{"title":"机器学习辅助物探记录的岩相分类技术现状","authors":"Bappa Mukherjee, Sohan Kar, Kalachand Sain","doi":"10.1007/s00024-024-03563-4","DOIUrl":null,"url":null,"abstract":"<div><p>In the E&P industry, accurate lithology classification is an essential task for successful exploration and production. Geophysical logs provide high-resolution petrophysical properties, but core logging is expensive and traditional techniques may not accurately classify lithologies. We demonstrated a comparative analysis of six ML algorithms: k-nearest neighbor (kNN), support vector machine (SVM), decision tree (DT), random forest (RF), extreme gradient boosting (XGBoost) and artificial neural network (ANN) for the prediction of lithologies from geophysical logs. Here we analysed the wireline logs of eight wells associated with the petroliferous Lakadong-Therria formation of the Bhogpara oil field of the Assam-Arakan Basin. This formation contains eight typical lithologies, namely clay stone, sand stone, calcareous sandstone, shale, calcareous shale, carbonaceous shale, coal and limestone. Performance of the ML algorithms were evaluated through accuracy, precision, recall, F1-score and receiver operating characteristic (ROC) curve. During the training and test phases, the computed overall accuracy of the predicted ML modes exceeded 82% and 71%, respectively. The model accuracy hierarchy was ANN > XGBoost > RF > SVM > DT > kNN during training, and ANN/XGBoost > kNN > DT/RF > SVM during testing. This approach allows interpreters to select the most accurate ML model based on training phase performance. This study provided a clear insight towards generating a supplement for litholog sequence and improving the accuracy and efficiency of lithology prediction in a geologically complex petroleum reservoir using pre-received core derived litholog information at few wells.</p></div>","PeriodicalId":21078,"journal":{"name":"pure and applied geophysics","volume":null,"pages":null},"PeriodicalIF":1.9000,"publicationDate":"2024-08-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Machine Learning Assisted State-of-the-Art-of Petrographic Classification From Geophysical Logs\",\"authors\":\"Bappa Mukherjee, Sohan Kar, Kalachand Sain\",\"doi\":\"10.1007/s00024-024-03563-4\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<div><p>In the E&P industry, accurate lithology classification is an essential task for successful exploration and production. Geophysical logs provide high-resolution petrophysical properties, but core logging is expensive and traditional techniques may not accurately classify lithologies. We demonstrated a comparative analysis of six ML algorithms: k-nearest neighbor (kNN), support vector machine (SVM), decision tree (DT), random forest (RF), extreme gradient boosting (XGBoost) and artificial neural network (ANN) for the prediction of lithologies from geophysical logs. Here we analysed the wireline logs of eight wells associated with the petroliferous Lakadong-Therria formation of the Bhogpara oil field of the Assam-Arakan Basin. This formation contains eight typical lithologies, namely clay stone, sand stone, calcareous sandstone, shale, calcareous shale, carbonaceous shale, coal and limestone. Performance of the ML algorithms were evaluated through accuracy, precision, recall, F1-score and receiver operating characteristic (ROC) curve. During the training and test phases, the computed overall accuracy of the predicted ML modes exceeded 82% and 71%, respectively. The model accuracy hierarchy was ANN > XGBoost > RF > SVM > DT > kNN during training, and ANN/XGBoost > kNN > DT/RF > SVM during testing. This approach allows interpreters to select the most accurate ML model based on training phase performance. This study provided a clear insight towards generating a supplement for litholog sequence and improving the accuracy and efficiency of lithology prediction in a geologically complex petroleum reservoir using pre-received core derived litholog information at few wells.</p></div>\",\"PeriodicalId\":21078,\"journal\":{\"name\":\"pure and applied geophysics\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":1.9000,\"publicationDate\":\"2024-08-31\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"pure and applied geophysics\",\"FirstCategoryId\":\"89\",\"ListUrlMain\":\"https://link.springer.com/article/10.1007/s00024-024-03563-4\",\"RegionNum\":4,\"RegionCategory\":\"地球科学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q2\",\"JCRName\":\"GEOCHEMISTRY & GEOPHYSICS\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"pure and applied geophysics","FirstCategoryId":"89","ListUrlMain":"https://link.springer.com/article/10.1007/s00024-024-03563-4","RegionNum":4,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"GEOCHEMISTRY & GEOPHYSICS","Score":null,"Total":0}
引用次数: 0
摘要
在勘探和开发行业,准确的岩性分类是成功勘探和生产的一项基本任务。地球物理测井可提供高分辨率的岩石物理特性,但岩心测井费用昂贵,而且传统技术可能无法对岩性进行准确分类。我们展示了六种 ML 算法的比较分析:k-近邻(kNN)、支持向量机(SVM)、决策树(DT)、随机森林(RF)、极梯度提升(XGBoost)和人工神经网络(ANN),用于从地球物理测井记录预测岩性。在此,我们分析了与阿萨姆-阿拉干盆地博格帕拉油田含油层拉卡东-特里亚地层相关的八口油井的有线测井记录。该地层包含八种典型岩性,即粘土岩、砂岩、钙质砂岩、页岩、钙质页岩、碳质页岩、煤和石灰岩。通过准确度、精确度、召回率、F1-分数和接收者操作特征曲线(ROC)对 ML 算法的性能进行了评估。在训练和测试阶段,计算得出的 ML 模式预测总体准确率分别超过 82% 和 71%。在训练阶段,模型准确率等级为 ANN > XGBoost > RF > SVM > DT > kNN;在测试阶段,模型准确率等级为 ANN/XGBoost > kNN > DT/RF > SVM。这种方法允许解释人员根据训练阶段的表现选择最准确的 ML 模型。这项研究提供了一个清晰的视角,有助于为岩性序列提供补充,并利用在少数油井中预先接收的岩心衍生岩性信息,提高地质复杂的石油储层中岩性预测的准确性和效率。
Machine Learning Assisted State-of-the-Art-of Petrographic Classification From Geophysical Logs
In the E&P industry, accurate lithology classification is an essential task for successful exploration and production. Geophysical logs provide high-resolution petrophysical properties, but core logging is expensive and traditional techniques may not accurately classify lithologies. We demonstrated a comparative analysis of six ML algorithms: k-nearest neighbor (kNN), support vector machine (SVM), decision tree (DT), random forest (RF), extreme gradient boosting (XGBoost) and artificial neural network (ANN) for the prediction of lithologies from geophysical logs. Here we analysed the wireline logs of eight wells associated with the petroliferous Lakadong-Therria formation of the Bhogpara oil field of the Assam-Arakan Basin. This formation contains eight typical lithologies, namely clay stone, sand stone, calcareous sandstone, shale, calcareous shale, carbonaceous shale, coal and limestone. Performance of the ML algorithms were evaluated through accuracy, precision, recall, F1-score and receiver operating characteristic (ROC) curve. During the training and test phases, the computed overall accuracy of the predicted ML modes exceeded 82% and 71%, respectively. The model accuracy hierarchy was ANN > XGBoost > RF > SVM > DT > kNN during training, and ANN/XGBoost > kNN > DT/RF > SVM during testing. This approach allows interpreters to select the most accurate ML model based on training phase performance. This study provided a clear insight towards generating a supplement for litholog sequence and improving the accuracy and efficiency of lithology prediction in a geologically complex petroleum reservoir using pre-received core derived litholog information at few wells.
期刊介绍:
pure and applied geophysics (pageoph), a continuation of the journal "Geofisica pura e applicata", publishes original scientific contributions in the fields of solid Earth, atmospheric and oceanic sciences. Regular and special issues feature thought-provoking reports on active areas of current research and state-of-the-art surveys.
Long running journal, founded in 1939 as Geofisica pura e applicata
Publishes peer-reviewed original scientific contributions and state-of-the-art surveys in solid earth and atmospheric sciences
Features thought-provoking reports on active areas of current research and is a major source for publications on tsunami research
Coverage extends to research topics in oceanic sciences
See Instructions for Authors on the right hand side.