机器学习算法在岩性识别上的应用对比研究  

Comparative study on the application of machine learning algorithms in lithology identification

作　　者：徐海寒 秦昊 张辉[2] 陈晓 XU Haihan;QIN Hao;ZHANG Hui;CHEN Xiao(Shanghai Jiao Tong University;Defense Engineering Institute,AMS,PLA)

机构地区：[1]上海交通大学 [2]军事科学院国防工程研究院

出　　处：《防护工程》2024年第6期54-61,共8页Protective Engineering

摘　　要：基于算法分析随钻测量数据与岩性分析是数字钻探技术的重要研究方向之一,采用便携式电驱动数字钻进设备进行随钻测量,收集随钻测量参数与岩石物理力学参数数据后运用岩性识别中常见的10种机器学习算法进行对比训练。通过对比筛选出表现最优的5种算法,对其进行精确地调参优化。结果显示,随机森林算法在岩性识别上的准确率高达95%,显著提升了识别的精确度和效率。The analysis of while-drilling measurement data and lithology analysis based on algorithms is one of the important research directions in digital drilling technology.Using portable electric-driven digital drilling equipment for while-drilling measurement,this paper collects while-drilling measurement parameters and rock physical and mechanical property data.Ten common machine learning algorithms for lithology identification are employed for comparative training.Through this comparison,the top five algorithms with the best performance are selected for precise parameter tuning and optimization.The results indicate that the random forest algorithm achieves an accuracy rate of up to 95%in lithology identification,significantly enhancing the precision and efficiency of the identification process.

关 键 词：数字钻进设备 随钻测量 岩性识别 机器学习算法 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程] P634[自动化与计算机技术—控制科学与工程]