西南石油大学学报(自然科学版)

• 地质勘探 • 上一篇    下一篇

概率法反演技术研究

范廷恩1,2 *,黄旭日1,胡光义2,王宗俊2,尹成1   

  1. 1. 西南石油大学地球科学与技术学院,四川成都610500
    2. 中海油研究总院,北京朝阳100028
  • 出版日期:2016-06-01 发布日期:2016-06-01
  • 通讯作者: 范廷恩,E-mail:fante@cnooc.com.cn
  • 基金资助:

    国家油气重大专项(2011ZX05024 001 005)。

A Study on Probability Inversion Technology

FAN Tingen1,2*, HUANG Xuri1, HU Guangyi2, WANG Zongjun2, Yin Cheng1   

  1. 1. School of Geoscience and Technology,Southwest Petroleum University,Chengdu,SiChuan 610500,China
    2. CNOOC Research Institute,Chaoyang,Beijing 100028,China
  • Online:2016-06-01 Published:2016-06-01

摘要:

针对常规反演技术(如叠后约束稀疏脉冲反演等)垂向分辨能力难以满足开发需求的问题,开展了概率法反演
技术的研究。研究中引入贝叶斯后验概率公式,将地震、测井、地质的不确定性转换为后验概率分布,利用蒙特卡洛算
法估算随机样本期望,并将随机样本作为空间储层解。该方法充分结合了井上分辨率高和地震横向趋势表征能力强
的优点,有效提高了地震反演的分辨率,降低了地震反演的多解性,使储层的不确定性分析更符合地质规律和地质认
识。实际资料测试表明,新方法在砂体叠置关系认识、砂体边界再认识、井网部署和井位优化等方面取得了很好的应
用效果。

关键词: 精细地层格架, 概率法反演, 不确定性描述, 井网部署, 井位优化

Abstract:

The limited vertical resolution of post-stack constraint sparse pulse inversion method has difficulty meeting the
needs of development. However,by using the probability inversion technique,we get the inversion results by introducing
Bayesian posterior probability to put the uncertainty of seismic,well logs and geology information into probability,and then
we draw random samples from the posterior probability distribution through Markov Chain,and finally make the optimal
expectation obtained by Monte Carlo method as the inversion results. The technology can satisfy the needs of oil development
as it combines the advantages of the high resolution of well with good representation ability of seismic lateral tendency,and can
effectively improve the seismic resolution. The multi-solution of the seismic inversion can be effectively reduced by realizing
lots of equal probability results. The uncertainty analysis of reservoir description is more suitable to geological regularity and
geological understanding when the probability thought is considered. During the research in meandering river sedimentary in
some offshore oilfield,this method has obtained good application effect in sand body superimposed relationship recognition,
sand body boundary recognition,well pattern deployment,and well location optimization.

Key words: fine stratigraphic framework, probability inversion, uncertain description, well pattern deployment, well location
optimization

中图分类号: