变差函数的参数和井数对随机反演精度影响的分析

2016年 55卷 第No. 5期
阅读:74
查看详情
Analysis on the influence of variogram and well number on the precision of seismic stochastic inversion
(1.中国石油天然气股份有限公司勘探开发研究院,北京100083;2.中国石油天然气股份有限公司华北油田分公司勘探开发研究院,河北任丘062552)
(1.PetroChina Research Institute of Petroleum Exploration & Development,Beijing 100083,China;2.Exploration and Development Research Institute,North China Oilfield Company,Renqiu 062552,China)

高含水后期陆相油藏可寻找的剩余油多分布在较薄的储层中,高效挖潜剩余油的重点和难点是井间储层的预测,因为现有井网仍然无法有效控制砂体边界,因此随机反演技术成为预测井间砂体边界的有效手段。以XSG油田典型区块为试验区,分析改变变差函数的参数和参与反演的井数对随机反演预测井间储层精度的影响。结果表明,垂向变差函数对随机反演结果的影响比水平变差函数更敏感;随着反演井数的增加,预测的砂岩储层空间分布趋于更加客观真实。在正确选择变差函数和密井网的条件下,随机反演可以大幅度提高薄储层反演的精度。因此,具有良好薄层预测能力的随机反演将是高含水后期薄储层预测的一项关键技术。

At the late stage with high water cut,the remaining oil to be discovered is usually distributed in the thin layers of continental reservoirs.The focus and difficulty to tap the potential of remaining oil is the crosswell reservoir prediction,because the existing well pattern may be unable to effectively control the sand body boundary.Stochastic inversion technique is an effective method to predict the crosswell sand body boundary.Taking the typical area in XSG oilfield as the inversion target,we analyzed the impact of variogram parameters and inversion wells on the crosswell reservoir prediction precision by stochastic inversion.The vertical variogram is more sensitive than the horizontal variogram on the results of stochastic inversion.The predicted distribution of sandstone reservoir tends to be more objective and real along with the increase of wells participating in inversion.Stochastic inversion can greatly improve the accuracy of thin reservoir inversion under the conditions of the correct choice of variogram and dense well pattern.Therefore,stochastic inversion,with good prediction ability for thin reservoirs,will be a key technique on thin reservoir prediction in the late stage with high water cut.

开发地震; 随机反演; 储层预测; 变差函数;
development seismic,; stochastic inversion,; reservoir prediction,; variogram;

国家科技重大专项(2011ZX05010-001)资助。

10.3969/j.issn.1000-1441.2016.05.015