Searcharxiv⌕ Search

arXiv subjects

Andrea Di Mascio

Publications and source records attributed to Andrea Di Mascio.

2 recordsLinked to original sources

Turbulent Mixing Dynamics of Under-Expanded Hydrogen Jets in Propulsion Systems

Underexpanded jets are present in various engineering applications; in recent years, they have gained special attention because of the development of gas-fueled propulsion systems. In these apparatuses, the direct injection of fuels such as hydrogen in innovative low-emission engines' chambers induces turbulent under-expanded jets. In this study, we performed high-fidelity Large Eddy Simulations of under-expanded hydrogen jets to investigate mixing characteristics and provide valuable insights for developing injectors suitable for hydrogen and, more generally, gaseous-fueled propulsion systems. We initially assessed the method's accuracy, evaluating the convergence and uncertainty of the numerical results and validating them against experimental particle image velocimetry and Schlieren data. The simulated jets, the Mach disc dimensions, and the resulting velocity field align closely with the experimental observations. Then, we analysed the jet structure for pressure ratios of 4 to 25 and examined the effects of the geometrical configuration of the nozzle on the characteristics of the air-fuel mixture obtained. We compared the jets resulting from a round-hole nozzle with annular ones resembling outward-opening injectors.

physics.flu-dyn↗

A Chimera method for high-fidelity simulation of turbulent flows

We develop a block-structured solver for high-fidelity simulation of flows in complex geometries, based on overlapping (Chimera) meshes. The key components of the algorithm are a baseline dissipation-free central discretization and selective high-order filtering, which ensure uniform accuracy and minimal numerical diffusion. These favorable properties are preserved through efficient interpolation across overlapping blocks. Numerical tests demonstrate that the method guarantees a uniform order of accuracy even for distorted, overlapping meshes. Tests conducted for turbulent flow in a pipe show no significant issues at the interfaces or overlapping blocks, highlighting the method's potential for direct numerical simulation (DNS). Finally, we show that the method performs satisfactorily for geometrically complex problems, such as flow past a ship propeller, where it accurately replicates experimental data.

physics.flu-dyn↗