基于GRNN振幅谱估计的井控提高地震分辨率技术

提高分辨率一直是地震勘探必不可少的处理过程,当前简单油气藏逐渐减少,薄层、薄互层等复杂地质体已成为油气藏勘探开发的主要目标,对勘探精度的要求也越来越高。传统提高分辨率的方法主要依据地震剖面信息,这种处理方法往往比较盲目,缺少判断依据,而井控地震处理技术能将井资料信息运用到地震勘探处理中。提出了一种与广义回归神经网络(GRNN)相结合,利用井资料的介入提高地震资料分辨率的方法。由于广义回归神经网络具有较强的自适应学习逼近能力,可将其作为修整和拓展地震数据频谱的手段,以井作为约束条件,提高地震数据的分辨率。模型测试和实际数据处理表明基于GRNN振幅谱估计的井控提高地震分辨率技术是有效可行的。

Well-controlled seismic resolution enhancement technology based on GRNN amplitude spectrum estimation

Resolution enhancement is always an essential process in seismic exploration. Complex geological bodies, such as thin layer and thin interbed, have become the main target of reservoir exploration with the decrease of simple reservoir, and exploration precision is required more higher. The traditional methods of resolution enhancement are mainly based on seismic profile information, tending to be blind and lack of judgment criterion. However, when applying well-controlled seismic processing technology, well data can be used for seismic exploration. This study proposes a method which introduces well information to improve seismic data resolution in combination with generalized regression neural network (GRNN). With strong self-adaptive learning and approaching ability, GRNN can be taken as a means to modify and expand seismic data spectrum, so as to improve the resolution of seismic data under the constraint of well. Model test and actual data processing indicate that the well-controlled seismic resolution enhancement technology based on GRNN amplitude spectrum estimation is effective and feasible.