基于粒子群优化算法的AVAF反演方法  

AVAF inversion method based on particle swarm optimization algorithm

作　　者：王震宇[1] 刘俊州[1] 时磊[1] 刘颖[2] 夏红敏[1] WANG Zhen-yu;LIU Jun-zhou;SHI Lei;LIU Ying;XIA Hong-min(Sinopec Petroleum Exploration and Production Research Institute,Beijing 100083,China;China University of Petroleum,Beijing 102249,China)

机构地区：[1]中国石化石油勘探开发研究院,北京100083 [2]中国石油大学(北京),北京102249

出　　处：《地球物理学进展》2019年第6期2346-2352,共7页Progress in Geophysics

基　　金：国家科技重大专项(2016ZX05002-005)资助

摘　　要：岩石物理实验和实际观测研究表明,纵波速度的频散现象通常都与地层的含气性有着密切的关系,它是纵波反射系数随频率变化所导致的.但是,传统的AVA反演方法忽略了这种速度频散现象,因此引入了误差,增加了含气预测的风险.本文我们提出了一种适用于频变反射系数和速度频散的反演方法,采用传播矩阵方程来进行正演模拟.而考虑到加入频散信息的AVAF反演问题具有高度的非线性特征,我们基于粒子群优化(PSO)算法来进行AVAF反演.经过模型与实际数据的测试证明,我们的反演方法适用于包含频散信息的地震数据,且具有一定的抗噪性,即使是在含噪数据下,也能够挖掘出数据中的纵波速度频散信息,为之后利用纵波速度的频散规律来解释储层含气性提供可靠的依据.Rock-physics experiment and field observation shown that the frequency-dependent P-wave velocity is often associated with hydrocarbon deposit,which results in P-wave reflection coefficients varying with frequency.However,this effects is often neglected in the conventional AVA/AVO(Amplitude Versus Angle or Offset)inversion method,and thus error is introduced to increases risk of detect hydrocarbon bearing formations.In this paper,we propose an inversion method that is suitable for frequency dependent reflection coefficients and velocity dispersion.We employ forward modeling based on propagator matrices that include frequency dependent elastic coefficients.Due to the AVAF inversion is a kind of high non-linear problem,we use Particle Swarm Optimization(PSO)for inversion of layer parameters.Synthetic and field data examples demonstrate the ability and usefulness of this method could digging the information that P-wave velocity dispersion at seismic frequencies,even if the seismic data is noisy.It can be utilized to favorably indicate gas reservoirs.

关 键 词：速度频散 AVAF反演 含气性解释 粒子群优化算法 

分 类 号：P631[天文地球—地质矿产勘探]