叠前地震随机反演方法及实际资料应用  

Pre-stack seismic stochastic inversion and application on real data

作　　者：王小丹 印兴耀[2] 金惠[1] 谢武仁[1] 苏楠 马石玉[1] 郝翠果[1] 孙爱[1] 武赛军[1] WANG Xiao-dan;YIN Xing-yao;JIN Hui;XIE Wu-ren;SU Nan;MA Shi-yu;HAO Cui-guo;SUN Ai;WU Sai-jun(Branch of Research Institute of Petroleum Exploration and Development,PetroChina,Hebei Langfang 065007,China;Shandong China University of Petroleum (Huadong),Shandong Qmgdao 266580,China)

机构地区：[1]中国石油勘探开发研究院,河北廊坊065007 [2]中国石油大学(华东),山东青岛266580

出　　处：《地球物理学进展》2018年第6期2471-2476,共6页Progress in Geophysics

基　　金：国家重点基础研究发展计划(973)(2013CB228604);国家重大科技专项(2016ZX05007-002)联合资助

摘　　要：常规的确定性反演存在垂向分辨率较低的缺陷,而随机反演方法可以充分利用井数据来提高分辨率.相比叠后地震反演方法,叠前地震反演能够获得更多的储层信息,有利于改善储层描述的精度.结合随机反演和叠前反演两种各具优势的反演方法,本文研究了一种基于贝叶斯理论构建的叠前地震随机反演方法.在算法的实现方面,首先应用快速傅里叶滑动平均模拟(FFT-MA)算法和逐渐变形算法(GDM)来获取先验信息,以此来减少计算量并提高效率.然后,应用基于快速模拟退火(VFSA)算法改进的粒子群优化(PSO)方法(简称VFSA-PSO)来得到P波速度、S波速度、密度等弹性参数的后验分布,以此来进一步提高计算效率,而不失反演精度.模型试算和和实际资料应用验证了本文方法是一种高效的反演方法,能够得到高精度的反演结果.Considering the drawback of low vertical resolution in conventional deterministic inversion,the stochastic inversion can improve the resolution by making better use of the well information. Compared with the post-stack seismic inversion,pre-stack seismic inversion method can obtain more reservoir information,and it is benefit to improve the accuracy of reservoir characterization.In this paper,we studied the pre-stack stochastic inversion by combining the two advantageous inversion methods together,which is formulated based on the Bayesian framework.For the implementation of algorithm, the Fast Fourier Transform-Moving Average (FFT-MA)algorithm and the Gradual Deformation Method (GDM)algorithm are firstly used to acquire the priori information,which can reduce the computing costs and enhance efficiency.Then,Particle Swarm Optimization (PSO)approach based on the Very Fast Simulated Annealing (VFSA)algorithm is chosen to obtain posterior distribution of the elastic parameters,such as P-wave velocity,S-wave velocity,density, etc.,which can increase computational efficiency further without losing the inversion accuracy.We demonstrate through model tests and real data analysis that the proposed method works very fast and provides high accuracy inversion results.

关 键 词：叠前随机反演 贝叶斯理论 FFT-MA GDM VFSA-PSO 先验信息 后验分布 

分 类 号：P631[天文地球—地质矿产勘探]