加权MPFI方法及其在三维连片处理中的应用  

Weighted MPFI method and its application in 3D joint processing

作　　者：徐兴荣[1] 苏勤[1] 王劲松 寇龙江[1] 文一华 王靖[1] XU Xingrong;SU Qin;WANG Jinsong;KOU Longjiang;WEN Yihua;WANG Jing(PetroChina Research Institute of Petroleum Exploration & Development-Northwest,Lanzhou 730020,China;Research Institute of Exploration and Development,PetroChina Tuha Oilfield Company,Hami 839009,Xinjiang,China)

机构地区：[1]中国石油勘探开发研究院西北分院,兰州730020 [2]中国石油吐哈油田分公司勘探开发研究院,新疆哈密839009

出　　处：《岩性油气藏》2019年第1期122-129,共8页Lithologic Reservoirs

基　　金：国家重点研发计划项目"面向E级计算的能源勘探高性能应用软件系统与示范"(编号:2017YFB0202905);中国石油天然气集团有限公司科技管理部项目"基于地震物理模拟实验的深层-超深层弱信号增强技术研究"(编号:2017-5307073-000008-01)联合资助

摘　　要：在对采集参数差异较大的多块三维区块开展连片处理时,传统的基于MPFI,ALFT,Low-Rank及MWNI等理论的规则化方法要么无法有效克服空间假频,要么无法满足其线性假设条件,不能真正实现保真规则化处理。针对这一问题,在传统MPFI算法的基础上,将不含假频的低频信息谱作为先验信息进行加权,实现了加权MPFI数据插值,并建立了规则化方法。实际测试结果表明,该方法与传统MPFI方法相比,一方面可以有效克服空间假频,实现多块三维区块的连片规则化处理;另一方面,该方法并未增加额外的地质信息,只是在一定程度上改变了观测系统的空间采样属性,具有比传统规则化方法更高的保真性。通过数据正演模型试算及野外地震数据实算,验证了该方法的准确性、有效性和对野外地震资料的适应性。Conventional regularization methods based on the theories of MPFI,ALFT,Low-Rank and MWNI either cannot effectively overcome spatial aliasing,or cannot satisfy the linear assumption,therefore,they are not effective to achieve fidelity regularization in three-dimensional multi-block joint processing with different acquisition parameters. On the basis of conventional MPFI algorithm,the low-frequency spectrum without false frequency was weighted as prior information,the weighted MPFI data interpolation was realized,and the regularization method was established. The results show that compared with the conventional MPFI method,this method can effectively overcome the spatial aliasing and realize three-dimensional multi-block regularization joint processing. Moreover,it does not add additional geological information,but improves the spatial sampling properties of the observation system to some extent,and has a higher fidelity than the conventional regularizationmethod. The accuracy,validity and adaptability to field seismic data of this method were verified by trial calculation of forward model and field seismic data.

关 键 词：匹配追踪 傅里叶变换 数据插值 规则化 加权MPFI 

分 类 号：P631.4[天文地球—地质矿产勘探]