论文部分内容阅读
本文应用多元统计方法，利用地震属性并辅以钻井、测井资料对某地区某层段砂岩的厚度和孔隙率进行了横向预测。在多元统计方法中，主要应用了有效数计算、逐步回归分析和Krig插值等方法。有效数可以定量判定哪些地震属性能进行储层预测，从而筛选出地震属性；逐步回归分析是利用已筛选后的地震属性计算储层参数，用它可以提高预测效率和精度，但计算结果仍与建模井点处的真值存在误差；再利用剩余误差的结构性进行Krig插值，可使误差进一步减小。用本方法对陆相薄互层的计算结果表明，其精度可以达到勘探甚至开发阶段储层描述的要求。
In this paper, the multivariate statistical method is used to predict the thickness and porosity of sandstone in a certain interval in a certain area by using seismic attributes and complemented by drilling and logging data. In multivariate statistical methods, the main methods of effective number calculation, stepwise regression analysis and Krig interpolation methods. The effective number can quantitatively determine which seismic attributes can predict the reservoir, so that the seismic attributes can be screened out. Stepwise regression analysis uses the filtered seismic attributes to calculate reservoir parameters, which can improve the prediction efficiency and accuracy. However, There is error in the true value at the modeling well point. Krig interpolation can be used to make the error further reduced by using the structure of residual error. The calculation results of interbedded continental facies by this method show that the precision can meet the requirements of reservoir description during exploration and even development.