Plasma spray deposition on surfaces with curvature

• Main Author: Ba, Te
• Other Authors: Ng Heong Wah
• Format: Theses and Dissertations
• Language: English
• Published: 2012
• Subjects: DRNTU::Engineering::Materials::Ceramic materials; DRNTU::Engineering::Computer science and engineering::Computing methodologies::Simulation and modeling; DRNTU::Engineering::Manufacturing::Quality control
• Online Access: https://hdl.handle.net/10356/48646
• Institution: Nanyang Technological University

Plasma spraying has been widely used to apply coatings on industrial components due to its high deposition rate, wide range of material compositions, large component size, etc. However, the complexity of the components shape leads to the deficiency of the plasma spraying process. In addition, the coating profile continues to change, which will affect the subsequent spraying. Hitherto, experimental and numerical investigations on continuously changing profiles on complex surfaces are imperative. Curved surface is one of the types of the complex surfaces. More complex component surface also can be considered to comprise of multiple flat surfaces and curved surfaces. A semi-empirical methodology is proposed to predict the deposit formation on curved substrates at different positions, which is composed of three vital steps: 1) Computational fluid dynamics (CFD) analysis using FLUENT V6.03© to obtain the spatial distribution of particles and their corresponding in-flight parameters. 2) Droplet impacting behavior analysis to establish correlations for spread factor, aspect ratio and elongation factor with respect to the impact velocity and angle. 3) Modeling of deposit growth with time, with the data acquired in the particle parameters simulation and the correlations for splat morphologies. The spraying process is modeled in Fluent© as a three-dimensional steady state plasma plume by the volumetric heating method. Particles are introduced into and interact with the plasma flow by a one-way coupling method. The heat and momentum equations are solved to obtain the spatial distributions of the particles and their corresponding in-flight parameters. SprayWatch© on-line diagnostics system is used to measure the particles velocity and flight angle for the case with no substrate inclusion. The simulated results captured by a plane at the distance 80 mm agree well with the SprayWatch© measurements. From the simulation it is found that the substrate inclusion and shape (concave or convex) significantly influence the plasma flow fields in the vicinity of the substrate. The particles parameters remain relatively unaffected if their size is larger than a threshold value (10 μm). The droplet impacting behaviors analysis is divided into two aspects: a droplet impacts onto (1) the flat substrate under normal impact and (2) the curved substrate at different impact angles. Individual splats are captured by the substrates with the help of a shutter system to avoid excessive particles impacting on thin stainless steel coupons. Curved substrate is formed by wrapping the substrate coupon around a cylindrical holder. After spraying, the curved substrate is flattened for characterisation. The splats are characterized by scanning electronic microscopy (SEM) for flat substrate and optical microscopy for curved substrate. The splat 3D profiles are measured by the confocal imaging profiler. Simulations are carried out by Flow-3D® to complement the experimental work. Spread factor and aspect ratio of the simulation results fall in the range of their experimental counterparts, which validate the numerical models. Formulae of spread factor, aspect ratio and elongation factor are derived to simplify the splat geometries, with respect to the impact velocity and impact angle. Combining the simulation and experiments, droplet impacting behavior such as fringe elevation, tiny and long fingering (characteristics produced where streams of molten material jets out from the splat periphery) are analyzed. Long fingering is found to occur at the early stage of the impacting, while tiny fingering occurs after the droplet flattening ceases. Having obtained the correlations of splat shapes with the impact velocity and angles, together with the particle parameters from the FLUENT© simulation, the deposition code written in C++ is implemented to predict the deposit growth procedure. The deposit profile is updated at preset time steps, which makes the simulation more realistic. The prediction reasonably mimics the deposit growth with time. Good agreement of the peak deposit thickness is found between the simulation and SprayWatch© measurements.