In order to predict the complex connected condition between channel sand bodies,according to the data of dense well network of the PI oil layer formation in western middle area of Saertu oilfield,considering the main factors that affecting the connectivity of sand body (such as physical parameters,sand body size parameters and interlayer characteristic parameters and so on) and connectivity level,a nonlinear mapping model was established,quantitative analysis of sand body connectivity was carried out through feedback error learning method.Based on the analysis of fluvial reservoir structure of PI2 and PI3,that was the contact pattern and distribution characteristics of single sand body,selecting the suitable parameters and representative learning samples that affecting connectivity,using BP neural network discriminant model,repeated learning and prediction was carried out.By comparing the error between the predicted output and the desired output,it could be obtained that the quantitative identification of sand body connectivity based on the BP neural network had a good application effect,and could be extended to apply to the connectivity prediction of the whole block.%为了更好地预测河道砂体间的复杂连通情况,根据萨尔图油田中区西部葡I油层组的密集井网资料,考虑将影响砂体连通性的主要因素(物性参数、砂体规模参数、隔层特征参数等)与连通级别建立一个非线性映射模型,通过反馈误差学习方法对砂体连通性进行定量分析.通过对PI2、PI3河流相储层结构即单砂体接触模式及展布特征的分析,选取影响连通性的合适参数和代表性学习样本,利用BP神经网络判别模型进行反复学习和预测.通过比较预测输出与期望输出的误差情况可知,基于BP神经网络的砂体连通性定量判别具有较好的应用效果,并且可以考虑延伸应用于对整个区块的连通性预测当中去.
展开▼
