Flow, Turbulence and Combustion最新文献

This study investigates the influence of forebody configuration on aerodynamic noise generation and radiation in standard squareback vehicles, employing a hybrid computational aeroacoustics approach. Initially, a widely used standard squareback body is employed to establish grid-independent solutions and validate the applied methodology against previously published experimental data. Six distinct configurations are examined, consisting of three bodies with A-pillars and three without A-pillars. Throughout these configurations, the reference area, length, and height remain consistent, while systematic alterations to the forebody are implemented. The findings reveal that changes in the forebody design exert a substantial influence on both the overall aerodynamics and aeroacoustics performance of the vehicle. Notably, bodies without A-pillars exhibit a significant reduction in downforce compared to their A-pillar counterparts. For all configurations, the flow characteristics around the side-view mirror and the side window exhibit an asymmetrical horseshoe vortex with high-intensity pressure fluctuations, primarily within the confines of this vortex and the mirror wake. Side windows on bodies with A-pillars experience more pronounced pressure fluctuations, rendering these configurations distinctly impactful in terms of radiated noise. However, despite forebody-induced variations in pressure fluctuations impacting the side window and side-view mirror, the fundamental structure of the radiated noise remains relatively consistent. The noise pattern transitions from a cardioid-like shape to a monopole-like pattern as the probing distance from the vehicle increases.

In this study, we have investigated the effects of inflow spurious periodicity on a turbulent boundary layer subjected to a weak pressure gradient due to a gently backward ramp utilizing wall-resolved large-eddy simulation. The spurious periodicity is generated through repeating segments of inflow data produced via a long boundary layer precursor simulation. Using a short segment of inflow data and recycling it while introduces a spurious periodicity, can help to reduce the computational cost of the precursor simulation as well as the storage needed to record the massive data. However, there is no quantitative analysis as to how far the results are affected by the spurious periodicity in case of a separated turbulent boundary layer. This study compares several cases with various inflow segment lengths with respect to a reference case with no periodicity. The inflow segments are created by truncating the reference case and thus an intrinsic disturbance is associated with them. An additional case without the disturbance is also considered to examine the latter’s influence on the results. First- and second-order flow statistics are assessed, and spectral analysis is conducted to scrutinize the impact of spurious periodicity. An additional second-order parameter is introduced as an indicator of the spurious periodicity influence. Finally, an especial configuration with active flow control using a synthetic jet actuator is investigated with and without the spurious periodicity. This will help to assess if the dominant frequency of the active flow control is affected by recycling inflow data.

The effects of different vortex breakdown states on the evaporation process characterizing air-acetone vapor swirling jets laden with liquid acetone droplets in the dilute regime are discussed based on results provided by direct numerical simulations. Adopting the point-droplet approximation, the carrier phase is solved using an Eulerian framework, whereas a Lagrangian tracking of the dispersed phase is used. Three test cases are investigated: one with fully-turbulent pipe inflow conditions and two with a laminar Maxworthy velocity profile at different swirl rates. Consequently, turbulent, bubble-type, and regular conical vortex breakdown states are established. Following phenomenological and statistical analyses of both phases, a significant enhancement of the overall droplet evaporation process due to the onset of the conical vortex breakdown is observed due to the strongest centrifugal forces driving the entire liquid drops towards the low-saturation mixing layer of the jet. The effects of droplet inertia on evaporation are isolated through an additional set of simulations where liquid droplets are treated as Lagrangian tracers. While it is found that inertial effects contribute to enhanced vaporization near the mixing layer under bubble vortex breakdown conditions, droplet inertia plays a secondary role under both turbulent and conical vortex breakdown due to intense turbulent mixing and high centrifugal forces, respectively.

The applicability of Damköhler’s hypotheses for homogenous mixture (i.e. constant equivalence ratio) moderate or intense low-oxygen dilution (MILD) combustion processes (with methane as the fuel) has been assessed using three-dimensional direct numerical simulation data with a skeletal mechanism. Two homogeneous MILD combustion cases with different levels of ({{text{O}}}_{2}) concentration (4.8% and 3.5% by volume) and different turbulence intensities have been investigated to analyse the influence of dilution level, turbulence intensity and the choice of the reaction progress variable definition (i.e. different choices of major species for turbulent burning velocity and flame surface area evaluations) on the applicability of Damköhler’s hypotheses in MILD combustion. It has been found that the normalized volume-integrated burning rate remains of the same order of magnitude as that of the normalized flame surface area only for the reaction progress variable definition based on a species mass fraction which has a Lewis number close to unity (e.g. ({{text{CH}}}_{4})) but the level of applicability deteriorates when the Lewis number of the species mass fraction, based on which the reaction progress variable is defined, deviates significantly from unity (e.g. ({{text{CO}}}_{2})). Moreover, it has been demonstrated that the flame surface area calculation from the OH mole fraction-based information can lead to significant departures from Damköhler’s first hypothesis. It is also found that the relative magnitudes of normalised volume-integrated burning rate and normalised flame surface area are significantly affected by the level of dilution and the choice of the reaction progress variable definition. Damköhler’s second hypothesis, which provides a relation between the normalised turbulent burning velocity and the ratio of turbulent to molecular diffusivities, has been found to hold in an order of magnitude sense in homogeneous mixture MILD combustion only for the reaction progress variable definition based on species that has a Lewis number close to unity (e.g. ({{text{CH}}}_{4})) but the level of disagreement increases as the Lewis number of the reaction progress variable deviates significantly from unity (e.g. ({{text{CO}}}_{2})).

Direct Numerical Simulations (DNS) of three-dimensional premixed turbulent hydrogen-air flames enriched with 19%, 36%, 44% and 57% of NH(_3) (in volume) are performed. Starting from an equivalence ratio of 0.44 for the case with 19% of NH(_3), richer mixtures of (phi =) 0.54, 0.69 and 0.95 are considered when increasing NH(_3) concentration to obtain comparable laminar flame speeds, i.e., 0.17 m/s for 19% and 36 % NH(_3) enriched case, and 0.30 m/s when NH(_3) concentration is increased to 44 and 57%. The composition and characteristics of the studied mixtures enable to investigate the effects of thermo-diffusivity in a turbulent flow and the role of chemistry and stretch effects in the development of the flames. Given a composition of ammonia and hydrogen and an equivalence ratio, a predictive method is described to identify compositions where thermo-diffusive effects impact the flame and predict the stretch factors. Two maps are proposed to achieve this: the first one is based on the Markstein number and the second one is based on the ratio of consumption speed of strained flames over the laminar unstretched flame speed. This prediction can guide model selection and help manufacturers and experimentalists identify relevant operating points based on desired energy output.

In order to sustain applications dealing with various liquid fuels in internal combustion engine (ICE), it is essential to make available prediction methodologies that allow an early evaluation of the potential usefulness of such fuels in terms of favourable mixture preparation process already in realistic configurations. Since the air-mixture formation and subsequent processes are predominantly governed by the fuel injection, a DNS based numerical investigation coupled with VOF as an interface tracking method is carried out in this paper to gain better insight on the fuel injection from an industrial injector "Spray G" configuration. Chosen from Engine Combustion Network (ECN), this is a gasoline direct injector (GDI) featuring 8-holes orifices and operating with high injection pressure (200 bar). Under consideration of the required computational cost associated with DNS, only the 1/8 of the nozzle geometry including one orifice is used. The numerical simulation is accomplished for the quasi-steady injection condition with nozzle needle fully opened. The obtained results are first validated with available experimental data for nozzle mass flow rate and spray spread angle showing a good agreement. Then, a detailed numerical analysis is provided for the in/near nozzle flow evolution especially for flow turbulence, primary and secondary atomization. Furthermore, droplet statistics in terms of droplet shape, and droplet size-velocity distribution together with a breakup regime map are reported. Finally, a 2-D data curation technique is proposed to extract the droplet statistics along selected planes and evaluated by direct comparison with three-dimensional droplet data, which may allow handling of the DNS data in more feasible and economical way especially for time series data with higher frequency. The comprehensive DNS data generated by this DNS-VOF approach enable not only to carry out detailed numerical analysis of in- and near-nozzle physical phenomena for which experimental data are still scarce, but also to provide a hint of more reliable injector boundary conditions useful for lower order spray injection method based on Lagrangian particle tracking.

A practical method is presented to generate inflow boundary conditions for the laminar but spatially inhomogeneous flow out of filters, packed beds, or sinter plates. A synthetic velocity field that statistically reproduces the outflow characteristics is created based on digitally filtered random noise by the integral lengthscale of the investigated flow at a low computational cost. The method is tested to reproduce a real flow downstream of a sinter plate, which has its velocity measured using a hot-wire anemometer. Results demonstrate that the present method can efficiently generate synthetic inflow fields.

Turbulent spray combustion in a generic kerosene-fueled single-cup combustor at typical idle and cruise conditions of an aeroengine are studied with Large Eddy Simulations (LES) using Lagrangian spray and finite-rate chemistry combustion modeling. Three reaction mechanisms of varying complexity are used to model the combustion chemistry. The choice of turbulence-chemistry interaction model is shown to affect the results significantly. The impact of the choice of chemical reaction mechanism and the difference in operating conditions are gauged in terms of time-averaged flow, spray, and combustion characteristics as well as unsteady behavior. Good agreement between LES predictions and experimental results are generally observed but with a notable dependence on the choice of chemical reaction mechanism. The mechanism specifically targeting Jet A displays the best agreement. The choice of reaction mechanism is further demonstrated to influence the flow and thermoacoustics in the combustor, resulting in different thermoacoustic modes dominating. The spray cone is found to be too narrow and thin, an inaccuracy which could be remedied by either making the injection method more empirical or by introducing additional models.