地震测井资料综合预测方法─-随机顺序指示条件模拟研究与应用

Comprehensive prediction using seismic and logging data: research and application of stochastic sequential-indicative condition simulation

聂凯轩，李承楚，王宏琳．地震测井资料综合预测方法─-随机顺序指示条件模拟研究与应用．石油地球物理勘探．1997，32（6）：842～851随机顺序指示条件模拟（SIS）作为空间变量的预测方法，在保持原空间值一致性的同时，还增加了空间结构场中的条件信息。使原始条件数据被转换为由不同门槛值定义的指示数据，利用这些指示数据，建立起指示克立金系数矩阵，从而求得预测点处的概率分布函数值cpdf。据此，便可直接采用蒙托卡罗方法得到相应点的模拟值。建立反映储层非均质性的储层地质模型是油藏描述的关键。本文以大王庄实际资料的应用结果表明，SIS方法不仅比常规的插值方法更能反映储层参数的变化，而且能够描述由于资料的缺乏而引起的模型的不确定性。

The stochastic sequential-indicative condition simulation, a method for predict-ing spatial variables, keeps original spatial data characteristic and increases condi-tion informations of a spatial structure field. The original condition data are con-verted into indicators which are defined by different thresholds; these indicators are then used to derive indication Kriging coefficient matrix, by which we obtain condi-tional probability distribution function (cpdf) at each prediction point. Thereafter, the simulated values at corresponding points can be estimated by directly usingMonte Carlo method. The key to reservoir description lies in the construction of reservoir geology model that exhibits heterogeneity of the reservoir. The application result of real da-ta in Dawangzhuang area says that the technique shows better parameter variation of reservoir than ordinary interpolation method, and predicts model uncertainty which results from the insufficiency of useful data.