分方位马尔科夫链叠前地质统计反演在Q区块碎屑岩薄储层预测中的应用

Q区块目的层为砂泥岩薄互层,常规方法预测的储层精度不能满足油田时下的需求。通过调研,认为薄互层具有各向异性,“两宽一高”地震数据的出现,为各向异性解释提供了数据基础。通过理论分析,采用分方位马尔科夫链地质统计反演预测储层展布规律。该方法将马尔科夫链转移概率与贝叶斯遗传概率相结合,利用地震波组特征进行约束,同时,方位信息的加入为储层预测精度的提高提供了选项。突破了以往只在碳酸盐岩及基岩各向异性反演的认知限制,使碎屑岩薄储层中的各向异性得以体现。通过理论与实践结合,证实了优势方位的储层预测精度优于其他方位。

Application of pre-stack geostatistical inversion of sub-azimuth Markov chain in block Q of an oilfield

The target layer of block Q in an oilfield is a thin interbed layer of sandstone and mudstone,where the reservoir accuracy predicted by the conventional method cannot meet the current requirement of the oilfield.Our investigation suggests that this thin interbed has anisotropy,and the appearance of“two widths and one high”seismic data provides a data basis for anisotropic interpretation.Through theoretical analysis,the geostatistical inversion of the sub-azimuth Markov chain is used to predict the distribution law of the reservoir.This method combines Markov chain transition probability with Bayesian genetic probability and constrains it with seismic wave group features.The addition of azimuth information provides an option for improving reservoir prediction accuracy.It breaks through the cognitive limitation of anisotropic inversion only in carbonate rock and bedrock,which makes the anisotropy be expressed in the thin reservoir of clastic rock.Combination of theory and practice proves that the reservoir prediction accuracy of the dominant azimuth is better than other azimuths.