火山岩储层地震预测

摘要

Seismic technology represented by seismic attribute and inversion is the universal and effective technology for reservoir prediction. Taking for example volcanic rocks in Honghaoershute Sag of Erlian Basin,by well-seismic calibration, we summarized reflection characters of volcanic rocks on the seismic section and horizontal time slice and then identified the volcanic rock. As volcanic reservoir has strong heterogeneity and conventional seismic inversion is too dependent on module, we tried to extrapolate the distributions of volcanic reservoir by applying AVF inversion and reservoir-parameter inversion based on BP neural network. In the end,we analyzed the two inversion results and pointed out the exact effective reservoir zones and non-reservoir zones,which accorded well with the real datum. We believe the findings can effectively guide the exploration of volcanic rocks. This technology and method on reservoir prediction can be used for reference to especial rocks,which has strong heterogeneity.%以地震属性及反演为代表的地震技术是进行储层预测的最为普遍且有效的技术，以二连盆地洪浩尔舒特凹陷白垩系阿尔善组火山岩为例，通过井震标定，总结火山岩在常规地震剖面及水平时间切片上的反射特征，对火山岩进行了识别，在此基础上，针对火山岩储层具有较强非均质性的特点以及常规地震反演对模型依赖性较大的缺陷，探讨性地采用分频属性反演及基于神经网络的储层参数反演分别对该区火山岩储层分布进行了预测，将两种预测结果综合分析，指出了该区较确切的火山岩有效储层分布区与非储层区，与实际钻井资料吻合较好，为下一步开展火山岩勘探生产工作起到了一定的指导作用。该技术方法对于非均质性较强的特殊岩体的储层预测有一定的借鉴意义。