基于人工智能的断层自动识别研究进展

Research progress of automatic fault recognition based on artificial intelligence

随着油气勘探开发工作的进行,构造圈闭的勘探难度不断提高,断层作为油气运移、聚集的主要通道之一,断层的识别精度很大程度上影响了油气藏的勘探开发.在断层识别的发展过程中,国内外学者们提出了许多切实可行的方案.近年来,人工智能领域的兴起,使得断层自动识别方法更加多样化.本文通过调研大量的国内外相关文献,对基于人工智能的断层自动识别方法进行了归纳总结,将其划分为两个大类:智能算法、机器学习,又对每个大类进行了精细划分,并阐述了通过蚁群算法、边缘检测、BP神经网络、支持向量机、生成对抗网络、聚类及卷积神经网络来自动识别断层的基本原理、发展现状及优缺点.此外,卷积神经网络拥有卷积层、池化层等特殊结构,可以直接通过学习输入与输出之间的映射关系来实现断层自动识别,具有很好的非线性表达及泛化能力,相较于其他人工智能方法,兼具了较高的效率与精度.因此本文对基于卷积神经网络来自动识别断层的一些关键技术及优化算法进行了重点介绍,且利用大量三维合成地震记录及断层标签作为训练样本,实现了基于卷积神经网络的三维地震资料断层自动识别并对最终结果进行了深入分析.最后对断层自动识别技术的提高,给出展望与相应的结论.

With the progress of oil and gas exploration and development, the exploration difficulty of structural traps continues to increase. Faults are one of the main channels for oil and gas migration and accumulation. The accuracy of fault identification greatly affects the exploration and development of oil and gas reservoirs. In the development process of fault recognition, scholars at home and abroad have put forward many practical solutions. In recent years, the rise of Artificial Intelligence(AI) has made automatic fault recognition methods more diverse. This paper summarizes the automatic fault recognition methods based on AI by investigating a large number of relevant domestic and foreign literatures, and divides them into two major categories: intelligent algorithms and machine learning, and the basic principles, development status, advantages and disadvantages of various methods to automatically identify fault through ant colony algorithm, edge detection, Back Propagation(BP) neural network, Support Vector Machine(SVM), Generative Adversarial Network(GAN), clustering and Convolutional Neural Network(CNN) are introduced. In addition, the CNN has special structures such as convolutional layer and pooling layer. It can directly realize automatic fault recognition by learning the mapping relationship between input and output, and has good nonlinear expression and generalization in fault recognition tasks. Compared with other AI methods, it has both higher efficiency and accuracy. Therefore, this paper focuses on some key technologies and optimization algorithms for automatic fault identification based on CNN, and uses a large number of 3 D synthetic seismic records and fault tags as training samples to realize 3 D seismic data fault automation based on CNN and analyzes the final result in depth. Finally, prospects and corresponding conclusions are given for improvement of fault recognition technology.