Design And Implementation Of Data Mining System For Oil And Gas Resources

As national important strategic resource,oil plays special significance to national economy and external relations.Under this circumstance,traditional analytical methods,in the face of massive oil-gas resource data with both spatial and attribute information,can no longer meet the demand for acquiring knowledge from data.Therefore,it is of theoretical and practical significance to design and develop data mining system for oil-gas resources.First of all,this paper introduces the domestic and overseas research status of digital oilfield and oil-gas resource data mining.After that,the causes of data growth in oil industry and solutions to spatial data mining are analyzed.Based on the data characteristics of oil-gas resources,the solution to data mining of oil-gas resources is designed.Firstly,the demand analysis is conducted according to the existing data of basins,oil-as fields and wells.Secondly,the detailed analysis as well as design of data mining system of oil-gas resources is performed.In this way,the B/S-based oil-gas resource data mining system including four types is realized.By means of exploratory data analysis method,the non-intuitive data features and abnormal data in oil-gas resources can be found.Meanwhile,data focus,the interested oil-gas resource data,can be selected.Finally,the hidden features and rules can be preliminarily discovered.In specific,the mining methods include data query,similarity analysis,spatial data exploration and analysis,and attribute data exploration and analysis.By combining spatial query and attribute query,the data query builds query conditions to screen out the data in accordance with researchers' set conditions from the massive oil-gas field data.The similarity analysis sorts the grades of the standard deviations of the attribute characteristics between oil-gas fields within the interested region and the reference,so as to find out the oil-gas field similar to users' interests.Besides,the trend and location of the spatial distribution of oil-gas fields are found through calculating the data center and standard deviation ellipse of oil-gas fields in spatial distribution.Taking the advantage of attribute statistics such as reservoir,entrapment and cap rocks,the characteristic rules of oil-gas resources are learned,including the maximum,minimum,mean,variance,and mode,etc.Spatial data analysis covers spatial interpolation analysis and spatial clustering.By using the density-based DBSCAN algorithm,spatial data clustering method clusters oil-gas fields within users' interested region,uncovers high capacity oil-gas field clusters,and summarizes the rules of spatial differentiation.In adoption of empirical Bayesian Kriging interpolation,spatial interpolation analysis calculates the reservoir characteristics of the known oil-gas fields,predicts the reservoir conditions of the unknown oil-gas fields,and then figures out the spatial distribution of characteristics.Through density analysis,Getis-OrdGi* hotspot analysis,and Anselin Local Moran's I outlier analysis,spatial distribution pattern analysis excavates the spatial distribution of oil-gas field data.Density analysis,based on the calculation of spatial density distribution of oil-gas fields,can conclude the relatively concentrated or scattered oil-gas fields,so as to find the regions with development potential.Hotspot analysis and outlier analysis are applied to identify hotspots,cold spots and outliers that are statistically significant in the spatial characteristics of oil-gas fields.Visual analysis helps reveal the status,inherent nature and regularity of oil-gas resource data.The statistical chart designed in this paper can be divided into visualization and thematic map visualization.In consideration of users' different expression requirements for oil-gas data attribute values,the statistical chart visualization mainly includes four kinds,which are histogram,pie chart,bar chart and line chart.By intuitively displaying the spatial data mining results on the map,the thematic map visualization,on the one hand,can display the distribution rules of its knowledge features.On the other hand,it can also visually explain the mining results to achieve the optimal analysis results.The oil-gas resource mining system is realized via ArcGIS API for JavaScript,and the four data mining methods are tested by global oil-gas resource data.The results show that the system function is normal,the test results are consistent with the actual situation,and the system achieves the expected results.