摘要
薄层、薄互层厚度预测是储层横向预测的一重要环节。常规计算薄层厚度的方法是在时间域或频率域实现，主要使用单参数计算。本文利用对薄层厚度敏感的地震特征参数之间的非线性关系，使用神经网络算法，建立了一套计算薄层及薄互层厚度方法。通过模型的正反演结果表明：该算法对薄层厚度及薄互层累积厚度的预测均有较好的效果，具有一定的抗噪声能力。对塔中DL92─04测线部分剖面的石炭系Ⅰ油组薄互层砂岩的累积厚度进行了预测，结果令人满意。
The prediction of thin bed and thin interbed thicknesses is an important link for the reservior lateral prediction. The conventional method to calculate the thicknesses is implemented in time or frequency domains, mainly using single parameter. In the papr, based on the nonlinear relation between seismic char acteristic parameters sensitive to the thin bed and thin interbed thicknesses, we construct a set of mehtods for calculating the thicknesses by means of a neural network algorithm. Throught the forward modeling and inversion, it is shown that the method is excellent to predict the thicknesses, and have a certain antinoise ability. The method has been used to predict the cumulative thickness of thin interbeded sandstone of a oil-bearing formation of the DL82-04 line. Tge result is satisfactory.
出处
《石油物探》
EI
CSCD
北大核心
1997年第4期92-98,共7页
Geophysical Prospecting For Petroleum
关键词
薄互层
厚度
薄层
储集层
多参量法
Characteristic Paramenter, Gradational Thin Interbed, Cumulative Thickness, Neural Network
