Direct Numerical and Large-Eddy Simulation of Primary Atomization in Complex Geometries

Desjardins, Olivier; McCaslin, Jeremy; Owkes, Mark; Brady, Peter

doi:10.1615/atomizspr.2013007679

Cited by 90 publications

(50 citation statements)

References 65 publications

(64 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The selected test-case is an air-assisted atomiser, presented and studied by [3] and [2]. The injector geometry consists of a straight jet of diameter d1 = 1.3335 · 10 −3 m, where n-dodecane flows.…”

Section: Results and Discussion Numerical Set-upmentioning

confidence: 99%

“…The gas-to-liquid momentum ratio is M = 2.53, which suggest a length core of 3.9 jet diameters (based on experimental correlations [15]) The simplicity of the geometry is well suited for numerical modelling and validation. Desjardins et al [3] performed Direct Numerical Simulation using an advanced level-set method of the same configuration, with a mesh of 512 × 256 × 256 on a domain of 16d1 × 8d1 × 8d1. The atomization DNS was performed on 1024 processors [3].…”

Section: Results and Discussion Numerical Set-upmentioning

confidence: 99%

“…Desjardins et al [3] performed Direct Numerical Simulation using an advanced level-set method of the same configuration, with a mesh of 512 × 256 × 256 on a domain of 16d1 × 8d1 × 8d1. The atomization DNS was performed on 1024 processors [3]. In the present work, the grid size is increased approximately three times in each direction respect to the base DNS, using a mesh of 180 × 90 × 90 grid cells and the simulations were performed in a single workstation.…”

Section: Results and Discussion Numerical Set-upmentioning

confidence: 99%

“…Local Weber distribution, W eκ, based on curvature radius as reference length Figure 5 shows the droplet distribution of the present simulations, compared to the experimental and DNS results presented in [3]. To present fair comparison a diameter normalisation d * = (d − dmin)/(dmax − dmin) has been performed.…”

Section: 321mentioning

confidence: 97%

“…The original DNS [3] showed the formation of the instabilities along the liquid core, which develop into liquid ligaments and then droplets. The present results show similar overall characteristics but with a large dependency of the compression ratio.…”

Section: 321mentioning

confidence: 99%

See 4 more Smart Citations

Detailed simulation of air-assisted spray atomization: effect of numerical scheme at intermediate Weber number

Tretola

Vogiatzaki

Navarro‐Martinez

2017

Proceedings ILASS–Europe 2017. 28th Conference on Liquid Atomization and Spray Systems

View full text Add to dashboard Cite

Numerical simulations are often used to understand spray atomisation and estimate the size of the liquid fragments. Several techniques (Level Set, Volume of Fluid, Smooth Particle Hydrodynamics, among others) exist to compute multiphase flows and potentially represent liquid-break-up. However, the complexity of the breakup process and the wide range of scales prevents the use of an unified approach to simulate the complete spray. Numerical techniques face different challenges depending on the spray characteristics. The incorrect representation of surface forces in capillary dominated flows, creates large parasitic currents that distort and in some cases destroy the interface. Methods that perform well in the capillary regime aim to capture the interface directly and the surface radius curvature is therefore larger than the mesh size. However, this creates large constrains on the mesh resolution and limits its applications to low Weber number flows, when there is no extensive atomization. Methods that simulate large Weber number flows (typical of industrial injectors) do not resolve the interface directly and the mesh is larger than the smallest radius of curvature. These models often have numerical or artificial diffusion that destroys small scale structures and alters the break-up. However, even at large Weber flows, the spray formation can be affected by errors due to the local imbalance between pressure and surface tension forces and interface curvature errors. Numerical schemes work around these problems by adjusting the amount of numerical diffusion of the scheme depending on the spray application. Intermediate Weber number sprays are well suited to study the performance of numerical methods as they exhibit hybrid behaviour between capillary flows and full atomization. In the present work an intermediate gas Weber of a laboratory air-blast atomiser is investigated using a volume of fluid approach. The amount of numerical diffusion is controlled by a compressive factor in the volume of fluid transport equation. The effect of the compressive term on spray atomization and droplet size distribution is explored. The results suggest that the optimal amount of diffusion depends on the local Weber number. Keywords VOF, OpenFOAM, Atomisation, intermediate Weber. IntroductionLiquid atomization is a complex phenomenon present in many engineering processes. The atomization process depends on complex interactions between aerodynamic and capillary forces. Liquid structures are shed from the dense spray core, forming ligaments that pinch-off and form droplets. The droplet break-up pattern itself is also very complex, a droplet may break into few large droplets (for example: vibrational break-up) or a myriad of small droplets (bag break-up), among other. For detailed reviews regarding the physics of the atomisation, see [4] and [5]. The atomisation process involves a wide range of length scales: from the nozzle diameter to the smallest droplets. The smallest scale in atomisation is not still characterised and this ...

show abstract

Section: Results and Discussion Numerical Set-upmentioning

confidence: 99%