薄层地震属性参数分析和厚度预测

Parameter analysis of seismic attributes and thickness prediction for thin bed

大庆石油学院, 安达 151400

Daqing Petroleum Institute, Anda 151400

薄层厚度预测一直是公认的难题之一, 其难度就在于如何准确地提取薄层的地震属性。传统的方法是利用砂厚和视振幅间的Widess原理, 即在调谐厚度之内, 视振幅与砂厚呈单调上升关系, 而波峰及波谷间的时差变化不大[1]；或是利用频率信息[2], 其基本原理是随着砂层厚度增大, 地震波主频变低。但是理论和实际资料都表明, 不同的砂厚和地层组合对地震波的动力学信息影响很大, 各种参数与砂厚都是非线性关系, 使用单一的信息不可能准确预测薄层厚度。因此我们采用了时域、频域多种参数, 利用解决非线性问题的有力武器──神经网络预测薄层厚度。另外, 我们还对资料进行了高保真、高信噪比、高分辨率处理, 取得了令人满意的效果。

The prediction of thin bed thickness is one of the difficult problems generally acknowledged. Its difficulty lies in how to exactly extract the seismic attributes of thin bed. The traditional method is to usethe Widess principle about sandstone thickness and apparent amplitude. That is to say, in the tuningthickness range, the relationship between the apparent amplitude and sandstone thickness is monotonically ascensive, while the time difference between the wave crest and trough is not seroitive to the thicknass. Or the fraqency information is used, which basic principle is that the master frequency of seismicwave will become low with the incrament of sandstone thickness. But both theoretical and practical dataindicate that different sandstone thickness and strata combinations have a great influence on the dynamicinformation of seismic wave. Each parameter depends nonlinearly on the sandstone thickness. Th thinbed thickness can not be predicted exactly by using single information. Therefore, we adopt several parameters in time and frequency domains, and use the neural network, which is a powerful tool for solvingnonlinear problems, to predict the thin bed thicknass. In addition, we carry out high-fidelity, high signal-to-noise ratio, and high-resolution processing for the data, and achieve a satistactory result.