Study On Automatic Recognition Model Of Sedimentary Microfacies Of PI2 Sand Formation Of The Middle West-2 Area In South Block Of Lamadian Oilfield

Lamadian Oilfield is located in the northern part of Daqing Oilfield in Songliao Basin.After several decades of development,it has entered a stage of high and extra-high water cut.At present,tapping potential remaining oil is the main task of exploration and development.Sedimentary microfacies research is the basic work of reservoir and remaining oil research.Conventional research on sedimentary microfacies is based on facies change law and sedimentary characteristics.Artificial recognition is often confronted with problems such as repetition of recognition process and huge workload.Therefore,it is of great practical significance to find an automatic recognition mode for sedimentary microfacies in the study area,which can replace manual recognition in order to reduce workload and improve recognition efficiency.The research object of this paper is the 2th sand group of Putaohua I reservoir group(PI2)in Lamadian Oilfield.Based on stratigraphic division and correlation,PI2 sand group is subdivided into 3 layers,6 single sand layers.Based on the previous research results,core and logging data,the braided fluvial facies of the PI2 in the study area can be divided into diara microfacies,braided channel microfacies,inter-channel floodplain microfacies and inter-channel mud microfacies.The distribution and evolution characteristics of sedimentary microfacies in the study area are analyzed by drawing the profiles and ichnography of sedimentary microfacies.On the basis of research on sedimentary microfacies,investigating the identification parameters of automatic recognition pattern of sedimentary microfacies,combining with the actual situation of the study area,three logging curves,spontaneous potential(SP),gamma ray(GR)and acoustic logging(AC),are selected.The average median,relative gravity center,variance,variable variance root are selected as parameters of the curves.In summary,12 parameters are determined as characteristic parameters to establish recognition model.According to the parameters,180 sets of the training sets and 48 sets of test sets of automatic identification model are formed.Support Vector Machine(SVM)algorithm is constructed by MATLAB software to process 180 sets of samples.After parameter optimization,an automatic identification model of PI2 microfacies in the study area is established.After verification test,the recognition accuracy of the optimal model can reach85.42%.On this basis,the model is optimized by principal component analysis(PCA),prediction accuracy can reach 91.67%.The results show that it is feasible to apply the SVM method to the identification of braided river sedimentary microfacies in the study area.The application of support vector machine method to establish the automatic identification model of sedimentary microfacies has good application prospects.