Flow, Turbulence and Combustion最新文献

{"title":"Aerodynamic Characteristics Analysis of High-Speed Train Pantograph Under Various Crosswind and Train Speed Conditions","authors":"Peilin Gong,&nbsp;Zhuojun Li,&nbsp;Yuhan Guo,&nbsp;Huapu Song,&nbsp;Chunjiang Chen,&nbsp;Fentian Zhu,&nbsp;Yang Chen,&nbsp;Kailong Jin,&nbsp;Sijun Huang,&nbsp;Huadong Yao,&nbsp;Shuoguo Zhang,&nbsp;Jiqiang Niu","doi":"10.1007/s10494-026-00787-3","DOIUrl":"10.1007/s10494-026-00787-3","url":null,"abstract":"<div><p>As train speeds increase, the aerodynamic challenges posed by the complex and variable characteristics of higher-speed pantographs under different operating conditions and crosswind speeds became more pronounced. These challenges significantly compromised the stability of the pantograph, jeopardizing the safety of train operations. This study aimed to investigate the impact of various operating conditions of higher-speed pantographs, train speeds, and crosswind conditions on the unsteady aerodynamic characteristics of pantographs. Numerical simulations were conducted using the improved delayed detached-eddy simulation (IDDES) method, and the results were validated against wind tunnel test data. The results indicated that the operational orientation of the pantograph exerted a profound influence on its aerodynamic behavior. While the pantograph generated a higher aerodynamic uplift force in the knuckle-downstream direction, this lift exhibited more pronounced fluctuations in the knuckle-upstream direction, accompanied by more intricate component load characteristics and the evolution of more complex vortex structures. As the train’s speed increased, the lift and drag coefficients of the lower framework of the pantograph demonstrated a pronounced sensitivity to speed variations, with its average drag coefficients surpassing that of the upper frameworkwork. Simultaneously, vortex shedding phenomena in the critical regions of the pantograph’s flow field exhibited a direct correlation with speed, with both the magnitude and intensity of the coherent structures escalating substantially. The flow field within the wake region beneath the pantograph emerged as the most complex due to the interference from the wake structures of multiple components. As crosswind velocity intensified, the lift and drag coefficients of each pantograph components underwent nonlinear, abrupt transitions with increasing wind speed. Notably, significant transitions occurred at a crosswind velocity of 10 m/s. At higher crosswind speeds, the wake vortex system in the pantograph’s flow field exhibited increasingly pronounced deflection and superposition effects, further complicating the overall flow structure.</p></div>","PeriodicalId":559,"journal":{"name":"Flow, Turbulence and Combustion","volume":"117 2","pages":""},"PeriodicalIF":2.2,"publicationDate":"2026-08-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148837931","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"PIV Analysis of Flow Dynamics and Crystal Motion During Solidification of an Aqueous Ammonium Chloride-Water Solution in a Large Containment","authors":"Gloshan Shayesteh,&nbsp;Andreas Ludwig,&nbsp;Mihaela Stefan-Kharicha,&nbsp;Menghuai Wu,&nbsp;Abdellah Kharicha","doi":"10.1007/s10494-026-00786-4","DOIUrl":"10.1007/s10494-026-00786-4","url":null,"abstract":"<div><p>A dual-phase Particle Image Velocimetry (DPVI) technology is employed to capture the velocity fields of both the liquid and equiaxed crystals, enabling a quantitative analysis of the solidification dynamics of an aqueous ammonium chloride-water solution within a containment system of unprecedented size, a scale that has not been previously reported in the literature. Due to the large dimensions of the containment, several key phenomena occur, including (i) flow channels in the columnar mushy layers, (ii) unsteadily sedimenting of equiaxed crystals, (iii) agglomerations of crystals that collapse when they grow too large, resulting in crystal avalanches, and (iv) flow stratification driven by double-diffusive convection. The last-mentioned phenomenon was observed not only at the end of the process, starting from the top of the containment, but also as an intermediate occurrence in the middle of the test cell. It terminates when the dynamics of equiaxed crystal become too intense, and the turbulent kinetic energy increases significantly. The DPIV measurements provide a quantitative description of these complex processes, which could contribute to the development of future numerical models.</p></div>","PeriodicalId":559,"journal":{"name":"Flow, Turbulence and Combustion","volume":"117 2","pages":""},"PeriodicalIF":2.2,"publicationDate":"2026-08-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1007/s10494-026-00786-4.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148837360","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Laminarization Mechanism of Near-Wall Turbulence in Premixed Natural Gas Combustion Investigated by Direct Numerical Simulation","authors":"Takashi Ohta,&nbsp;Yuga Nagai,&nbsp;Yasuyuki Sakai,&nbsp;Genki Ueno","doi":"10.1007/s10494-026-00783-7","DOIUrl":"10.1007/s10494-026-00783-7","url":null,"abstract":"<div><p>In this study, we clarified the characteristics of flame propagation during combustion and the mechanism of laminarization resulting from the attenuation of turbulence structures in the flow field under the influence of the flame. We performed three-dimensional simulations of the wall turbulence in the combustion of premixed natural gas and premixed hydrogen. We adopted a direct numerical simulation method for compressible flow with combustion using a reaction model. Premixed natural gas and premixed hydrogen combustion have different laminarization mechanisms in turbulent combustion fields. The heat of the flame affects the velocity streaks and vortices of the turbulence structures differently, and changes in the wall shear stress and wall heat flux exhibit different tendencies. Depending on the thermal Péclet number, defined using the maximum local intensity of turbulence as a parameter for classifying the characteristics of the laminarization mechanism of the flow field, the wall shear stress may change before the wall heat flux, or both may change simultaneously, under the influence of combustion. The aforementioned results provide insights into the prediction and control of turbulent combustion for future applications.</p></div>","PeriodicalId":559,"journal":{"name":"Flow, Turbulence and Combustion","volume":"117 2","pages":""},"PeriodicalIF":2.2,"publicationDate":"2026-08-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1007/s10494-026-00783-7.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148751407","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Increased Flashback Propensity in Simple, Perfectly Premixed, Laminar Thermoacoustic Hydrogen Systems","authors":"Adriano Bordoni,&nbsp;Antonio J. Torregrosa,&nbsp;Alberto Broatch,&nbsp;Jorge García-Tíscar","doi":"10.1007/s10494-026-00785-5","DOIUrl":"10.1007/s10494-026-00785-5","url":null,"abstract":"<div><p>The use of hydrogen to enrich or replace hydrocarbon-based flames represents an appealing perspective to reduce the emission of pollutants in gas turbine combustors. This implementation is, however, not straightforward, as the properties of hydrogen consistently differ from those of the most widely used fuels, resulting in a change in the combustion stability of the mixture. This work investigates the non-linear thermoacoustic stability of a simple Rijke tube containing a perfectly premixed, laminar, conical flame fed by either a pure methane-based mixture or a hydrogen-enriched one. The mean heat release rate and the geometry of the system were kept constant for both mixtures. Flame dynamics were modelled using a Level Set Method, from which a Flame Describing Function was derived for each mixture and coupled with a one-dimensional acoustic model within a low-order thermoacoustic framework to asses the stability of the system. The hydrogen-enriched flame was found to exhibit a weaker non-linear response than the methane-based flame, in agreement with previous observations reported in the literature. This behaviour was attributed to the higher unstretched laminar flame speed of the hydrogen-enriched mixture, which enhances the resilience of the flame front to develop geometric non-linearities following an acoustic perturbation, reducing its tendency towards heat release rate saturation. As a consequence, the hydrogen-enriched system generated larger thermoacoustic oscillations and ultimately experienced flashback, whereas the methane-based configuration remained bounded by more consistent non-linear saturation effects.</p></div>","PeriodicalId":559,"journal":{"name":"Flow, Turbulence and Combustion","volume":"117 2","pages":""},"PeriodicalIF":2.2,"publicationDate":"2026-08-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148751113","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Liquid Water Addition as an Emission Control Strategy for Ammonia Flames: A High-Fidelity Numerical Study","authors":"Riccardo Concetti,&nbsp;Maximilian Bambauer,&nbsp;Federica Ferraro,&nbsp;Josef Hasslberger,&nbsp;Markus Klein","doi":"10.1007/s10494-026-00780-w","DOIUrl":"10.1007/s10494-026-00780-w","url":null,"abstract":"<div><p>The mitigation of nitrogen oxide (<span>(textrm{NO}_x)</span>) emissions remains a key challenge for the use of ammonia as a carbon-free energy carrier. In contrast to hydrogen, ammonia combustion is inherently prone to fuel-<span>(textrm{NO}_x)</span> formation due to the nitrogen content of the fuel, limiting the effectiveness of purely temperature-based mitigation strategies. In this work, liquid water addition is investigated as a potential emission control approach for premixed ammonia/air flames using high-fidelity numerical simulations. A hybrid Eulerian–Lagrangian framework is employed to resolve the interaction between a spherically expanding quasi-laminar ammonia flame and evaporating water droplets. The influence of droplet sizes on flame propagation, flame thickness, and emission behavior is examined under stoichiometric conditions. Water addition is found to reduce the burning rate per unit area and increase flame thickness through combined thermal and dilution effects, while having a limited impact on flame surface wrinkling. Both small and large droplets lead to a significant reduction in NO formation, with large droplets producing localized regions of strong suppression and small droplets yielding a more spatially uniform reduction. Analysis of intermediate nitrogen species indicates a minor contribution of the <span>(textrm{N}_{2}textrm{O})</span> pathway under the present conditions, while thermal and fuel-<span>(textrm{NO}_x)</span> pathways remain sensitive to water-induced temperature changes. These results provide physical insight into the coupling between water evaporation, flame structure, and <span>(textrm{NO}_x)</span> formation in premixed ammonia flames.</p></div>","PeriodicalId":559,"journal":{"name":"Flow, Turbulence and Combustion","volume":"117 2","pages":""},"PeriodicalIF":2.2,"publicationDate":"2026-08-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1007/s10494-026-00780-w.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148750547","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Statistical Analysis and RSM Optimization of Combustion and Emission Characteristics in an RCCI Engine Fueled with Macroalgal Biodiesel and Iso-butanol","authors":"Ajmeera Naresh,&nbsp;Ravi Kumar Puli","doi":"10.1007/s10494-026-00784-6","DOIUrl":"10.1007/s10494-026-00784-6","url":null,"abstract":"<div><p>This study presents a robust multi-objective optimization framework for Reactivity Controlled Compression Ignition (RCCI) engines using a ternary blend of Ulva fasciata macroalga biodiesel and isobutanol. Employing Response Surface Methodology (RSM) integrated with Design of Experiments (DoE), the framework elucidates the nonlinear, coupled effects of isobutanol fraction, Exhaust Gas Recirculation (EGR) rates, and engine load on dual-fuel reactivity. Predictive models were rigorously validated via Analysis of Variance (ANOVA), enabling high-fidelity mapping of the engine’s operational envelope and thermodynamic behaviors. Optimized conditions (10–30%) isobutanol fraction, 5% EGR, and 12 kg load yielded a Brake Thermal Efficiency (BTE) of 31.1% (7.3% improvement over baseline diesel) and a 33% reduction in Brake Specific Fuel Consumption (BSFC). Emissions analysis revealed a 35.2% drop in <i>NOx</i>, driven by EGR’s thermal dilution and heat capacity effects, despite marginal increases in unburned Hydrocarbons 1.9% and Carbon Monoxide 2.3% from low-temperature combustion. These results highlight the potential of Ulva fasciata biodiesel-isobutanol blends for efficient, low-emission RCCI operation, advancing sustainable engine technologies.</p></div>","PeriodicalId":559,"journal":{"name":"Flow, Turbulence and Combustion","volume":"117 2","pages":""},"PeriodicalIF":2.2,"publicationDate":"2026-07-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148614685","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Effect of Nanosecond Plasma Preconditioning on Detonation in High-Pressure H₂/O₂ Mixtures: A Multiscale Numerical Study","authors":"Giacomo Cinieri,&nbsp;Zubair Ali Shah,&nbsp;Ghazanfar Mehdi,&nbsp;Maria Grazia De Giorgi","doi":"10.1007/s10494-026-00782-8","DOIUrl":"10.1007/s10494-026-00782-8","url":null,"abstract":"<div><p>This work investigates the effect of nanosecond-plasma preconditioning on detonation in high-pressure hydrogen-oxygen mixtures through a multiscale numerical framework coupling zero-dimensional plasma kinetics, one-dimensional ZND detonation analysis, and two-dimensional reactive CFD. ChemPlasKin is used to compute the post-discharge thermochemical state of hydrogen oxygen mixtures, which is then mapped at the end of the afterglow as the initial condition for both ZND and CFD calculations. Parametric analyses are performed at 10 bar over reduced electric field (140–260 Td), deposited energy (25–150 mJ/cm³), and equivalence ratio (0.2–1.2). The ZND results show that non-equilibrium plasma discharge has little effect under lean conditions, whereas for ϕ ≥0.8 it shortens the induction length and reduces the predicted detonation cell size by up to 30%. In addition, the threshold reduced electric field required to produce a measurable plasma effect increases by approximately 40 Td for every 0.4 decrease in equivalence ratio. A complementary multi-pulse analysis shows that pulse repetition is effective mainly when the deposited energy per pulse is kept constant, while redistributing the same total energy over a larger number of weaker pulses yields only a limited additional benefit. The sensitivity analysis shows that the dominant control of detonation cell size remains associated with the main hydrogen branching reaction H + O2 ≤&gt;O + OH in both clean and plasma-treated mixtures. At higher pulse energy, the sensitivity pattern changes markedly, with the HO₂/H₂O₂ submechanism becoming significantly more influential. Two-dimensional DDT simulations in a 0.9 mm channel confirm that these thermochemical modifications strongly accelerate the transition, reducing the run-up distance and onset time from 68.5 mm and 136 μs in the untreated case to 9.7 mm and 17.5 μs in the most activated configuration. Overall, the results show that nanosecond plasma preconditioning can substantially compress detonation scales and promote faster DDT for high-pressure hydrogen-oxygen microdevices.</p></div>","PeriodicalId":559,"journal":{"name":"Flow, Turbulence and Combustion","volume":"117 2","pages":""},"PeriodicalIF":2.2,"publicationDate":"2026-07-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1007/s10494-026-00782-8.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148614687","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Combustion Model Validation Framework for Turbulent Premixed Hydrogen-Air Flames Utilizing OpenFOAM","authors":"Svenja Nerzak,&nbsp;Ludovico Nista,&nbsp;Marco Vivenzo,&nbsp;Michael Gauding,&nbsp;Heinz Pitsch","doi":"10.1007/s10494-026-00745-z","DOIUrl":"10.1007/s10494-026-00745-z","url":null,"abstract":"<div><p>Developing accurate and robust combustion models is challenging, and their validation can be equally demanding. Even when suitable experimental or numerical datasets are available, setting up and performing validation simulations often requires substantial resources. As a result, most validation studies focus on a single dataset from experiments or numerical simulations, which is generally insufficient for a comprehensive model assessment. Systematic model evaluation, therefore, benefits from a comprehensive, standardized, and easy-to-use validation framework containing multiple validation datasets. This paper introduces a validation framework for combustion models and demonstrates its application to a state-of-the-art flamelet model for large-eddy simulations. The framework deploys the open-source <i>OpenFOAM</i> toolbox to provide an openly accessible simulation standard allowing straightforward integration of newly developed models and facilitating the adoption by the research community. It includes eight validation cases based on both experimental data and direct numerical simulations, and is designed to be readily expandable by the community to accommodate an arbitrarily large number of additional cases. For all cases, the reference data, together with the required configurations and mesh files, are provided. The framework facilitates the systematic assessment of combustion models by automatically performing simulations for all cases and providing comparisons between model predictions and reference data. As a demonstration, a flamelet-based model for lean premixed hydrogen combustion is implemented in <i>OpenFOAM</i> and evaluated within the proposed framework, providing insights into the model’s capabilities and limitations across different flow and combustion conditions.</p></div>","PeriodicalId":559,"journal":{"name":"Flow, Turbulence and Combustion","volume":"117 2","pages":""},"PeriodicalIF":2.2,"publicationDate":"2026-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1007/s10494-026-00745-z.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148614066","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Turbulence Modification Under the Influence of Pressure Gradients","authors":"Ali Uzun,&nbsp;Mujeeb R. Malik","doi":"10.1007/s10494-026-00781-9","DOIUrl":"10.1007/s10494-026-00781-9","url":null,"abstract":"<div><p>Reynolds stress budgets computed from a direct numerical simulation of the spanwise-periodic flow over a Gaussian bump are analyzed to gain insight into turbulence modification under different pressure-gradient regimes in three subregions. The four nonzero Reynolds stress components are defined in the local orthogonal coordinate system at a given location on the bump surface. The first subregion starts upstream of the bump under a mild adverse pressure gradient (APG) that becomes progressively stronger toward the bump. The second subregion succeeds the first subregion near the bump foot and contains a strong favorable pressure gradient (FPG) until very near the bump apex. The third subregion is the strong APG section that begins near the apex and lasts until the flow separation point. The mild/strong APG reduces the mean shear rate near the wall, which has a dampening effect on turbulence production while the strong FPG leads to an order of magnitude increase in near-wall turbulence production as a result of the much-enhanced mean shear rate. The mean shear rate of the buffer zone between inner and outer parts of the decelerated flow past the apex also gives way to enhanced production. The turbulent kinetic energy (TKE) gets redistributed internally among the normal stress components via the pressure-strain correlation. For the shear stress, pressure strain is generally a consuming term. The reduction in near-wall TKE production in the APG sections leads to decreases in pressure-strain amplitudes for all normal stresses. For the shear stress, a similar correlation between production and pressure strain also exists in the mild APG section, but a more intricate behavior is found in the strong APG section. In the strong FPG region, the much-amplified near-wall production of turbulence leads to substantial enhancements of the pressure-strain correlation for all components. As the flow goes through different pressure-gradient regimes, the decreasing/increasing deficits or surpluses between energizing (production and/or pressure strain) and consuming (dissipation and/or pressure strain) terms near the wall are balanced by the relevant spatially-redistributive terms that proportionally either diminish or amplify in magnitude. The budget surpluses resulting from the much-amplified near-wall turbulence production in the strong FPG section lead to the formation of an internal layer beneath the accelerated boundary layer, where new near-wall peaks emerge for the wall-normal and shear stresses as the original peaks of the other components strengthen substantially. The paper describes the rather complex flow dynamics that eventually leads to boundary-layer separation.</p></div>","PeriodicalId":559,"journal":{"name":"Flow, Turbulence and Combustion","volume":"117 2","pages":""},"PeriodicalIF":2.2,"publicationDate":"2026-07-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148613687","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Data-Driven Calibration of Transition Models for Natural Convection Flows","authors":"Ioannis Kyritsopoulos,&nbsp;Saleh Rezaeiravesh,&nbsp;Vladimir Duffal,&nbsp;Sofiane Benhamadouche,&nbsp;Alistair Revell","doi":"10.1007/s10494-026-00770-y","DOIUrl":"10.1007/s10494-026-00770-y","url":null,"abstract":"<div><p>This study investigates the physical mechanism of laminar-to-turbulent transition in natural convection by developing algebraic Local-Correlation-based Transition Models (LCTMs) through a physics-based, data-driven framework combining Bayesian optimisation and Symbolic Regression. Accurate transition prediction is critical for high-Rayleigh-number natural convection applications, where standard RANS models and existing transition correlations, calibrated for flows without buoyancy effects, fail to capture the transition onset and boundary layer development. We employ DNS data from differentially heated rectangular cavities (Rayleigh number (Ra) = <span>(10^{10})</span> to <span>(10^{11})</span>) to calibrate model parameters by Bayesian optimisation, then use Symbolic Regression with buoyancy and transition-specific invariants to discover explicit algebraic expressions that generalise across flow conditions. Two models are developed: ML-1, optimised for high Rayleigh-number accuracy, and ML-2, designed for broader generalisation across laminar and turbulent regimes. Validation against a natural convection vertical boundary layer, tall rectangular cavities, and a square cavity demonstrates substantial improvements over baseline RANS models, with the data-driven LCTMs accurately predicting transition location, Nusselt number, and velocity and temperature profiles, showing promising capability for enhancing natural convection predictions while maintaining RANS efficiency.</p></div>","PeriodicalId":559,"journal":{"name":"Flow, Turbulence and Combustion","volume":"117 2","pages":""},"PeriodicalIF":2.2,"publicationDate":"2026-07-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://link.springer.com/content/pdf/10.1007/s10494-026-00770-y.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148614062","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}