摘要
针对反演近地表物性参数的地震层析成像方法的发展历程开展了论述。首先,阐述了传统的联合代数重建法(SIRT)、最小二乘QR分解(LSQR)等求解方程组的方法以及最新发展的局部最优化算法(包括伴随状态法与改进的散射积分法)的模型更新方式;其次,从简单到复杂论述了层析成像方法在射线理论、有限频理论、波动理论下的发展历程;再次,从初至波到时频率依赖的初至波到时初至波波形反演方面阐述了反演精度逐步提升的技术关键;然后,介绍了目前对地下物性参数要求提升的情况下,地震层析成像方法从单参数速度到各向异性等多参数反演的最新研究动态;最后,分析了地震层析成像这种高度非线性的反演过程的模型参数化、反演策略、Hessian矩阵利用等问题。综上所述表明:地震层析成像在实际应用中仍具有不可替代的作用;通过不断挖掘数据中更多的信息,其理论方法与实现方式正在向全波形反演靠拢。
In this study,the development of tomographic methods is summarized,which are responsible for inverting near-surface models.In terms of model improvements,this study compares conventional methods that are used for solving equations,such as the SIRT and the LSQR,with recently developed local optimal algorithms,such as the adjoint-state and the improved scattering-integral method.In terms of theoretical advances,this study illustrates the developments of tomographic techniques;this is achieved by incorporating simple ray and finite-frequency to the complex wave-equation theory.Furthermore,this study analyzes the key factor for improving the inversion accuracy,from the first-arrival ray-based travel time and the frequency-dependent first-arrival travel time to the early-arrival waveform.Subsequently,the latest development of seismic tomography is presented from the monochromatic velocity inversion to multi-parameter inversion;this is aimed at meeting the higher requirements of advanced subsurface models.Further,model parameterization,inversion strategy,and Hessian matrix utilization of seismic tomography in a highly nonlinear inversion process are discussed.In summary,it is revealed that seismic tomography is still an irreplaceable technique in practical applications.By utilizing an increasing amount of information retrieved from seismic data,the performance of seismic tomography can be compared to that of a complete waveform inversion.
作者
刘玉柱
吴世林
刘伟刚
黄鑫泉
伍正
LIU Yuzhu;WU Shilin;LIU Weigang;HUANG Xinquan;WU Zheng(State Key Laboratory of Marine Geology,Tongji University,Shanghai 200092,China;School of Ocean and Earth Science,Tongji University,Shanghai 200092,China)
出处
《石油物探》
EI
CSCD
北大核心
2020年第1期1-11,共11页
Geophysical Prospecting For Petroleum
基金
中国科学院战略性先导科技专项(XDA14010203)
国家自然科学基金项目(41774122,41630964,41674115)共同资助。