Flow, Turbulence and Combustion最新文献

The flame stabilisation mechanism for single and multiple reacting jets in cross-flow has been investigated using Large Eddy Simulation (LES) with the Conditional Moment Closure (CMC) as the sub-grid combustion model and a detailed chemical mechanism for pure hydrogen fuel. It has been found that a single jet in cross-flow (SJICF) has higher jet penetration depth compared to multiple jets in cross-flow (MJICF). This behaviour is attributed to the proximity of the counter-rotating vortex pairs of the three jets, which induces a downward negative velocity component, thereby influencing the jet stem. The flame stabilisation mechanism has been investigated using the budget of individual terms in the CMC equation. The CMC budget analysis reveals that upstream of the reactive zone, a premixed flame structure is observed with a convection-diffusion balance, whereas further downstream, a non-premixed flame structure prevails with a balance between micromixing and chemical reactions.

Wind tunnel experiments using Particle Image Velocimetry method were conducted to compare the flow characteristics induced by slender square and circular cylinders fully immersed in the boundary layer at gap ratios from 0 to 2. The results indicate that the shedding vortex is suppressed by the presence of the bottom wall in both flows if the wall-cylinder gap distance is lower than a threshold. However, the threshold gap for the square cylinder is lower than that for the circular cylinder. Although the scales of the large-scale turbulent structures are significantly reduced by the presence of cylinder, the two types of disturbed boundary layer have a similar downstream recovery process, without significant scale difference due to the small characteristic size of the disturbing cylinder. Since the near-wake flow behind the square cylinder is larger than that behind the circular cylinder for the same gap, the shedding vortex exhibits noticeably different contributions to the energy fraction of various turbulent modes. For boundary layer disturbed by the square cylinder, the Kármán vortex street becomes the dominant structure in the third and fourth order modes, whereas the effects are visible in the 4-10th order turbulent modes in the circular-cylinder disturbed boundary layer if the gap is high enough.

Turbulent nonpremixed combustion is widely used in the gas turbine combustors and in combustors attached to thermal power plant boilers. Improving the performance of these combustors is essential for reducing their CO₂ emissions. Understanding the effects of flame curvature on the local structure of turbulent nonpremixed flames is necessary for improving the accuracy of the laminar flamelet model and determining how the combustor performance can be improved. This study experimentally clarified the effects of the Lewis number of the fuel flow (Le_F) on curved counterflow nonpremixed flames using methane or propane as the fuel and air as the oxidizer. The effects of Le_F were isolated by comparing observations of methane–nitrogen fuel flows, for which Le_F is unaffected by the dilution rate, with observations of propane–nitrogen fuel flows, for which Le_F changes substantially with the dilution rate. Although it was previously believed that flame curvature effects could be neglected in hydrocarbon–nitrogen flames, this study revealed that even with hydrocarbon fuels, flame curvature can induce Lewis number effects when the flame radius is small. The effects of flame curvature on extinction were observed when the ratio of the flame zone thickness to the flame radius was on the order of 10^− 1, which is expected to occur in actual combustors. These results indicate that the effects of flame curvature should be considered when the laminar flamelet model is used to analyze turbulent nonpremixed flames at Le_F ≠ 1.

This study presents a compact data-driven Reynolds-averaged Navier-Stokes (RANS) model for wind turbine wake prediction, built as an enhancement of the standard (k)-(varepsilon) formulation. Several candidate models were discovered using the symbolic regression framework Sparse Regression of Turbulent Stress Anisotropy (SpaRTA), trained on a single Large Eddy Simulation (LES) dataset of a standalone wind turbine. The leading model was selected by prioritizing simplicity while maintaining reasonable accuracy, resulting in a novel linear eddy viscosity model. This selected leading model reduces eddy viscosity in high-shear regions—particularly in the wake—to limit turbulence mixing and delay wake recovery. This addresses a common shortcoming of the standard (k)-(varepsilon) model, which tends to overpredict mixing, leading to unrealistically fast wake recovery. Moreover, the formulation of the leading model closely resembles that of the established (k)-(varepsilon)-(f_P) model. Consistent with this resemblance, the leading and (k)-(varepsilon)-(f_P) models show nearly identical performance in predicting velocity fields and power output, but they differ in their predictions of turbulent kinetic energy. In addition, the generalization capability of the leading model was assessed using three unseen six-turbine configurations with varying spacing and alignment. Despite being trained solely on a standalone turbine case, the model produced results comparable to LES data. These findings demonstrate that data-driven methods can yield interpretable, physically consistent RANS models that are competitive with traditional modeling approaches while maintaining simplicity and achieving generalizability.

Combustion noise is emerging as a significant contributor to an aircraft’s overall noise signature, making it essential to understand the factors influencing it. This study investigates numerically combustion noise mechanisms in a real helicopter engine. A method combining Large Eddy Simulations (LES), Actuator Disk Theory, and a Helmholtz solver is employed to analyze combustion noise in the combustion chamber, turbine, and far-field, respectively. Two LES databases are compared against experimental data to assess their effects on the predicted noise: (i) the first includes a liquid-fuel injection strategy in the combustion chamber coupled with a realistic first-stage stator downstream, (ii) the second relies on a gaseous-fuel injection in the combustor with a simplified nozzle at the exit. Compared to the gaseous injection, the presence of a liquid spray leads to more pronounced differences in the mechanisms of acoustic generation. Adopting the liquid injection and the stator, combustion noise predictions match more accurately experimental results at the turbine exit and in the far field, displaying a low-frequency resonant peak which is not accounted for in the gaseous case. Direct noise dominates at high frequencies, while indirect entropy noise is more important at low frequencies. The most significant discrepancy between both cases concerns the azimuthal noise distribution at the turbine exit. In the gaseous case, noise is evenly distributed across azimuthal modes, whereas the planar and first azimuthal modes predominantly drive the noise in the liquid injection configuration. This study provides insights into the complex interplay between combustion noise mechanisms, their modes, and the resulting far-field noise.

The Lattice Boltzmann Method (LBM), a weakly compressible approach, has attracted considerable attention for turbulent flow simulations but remains limited in high-Reynolds-number applications due to its reliance on uniform meshes. This study investigates the interpolation-based Lattice Boltzmann Method (IBLBM) for direct numerical simulation (DNS) of turbulent channel and duct flows using nonuniform grids. Simulations are conducted for turbulent Poiseuille channel flows at (boldsymbol{Re_{tau} = 180, 395, 640},) and 1000, as well as for turbulent duct flow at (boldsymbol{Re_{tau} = 360}). The predicted mean velocity profiles, Reynolds stresses, vorticity distributions, and turbulent kinetic energy budgets show excellent agreement with benchmark DNS data. The IBLBM scheme accurately resolves near-wall turbulence and secondary flow structures in ducts while maintaining similar computational overhead across different interpolation orders. For the 7- and 200-million-grid cases at (boldsymbol{Re_{tau} = 180}) and 640, LBM with (boldsymbol{Delta^+ sim 3}) is approximately 1.9 and 3.7 times faster, respectively, than sixth-order IBLBM; however, to match IBLBMs wall resolution ((boldsymbol{Delta^+ sim 0.3})) using (boldsymbol{Delta^+ sim 1}), LBM becomes four to six times slower. These results demonstrate the robustness and accuracy of IBLBM for high-Reynolds-number wall-bounded turbulence, offering an optimal balance between computational efficiency and physical fidelity through the use of nonuniform grids.

Flow control has attracted research for its potential role in reducing drag, suppressing turbulence, and enhancing mixing in fluid systems. The emergence of data-driven modeling and machine learning techniques has sparked new interest in designing control strategies that can adapt in real time to complex, high-dimensional flow environments. However, fluid systems remain particularly challenging testbeds for control design due to their nonlinear and convective nature, which introduces large time delays. In active control, additional difficulties arise from practical constraints, such as the use of localized sensors in limited number. In this work, we investigate a reinforcement learning framework based on a suitable actor–critic algorithm designed to address these challenges. Two test cases representative of transitional shear flows are considered: a linearized version of the Kuramoto–Sivashinsky equation and the control of instabilities in a two-dimensional boundary-layer flow over a flat plate, using a minimal but realistic sensor–actuator configuration. This choice reflects our focus on the limitations that arise from plants of experimental interest. Time delays are identified during a pretraining stage, while the control algorithm employs multistep returns during value iteration. This approach improves both the convergence rate and stability of learning. Furthermore, we show that the look-ahead in the multistep formulation provides a non-trivial beneficial effect in plants where the control task is characterized by a severe credit-assignment issue.

Air pollution modeling as a problem of continuum mechanics provides a rigorously predictive approach invaluable for industrial control and engineering. To address the limitations of classical atmospheric Eulerian transport models, such as CHIMERE, this study introduces a new numerical method: The Unsteady Mass Conservation Approach. Rooted in the principle of mass conservation defined by the continuity equation, UMCA incorporates dynamic correction terms both instantaneous and advective to improve the simulation under non-stationary conditions. Pollutant mass concentration is modeled through a system of coupled differential equations that account for key physical processes, including advection, turbulence, atmospheric reactions, emissions, and deposition. Several important variables are analyzed to understand and evaluate the model’s behavior, such as the Reynolds number, temperature, pressure, wind speed, momentum, and the zonal and meridional wind components. The accuracy, stability, and adaptability of the UMCA method are assessed using multiple performance metrics: Pearson correlation coefficient, root mean square error, logarithmic residual gain, and residual norm. A series of validation tests is conducted, including wind-driven dispersion, short-term forecasting, spatial consistency, solver stability, and comparison with standard data assimilation techniques. The results show that UMCA is both practical and precise, with benefits in simplicity, portability, and performance. It’s also flexible enough to be applied to other complex flow problems, including ongoing work in turbulence and nonlinear atmospheric modeling.

This study presents an experimental and numerical investigation of the stability of an unconfined annular premixed swirl burner operating with a premixed n-butane-air mixture at an equivalence ratio (ϕ) = 1, 1 bar, and 300 K. The key variable, the annular exit area, was varied while maintaining constant Reynolds number, exit velocity, and equivalence ratio. Three burner heads with exit areas of 98 mm², 157 mm², and 216 mm² were analyzed. Particle image velocimetry (PIV), formaldehyde planar laser induced fluorescence (CH₂O PLIF) and CH* chemiluminescence were utilized to characterize the turbulent reactive flow field. A numerical simulation using the Delayed Detached Eddy Simulation (DDES) turbulence model and the Flamelet Generated Manifold (FGM) model, coupled with a reduced USC Mech, was performed to capture turbulence-chemistry interactions. Q-criterion was used to visualize the 3D flow structures. Results indicate that the burner with an annular area of 157 mm² exhibited higher entrainment and strain rates, leading to increased turbulence intensity and vortex stretching. This resulted in a more unstable flame root, frequent extinction-reignition events, and a narrower lean blowoff (LBO) limit. These findings highlight the impact of exit annular area modification on key combustion dynamics.

To improve the understanding of combustion dynamics in large industrial systems, two hot blast stove designs, C1 and C2, are analysed on their linear thermoacoustic stability. In the first part, the Flame Transfer Function (FTF) is computed using Scale Adaptive Simulations, with a flamelet generated manifold model for the combustion. The fuel mass flows of the non-premixed flames are forced using a superposition of sine waves, with 16 points throughout the frequency range. The results show large differences in flame shape between the two burner designs. Most notably, the larger number of reactant ports reduces the flame length in C2. The convective time delay describes the lag between forcing at the fuel inlet and response in the total flame. Due to the smaller flame length this reduces form 95 ms in C1 to 40 ms in C2. The gain generally shows low-pass behaviour with local maxima and minima. It is shown the response of flame surface area quickly vanishes for higher frequencies, while mass burning rate fluctuations are thought to govern the overall response. Next, the FTF is applied in an acoustic network model to compute the linear stability of the complete system. The model predicts unstable modes for C1 but (marginally) stable modes for C2. This mode prediction is in good agreement with experimental pressure measurements.