[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-117518-en":3,"doc-seo-117518-105":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":4,"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},117518,687197207057,"Sage","https://ap-avatar.wpscdn.com/davatar_29158cc5080c5b710cf443261637dec0",8,"Research & Report","Machine learning-based prediction of species mass fraction and flame characteristics in partially premixed turbulent jet flame","This study leverages machine learning (ML) together with large eddy simulation (LES) to predict species mass fractions and key flame characteristics in partially premixed turbulent jet flames. High-fidelity LES simulations using a flamelet-based chemistry technique generate a benchmark dataset for surrogate training. Three ML surrogates (NN, Linear Regression, Decision Tree Regression) are compared, and the NN model reaches extremely high accuracy (R² > 0.9998) with very low MAE across species. Temperature and progress variable dominate sensitivity, while uncertainty is highest for OH in turbulent flame fronts. The NN surrogate supports targeted optimization using multiple fitness functions and reduces total computing time by 17.25× versus LES.","1 Machine learning-based prediction of species mass fraction and flame  \n2 characteristics in partially premixed turbulent jet flame  \n3 Amirali Shateri, Zhiyin Yang, Jianfei Xie *  \n4 School of Engineering, University of Derby, DE22 3AW, UK  \n5 Abstract  \n6 This study uses machine learning (ML) and large eddy simulation (LES) to predict the mass  \n7 fractions of species and flame characteristics in partially premixed turbulent jet flames. A  \n8 flamelet-based chemistry technique was used to perform high-fidelity LES simulations of  \n9 Sandia Flame D. The resulting dataset was used to train three ML models—Neural Networks  \n10 (NN), Linear Regression (LR), and Decision Tree Regression (DTR)—for surrogate prediction.  \n11 Among them, the NN model achieved the highest accuracy, with R-squared (R²) values  \n12 exceeding 0.9998 and Mean Absolute Error (MAE) values below 1.0×10⁻⁴ across all species.  \n13 Sensitivity analysis identified temperature and progress variable as dominant input features.  \n14 Uncertainty quantification confirmed high model confidence in stable regions, while elevated  \n15 uncertainty was observed for the hydroxyl radical due to its short-lived, highly reactive nature  \n16 in turbulent flame fronts. The NN surrogate was also used for targeted optimisation, enabling  \n17 to find ten combustion states with species compositions that were consistent with experimental  \n18 data within the reported range of uncertainty. Four fitness functions—Euclidean distance, 19 Manhattan distance, Collinearity coefficient, and Amplitude correlation coefficient—were  \n20 applied to guide the optimisation process. Manhattan distance consistently demonstrated the  \n21 lowest absolute errors for key species such as CH₄(0.0014), OH(0.0002), and O₂(0.0072), 22 indicating its superior accuracy and compatibility with the LES benchmark data. Additionally, 23 the ML surrogate achieved a 17.25× reduction in total computing time compared to LES solver, 24 promising efficient parametric exploration and rapid predictive capability. These findings  \n25 demonstrate the potential of ML-based surrogates to support real-time combustion diagnostics, 26 optimization, and design.  \n27 Keywords: Partially premixed combustion; Turbulent jet flow; Flame pattern; Species mass  \n28 fraction; Machine learning  \n29 1. Introduction  \n30 Combustion is a fundamental process with wide-ranging applications in various industrial  \n31 sectors. Accurate modelling and prediction of turbulent combustion systems are essential for  \n32 optimizing performance, improving efficiency, and reducing emissions [1] . Combustion  \n33 processes, particularly those involving turbulent flows, are inherently complex due to the  \n34 interplay of fluid dynamics, chemical reactions, and heat transfer. The accurate prediction and  \n35 modelling of such processes are crucial for the design and optimization of combustion systems, 36 which are pivotal in energy production, propulsion, and manufacturing industries. This  \n37 complexity is further amplified in the context of partially premixed flames, such as those found  \n38 in methane/air combustion systems, where the non-homogeneous mixture of fuel and oxidizer  \n39 introduces additional variability in flame behaviour and emissions [2] . Recent advancements  \n40 in computational power and numerical methods have enabled significant progress in the field  \n41 of turbulent combustion modelling. However, the development of models that can accurately  \n42 predict species mass fraction, flame characteristics, and emissions in turbulent jet flows with  \n43 partially premixed methane/air flames remains a challenging endeavour. This is due to the  \n44 intricate coupling between the turbulence and chemical kinetics, which requires detailed  \n45 representation in computational models to achieve predictive accuracy [3-6] . Flamelet  \n46 combustion is characterized by partially mixed reactants, resulting in a flame structure with  \n47 distinct reg","cbCaim7NkjXixZuU","https://ap.wps.com/l/cbCaim7NkjXixZuU","pdf",5661477,1,48,"English","en",105,"# Abstract\n## Introduction\n## Flamelet combustion and modelling assumptions\n## Large eddy simulation (LES) in turbulent combustion\n## Related work and limitations","[{\"question\":\"What data and simulation approach does the study use to train machine learning models?\",\"answer\":\"High-fidelity LES simulations are performed using a flamelet-based chemistry technique, producing a benchmark dataset. This dataset is then used to train surrogate ML models to predict species mass fractions and flame characteristics.\"},{\"question\":\"Which ML model performs best and what accuracy level is reported?\",\"answer\":\"The neural network (NN) surrogate achieves the highest accuracy, with R-squared values exceeding 0.9998 and mean absolute error (MAE) below 1.0×10⁻⁴ across all species.\"},{\"question\":\"How does uncertainty behave across species, and what drives sensitivity in the models?\",\"answer\":\"Sensitivity analysis identifies temperature and the progress variable as dominant input features. Uncertainty quantification shows high confidence in stable regions, while elevated uncertainty appears for the hydroxyl radical (OH) due to its short-lived, highly reactive behavior at turbulent flame fronts.\"}]","Machine learning-based prediction of species mass fraction and flame characteristics in partially premixed turbulent jet flame | PDF",1785676568,121,{"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},"machine-learning-based-prediction-of-species-mass-fraction-and-flame-characteristics-in-partially-premixed-turbulent-jet-flame","",{"@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/machine-learning-based-prediction-of-species-mass-fraction-and-flame-characteristics-in-partially-premixed-turbulent-jet-flame/117518/",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","2026-08-02",true,{"@type":65,"interactionType":66,"userInteractionCount":4},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What data and simulation approach does the study use to train machine learning models?","Question",{"text":75,"@type":76},"High-fidelity LES simulations are performed using a flamelet-based chemistry technique, producing a benchmark dataset. This dataset is then used to train surrogate ML models to predict species mass fractions and flame characteristics.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"Which ML model performs best and what accuracy level is reported?",{"text":80,"@type":76},"The neural network (NN) surrogate achieves the highest accuracy, with R-squared values exceeding 0.9998 and mean absolute error (MAE) below 1.0×10⁻⁴ across all species.",{"name":82,"@type":73,"acceptedAnswer":83},"How does uncertainty behave across species, and what drives sensitivity in the models?",{"text":84,"@type":76},"Sensitivity analysis identifies temperature and the progress variable as dominant input features. Uncertainty quantification shows high confidence in stable regions, while elevated uncertainty appears for the hydroxyl radical (OH) due to its short-lived, highly reactive behavior at turbulent flame fronts.","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"]