首页|基于α-shape和多边形偏置的断层自动提取方法

基于α-shape和多边形偏置的断层自动提取方法

扫码查看
断层解释在储层预测和地质建模中至关重要,目前采用人工方式解释断层工作量大、效率低,不能满足油田勘探开发的精细要求.为此,以传统相干属性或断层智能预测数据体分析、优选为基础,提出一种基于α-shape和多边形偏置的断层自动提取方法.首先通过人机交互确定目标断层位置;其次研发并利用α-shape断点边缘提取算法和倾角约束多边形偏置算法,实现对断层空间边界的精确刻画;最后以该边界为约束条件设置目标断层重合度阈值,约束区域生长算法的延展范围,完成三维断层自动提取.所提方法在胜利油田多个地区应用均取得较好效果,在征6地区刻画断层空间展布形态,断层完整性及精确度明显优于某商业软件;在辛50地区开展断层自动提取,大幅提高了工作效率.
Automatic fault extraction based on α-shape and polygon offset
Fault interpretation is crucial in reservoir prediction and geological modeling.Currently,fault inter-pretation is mostly conducted manually,with a large workload and low efficiency,which fails to meet the pre-cise requirements of oilfield exploration and development.Therefore,based on the analysis and optimization of the traditional coherent attributes or intelligent fault prediction data volumes,this paper proposes an automatic fault extraction method based on α shape and polygon offset.Firstly,it determines the location of the target fault through human-computer interaction.Secondly,it develops an α-shape breakpoint edge extraction algo-rithm and an inclination constrained polygon offset algorithm to accurately depict the spatial boundary of faults.Finally,by using the spatial boundary as a constraint condition and setting a target fault coincidence threshold,the extension range of the region growth algorithm is constrained to complete the automatic extraction of 3D faults.The application of the proposed method has achieved great results in many areas of Shengli Oilfield.The spatial distribution pattern of faults is depicted in Zheng 6 area,and the fault integrity and accuracy are sig-nificantly better than a certain commercial software.The automatic extraction of faults has greatly improved work efficiency in Xin 50 area.

breakpoint edge extractioninclination constraintpolygon offset algorithmfault coincidence threshold

颜世翠、王树华、张娟

展开 >

中国石化胜利油田分公司勘探开发研究院,山东东营 257015

断点边缘提取 倾角约束 多边形偏置算法 断层重合度阈值

中国石化科技攻关项目&&

P20071-1P22161

2024

石油地球物理勘探
东方地球物理勘探有限责任公司

石油地球物理勘探

CSTPCD北大核心
影响因子:1.766
ISSN:1000-7210
年,卷(期):2024.59(1)
  • 21