关键词: 灰色关联分析, 支持向量机, 河道, 大庆长垣

Abstract: It is difficult to predict the accuracy of the river by using the conventional seismic attributes, due to the complex conditions of the terrigenous clastic basin reservoirs and the rapid change of the rock facies. In this paper, XII7-12 layer in N area of S oilfield is taken as an example. Through the seismic forward analysis, the seismic section reflection characteristics of different types of reservoirs are different, and the seismic attributes of the combined layer can be used as the effective dimension for the main channel prediction in this area. On this basis, using the method of the combination of Gray Relational Analysis and Support Vector Machine, the N zone seismic attribute prediction based on Gray Relational Support Vector Machine is completed. Drilling confirmed that based on Gray Correlation and Support Vector Machine attribute prediction, the drilling coincidence rate is higher. By using the advantage of seismic inversion to predict the boundary of the channel sand, the comprehensive analysis of dynamic and static data of the polymer flooding well group effectively solved the contradiction between the injection and production system of the XII7-12 strata system in the N area of the S oilfield, thus, it further confirms the accuracy of attribute prediction based on GRA-SVM. Comprehensive research shows that this method is suitable for high accuracy of the river prediction and can be used as a better river prediction method under complex geological conditions.

Key words: Grey relational analysis, Support vector machine, Channel, Daqing placanticline