OBRABOTKAMETALLOV MATERIAL SCIENCE Vol. 25 No. 4 2023 computational domain was instituted within the accelerator, as well as between the nozzle slice of the tube and the eroded surface. Fig. 2 displays a general view of the accelerator tube with the nozzle. A mixture of particles and air enters the Laval nozzle, accelerates, and exits through the tube onto a sample comprised of titanium alloy Ti6Al4V. Due to the axisymmetric nature of the problem, the two-phase flow area can be depicted in a twodimensional axisymmetric format, which boosts the accuracy of the calculation and reduces computational resources. The computational domain was entirely modelled using two mesh regions, namely the accelerator, and the flow area between the accelerator and the sample. A block mesh with a structured design and a high dimensionless distance y+ near the erodible surface was created in ICEM CFD software. This was due to the utilization of a scaled wall function for boundary layer modelling. The labelled schematic diagram with the designation of the types of boundary conditions (BCs) is shown in fig. 3. Fig. 2. Flow accelerator model: mixer (1), converging part (2), diverging part (3) Fig. 3. 2d axisymmetrical schematic diagram and boundary conditions: accelerator area (1); outflow from accelerator to sample (2); inlet boundary condition (air + particle initialization area) (3); wall boundary condition (4); sample wall BC (4.1); pressure outlet boundary condition (5) Physical Models/ Grid Convergence Study The model in question is based on using Reynolds-averaged Navier-Stokes equations to describe the movement of the carrier phase – air (ideal gas). To average the results, considering turbulent phenomena through a turbulence model is necessary, the choice of which can substantially affect the outcomes. A specific evaluation of both models and its coefficients’ sensitivity is required. Next, we will discuss the impact of models founded on equations for turbulent kinetic energy (k), its dissipation rate (e), and models founded on k and specific dissipation rate (ω): k-epsilon standard, k-epsilon RNG, and Generalised equation k-omega (GEKO) [29–31]. The k-epsilon standard model serves as the foundation for numerous turbulence models intended to explain phenomena within the flow core. RNG is deemed to provide increased precision for high velocity gradient, swirling flows [30]. GEKO is a new model, based on k and ω, which uniquely
RkJQdWJsaXNoZXIy MTk0ODM1