Intelligent identification method and application of seismic faults based on a balanced classification network  

基于平衡分类网络的地震断层智能识别方法及应用

作　　者：Yang Jing Ding Ren-Wei Wang Hui-Yong Lin Nian-Tian Zhao Li-Hong Zhao Shuo Zhang Yu-Jie 杨晶;丁仁伟;王惠勇;林年添;赵俐红;赵硕;张玉洁(山东省沉积成矿作用与沉积矿产重点实验室,山东科技大学地球科学与工程学院,山东青岛266590;海洋矿产资源评价与探测技术功能实验室,青岛海洋科学与技术国家实验室,山东青岛266237;中石化石油勘探开发研究院信息资料中心,北京100083)

机构地区：[1]Affiliation and detailed correspondence:College of Earth Science and Engineering,Shandong University of Science and Technology,Qingdao 266400,China [2]Laboratory of Marine Mineral Resources Evaluation and Prospecting Technology,Qingdao National Laboratory for Marine Science and Technology,Qingdao 266237,China [3]Research Institute of Petroleum Exploration and Development,Sinopec,Beijing 100083,China

出　　处：《Applied Geophysics》2022年第2期209-220,307,共13页应用地球物理（英文版）

基　　金：supported by the Natural Science Foundation of Shandong Province(ZR202103050722).

摘　　要：This study combined fault identification with a deep learning algorithm and applied a convolutional neural network(CNN)design based on an improved balanced crossentropy(BCE)loss function to address the low accuracy in the intelligent identification of seismic faults and the slow training speed of convolutional neural networks caused by unbalanced training sample sets.The network structure and optimal hyperparameters were determined by extracting feature maps layer by layer and by analyzing the results of seismic feature extraction.The BCE loss function was used to add the parameter which is the ratio of nonfaults to the total sample sets,thereby changing the loss function to find the reference of the minimum weight parameter and adjusting the ratio of fault to nonfault data.The method overcame the unbalanced number of sample sets and improved the iteration speed.After a brief training,the accuracy could reach more than 95%,and gradient descent was evident.The proposed method was applied to fault identification in an oilfield area.The trained model can predict faults clearly,and the prediction results are basically consistent with an actual case,verifying the effectiveness and adaptability of the method.针对训练样本集不均衡造成的地震断层智能识别精度不高,卷积神经网络(CNN)训练速度慢的问题,本文将断层识别与深度学习算法相结合,设计了一种基于改进的平衡交叉熵(BCE)损失函数的CNN的地震断层智能识别方法。通过自编算法逐层提取特征图,分析地震特征提取结果,从而确定网络结构与最优参数,进而修改CNN以优化模型。利用BCE损失函数,添加非断层与总样本集的比率参数,从而改变损失函数寻找最小权重参数的基准,调优断层与非断层的数据比例。该方法克服了样本集类别数量不均衡的问题,提高了迭代速度,经过少量的训练即可达到95%以上的精度,梯度下降明显。将本文方法应用于某油田地区的断层识别,所训练的模型预测断层较为清晰,预测结果与实际情况基本吻合,因此该方案具有有效性和适应性。

关 键 词：convolutional neural network seismic fault identification balanced cross-entropy loss function feature map 

分 类 号：P631.4[天文地球—地质矿产勘探]