由两点到多点的地质统计学储层建模

Reservoir stochastic modeling using geostatistics from two-point to multiple-point

陈培元¹ 姜楠¹ 杨辉廷¹ 刘学利²

西南石油大学资源与环境学院，四川 成都 610500 中国石化西北油田分公司，新疆 乌鲁木齐 830011

School of Resources and Environment, Southwest Petroleum University, Chengdu 610500, China Northwest Oilfield Company, SINOPEC, Urumqi 830011, China

传统的两点地质统计学建模方法，以象元为空间赋值单元、变差函数为工具建立确定性的模型，或者应用各种随机模拟方法建立可选的模型，在精确表征复杂的空间结构及目标体几何形态方面有一定的局限性。有别于两点地质统计学的多点地质统计学，可有效地解决更广泛的地质模拟问题。然而，在实际应用过程中，受岩?鄄相模型及与之相对应的训练图像可靠性的影响，结果变得比较复杂。因此，选择合适的训练图像及恰当的算法可有助于提高储层建模的精度和效率。以××油田曲流河沉积为例，采用两点和多点统计学方法构建模型。对比发现，基于多点地质统计学的地质建模方法真实可再现河流相的沉积形态，还降低随机建模的不确定性。尽管模拟结果与井点真实数据之间存在误差，但通过调整随搜索半径、训练图像大小及概率计算中临近点个数限制，可显著提高模型精度。

Conventional two-point geostatistics modeling mainly uses the pixel-based method and variogram to establish the deterministic model, or uses all kinds of stochastic simulation methods to establish optional model. But it can not fully reflect the variability of the space structure and geometric shape of object. Unlike the two-point geostatistics, the multiple-point geostatistics can solve the problems of geologic simulation widely. In actual application process, due to the lithofacies and the reliability of corresponding training image, it is necessary choosing the suitable training image and appropriate algorithm to improve the accuracy and efficiency of simulation. Taking the meandering river sedimentation of some oilfield as an example, the reservoir model is built by two-point and multiple-point geostatistics methods. Comparison results of two models show that the method based on the multiple-point not only represents the real sedimentary form of fluvial facies, but also reduces the uncertainty of stochastic modeling effectively and improves the modeling accuracy at maximum. Although the simulation results do not agree with the real well date completely, the prediction accuracy can be improved through adjusting the search radius, the size of training image and the number limit of nearest-neighbor points used in probability calculation.

10.6056/dkyqt201205012