密集井网下随机地震反演方案及砂体预测

在油田深度开发阶段的高密度井网条件下，单纯依靠井点信息或地震资料很难预测河道砂体。为此，以中国东部油田某断块为例，研究了密集井网条件下随机地震反演的关键技术环节，分析了不同井网密度下反演结果对砂体的识别和预测能力。得出以下认识：①影响随机反演效果的关键技术环节主要有储层敏感曲线的选取与提高分辨率为目的的重构处理、地震垂向采样率加密和符合研究区地质、地震资料情况的变差函数的优化等；②在密集井网条件下变差函数的垂向变程由井曲线拟合所得，横向变程的选取要统筹考虑区域沉积相带几何构型特征及井曲线和地震数据的拟合结果等，通过对比实验找到井、震两类数据空间构型的最佳融合方案——变程参数的设置；③随机地震反演在研究区的砂体预测结果表明，厚度为2~3m、1~2m、0~1m的单层砂体的预测精度分别为73%、56%和33%，互层砂体的预测精度分别为79%、59%和41%。

Stochastic seismic inversion scheme and sand body prediction in dense well pattern areas

For dense well pattern in the deep oilfield development stage, it is difficult to predict channel sand bodies with only well-point information or seismic data.Therefore, taking a block in an eastern oilfield as an example, key technical procedures of stochastic seismic inversion in dense well pattern are studied, and the ability to recognize and predict sand bodies in different well pattern densities is analyzed.The following conclusions can be drawn:A.Key techniques of the stochastic inversion are reservoir sensitivity curve selection, reconstructing process to improve the resolution, seismic data resampling, and variation function optimization; B.The vertical range of the variogram under the condition of dense well pattern is obtained from well curve fitting.For the lateral range selection, we should consider geometric characteristics of sedimentary facies and the fitting results of well curves and seismic data.With experiment comparison, the best fusion scheme of well and seismic data space configurations (variable range parameter setting) is obtained; C.The results of stochastic seismic inversion show that the prediction accuracy of single sand body with thickness of 2~3m, 1~2m, and 0~1m are 73%, 56%, and 33%, the prediction accuracy of inter-bed sands are 79%, 59%, and 41% respectively.