Pub Date : 2024-08-01DOI: 10.1016/j.proci.2024.105652
C. Russo, A. Ciajolo, M.M. Oliano, B. Apicella, M. Sirignano
The paper reports a detailed study on carbon particulate matter (PM) sampled in ethylene flames stabilized on a burner-stabilized stagnation (BSS) system, aiming to give more insights on the characteristics of particles produced in this peculiar flame configuration. The study employs various diagnostic tools to analyze PM collected on the stagnation plate of flames at a constant equivalence ratio (Φ =2.07) and different flame temperatures obtained by varying the cold gas flow velocity. The carbon network of PM was analyzed by Raman and UV–Visible spectroscopy verifying the strong temperature effect on the nanostructure. The FTIR analysis allowed to quantitatively follow the temperature effect on the aromatic and aliphatic CH bonds, also evaluating the H/C atomic ratio that was found to be rather high (ranging from 0.3 to 0.5) initially decreasing and finally re-increasing as the flame temperature rises. The initial hydrogen loss with the rise of temperature was due to the loss of aromatic hydrogen, followed at higher temperature by the relevant enrichment of hydrogen bonded to aliphatic carbon. This observation is in contradiction with the expectation that higher flame temperatures would lead to an enhanced dehydrogenation of carbon particles, thereby reducing also aliphatic hydrogen. It was suggested that the enrichment in aliphatic hydrogen could be due to the small size of particles having higher radical character and surface area. Indeed, the peculiar features of such carbon particles deserve further work for understanding soot formation and growth and the relevance of BSS carbon material for optical and electronic applications.
{"title":"Deepening the knowledge of carbon particulate matter features in the BSS flame configuration","authors":"C. Russo, A. Ciajolo, M.M. Oliano, B. Apicella, M. Sirignano","doi":"10.1016/j.proci.2024.105652","DOIUrl":"https://doi.org/10.1016/j.proci.2024.105652","url":null,"abstract":"The paper reports a detailed study on carbon particulate matter (PM) sampled in ethylene flames stabilized on a burner-stabilized stagnation (BSS) system, aiming to give more insights on the characteristics of particles produced in this peculiar flame configuration. The study employs various diagnostic tools to analyze PM collected on the stagnation plate of flames at a constant equivalence ratio (Φ =2.07) and different flame temperatures obtained by varying the cold gas flow velocity. The carbon network of PM was analyzed by Raman and UV–Visible spectroscopy verifying the strong temperature effect on the nanostructure. The FTIR analysis allowed to quantitatively follow the temperature effect on the aromatic and aliphatic CH bonds, also evaluating the H/C atomic ratio that was found to be rather high (ranging from 0.3 to 0.5) initially decreasing and finally re-increasing as the flame temperature rises. The initial hydrogen loss with the rise of temperature was due to the loss of aromatic hydrogen, followed at higher temperature by the relevant enrichment of hydrogen bonded to aliphatic carbon. This observation is in contradiction with the expectation that higher flame temperatures would lead to an enhanced dehydrogenation of carbon particles, thereby reducing also aliphatic hydrogen. It was suggested that the enrichment in aliphatic hydrogen could be due to the small size of particles having higher radical character and surface area. Indeed, the peculiar features of such carbon particles deserve further work for understanding soot formation and growth and the relevance of BSS carbon material for optical and electronic applications.","PeriodicalId":408,"journal":{"name":"Proceedings of the Combustion Institute","volume":"34 1","pages":""},"PeriodicalIF":3.4,"publicationDate":"2024-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141886574","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-01DOI: 10.1016/j.proci.2024.105655
Tianjiao Li, Zhichao Hu, Weijie Yan, Chun Lou, Dong Liu, Li Sun, Huaichun Zhou
This study introduces flame image processing techniques to extract both the temperature and radiation parameters in the furnace. Additionally, a two-dimensional rectangular furnace system is established with emitting and reflecting walls and emitting and scattering spatial media. The radiation imaging model, developed through the distributions of ratios of energy scattered or reflected method, establishes a quantitative functional relationship between monochromatic radiation intensity images of the flame at two wavelengths and internal furnace temperature and radiation parameters. The Tikhonov regularization algorithm is used to reconstruct the radiation source terms within the furnace. An optimization algorithm is used to reconstruct the temperature and radiation parameters within the furnace, assuming uniform absorption and scattering coefficients. Despite the non-uniform distribution of internal radiation parameters, reconstructing the furnace temperature distribution using uniform radiation parameters remains feasible. The maximum relative error in temperature reconstruction is 2.28 %, which meets industrial temperature measurement requirements. Moreover, experimental studies are conducted on a coal-fired boiler to simultaneously detect both furnace cross-sectional temperature and radiation parameters. A single detector is used to obtain data sequentially from eight observation ports. During this process, flame images are captured under stable boiler operating conditions. These data are used to reconstruct the cross-sectional temperature distribution and radiation parameters in the burnout air zone of the boiler under different load conditions. Experimental results indicate that as the boiler load increases from 147 to 159 MW, the furnace temperature, absorption coefficient, and scattering coefficient all increase. Notably, the flame imaging processing method serves as a reliable method for monitoring the cross-sectional temperature field and radiation parameters in the large coal-fired boilers and is crucial for obtaining the data required for numerical simulations of combustion in large furnaces.
{"title":"In situ measurement of cross-section temperature field of pulverized coal boiler based on solving radiative transfer equation using a single image sensor","authors":"Tianjiao Li, Zhichao Hu, Weijie Yan, Chun Lou, Dong Liu, Li Sun, Huaichun Zhou","doi":"10.1016/j.proci.2024.105655","DOIUrl":"https://doi.org/10.1016/j.proci.2024.105655","url":null,"abstract":"This study introduces flame image processing techniques to extract both the temperature and radiation parameters in the furnace. Additionally, a two-dimensional rectangular furnace system is established with emitting and reflecting walls and emitting and scattering spatial media. The radiation imaging model, developed through the distributions of ratios of energy scattered or reflected method, establishes a quantitative functional relationship between monochromatic radiation intensity images of the flame at two wavelengths and internal furnace temperature and radiation parameters. The Tikhonov regularization algorithm is used to reconstruct the radiation source terms within the furnace. An optimization algorithm is used to reconstruct the temperature and radiation parameters within the furnace, assuming uniform absorption and scattering coefficients. Despite the non-uniform distribution of internal radiation parameters, reconstructing the furnace temperature distribution using uniform radiation parameters remains feasible. The maximum relative error in temperature reconstruction is 2.28 %, which meets industrial temperature measurement requirements. Moreover, experimental studies are conducted on a coal-fired boiler to simultaneously detect both furnace cross-sectional temperature and radiation parameters. A single detector is used to obtain data sequentially from eight observation ports. During this process, flame images are captured under stable boiler operating conditions. These data are used to reconstruct the cross-sectional temperature distribution and radiation parameters in the burnout air zone of the boiler under different load conditions. Experimental results indicate that as the boiler load increases from 147 to 159 MW, the furnace temperature, absorption coefficient, and scattering coefficient all increase. Notably, the flame imaging processing method serves as a reliable method for monitoring the cross-sectional temperature field and radiation parameters in the large coal-fired boilers and is crucial for obtaining the data required for numerical simulations of combustion in large furnaces.","PeriodicalId":408,"journal":{"name":"Proceedings of the Combustion Institute","volume":"220 1","pages":""},"PeriodicalIF":3.4,"publicationDate":"2024-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141886575","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-01DOI: 10.1016/j.proci.2024.105646
Lei Liu, Kexin Li, Hanzi Liu, Zhiqiang Sun
Thermochemical redox reactions in a bubbling-fluidized bed reactor involve the surface→grain→particle→reactor scales from the microscope to the macroscope, and the reaction contains some physical and chemical steps. There is a requirement to develop a comprehensive and precise rate equation for the redox processes. This work established a density functional theory (DFT)-based multi-scale model to simulate the kinetic behaviors of the thermochemical reactions. The model was applied to predict the oxidation and reduction kinetics of a manganese oxygen carrier in chemical looping. The reaction mechanisms of the manganese oxygen carrier with H and O were firstly calculated by the DFT methods at the surface scale, both showing two-step reaction paths with the rate-limiting step energy barrier of 0.96 eV and 0.63 eV respectively. The reaction rate constants were 0.19/Pa/s for the oxidation and 3.50 × 10/Pa/s for the reduction at 900 °C, obtained by the transition state theory (TST). The DFT and TST results were introduced to establish a microkinetic rate equation at the grain scale, which realizes the coupling of the surface reactions and the O anion diffusion in the grain bulk. The rate equation was implemented in the mass transfer models, and the influences of the gas diffusion at the particle scale and the reactor scale were further considered, including the internal, external and interphase gas diffusion. The theoretical prediction results are validated by the experimental data from the micro-fluidized bed thermogravimetric analysis. It is demonstrated that the DFT-based model could realize an accurate prediction of the reaction kinetics of the manganese oxygen carrier in bubbling-fluidized bed reactor at a wide range of reaction temperatures and gas partial pressures. The developed DFT-based rate equation solves the theoretical problem of scale-span phenomenon for the thermochemical redox reactions, i.e. oxidation and reduction steps of an oxygen carrier in chemical looping.
鼓泡流化床反应器中的热化学氧化还原反应涉及从微观到宏观的表面→颗粒→反应器尺度，反应包含一些物理和化学步骤。这就需要为氧化还原过程建立一个全面而精确的速率方程。这项研究建立了一个基于密度泛函理论（DFT）的多尺度模型来模拟热化学反应的动力学行为。该模型被用于预测锰氧载体在化学循环中的氧化和还原动力学。首先在表面尺度上用 DFT 方法计算了锰氧载体与 H 和 O 的反应机理，两者均呈现两步反应路径，限速步能障分别为 0.96 eV 和 0.63 eV。通过过渡态理论（TST）计算得出，在 900 °C 下，氧化反应的速率常数为 0.19/Pa/s，还原反应的速率常数为 3.50 × 10/Pa/s。通过引入 DFT 和 TST 结果，建立了晶粒尺度的微动力学速率方程，该方程实现了表面反应与 O 阴离子在晶粒体中扩散的耦合。在传质模型中实现了该速率方程，并进一步考虑了颗粒尺度和反应器尺度上气体扩散的影响，包括内部、外部和相间气体扩散。微流床热重分析的实验数据验证了理论预测结果。结果表明，基于 DFT 的模型可以在较宽的反应温度和气体分压范围内准确预测锰氧载体在鼓泡流化床反应器中的反应动力学。所建立的基于 DFT 的速率方程解决了热化学氧化还原反应（即氧载体在化学循环中的氧化和还原步骤）的尺度跨度现象理论问题。
{"title":"DFT-based rate equation for thermochemical redox kinetics in a bubbling-fluidized bed reactor and its application to a manganese oxygen carrier in chemical looping","authors":"Lei Liu, Kexin Li, Hanzi Liu, Zhiqiang Sun","doi":"10.1016/j.proci.2024.105646","DOIUrl":"https://doi.org/10.1016/j.proci.2024.105646","url":null,"abstract":"Thermochemical redox reactions in a bubbling-fluidized bed reactor involve the surface→grain→particle→reactor scales from the microscope to the macroscope, and the reaction contains some physical and chemical steps. There is a requirement to develop a comprehensive and precise rate equation for the redox processes. This work established a density functional theory (DFT)-based multi-scale model to simulate the kinetic behaviors of the thermochemical reactions. The model was applied to predict the oxidation and reduction kinetics of a manganese oxygen carrier in chemical looping. The reaction mechanisms of the manganese oxygen carrier with H and O were firstly calculated by the DFT methods at the surface scale, both showing two-step reaction paths with the rate-limiting step energy barrier of 0.96 eV and 0.63 eV respectively. The reaction rate constants were 0.19/Pa/s for the oxidation and 3.50 × 10/Pa/s for the reduction at 900 °C, obtained by the transition state theory (TST). The DFT and TST results were introduced to establish a microkinetic rate equation at the grain scale, which realizes the coupling of the surface reactions and the O anion diffusion in the grain bulk. The rate equation was implemented in the mass transfer models, and the influences of the gas diffusion at the particle scale and the reactor scale were further considered, including the internal, external and interphase gas diffusion. The theoretical prediction results are validated by the experimental data from the micro-fluidized bed thermogravimetric analysis. It is demonstrated that the DFT-based model could realize an accurate prediction of the reaction kinetics of the manganese oxygen carrier in bubbling-fluidized bed reactor at a wide range of reaction temperatures and gas partial pressures. The developed DFT-based rate equation solves the theoretical problem of scale-span phenomenon for the thermochemical redox reactions, i.e. oxidation and reduction steps of an oxygen carrier in chemical looping.","PeriodicalId":408,"journal":{"name":"Proceedings of the Combustion Institute","volume":"33 1","pages":""},"PeriodicalIF":3.4,"publicationDate":"2024-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141886578","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-07-31DOI: 10.1016/j.proci.2024.105676
Yu Yang, Shu Zheng, Mingxin Xu, Bing Liu, Shaohua Zhu, Ran Sui, Qiang Lu
Simultaneous blending of hydrogen (H) and ammonia (NH) to hydrocarbon fuels can tackle the safety issues of H and improve burning efficiency of NH. While this strategy brings challenges for soot prediction due to the promotion effect of H and the suppression effect of NH, and the interactions between H and NH. In this study, the simultaneous addition of NH and H on soot formation was experimentally and numerically investigated in a co-flow diffusion CH flame. The interactions between NH and H, and how they impacted different soot formation processes were comprehensively revealed using a detailed soot sectional model. The decrease of peak SVF in CH flame caused by NH was 0.013 ppm, about 31.6 % smaller than that in CH/H flame (0.019 ppm), indicating that the inhibitive effect of NH on soot formation was promoted by H. The existence of H promoted the suppression effect of NH on soot nucleation, condensation and HACA processes in the CH flame. Compared with CH/NH flame, the pyrolysis rates of NH, NH and NH in the CH/NH/H flame were higher since more H and OH radicals were generated via H decomposition. This led to a larger consumption rate of H and OH radicals, which decreased the reaction rates of CH+OH=CH+HO and CH+OH=CH+HO, and promoted the combination of NO and CH. Both factors accounted for a stronger suppression effect of NH on the formation of A1 in CH/H flame than that in CH flame, and thus a stronger inhibitive effect on soot inception and condensation. Compared with the CH flame, NH resulted in a larger decline of H and OH radicals mole fractions in the CH/H flame, which explained the stronger suppression effect of NH on the HACA surface growth process in the CH/H flame.
在碳氢化合物燃料中同时掺入氢气（H）和氨气（NH）可解决 H 的安全问题并提高 NH 的燃烧效率。但由于氢气的促进作用和氨气的抑制作用，以及氢气和氨气之间的相互作用，这种策略给烟尘预测带来了挑战。本研究在同流扩散 CH 火焰中对同时添加 NH 和 H 对烟尘形成的影响进行了实验和数值研究。通过详细的烟尘断面模型，全面揭示了 NH 和 H 之间的相互作用以及它们如何影响不同的烟尘形成过程。在CH火焰中，NH导致的SVF峰值下降为0.013 ppm，比CH/H火焰中的SVF峰值（0.019 ppm）小约31.6%，表明H促进了NH对烟尘形成的抑制作用。与 CH/NH 火焰相比，CH/NH/H 火焰中 NH、NH 和 NH 的热分解率更高，因为 H 分解产生了更多的 H 和 OH 自由基。这导致 H 和 OH 自由基的消耗率增大，从而降低了 CH+OH=CH+HO 和 CH+OH=CH+HO 的反应速率，并促进了 NO 和 CH 的结合。这两个因素导致 NH 在 CH/H 火焰中对 A1 形成的抑制作用比在 CH 火焰中更强，从而对烟尘的萌发和凝结有更强的抑制作用。与 CH 火焰相比，NH 导致 CH/H 火焰中 H 和 OH 自由基摩尔分数的下降幅度更大，这解释了 NH 对 CH/H 火焰中 HACA 表面生长过程的抑制作用更强。
{"title":"Effect of simultaneous H2 and NH3 addition on soot formation in co-flow diffusion CH4 flame","authors":"Yu Yang, Shu Zheng, Mingxin Xu, Bing Liu, Shaohua Zhu, Ran Sui, Qiang Lu","doi":"10.1016/j.proci.2024.105676","DOIUrl":"https://doi.org/10.1016/j.proci.2024.105676","url":null,"abstract":"Simultaneous blending of hydrogen (H) and ammonia (NH) to hydrocarbon fuels can tackle the safety issues of H and improve burning efficiency of NH. While this strategy brings challenges for soot prediction due to the promotion effect of H and the suppression effect of NH, and the interactions between H and NH. In this study, the simultaneous addition of NH and H on soot formation was experimentally and numerically investigated in a co-flow diffusion CH flame. The interactions between NH and H, and how they impacted different soot formation processes were comprehensively revealed using a detailed soot sectional model. The decrease of peak SVF in CH flame caused by NH was 0.013 ppm, about 31.6 % smaller than that in CH/H flame (0.019 ppm), indicating that the inhibitive effect of NH on soot formation was promoted by H. The existence of H promoted the suppression effect of NH on soot nucleation, condensation and HACA processes in the CH flame. Compared with CH/NH flame, the pyrolysis rates of NH, NH and NH in the CH/NH/H flame were higher since more H and OH radicals were generated via H decomposition. This led to a larger consumption rate of H and OH radicals, which decreased the reaction rates of CH+OH=CH+HO and CH+OH=CH+HO, and promoted the combination of NO and CH. Both factors accounted for a stronger suppression effect of NH on the formation of A1 in CH/H flame than that in CH flame, and thus a stronger inhibitive effect on soot inception and condensation. Compared with the CH flame, NH resulted in a larger decline of H and OH radicals mole fractions in the CH/H flame, which explained the stronger suppression effect of NH on the HACA surface growth process in the CH/H flame.","PeriodicalId":408,"journal":{"name":"Proceedings of the Combustion Institute","volume":"51 1","pages":""},"PeriodicalIF":3.4,"publicationDate":"2024-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141886424","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-07-31DOI: 10.1016/j.proci.2024.105513
Constantinos Moularas, Philip Demokritou, Georgios A. Kelesidis
The light absorption dynamics of brown carbon (BrC) particles emitted during combustion or pyrolysis of pinewood are elucidated here using an integrated thermal incineration platform which enables pyrolysis of wood at controlled conditions. This platform is coupled with a variety of real-time aerosol instrumentation and time-integrated sampling systems. The BrC particles emitted from pinewood combustion contain about 80 % of condensed volatile organic compounds (VOCs), regardless of the O concentration, [O]. Removing the condensed VOCs by thermal denuding reveals that BrC nanoparticles from wood pyrolysis ([O] = 0 vol%) have a bi-modal size distribution containing 95 % of nanoscale particles with a mean mobility diameter, = 37 nm and 5 % of large particles with mean = 107 nm. Increasing [O] from 0 to 20 vol%, increases the fraction of large BrC nanoparticles up to 29 % and decreases their mean to 78 nm. In this regard, the average mass absorption cross-section, , of BrC increases from 0.1 to 0.27 m/g with increasing [O]. This indicates that the light absorption of BrC from wood combustion and pyrolysis is determined by the fraction of large particles with mean = 78–107 nm. The BrC measured here can be interfaced with global climate models to estimate the contribution of particulate emissions from biomass combustors and wildfires to global warming.
{"title":"Light absorption dynamics of brown carbon particles during wood combustion and pyrolysis","authors":"Constantinos Moularas, Philip Demokritou, Georgios A. Kelesidis","doi":"10.1016/j.proci.2024.105513","DOIUrl":"https://doi.org/10.1016/j.proci.2024.105513","url":null,"abstract":"The light absorption dynamics of brown carbon (BrC) particles emitted during combustion or pyrolysis of pinewood are elucidated here using an integrated thermal incineration platform which enables pyrolysis of wood at controlled conditions. This platform is coupled with a variety of real-time aerosol instrumentation and time-integrated sampling systems. The BrC particles emitted from pinewood combustion contain about 80 % of condensed volatile organic compounds (VOCs), regardless of the O concentration, [O]. Removing the condensed VOCs by thermal denuding reveals that BrC nanoparticles from wood pyrolysis ([O] = 0 vol%) have a bi-modal size distribution containing 95 % of nanoscale particles with a mean mobility diameter, = 37 nm and 5 % of large particles with mean = 107 nm. Increasing [O] from 0 to 20 vol%, increases the fraction of large BrC nanoparticles up to 29 % and decreases their mean to 78 nm. In this regard, the average mass absorption cross-section, , of BrC increases from 0.1 to 0.27 m/g with increasing [O]. This indicates that the light absorption of BrC from wood combustion and pyrolysis is determined by the fraction of large particles with mean = 78–107 nm. The BrC measured here can be interfaced with global climate models to estimate the contribution of particulate emissions from biomass combustors and wildfires to global warming.","PeriodicalId":408,"journal":{"name":"Proceedings of the Combustion Institute","volume":"75 1","pages":""},"PeriodicalIF":3.4,"publicationDate":"2024-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141886427","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-07-31DOI: 10.1016/j.proci.2024.105636
Michael Geuking, Pavan Prakash Duvvuri, Agnes Jocher
Reversible dimerization of coronene is implemented into a hybrid method of moment based soot model and used with a newly generated reduced mechanism for ethylene combustion. The employed semi-automated mechanism reduction approach includes a novel error function using curve matching of species mole fraction, temperature, and heat release profiles of a counterflow diffusion flame, which was added to a path flux analysis with subsequent sensitivity analysis. The generated reduced mechanism, which maintains predictability of selected higher aromatics, while drastically reducing required computational resources, was validated for species concentration of lower hydrocarbons and aromatics for laminar premixed and counterflow diffusion ethylene flames. It was then used to model reversible dimerization of coronene and to predict soot volume fraction for several laminar premixed flames. For the analyzed cases, the combination of the newly reduced mechanism with the enhanced soot model, including reversible dimerization, was able to enhance the prediction of soot concentration trends. Finally, a discussion on uncertainties related to the equilibrium constant for dimerization is presented.
{"title":"Modeling reversible soot nucleation with a reduced kinetic mechanism including coronene","authors":"Michael Geuking, Pavan Prakash Duvvuri, Agnes Jocher","doi":"10.1016/j.proci.2024.105636","DOIUrl":"https://doi.org/10.1016/j.proci.2024.105636","url":null,"abstract":"Reversible dimerization of coronene is implemented into a hybrid method of moment based soot model and used with a newly generated reduced mechanism for ethylene combustion. The employed semi-automated mechanism reduction approach includes a novel error function using curve matching of species mole fraction, temperature, and heat release profiles of a counterflow diffusion flame, which was added to a path flux analysis with subsequent sensitivity analysis. The generated reduced mechanism, which maintains predictability of selected higher aromatics, while drastically reducing required computational resources, was validated for species concentration of lower hydrocarbons and aromatics for laminar premixed and counterflow diffusion ethylene flames. It was then used to model reversible dimerization of coronene and to predict soot volume fraction for several laminar premixed flames. For the analyzed cases, the combination of the newly reduced mechanism with the enhanced soot model, including reversible dimerization, was able to enhance the prediction of soot concentration trends. Finally, a discussion on uncertainties related to the equilibrium constant for dimerization is presented.","PeriodicalId":408,"journal":{"name":"Proceedings of the Combustion Institute","volume":"82 1","pages":""},"PeriodicalIF":3.4,"publicationDate":"2024-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141886429","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-07-31DOI: 10.1016/j.proci.2024.105512
Runze Mao, Min Zhang, Yingrui Wang, Han Li, Jiayang Xu, Xinyu Dong, Yan Zhang, Zhi X. Chen
Recent progress in machine learning (ML) and high-performance computing (HPC) have brought potentially game-changing opportunities in accelerating reactive flow simulations. In this study, we introduce an open-source computational fluid dynamics (CFD) framework that integrates the strengths of ML and graphics processing unit (GPU) to demonstrate their combined capability. Within this framework, all computational operations are solely executed on GPU, including ML-accelerated chemistry integration, fully-implicit solving of fluid transport PDEs, and computation of thermal and transport properties, thereby eliminating the CPU–GPU memory copy overhead. Optimisations both within the kernel functions and during the kernel launch process are conducted to enhance computational performance. Strategies such as static data reorganisation and dynamic data allocation are adopted to reduce the GPU memory footprint. The computational performance is evaluated in two turbulent flame benchmarks using quasi-DNS and LES modelling, respectively. Remarkably, while maintaining a similar level of accuracy to the conventional CPU/implicit ODE-based solver, the GPU/ML-accelerated approach shows an overall speedup of over two orders of magnitude for both cases. This result highlights that high-fidelity turbulent combustion simulation with finite-rate chemistry that requires normally hundreds of CPUs can now be performed on portable devices such as laptops with a medium-end GPU.
机器学习（ML）和高性能计算（HPC）的最新进展为加速反应流模拟带来了可能改变游戏规则的机会。在本研究中，我们介绍了一个开源计算流体动力学（CFD）框架，该框架整合了 ML 和图形处理器（GPU）的优势，展示了它们的综合能力。在这一框架内，所有计算操作都完全在 GPU 上执行，包括 ML 加速的化学集成、流体传输 PDE 的全隐式求解以及热和传输属性的计算，从而消除了 CPU-GPU 内存拷贝的开销。在内核函数内部和内核启动过程中都进行了优化，以提高计算性能。采用静态数据重组和动态数据分配等策略来减少 GPU 内存占用。在两个湍流火焰基准中分别使用准 DNS 和 LES 模型对计算性能进行了评估。值得注意的是，在保持与传统的基于 CPU/implicit ODE 的求解器相似的精度水平的同时，GPU/ML 加速方法在两种情况下都显示出超过两个数量级的整体速度提升。这一结果表明，通常需要数百个 CPU 才能完成的高保真湍流燃烧有限速率化学模拟，现在只需一台中端 GPU 就能在笔记本电脑等便携设备上完成。
{"title":"An integrated framework for accelerating reactive flow simulation using GPU and machine learning models","authors":"Runze Mao, Min Zhang, Yingrui Wang, Han Li, Jiayang Xu, Xinyu Dong, Yan Zhang, Zhi X. Chen","doi":"10.1016/j.proci.2024.105512","DOIUrl":"https://doi.org/10.1016/j.proci.2024.105512","url":null,"abstract":"Recent progress in machine learning (ML) and high-performance computing (HPC) have brought potentially game-changing opportunities in accelerating reactive flow simulations. In this study, we introduce an open-source computational fluid dynamics (CFD) framework that integrates the strengths of ML and graphics processing unit (GPU) to demonstrate their combined capability. Within this framework, all computational operations are solely executed on GPU, including ML-accelerated chemistry integration, fully-implicit solving of fluid transport PDEs, and computation of thermal and transport properties, thereby eliminating the CPU–GPU memory copy overhead. Optimisations both within the kernel functions and during the kernel launch process are conducted to enhance computational performance. Strategies such as static data reorganisation and dynamic data allocation are adopted to reduce the GPU memory footprint. The computational performance is evaluated in two turbulent flame benchmarks using quasi-DNS and LES modelling, respectively. Remarkably, while maintaining a similar level of accuracy to the conventional CPU/implicit ODE-based solver, the GPU/ML-accelerated approach shows an overall speedup of over two orders of magnitude for both cases. This result highlights that high-fidelity turbulent combustion simulation with finite-rate chemistry that requires normally hundreds of CPUs can now be performed on portable devices such as laptops with a medium-end GPU.","PeriodicalId":408,"journal":{"name":"Proceedings of the Combustion Institute","volume":"31 1","pages":""},"PeriodicalIF":3.4,"publicationDate":"2024-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141886428","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-07-31DOI: 10.1016/j.proci.2024.105681
Shangyong Zhou, Jianjun Xiao, Zhenmin Luo, Mike Kuznetsov, Zheng Chen, Thomas Jordan, Daniel T. Banuti
This study investigates the spontaneous ignition of high-pressure hydrogen-enriched methane in air within a rectangular tube. A computationally efficient approach has been adopted, utilizing a reduced reaction mechanism and ignition delay model within a 3D Large Eddy Simulation (LES) framework. This approach overcomes the limitations of traditional 1D and 2D simulations with detailed chemistry models, which are unable to accurately reproduce the complex 3D shock wave structures within the tube. The simulated shock wave behavior during 9 MPa hydrogen leakage (case 1) and 11 MPa 90 vol% hydrogen/10 vol% methane mixture leakage (case 2) are found to agree well with experimental observations. In case 2, the hot spots generated by reflected shock waves and Mach reflections ignite the hydrogen/methane-air mixture, resulting in three sequential spontaneous ignitions. The flame is observed to primarily propagate along the tube corners and wall centers, with the central ignition spreading across the entire cross-section. For the 25 MPa 24 vol% hydrogen/76 vol% methane mixture leakage (case 6), the shock intensity is significantly reduced due to the increased methane proportion, leading to spontaneous ignition only at the tube corners when the hemispherical shock wave reflects from the wall. The flame predominantly forms downstream along the tube corner, gradually spreading along the tube wall. It is indicated that while the probability of spontaneous ignition decreases with increasing methane content, the risk remains significant under sufficiently high pressures. To the best our knowledge, this study represents the first 3D large eddy simulation of spontaneous ignition for high-pressure hydrogen-enriched methane leakage into air, providing valuable insights into the underlying physical phenomena.
{"title":"Analysis of spontaneous ignition of hydrogen-enriched methane in a rectangular tube","authors":"Shangyong Zhou, Jianjun Xiao, Zhenmin Luo, Mike Kuznetsov, Zheng Chen, Thomas Jordan, Daniel T. Banuti","doi":"10.1016/j.proci.2024.105681","DOIUrl":"https://doi.org/10.1016/j.proci.2024.105681","url":null,"abstract":"This study investigates the spontaneous ignition of high-pressure hydrogen-enriched methane in air within a rectangular tube. A computationally efficient approach has been adopted, utilizing a reduced reaction mechanism and ignition delay model within a 3D Large Eddy Simulation (LES) framework. This approach overcomes the limitations of traditional 1D and 2D simulations with detailed chemistry models, which are unable to accurately reproduce the complex 3D shock wave structures within the tube. The simulated shock wave behavior during 9 MPa hydrogen leakage (case 1) and 11 MPa 90 vol% hydrogen/10 vol% methane mixture leakage (case 2) are found to agree well with experimental observations. In case 2, the hot spots generated by reflected shock waves and Mach reflections ignite the hydrogen/methane-air mixture, resulting in three sequential spontaneous ignitions. The flame is observed to primarily propagate along the tube corners and wall centers, with the central ignition spreading across the entire cross-section. For the 25 MPa 24 vol% hydrogen/76 vol% methane mixture leakage (case 6), the shock intensity is significantly reduced due to the increased methane proportion, leading to spontaneous ignition only at the tube corners when the hemispherical shock wave reflects from the wall. The flame predominantly forms downstream along the tube corner, gradually spreading along the tube wall. It is indicated that while the probability of spontaneous ignition decreases with increasing methane content, the risk remains significant under sufficiently high pressures. To the best our knowledge, this study represents the first 3D large eddy simulation of spontaneous ignition for high-pressure hydrogen-enriched methane leakage into air, providing valuable insights into the underlying physical phenomena.","PeriodicalId":408,"journal":{"name":"Proceedings of the Combustion Institute","volume":"71 1","pages":""},"PeriodicalIF":3.4,"publicationDate":"2024-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141886425","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-07-30DOI: 10.1016/j.proci.2024.105599
B. Franzelli, J. Bonnety, J. Yi, Y. Ogata, A. Cuoci, C. Betrancourt
Metal-oxide nanoparticles are paving the way for the development of new materials, and flame spray pyrolysis (FSP) systems are gaining attention for their large-scale production. On an industrial level, precise control of particle characteristics is needed while guaranteeing an almost zero-emission process. In this context, computational fluid dynamic (CFD) simulations of nanoparticle production in flames are sought to optimize the design of FSP systems. In this work, numerical simulations of TiO nanoparticles production from Titanium(IV) isopropoxide (TTIP) are performed for a laminar coflow H/Ar flame as a first step towards this objective. To lower the CPU cost for 2-D simulations, reduced descriptions for the gas phase and for nanoparticles are considered. For H combustion, a 10-species kinetic mechanism is retained. Five different submechanisms are tested for the description of TTIP conversion into Ti(OH), considered as the TiO precursor. The description of the solid phase relies on a classical three-equation monodisperse formulation. The objective of this work is not to validate the considered CFD strategy, for which a more extensive database would be required, but to identify the most relevant processes for flame synthesis in a diffusion flame by performing a parametric sensitivity study. The originality of this investigation relies on the study of particle characteristics along an H laminar flame in a non-premixed configuration. Thus, the focus of the parametric study is on the effect on nanoparticle characteristics of: (1) diffusion processes of gaseous phase and nanoparticles; (2) aerosol processes. Numerical results are compared to experimental data in terms of conversion rate, volume fraction, and primary particle diameter along the flame height. Trends from the literature on the effect of aerosol process parameters are retrieved. Results highlight the key role of diffusion processes on nanoparticle production in non-premixed flames and the need for future improvements of TTIP conversion kinetics.
金属氧化物纳米粒子正在为新材料的开发铺平道路，而火焰喷射热解（FSP）系统在其大规模生产方面正受到越来越多的关注。在工业层面，需要精确控制颗粒特性，同时保证过程几乎零排放。在这种情况下，人们寻求对火焰中纳米粒子的产生进行计算流体动力学（CFD）模拟，以优化 FSP 系统的设计。作为实现这一目标的第一步，本研究对异丙醇钛（TTIP）在层流 H/Ar 共流火焰中生成 TiO 纳米粒子的过程进行了数值模拟。为了降低二维模拟的 CPU 成本，考虑减少对气相和纳米颗粒的描述。对于 H 燃烧，保留了 10 种动力学机制。在描述 TTIP 转化为 Ti（OH）（被视为 TiO 前体）时，测试了五种不同的子机制。固相的描述依赖于经典的三方程单分散公式。这项工作的目的不是验证所考虑的 CFD 策略（为此需要更广泛的数据库），而是通过进行参数敏感性研究，确定与扩散火焰中火焰合成最相关的过程。这项研究的独创性在于研究了非预混合配置下 H 层流火焰的粒子特性。因此，参数研究的重点是以下因素对纳米粒子特性的影响：(1) 气相和纳米粒子的扩散过程；(2) 气溶胶过程。数值结果与实验数据在转化率、体积分数和沿火焰高度的主颗粒直径方面进行了比较。从文献中检索了气溶胶过程参数的影响趋势。结果凸显了扩散过程对非预混火焰中纳米粒子生成的关键作用，以及未来改进 TTIP 转化动力学的必要性。
{"title":"Numerical simulations of TiO[formula omitted] production in a laminar coflow H[formula omitted]/Ar/TTIP diffusion flame: Comparison with experiments and parametric sensitivity study","authors":"B. Franzelli, J. Bonnety, J. Yi, Y. Ogata, A. Cuoci, C. Betrancourt","doi":"10.1016/j.proci.2024.105599","DOIUrl":"https://doi.org/10.1016/j.proci.2024.105599","url":null,"abstract":"Metal-oxide nanoparticles are paving the way for the development of new materials, and flame spray pyrolysis (FSP) systems are gaining attention for their large-scale production. On an industrial level, precise control of particle characteristics is needed while guaranteeing an almost zero-emission process. In this context, computational fluid dynamic (CFD) simulations of nanoparticle production in flames are sought to optimize the design of FSP systems. In this work, numerical simulations of TiO nanoparticles production from Titanium(IV) isopropoxide (TTIP) are performed for a laminar coflow H/Ar flame as a first step towards this objective. To lower the CPU cost for 2-D simulations, reduced descriptions for the gas phase and for nanoparticles are considered. For H combustion, a 10-species kinetic mechanism is retained. Five different submechanisms are tested for the description of TTIP conversion into Ti(OH), considered as the TiO precursor. The description of the solid phase relies on a classical three-equation monodisperse formulation. The objective of this work is not to validate the considered CFD strategy, for which a more extensive database would be required, but to identify the most relevant processes for flame synthesis in a diffusion flame by performing a parametric sensitivity study. The originality of this investigation relies on the study of particle characteristics along an H laminar flame in a non-premixed configuration. Thus, the focus of the parametric study is on the effect on nanoparticle characteristics of: (1) diffusion processes of gaseous phase and nanoparticles; (2) aerosol processes. Numerical results are compared to experimental data in terms of conversion rate, volume fraction, and primary particle diameter along the flame height. Trends from the literature on the effect of aerosol process parameters are retrieved. Results highlight the key role of diffusion processes on nanoparticle production in non-premixed flames and the need for future improvements of TTIP conversion kinetics.","PeriodicalId":408,"journal":{"name":"Proceedings of the Combustion Institute","volume":"78 1","pages":""},"PeriodicalIF":3.4,"publicationDate":"2024-07-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141886434","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-07-30DOI: 10.1016/j.proci.2024.105605
Marius Schmidt, Jannick Erhard, Lars Illmann, Cooper Welch, Andreas Dreizler, Benjamin Böhm
Liquid fuel wall films are a known source of hydrocarbon and soot emissions in direct-injection spark-ignition (DISI) engines. Therefore, a comprehensive understanding of the evaporation, mixing, and combustion processes above wall films is desirable. In this study, laser-induced fluorescence (LIF) of acetone excited at 315nm is used to measure the fuel mole fraction in the gas phase above a wall film in an optically accessible DISI engine. To this end, acetone and 3-pentanone are characterized at excitation wavelengths from 305 to 316nm in a heated jet experiment under atmospheric conditions. It is shown that the excitation of acetone at 315nm results in a signal that is sufficiently temperature-independent under engine-relevant conditions. In addition, simultaneous high-speed particle image velocimetry (PIV) and Mie-scattering capture the flow field and cross-sectional flame development. The formation of soot is characterized by natural luminosity. A late injection of acetone during the compression stroke from a single-hole Spray M injector is used to add approximately 8% of the fuel to the homogeneously premixed isooctane-air mixture and form a fuel film on the piston surface. Heavy soot formation occurs when the engine is operated under cold start conditions. After combustion, incandescent soot structures form and persist until the exhaust phase. These soot structures are attributed to the pyrolysis of the fuel as it evaporates into the oxygen-depleted, high-temperature burnt gas. Increasing wall temperatures during cold-start cycles significantly reduces soot formation. However, even at similar temperature levels, strong variations occur. A multi-parameter analysis revealed a strong correlation of the projected soot area with the flow field at ignition and the acetone mole fraction above the film. It is shown that delayed flame-film contact reduces soot formation since it increases the time for evaporation and promotes mixing of acetone-rich regions. Acetone mole fractions in the bulk flow indicate strong turbulent mixing, with fuel-rich regions contributing to soot formation during combustion being typically limited to within 3 mm of the wall.
{"title":"Soot formation as a function of flow, flame and mixing field above evaporating fuel films in an optically accessible engine","authors":"Marius Schmidt, Jannick Erhard, Lars Illmann, Cooper Welch, Andreas Dreizler, Benjamin Böhm","doi":"10.1016/j.proci.2024.105605","DOIUrl":"https://doi.org/10.1016/j.proci.2024.105605","url":null,"abstract":"Liquid fuel wall films are a known source of hydrocarbon and soot emissions in direct-injection spark-ignition (DISI) engines. Therefore, a comprehensive understanding of the evaporation, mixing, and combustion processes above wall films is desirable. In this study, laser-induced fluorescence (LIF) of acetone excited at 315nm is used to measure the fuel mole fraction in the gas phase above a wall film in an optically accessible DISI engine. To this end, acetone and 3-pentanone are characterized at excitation wavelengths from 305 to 316nm in a heated jet experiment under atmospheric conditions. It is shown that the excitation of acetone at 315nm results in a signal that is sufficiently temperature-independent under engine-relevant conditions. In addition, simultaneous high-speed particle image velocimetry (PIV) and Mie-scattering capture the flow field and cross-sectional flame development. The formation of soot is characterized by natural luminosity. A late injection of acetone during the compression stroke from a single-hole Spray M injector is used to add approximately 8% of the fuel to the homogeneously premixed isooctane-air mixture and form a fuel film on the piston surface. Heavy soot formation occurs when the engine is operated under cold start conditions. After combustion, incandescent soot structures form and persist until the exhaust phase. These soot structures are attributed to the pyrolysis of the fuel as it evaporates into the oxygen-depleted, high-temperature burnt gas. Increasing wall temperatures during cold-start cycles significantly reduces soot formation. However, even at similar temperature levels, strong variations occur. A multi-parameter analysis revealed a strong correlation of the projected soot area with the flow field at ignition and the acetone mole fraction above the film. It is shown that delayed flame-film contact reduces soot formation since it increases the time for evaporation and promotes mixing of acetone-rich regions. Acetone mole fractions in the bulk flow indicate strong turbulent mixing, with fuel-rich regions contributing to soot formation during combustion being typically limited to within 3 mm of the wall.","PeriodicalId":408,"journal":{"name":"Proceedings of the Combustion Institute","volume":"100 1","pages":""},"PeriodicalIF":3.4,"publicationDate":"2024-07-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141886431","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：