Design And Development Of Reservoir Data Analysis And The Evaluation Of Injection-production Effect System Based On Information Mining

Currently, although small amounts of oil field test data management software in oilfields, single function and most of the test data management and analysis are still in themanual to search and analysis information, the low degree of computer automationapplications. To analyze a single well needs to view more than10related data whichinclude production data, the monitoring data, the production of auxiliary laboratory data,geological analysis chart. To analyze a test block or oil field needs to gather lots ofhistorical production data about oil wells and other relevant information. Those not onlyspend a lot of time and manpower, but also produce a lot of business data on the productiondecision-makers and the information provided is not useful. Thus, as soon as possible tosolve this problem, data mining technology came into being. The key to the success of datamining is to determine the appropriate mining goals. Established mining goals are necessaryto consider the needs of the business management, but also consider the characteristics ofthe existing data. Ensure that the target can be achieved on the basis of existing data.This paper conducts the analysis and design of the system, data mining technologyapplied to the information mining of system to support production managers.The specificresearch topics include the following aspects.Firstly, in accordance with the data miningprocess, data and information collected to meet the needs of mining theme. Then we findout the oil reservoir data and production of secondary containing laboratory data and otherbusiness data as the mining goals. Secondly, according to the characteristics of database is asubject-oriented, integrated relatively stable, reflects the historical changes and for to adaptto the environment of the oil field business data to achieve the establishment of oilfieldinformation database model. System of excavated goals after cleaning, classification,summary, and organized loaded to the oil fields in the database, in order to facilitate thebusiness data multidimensional queries, and thus support the extracted data from the massof data and information that is conducive to production decisions. Thirdly, it selects theappropriate mining algorithms to meet the system requirements for the excavated goals.System conducts a preliminary study and exploration of the BP Neural Networks, the GrayCorrelation Analysis method and the Multiple Regression method. Mining the results obtainedwith the actual results are compared to determine the meaning of its representative todetermine whether practical significance. Finally, system specific business processes andsystem functional requirements are analyzed, in accordance with the relevant requirements of the system to design the structure of software and visualization system based on B/S.