[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-122378-en":3,"doc-seo-122378-105":30,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":27,"seo_description":14,"update_tm":28,"read_time":29},122378,962075114765,"Quinn","https://ap-avatar.wpscdn.com/davatar_a8503ba1806abce46bf441b54a3ca4cd",8,"Research & Report","MWD Data-Based Rock Mass Classification Using Machine Learning Techniques","Rock mass quality classification is critical in tunnel design and construction, yet conventional empirical methods rely on subjective assessments. This study presents an objective machine learning framework using Measurement While Drilling (MWD) sensor data from a Norwegian Drill and Blast tunnelling project. Large MWD datasets were filtered and normalized into ML-ready metrics, with rock mass signatures aggregated per grout hole and mapped to corresponding Q-values. Class imbalance was addressed using SMOTE, and logistic regression and random forest models were trained and evaluated using accuracy, precision, recall, and F1-score, with random forest delivering superior performance.","MWD Data-Based Rock Mass Classification Using Machine Learning Techniques  \nT. B. Katuwal; A. Bruland & K. K. Panthi  \nNorwegian University of Science and Technology (NTNU), Trondheim, Norway [tek.b.katuwal@ntnu.no](tek.b.katuwal@ntnu.no)  \nA. H. Høien  \nNorwegian Public Roads Administration, Oslo, Norway  \nAbstract  \nThe classification of rock mass quality plays an important role in tunnel design and construction, and it is often a challenging task. The traditional empirical approaches for identifying rock mass quality are subjective, and this study explores an objective approach using various Machine Learning (ML) techniques to classify the rock mass quality. The data used as a basis is Measurement While Drilling (MWD) data collected from a Drill and Blast tunnelling project in Norway. Automated filtering and normalization techniques were employed to transform the large number of MWD sensor data into applicable metrics for ML based data-driven technique. The rock mass signature was established by capturing the mean (average) of filtered and normalized MWD dataset in each single grout hole and aggregating them for each tunnel grout section with corresponding mapped Q-value. The imbalanced rock mass distribution condition was handled by using the Synthetic Minority Oversampling Technique (SMOTE) . After that, more than 6 000 MWD dataset were randomly split into training (80%) and testing sets (20%) . The logistic regression (LR) and random forest (RF) classification model were trained with training dataset with their optimal hyperparameters. The output of the trained model was evaluated by using the testing dataset. The prediction performance of these models was evaluated by comparing different performance metrics such as accuracy, precision, recall, and F1-score. The RF exhibits better performance compared to LR model. Eventually, the ML techniques capable to capture the surrounding rock mass quality class leveraging MWD sensor data in Drill and Blast tunnel project, thereby reducing the inherent uncertainties and subjectivity of traditional rock mass classification systems.  \nKeywords  \nTunnelling, Measurement while drilling (MWD), Rock mass quality, Machine learning (ML)  \n1 Introduction  \nThe Drill and Blast (D&B) excavation method is a widely employed tunnelling technique in the Scandinavian hard rock mass conditions (Van Eldert et al. 2020) . The methods advantages compared to other methods is e.g. high flexibility, low capital investment, short start-up times, and adaptability to varying geological conditions and tunnel cross-section requirements (Macias and Bruland 2014; Katuwal and Panthi 2024) . The efficiency and effectiveness of this construction technique are greatly  \ninfluenced by the existing geological conditions. Therefore, accurate rock mass characterization is essential for successful planning and execution of tunnel projects.  \nThe quality of the rock mass is most widely evaluated using various rock mass classification systems, including the Rock Mass Rating (RMR) introduced by Bieniawski (1973), the Q-system introduced by Barton et al. (1974), and Geological Strength Index (GSI) suggested by Hoek (1994) (Panthi 2006; Erharter et al. 2024) . These systems are based on manual registration and is therefore subjected to inconsistent and biased results due to the subjectivity of the persons that are mapping (Elmo and Stead 2021) . Consequently, traditional rock mass classification systems may have some degree of uncertainty to represent the actual rock mass conditions, leading to potential overestimation or underestimation of rock support requirements.  \nTo overcome the issue associated to the subjectivity, researchers and practitioners are trying to use data from measurement while drilling (MWD) for rock mass characterization and rock support design (Høien and Nilsen 2014; Van Eldert et al. 2020) . MWD technology captures high-resolution sensor data associated with drilling parameters such as penetration rate","cbCaimPkyUYnKToi","https://ap.wps.com/l/cbCaimPkyUYnKToi","pdf",968636,1,7,"English","en",105,"# Abstract\n# 1 Introduction\n## Rock mass classification systems and limitations\n## MWD-based rock mass characterization and related challenges\n# 2 Background of dataset\n## Project and geological context of the Løren road tunnel","[{\"question\":\"Why is rock mass quality classification important for tunnel projects?\",\"answer\":\"Rock mass quality strongly influences planning and execution. Accurate characterization helps ensure suitable design decisions such as rock support requirements and grouting.\"},{\"question\":\"How are MWD sensor data used to build the machine learning inputs?\",\"answer\":\"MWD measurements are filtered and normalized, then aggregated per grout hole. These aggregated signatures are mapped to Q-values from the Q-system to create labeled data for ML.\"},{\"question\":\"Which models are trained, and how is performance compared?\",\"answer\":\"Logistic regression and random forest classifiers are trained using an 80/20 train-test split after handling class imbalance with SMOTE. Performance is assessed with accuracy, precision, recall, and F1-score, where random forest performs better.\"}]","MWD Data-Based Rock Mass Classification Using Machine Learning Techniques | PDF",1785810318,18,{"code":4,"msg":31,"data":32},"ok",{"site_id":24,"language":23,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"mwd-data-based-rock-mass-classification-using-machine-learning-techniques","",{"@graph":36,"@context":85},[37,54,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/mwd-data-based-rock-mass-classification-using-machine-learning-techniques/122378/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":62,"encodingFormat":61,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-04",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"Why is rock mass quality classification important for tunnel projects?","Question",{"text":75,"@type":76},"Rock mass quality strongly influences planning and execution. Accurate characterization helps ensure suitable design decisions such as rock support requirements and grouting.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How are MWD sensor data used to build the machine learning inputs?",{"text":80,"@type":76},"MWD measurements are filtered and normalized, then aggregated per grout hole. These aggregated signatures are mapped to Q-values from the Q-system to create labeled data for ML.",{"name":82,"@type":73,"acceptedAnswer":83},"Which models are trained, and how is performance compared?",{"text":84,"@type":76},"Logistic regression and random forest classifiers are trained using an 80/20 train-test split after handling class imbalance with SMOTE. Performance is assessed with accuracy, precision, recall, and F1-score, where random forest performs better.","https://schema.org",{"og:url":52,"og:type":87,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":89,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,119,122,127,130,134],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":21,"doc_module":4,"doc_module_name":46,"category_name":116,"show_sort_weight":117,"slug":118},"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":120,"slug":121},30,"research-report",{"id":123,"doc_module":4,"doc_module_name":46,"category_name":124,"show_sort_weight":125,"slug":126},9,"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":46,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":46,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":46,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]