摘要
在油气地震勘探中,速度分析是关键的地震数据处理步骤.但是在传统地震处理中,人工速度分析效率低、耗费时长且人为因素影响大.随着对油气资源的需求日益增加,人工速度分析已经不能满足当前生产的需要.目前,各种人工智能方法已经应用于地震速度分析的研究之中.本文分析了主流的人工智能速度分析方法的原理和应用效果,有普通神经网络方法,卷积神经网络方法,递归神经网络方法,卷积和递归组合网络方法,聚类机器学习方法.根据各种方法的表现,最后对人工智能速度分析方法做了总结以及进一步的展望.
Velocity analysis is the key step of seismic data processing in oil and gas seismic exploration. However, in traditional seismic processing, artificial velocity analysis is inefficient, time-consuming and greatly affected by human factors. With the increasing demand for oil and gas resources, manual speed analysis can not meet the needs of current production. At present, various artificial intelligence methods have been applied to the research of seismic velocity analysis. This paper analyzes the principle and application effect of the mainstream artificial intelligence speed analysis methods, including ordinary neural network method, Convolution Neural Network(CNN) method, Recursive Neural Network(RNN) method, convolution and recursive combined network method, clustering machine learning method. According to the performance of various methods, finally, the speed analysis methods of artificial intelligence are summarized and further prospected.
作者
彭冬冬
李振春
孙小东
王伟奇
PENG DongDong;LI ZhenChun;SUN XiaoDong;WANG WeiQi(School of Earth Science and Technology,China University of Petroleum(East China),Qingdao 266500,China;Key Laboratory of Deep Oil And Gas,Qingdao 266500,China;Functional Laboratory of Marine Mineral Resources Evaluation and Testing Technology,National Laboratory of Oceanography,Qingdao 266071,China)
出处
《地球物理学进展》
CSCD
北大核心
2022年第5期2010-2023,共14页
Progress in Geophysics
基金
塔里木盆地深层复杂高陡构造与碳酸盐岩储层地震速度建模及成像关键技术研究项目(ZD2019-183-003)资助。
关键词
人工智能
机器学习
深度学习
速度分析
卷积神经网络
递归神经网络
聚类
Artificial intelligence
Machine learning
Deeping learning
Speed analysis
Convolution Neural Network(CNN)
Recursive Neural Network(RNN)
Clustering
