基于角度域共成像点道集的微分相似优化法偏移速度分析  

Differential semblance optimization migration velocity analysis based on angle domain CIGs

作　　者：任芳 李振春[1] 张敏[1] 杨国权[1] 陈飞旭 

机构地区：[1]中国石油大学(华东)地球科学与技术学院,山东青岛266580

出　　处：《物探与化探》2017年第5期872-880,共9页Geophysical and Geochemical Exploration

基　　金：山东省自然科学基金(ZR2013DL012);高等学校博士学科点专项科研基金(20120133120012);国家自然科学基金(41504100)

摘　　要：微分相似优化(Differential Semblance Optimization,DSO)法是一种可以有效避免非线性最小二乘反演过程中不聚焦问题的速度反演方法。基于DSO原理的波动方程偏移速度分析将成像道集的聚焦或者拉平程度作为速度判定的准则,其目标泛函具有良好的凸性,可以有效避免局部极值问题,目标泛函的梯度较为光滑,并且对地震数据中低频成分的要求较低,对速度模型中背景速度的估计较为准确。一般情况下,DSO方法利用偏移距域共成像点道集建立目标泛函,文中利用角度域共成像点道集建立目标泛函对速度的准确性进行判定。模型试算结果表明,通过角道集可以更加直观地对速度的准确性进行判定,并且准确反映速度与深度的对应关系。Differential semblance optimization （DSO） is an approach to the inversion of the velocity which avoids the severe convergence associated with nonlinear least-squares inversion.DSO method-based wave-equation migration velocity analysis measures the focusing or flatness of image gathers and uses the deviation as the criterion to update the velocity.The DSO objective function has a decent global convexity property, therefore it can avoid the local minima problem, the gradient of objective function is smooth, it can inverse the background velocity accurately when there lacks low frequency data.Generally, ODCIGs is used to construct the objective function.In this paper,the authors use ADCIGs to construct the objective function,and the result of the test shows that DSO method with ADCIGs can estimate the correctness of velocity model directly and reflect the coupling relation between velocity and depth accurately.

关 键 词：偏移速度分析 微分相似优化法 共成像点道集 目标泛函 梯度 

分 类 号：P631.4[天文地球—地质矿产勘探]