密井网条件下地质统计学反演初始模型的构建及其对反演结果的影响——以大庆长垣油田密井间开发区为例

在断层附近区域,不加入断层信息的简化初始模型对地震反演结果的预测精度差,不能准确刻画砂体的分布规律。以大庆长垣油田密井网开发区为例,利用井震结合的方法构建了高精度的层位模型和构造模型,建立加入准确构造信息的反演初始模型,并在该初始模型约束下进行地质统计学反演,对比了不同初始模型约束下的地震反演结果及精度。研究结果表明,加入断层信息的初始模型能够有效地提高地质统计学反演的砂岩预测精度,厚度为3 m以上砂岩的预测符合率可提高约7%,厚度为1~3 m砂岩的预测符合率可提高约11%,厚度为1 m以下砂岩的预测符合率可提高约12%。

Inversion of initial model construction by geological statistics and its influence on inversion results under dense well pattern

The simplified initial model without adding fault information often makes the prediction accuracy of seismic inversion result worse near faults and then the distribution of sand bodies cannot be described accurately.A case study of the development zone with dense well pattern in Daqing Changyuan oilfield,a high precision layer model and structure model was constructed by integrating well log and seismic information,and an initial model of inversion with accurate structure information were constructed.Under the initial model,the geological statistics inversion was carried out,the accuracy of seismic inversion under different starting models was compared.The results show that the initial model with adding fault information can effectively improve the prediction accuracy of sandstone prediction,which is based on the results of the geological statistics inversion.The pre-diction accuracy of sandstone with thickness greater than 3m can be increased by about 7%.The prediction accuracy of sandstone with the thickness range from1 m to 3m can be increased by about 11%.The prediction accuracy of sandstone with the thickness less than 1m can be increased by about12%.