摘要
该文阐述了叠前AVA反演与储层预测软件系统的核心技术、设计实现及主要功能。该系统应用现代软件工程理论.针对大型物探软件特点.采用结构化数据管理及构件化软件架构方案设计而成,是一套面向油气储集层岩性、物性、合油气性参数预测的叠前反演及储层预测软件系统。该系统采用模拟退火叠前AVA同时反演方法、地震属性优选与融合方法、Kohonen自组织储层聚类分析与MLP多层感知器的非线性储层物性参数预测等方法,实现了储集层空间展布的精细刻画与物性参数的定量分析,因此可用来解决非均质储层预测问题。该系统已在多个研究区得到应用。预测结果与钻井结果吻合率较高。系统提供的界面开发库与基础算法库也为新技术、新方法的集成提供了基础平台。
The paper introduces the core theories, architecture and functions of pre-stack AVA inversion and reservoir prediction software system. This system is based on modem software engineering theory and was designed by structured data management and componentization software framework. It is a pre-stack inversion and reservoir prediction system used for lithology prediction, physical parameters and hydrocarbon potential prediction. By using simulated annealing pre-staek simultaneous AVA inversion, seismic attributes optimization and integration, Kohonen self-organizing reservoir cluster analysis and MLP muhilayer preceptor non-linear reservoir parameters prediction, reservoir spatial distribution characteristics can be delineated more accurately and reservoir parameters can be estimated quantitatively, therefore, the heterogeneous reservoir prediction problem was solved. This system has been applied to many work areas and achieved high consistency between prediction and actual drilling results. The interface development database and basic algorithm database also provide a basic platform for the integration of new techniques and new methods.
出处
《岩性油气藏》
CSCD
2010年第F07期163-167,共5页
Lithologic Reservoirs
关键词
模拟退火
AVA反演
储层参数预测
神经网络
属性融合
聚类分析
地学软件研发
simulated annealing
AVA inversion
reservoir parameter prediction
neural network
attributes integration
cluster analysis
geoseienees software development