[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-125394-en":3,"doc-seo-125394-105":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},125394,1374391974564,"Clementine","https://ap-avatar.wpscdn.com/avatar/14000253aa45c000a9e?x-image-process=image/resize,m_fixed,w_180,h_180&k=1779874745381141002",8,"Research & Report","Hybrid Machine Learning for CNC Process Monitoring","The transition to highly customized, one-off production demands advanced process monitoring to reduce waste, limit downtime, and lessen operator workload. CNC axes enable accessible monitoring through power supply data. By predicting reference signals and comparing them with real-time measurements, deviations support model-based monitoring and anomaly detection. The study evaluates hybrid ML models using features from a physical model and soft sensors for hard-to-measure quantities like forces and MRR. Tree-based models (RF, GB) are more accurate and robust than deep learning, especially with limited data, while DL improves with larger datasets but remains inferior.","Received 14 April 2025, accepted 15 May 2025, date of publication 26 May 2025, date of current version 2 June 2025. Digital Object Identifier 10.1109/ACCESS.2025.3573400  \nHybrid Machine Learning for CNC Process Monitoring  \nROBIN STRÖBEL1, SAMUEL DEUCKER 1, HANLIN ZHOU 1,  \nHAFEZ KADER2,(Graduate Student Member, IEEE), ALEXANDER PUCHTA1  \n,  \nBENJAMIN NOACK2,(Senior Member, IEEE), AND JÜRGEN FLEISCHER 1  \n1wbk Institute of Production Science, Karlsruhe Institute of Technology (KIT), 76131 Karlsruhe, Germany  \n2AMS–Autonomous Multisensor Systems, Otto von Guericke University Magdeburg, 39106 Magdeburg, Germany Corresponding author: Robin Ströbel ([robin.stroebel@kit.edu](robin.stroebel@kit.edu))  \nThis work was supported by the Federal Ministry for Economic Affairs and Climate Action (BMWK), based on a decision by German Bundestag via Gesellschaft zur Förderung angewandter Informatik e.V.—GFaI under Grant 22849 BG/2 .  \nABSTRACT The transition to highly customized, one-off production in modern manufacturing necessitates sophisticated process monitoring to reduce waste, minimize downtime, and alleviate operator burden. Computer Numerically Controlled (CNC) axes represent a fundamental component of automated manufacturing and offer a universal and accessible monitoring option through power supply data. By accurately predicting reference signals and comparing them with real-time measurements, deviations can be used for effective model-based process monitoring and anomaly detection. This study explores the efficacy of hybrid machine learning (ML) models in predicting reference signals for CNC axes using features derived from a physical model. Furthermore, relevant but difficult-to-measure features such as process forces and material removal rate (MRR) were made accessible through soft sensors. Various ML models were evaluated, including tree-based models (e.g. random forest (RF) and gradient boosting (GB)) and deep learning (DL) models (e.g. feed-forward neural networks (FNN), long short-term memory (LSTM) and transformers-based models (TF)). A feature importance analysis was performed to gain a better understanding of the influencing factors, which revealed that velocity, acceleration, process forces, spindle torque, and MRR are relevant. Tree-based models, particularly RF and GB, have been shown to be more accurate and robust than DL approaches, particularly when data is sparse and processes are complex. Although DL models improved with larger data sets, their performance remained inferior to that of tree-based methods. This study emphasizes the advantages of incorporating physical knowledge into hybrid ML models to improve model-based process monitoring.  \nINDEX TERMS Machine tool, CNC, process monitoring, signal prediction, machine learning.  \nI. INTRODUCTION  \nThe miniaturisation of sensors and enhanced data-processing capabilities enabled companies to collect and analyse production data with higher efficiency. At the same time, the growing demand for customised products is driving the need for greater flexibility in production [1], resulting in a shift towards one-off production. This trend also affects individual aspects of production, such as process monitoring. It must  \nThe associate editor coordinating the review of this manuscript and  \napproving it for publication was Szidonia Lefkovits  .  \nbecome more flexible and adaptable to produce customised products with high quality and low cost.  \nComputer numerical control (CNC) milling machines are essential for modern manufacturing, playing a pivotal role in the production of complex components with high precision. The functionality of machine tools is contingent upon the accurate and precise control of their CNC axes. Consequently, the power supply is a valuable source of data for the monitoring of machine tool operations providing real-time insight into the system’s behaviour. This provides an accessible and cost-effective option for process monitoring.  \n􀀊 2025 The Auth","cbCaitb6uMK0zF24","https://ap.wps.com/l/cbCaitb6uMK0zF24","pdf",2600678,1,14,"English","en",105,"# Introduction\n## Motivation: customized one-off production\n## Data sources: CNC power supply signals\n## Background: statistical and physical monitoring vs ML\n## Hybrid modeling approach and evaluation","[{\"question\":\"Why is process monitoring critical for CNC in customized one-off production?\",\"answer\":\"Customized production increases variability and requires flexible, accurate monitoring to reduce waste and downtime while lowering operator burden.\"},{\"question\":\"How does the method use power supply data for monitoring?\",\"answer\":\"Power supply data enables real-time insight into machine behavior. Predicted reference signals are compared with measurements so deviations can drive model-based monitoring and anomaly detection.\"},{\"question\":\"Which modeling approach performs best and under what conditions?\",\"answer\":\"Tree-based models, especially random forest and gradient boosting, show higher accuracy and robustness than deep learning when data are sparse and processes are complex; deep learning improves with larger datasets but still underperforms.\"}]","Hybrid Machine Learning for CNC Process Monitoring | PDF",1785898648,35,{"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},"hybrid-machine-learning-for-cnc-process-monitoring","",{"@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/hybrid-machine-learning-for-cnc-process-monitoring/125394/",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","2026-08-05",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 process monitoring critical for CNC in customized one-off production?","Question",{"text":75,"@type":76},"Customized production increases variability and requires flexible, accurate monitoring to reduce waste and downtime while lowering operator burden.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does the method use power supply data for monitoring?",{"text":80,"@type":76},"Power supply data enables real-time insight into machine behavior. Predicted reference signals are compared with measurements so deviations can drive model-based monitoring and anomaly detection.",{"name":82,"@type":73,"acceptedAnswer":83},"Which modeling approach performs best and under what conditions?",{"text":84,"@type":76},"Tree-based models, especially random forest and gradient boosting, show higher accuracy and robustness than deep learning when data are sparse and processes are complex; deep learning improves with larger datasets but still underperforms.","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},[93,97,101,105,110,115,120,123,128,131,135],{"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":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]