用改进的共轭梯度算法解波阻抗反演问题  

ACOUSTIC IMPEDANCE INVERSION USING IMPROVED CONJUGATE GRADIENT ALGORITHM

作　　者：张赛民[1] 周竹生[1] 刘海飞[1] 

机构地区：[1]中南大学信息物理工程学院,长沙410083

出　　处：《物探化探计算技术》2009年第4期338-343,共6页Computing Techniques For Geophysical and Geochemical Exploration

基　　金：国家自然科学基金(40804027)

摘　　要：研究快速有效的波阻抗反演算法,是地震反演中一直追求的目标。这里从目标函数出发,推导的反演方程,在常规共轭梯度算法的基础上,对算法搜索的共轭梯度方向提出了具体的改进措施,并在阻尼项及松弛因子的给定上实现了自适应化。另外,在整个算法过程中实现了向量运算,无需额外的内存。通过合成记录及实测资料试算对比发现,该算法具有精度高,运算速度快及抗病态能力强的特点。A fast and effective algorithm of acoustic impedance inversion is always our goal in seismic inversion.Inversion equation is established from the objective function.A new conjugate gradient algorithm has been put forward to improve impedance inversion,which is based on classical conjugate gradient algorithm and made improvements in conjugate gradient direction,adapted automatically in damping and relaxation constant.In addition,whole recursion is finished by vector operation,which greatly saved computer's RAM.The inversion result of synthesizing traces and measured seismic section shows that the new method has the characteristic of higher precision,rapider speed and stronger stability.

关 键 词：波阻抗反演 共轭梯度算法 随机逆 自适应算法 

分 类 号：P631.4[天文地球—地质矿产勘探]