地震属性优化及物性参数反演的一种非线性处理方法

A non-linear modlling method for seise data optimization and forecasting physics property parameter

常规地震资料解释方法误差较大,越来越不能满足现代地震勘探解释的需要.本文在对地震属性进行模式分类的基础上,利用LLE算法对地震属性进行非线性降维,然后利用小波BP网络对物性参数进行非线性反演,建立了一个具有自适应、复杂非线性的地震属性储层反演模型.以某工区为例进行计算,并对其精度进行检验,获得明显优于常规方法的地质效果,具有一定的实用价值.

The error of the method of general quake data explaination is very big,which can＇t satisfy the needs of the modern quake data explaination. This article bring forward a non-linear method of quake data explaination based on mode recognision, which carry through non-linear lower dimension（LLE） based on mode recognision, and forecast physics property parameter with non-linear method（WaveNet）, which is complicated nonlinearities and self organized, the reservoir charactors are forecasted,and the precision has been tested. The geographic effecieny obviously superior to traditional explanation methods is obtained.