse is a technique used for analysis and decision, while data-mining is a technique used to dig out interesting information from huge data. In the fields of finance, stock and market, many firms have gained huge benefits from their own data warehouse. So, the paper puts up with a systematic solution, using heating-meter data from a co-operating project between Jinan Precise Ltd and Jinan Energy, aiming at studying heating conditions of users and finding the reasons of abnormal values.Firstly, the paper identified the research meaning at the beginning, then we study the theory of data warehouse and data mining to make sure the feasibility of application in Centralized-Heating field.After designing the structure of data warehouse, the paper determined the research focus, the size of data and designed logic and physical model with modeling toolkit.Secondly, the paper designed dimension table and fact table by ETL manipulations of SSIS for its excellent performance, transforming the poor organized data into the same data warehouse. Then the paper solved and optimized the technique bottleneck problem and changed the relational data in data warehouse into Multidimensional-Cube style; the paper gave two ways to show the results of Multidimensional Analysis such as EXCEL pivot table and MDX language query.Lastly, the paper compared K-Means and EM mining results on the same data, then it chose EM mining result to analyze from four different views.Data warehouse is not a new technique, and there are a lot of successful applications in some companies. So setting up a data warehouse platform in heating meter field has strategic value, the enterprise can conserve human, physical and financial resources, solving problem mentioned in a scientific way. What’s more, papers about applying data warehouse and data mining technique in Centralized-Heating field is so few that the paper is a good reference for the researchers behind.
Keywords/Search Tags:Data Warehouse, Heat-Metering, Cube, Cluster-Mining
PDF Full Text Request
Related items