Modelling of Ceramic Particle Impact Behaviour in Aerosol Deposition

Abstract

Aerosol deposition (AD) is a coating technique developed from cold spray coatings, eliminating the need for preheating and precompression devices by using a vacuum pump and chamber. Ceramic coatings enhance substrate properties, such as wearing and electrical resistance, but face challenges, such as low deposition efficiency (DE) and the need for precise control of film formation to achieve low porosity, strong bonding, and uniformity. These challenges arise from numerous independent variables affecting particle position, velocity, and temperature just before impact, influencing DE, adhesive strength, and film uniformity. This thesis aims to develop parametric relations for the three influential factors: particle position, velocity, and temperature just before impact, to control the deposition process. Previous numerical simulations provided insights into increasing particle impact velocity and improving uniformity, but the effect of substrate standoff distance remains unclear due to conflicting conclusions. Five independent variables: particle injection position, particle diameter, nozzle expansion ratio, nozzle divergent part length, and substrate standoff distance, are studied using computational fluid dynamics (CFD). An intrinsic relationship between gas and particle properties and the independent variables is developed to explain the governing flow physics. Additionally, an equivalent conversion relation was proposed to predict the velocity and temperature just before the impact of other material particles based on alumina data, verified with copper particles with high accuracy (relative errors within 1%). Since no reliable and efficient method exists for predicting particle impact properties, two surrogate models were proposed. An empirical model predicts particle impact position with a relative error of 6%. A rapid semi-analytical method combines analytical solutions within the nozzles and empirical equations from CFD results to predict particle velocity and temperature just before impact, verified with high-fidelity numerical simulation data, showing relative errors of 9.0%±5.5% for velocity and 4.6%±2.2% for temperature. This new method significantly reduces calculation time from several hours or days to seconds. Finally, a particle-sticking probability map is generated to describe the relationship between independent variables and DE. This map, based on the new models and a sticking model, predicts and maximizes DE with independent variables, demonstrating the usefulness of surrogate models for more accurate control of film formation in AD, focusing on maximizing DE and achieving uniform film thickness.

Date of Award	20 Jun 2024
Original language	English
Awarding Institution	The University of Manchester
Supervisor	Hector Iacovides (Supervisor) & Nick Bojdo (Supervisor)