Evaluation Method Of Sand Body Connectivity And Its Application In Saertu Oilfield

Most oilfields in eastern China have poured into the middle or high water cut stage of water-flooding,and tapping the remaining oil is the main task of exploration.Most of the main reservoirs in the continental oilfields are fluvial reservoirs.The internal architecture of the reservoirs is complex and the heterogeneity is strong.It is necessary to get a clear understanding of the connectivity of sand bodies in underground reservoirs.The research object of this paper is the fluvial sand body of Putaohua oil layer in the central west of Saertu Oilfield.Based on the subdivision to single channel sand body and the research on overlapping way between different sand body,the connection way between sand bodies was described from three angles: 1)different single sand body over the same period,2)different period of the adjacent single sand body,3)single sand body internal.On the basis of the above analysis,the influence factors of the sand body connectivity are analyzed,and the dynamic interpretation of the sand body connectivity is analyzed.The known samples of connectivity evaluation were established.The method of support vector machine(SVM)is used to establish the evaluation model which can evaluate sand body connectivity.It provides a fast and accurate method for quantitative evaluation of reservoir connectivity.The sand body connectivity is divided into three types: horizontal connectivity,vertical connectivity and internal connectivity.The geological parameters,such as ratio of sand to mud,well spacing,permeability,interlayer frequency and permeability variation coefficient were chosen as connectivity evaluation parameters.Based on the selection of geological parameters and the optimization of the model parameters,the evaluation models of the horizontal,vertical and internal connectivity in sand bodies were established by using support vector machine(SVM).The prediction accuracy is 87.8%,91.7% and 92.4% respectively.Under the same conditions,the method was switched to BP neural network and the average prediction accuracy is about 86%.These results show that it is feasible to apply the pattern recognition method to the evaluation of sand body connectivity,and the support vector machine has more advantages than the BP neural network.