基于深度强化学习的测井曲线自动深度校正方法  

Automatic depth matching method of well log based on deep reinforcement learning

作　　者：熊文君 肖立志[1] 袁江如[2] 岳文正[1] XIONG Wenjun;XIAO Lizhi;YUAN Jiangru;YUE Wenzheng(China University of Petroleum,Beijing 102249,China;Research Institute of Petroleum Exploration&Development,PetroChina,Beijing 100083,China)

机构地区：[1]中国石油大学(北京),北京102249 [2]中国石油勘探开发研究院,北京100083

出　　处：《石油勘探与开发》2024年第3期553-564,共12页Petroleum Exploration and Development

基　　金：中国石油天然气集团有限公司-中国石油大学(北京)战略合作科技专项(ZLZX2020-03)。

摘　　要：针对传统测井曲线深度校正需要手动调整曲线,而对于多口井的深度校正工作量巨大,需要大量人工参与,且工作效率较低的问题,提出一种多智能体深度强化学习方法(MARL)来实现多条测井曲线自动深度匹配。该方法基于卷积神经网络(CNN)定义多个自上而下的双滑动窗口捕捉测井曲线上相似的特征序列,并设计一个智能体与环境的互动机制来控制深度匹配过程。通过双深度Q学习网络(DDQN)选取一个动作来平移或缩放测井特征序列,并利用反馈的奖励信号来评估每个动作的好坏,以学习到最优的控制策略达到提升深度校正精度的目的。研究表明,MARL方法可以自动完成多口井、不同测井曲线的深度校正任务,减少人工干预。在油田实例应用中,对比分析了动态时间规整(DTW)、深度Q学习网络(DQN)和DDQN等方法的测试结果,DDQN算法采用双网络评估机制有效改进了算法的性能,能够识别和对齐测井曲线特征序列上更多的细节,具有较高的深度匹配精度。In the traditional well log depth matching tasks,manual adjustments are required,which means significantly labor-intensive for multiple wells,leading to low work efficiency.This paper introduces a multi-agent deep reinforcement learning(MARL)method to automate the depth matching of multi-well logs.This method defines multiple top-down dual sliding windows based on the convolutional neural network(CNN)to extract and capture similar feature sequences on well logs,and it establishes an interaction mechanism between agents and the environment to control the depth matching process.Specifically,the agent selects an action to translate or scale the feature sequence based on the double deep Q-network(DDQN).Through the feedback of the reward signal,it evaluates the effectiveness of each action,aiming to obtain the optimal strategy and improve the accuracy of the matching task.Our experiments show that MARL can automatically perform depth matches for well-logs in multiple wells,and reduce manual intervention.In the application to the oil field,a comparative analysis of dynamic time warping(DTW),deep Q-learning network(DQN),and DDQN methods revealed that the DDQN algorithm,with its dual-network evaluation mechanism,significantly improves performance by identifying and aligning more details in the well log feature sequences,thus achieving higher depth matching accuracy.

关 键 词：人工智能 机器学习 深度校正 测井曲线 多智能体深度强化学习 卷积神经网络 双深度Q学习网络 

分 类 号：P631.4[天文地球—地质矿产勘探]