如何使地质数据更具代表性

DECLUSTERING EFFECT OF GEOLOGICAL DATA

由于地质过程的复杂性,导致了油气储层分布的非均质性。钻井总是被优先设计在有利储层的部位,从而导致了钻井分布的不均匀,进而使得获取的地质数据在空间上的分布也是不均匀的。如果不考虑这种地质数据分布的不均匀性,可能会导致数据统计结果的失真。这里介绍了二种方法解决这个问题:网格去丛聚方法和多边形去丛聚方法。这二种方法的基本思想都是给密集的数据以较低的权值,而给稀疏的数据较大的权值。应用这二种方法分别对石南油田某井区的微相类型比例及孔隙度分布进行了分析,结果表明,经过处理后的地质数据更能真实地反映实际情况。

The various complex geological processes result in the reservoir heterogeneity. Wells are always drilled in areas with a greater probability of good reservoir quality. Core measurements are taken preferentially from good quality reservoir rock. The data from these sources distribute asymmetrically in the interesting area. Usually, some data are spatially clustered. It needs to correct this preferential clustering before making statistic. Two methods are introduced to obtain a representative statistics of the entire area of interest: one is cell declustering, and the other is polygonal declustering. The basic idea of the two methods is to assign greater weights to closely spaced data and to assign lesser weights to widely spaced data. A case study in Shinan oil field is presented. The research result shows that the statistic results using these two declustering methods are more reliable.