利用地震几何属性和自组织神经网络进行地震相的自动识别

Method of Seismic Facies Auto-classification by the Seismic Geometrical Attribute and the Self-organizing Neural Network

阐述了2类地震几何属性①反映地震道空间变化量的属性;②反映地震属性空间相似程度及其衍生的各类参数。运用这些属性来判识地层或反射界面的空间连续性和倾向倾角等地质体的空间几何特征,采用自组织神经网络方法对这些属性进行自动分类来判别地震相。在山东东营凹陷永921井区的应用证明,该方法具有较好的应用效果。

This paper demonstrates lwo kinds of seismic geometrical altributes, The first one reflecls tile space variance of seismic altributes,and the second,lhe space semblance of the attributes among the adjacent seismic traces. Various parameters can be derived from them, The space geometrical characters of geol ogy such as the space continuity and dip and incline of strata or reflective inlerface can be recognized, Seismic facies can be auto classified by the seismic geometrical attribule and the method of self-organizing neu ral network. This technology can be applied to classify seismic facies in the area of Y921 well in Dongying Depression,Shandong Province, The result shows that this lechnology is effective in the applicalion.