|
5378 |
|
Friday, June 09, 2017, Hall 27 2:00 PM Plasma Spraying II |
|
On the validity of continuum computational fluid dynamics approach in very low pressure plasma spray conditions |
|
Dmitrii Ivchenko* / University of Limoges, France Tao Zhang / Xian Jiaotong University, China Gilles Mariaux/ University of Limoges, France Armelle Vardelle/ University of Limoges, France Chang-Jiu Li/ Xian Jiaotong University, China Simon Goutier/ University of Limoges, France Tatiana Itina/ Jean Monnet University, France |
|
Plasma Spray Physical Vapor Deposition is operated at pressure ranging from 50 to 200 Pa. It aims to substantially evaporate the processed powder and makes it possible to produce coatings with various microstructures, from lamellar to columnar, through the deposition of fine melted powder particles and nanoclusters and/or the condensation of vapor. Modeling can be a helpful tool to better understand the process and optimize the operating conditions. However, the combination of high temperature and low pressure with the appearance of shock waves resulting from hot gas expansion calls into question the suitability of the continuum approach for modeling of such a process. The continuum approach assumes that a small control volume can be defined to average the microscopic properties of gas particles (kinetic and internal energy, velocities) in macroscopic properties (pressure, temperature&) that vary continuously in space and are compatible with the scale of the system. This approach breaks down when the mean free path of the gas particles is large and thus particles do not undergo a sufficient number of collisions to maintain the equilibrium velocity distribution. The transition between continuum and rarefied flows can be drawn from the Knudsen number (Kn), defined as the ratio between the molecular mean free path and a characteristic physical length scale of the flow. In this work the validity of the continuum approach for thermal plasma flow simulation under very low pressure conditions was investigated. The methodology consisted in i) comparing the flow fields predicted with the ANSYS Fluent CFD code and with the Direct Simulation Monte Carlo (DSMC) code SPARTA which utilizes a particle-based kinetic method and ii) evaluating the Knudsen number to determine a breakdown criterion. An attempt of explanations of the sources of discrepancy between the results obtained by the two approaches was made. |