摘要
为了优化提高应用水平井数据建模的精度,本文针对油气藏精细建模中水平井采样统计偏差的问题,分析了水平井数据采样偏差产生的原因;应用Petrel中Declustering 3D内核去丛聚方法,分析了3D去丛聚方法的原理、适用条件以及水平井建模的难点。以河流相模型为例,对比分析了实际河流相模型与采样模型中河道相、天然堤相、泥岩相的分别占比,发现应用三维内核去丛聚方法使得沉积相模型更符合实际的沉积相分布,河道相和天然堤相占比分别由38.0%和13.0%降低到26.9%和9.4%,与实际沉积相比例15.7%和9.5%更加贴近。应用目标模拟河流相建模方法和序贯指示模拟方法对比去丛聚和不应用去丛聚的沉积相模拟结果,发现其应用去丛聚模拟结果可以大大提高应用水平井油藏建模的精度。
To optimize and improve the accuracy of reservoir modeling result using horizontal wells’data,focus on how to deal with the statistical distribution bias problem using horizontal wells in reservoir modeling,this article analyzed the reason of sample distribution bias,and using 3D kernel declustering method in Petrel demonstrate the theory of 3D declustering,suitable condition for using this method,and the key points during using declusering.And take fluvial facies as an example,compared the facies proportion of shale,channel,and levee which were from actual facies model and from the sampled facies model result.Find that by using 3D kernel declustering method can make facies modeling result match the actual facies proportion better,the channel proportion decrease from 38.0%to 26.9%,and levee facies proportion decrease from 13.0%to 9.4%,the result match the actual facies proportion better(15.7%and 9.5%).And using object method and sequential indictor simulation method to do the facies model,comparing the results by using declustering and without using declustering method,the results show that by using declustering method can make the facies model more accurate and to match the actual facies distribution.
作者
张改革
佟彦明
Zhang Gaige;Tong Yanming(Schlumberger China, Beijing, 10015, China)
出处
《非常规油气》
2020年第6期57-64,共8页
Unconventional Oil & Gas
关键词
水平井
油藏建模
PETREL
去丛聚
统计偏差
horizontal wells
reservoir modeling
Petrel
declustering
statistical bias
