论文部分内容阅读
本文介绍一种随机模拟方法,我们开发了这一方法来建立阿拉斯加北部部分斜坡的区域平均速度层面和确定主要地质标志层顶部的深度层面。该方法将井资料计算出的平均速度和用二维及三维地震数据的速度分析(“VELAN”)求出的叠加速度综合在一起。该方法的实现过程是:首先将速度数据体模拟成一个由若干统计上均一(平稳)群体组成的混合体,这些群体主要与近地表永冻层厚度以及其它已知的地质因素的变化相一致;其次,根据井速度对VELAN速度进行统计标定,软反演(soft inversion)算法用于选定位置(远离已知井点)上推断平均速度的范围;最后,使用(混合的)随机成像算法求出速度和深度随机成像图,这种速度图和深度图均满足于测井速度和软反演的速度范围以及数据的空间相关模型。经过30口井的“盲”试验,其结果说明,与常规内插方法相比,该方法的速度估计精度有较明显的提高(平均速度估计误差约为50ft/s),同时,这些井的检验结果也证实了局部估计风险值(置信区间)的正确性,这种风险值是该方法固有的副产品。
This article presents a stochastic simulation method that we developed to create a regional average velocity profile for part of a slope in northern Alaska and to determine the depth profile at the top of a major geological marker. The method combines the average velocity calculated from the well data with the stacking velocity obtained from the velocity analysis of 2D and 3D seismic data (“VELAN”). This method is implemented by first simulating the velocity data volume as a mixture of several statistically homogeneous (stationary) populations that are primarily consistent with changes in near-surface permafrost thickness and other known geologic factors Secondly, the VELAN velocities are statistically calibrated according to the well velocity. The soft inversion algorithm is used to infer the range of the average velocities from the selected positions (far away from the known well points). Finally, using the (mixed) stochastic imaging algorithm Stochastic maps of velocities and depths are obtained. Both the velocity and depth maps are satisfied with the velocity range of the logging and soft inversion and the spatial correlation of the data. After 30 wells’ “blind” test, the results show that compared with the conventional interpolation method, the speed estimation accuracy of the method is obviously improved (the average speed estimation error is about 50ft / s). At the same time, these The well test results also confirm the correctness of the locally estimated risk value (confidence interval), which is an inherent by-product of the method.