地球物理学进展 ›› 2019, Vol. 34 ›› Issue (5): 1917-1923.doi: 10.6038/pg2019CC0429

• 应用地球物理学Ⅰ(油气及金属矿产地球物理勘探) • 上一篇    下一篇

特征值相干理论诠释及效果比较

王静1,张军华1,*(),王延光2,杨勇3,冯德永2,武刚3,黄德峰1   

  1. 1. 中国石油大学(华东)地球科学与技术学院,山东青岛 266580
    2. 中石化胜利油田股份公司物探研究院,山东东营 257022
    3. 中石化胜利油田股份公司勘探开发研究院,山东东营 257015
  • 收稿日期:2018-12-03 修回日期:2019-06-05 出版日期:2019-10-28 发布日期:2019-10-28
  • 通讯作者: 张军华 E-mail:zjh_upc@163.com
  • 作者简介:王静,女,1985年生,在读博士,主要从事地震资料解释研究.(E-mail: r_wj@163.com)
  • 基金资助:
    国家科技重大专项(2017ZX05009-001);国家科技重大专项(2016ZX05006-002);国家科技重大专项(2016ZX05011-002)

Annotation and effect comparison of eigenvalue coherence

WANG Jing1,ZHANG Jun-hua1,*(),WANG Yan-guang2,YANG Yong3,FENG De-yong2,WU Gang3,HUANG De-feng1   

  1. 1. School of Geosciences, China University of Petroleum (East China), Shandong Qingdao 266580, China
    2. Geophysical Research Institute of Shengli Oil Field Branch, Sinopec, Shandong Dongying 257022, China
    3. Shengli Oilfield Exploration and Development Research Institute of Sinopec, Shandong Dongying 257015, China;
  • Received:2018-12-03 Revised:2019-06-05 Online:2019-10-28 Published:2019-10-28
  • Contact: Jun-hua ZHANG E-mail:zjh_upc@163.com

摘要:

相干体技术是目前应用最为广泛的地震解释技术之一,已由第一代互相关算法(C1)、第二代多道相似算法(C2)发展到第三代多道特征值相干算法(C3).常规的基于特征值的第三代相干(C31)定义为第一特征值与所有特征值之和的比.当采用最主要的3道算法时,也可以用Randen等(C32)、Bakker(C33)、Donias等(C34)、Wu(C35)给出的相干公式进行计算.本文比较了这5种表征方法,模型测试和实际应用结果表明:(1)C31、C34、C35都可以较好地表征相干,当储层较平时C35有不错的抗噪性,对于复杂断块储层,C31即俗称的第三代相干应是首选;(2)储层简单时C32特征值计算结果为负值,储层复杂时其特征值有正值也有负值,物理含义不明确,应用效果也不好;(3)C33计算的是不相干属性,抗噪性一般.在特征值相干的计算过程中,数据道计算窗长度的选取很重要,对于精细勘探而言,拟根据目标体的大小,在1/2波长到3/2波长时窗范围内,选取不同大小时窗,进行分级对比研究.本文的研究成果对于广大解释人员如何用好相干体这一实用技术,有一定的指导意义和借鉴作用.

关键词: 第三代相干, 特征值, 抗噪性, 三道算法, Qdome模型, 断层

Abstract:

Coherence is one of the most widely used seismic interpretation techniques, which has been developed from the first generation cross correlation algorithm (C1), the second generation multi-channel semblance-based algorithm (C2) to the third generation multi-channel eigenvalue coherence algorithm (C3). The conventional third generation coherence (C31) based on eigenvalues is defined as the ratio of the largest eigenvalue to the sum of all eigenvalues. When three traces algorithm is adopted, the formulas given by Randen et al. (C32), Bakker (C33), Donias et al. (C34) and Wu (C35) can also be used for representing the coherence. In this paper, the five characterization methods are used and compared, and the results of model test and practical application show that: (1) C31, C34 and C35 can better characterize coherence. C35 can denoise better when the reservoir is relatively flat and C31 commonly known as the third-generation coherence should be the first choice for complex fault block reservoirs. (2) The calculating eigenvalues of C32 are all negative when the reservoir is simple, and positive or negative when the reservoir is complex. The physical meaning of C32 is not clear and the application effect is not good. (3) The calculated result of C33 is non-coherence attribute, and its anti-noise ability is not so good. For eigenvalue coherence, it is of great importance to select the length of calculation window. For fine seismic exploration, time windows of different size should be selected in the window range from 1/2 wavelength to 3/2 wavelength according to the size of the target body to conduct a graded comparative study. The research results of this paper are of the general guiding significance and referential effect for explainers how to make good use of coherence technique.

Key words: The third generation coherence, Eigenvalue, Anti-noise property, Algorithm of 3 traces, Qdome model, Fault

中图分类号: