Modelling the EB-PVD thermal barrier coating process: Component clusters and shadow masks
Date published
Free to read from
Supervisor/s
Journal Title
Journal ISSN
Volume Title
Publisher
Department
Type
ISSN
Format
Citation
Abstract
Electron beam-physical vapour deposition (EB-PVD) is a commonly employed process for the production of thermal barrier coatings (TBCs), used in high performance applications such as gas turbines high-pressure aerofoil blades for the aerospace industry. Computer modelling can contribute to improved control of the coating process, important to support a continuous drive for increased efficiency. This paper considers two aspects associated with the EB-PVD coating of TBCs for commercial application: firstly, that clusters of blades are coated simultaneously in commercial coaters and, secondly, that these parts possess a complex geometry, such that shadow masks need to be taken into account. In this context, a computer model that calculates coating thickness distribution along the surface of different engine components, based on the analysis of the vapour deposition flux around complex geometries, is presented. To validate the predictive capability of the computer model two deposition trials were performed. Firstly, a cluster of components was simulated using three rotating cylinders, as a simple representation of coating multiple blades. Secondly, the effect of shadow masks was studied with an arrangement in which flat plates were welded, in the form of a U-shaped component, but with one side shorter than the other. The predicted results generated by the computer model compare favourably with those measured in the experimental runs presented. For the cluster of three cylinders, an error of 4% was obtained whilst the divergence was around 20% for the simulated shadow mask due to the fact that overall coating thickness was significantly reduced. In spite of this, the results obtained from the model were promising with respect to the degree of fit of the inverse square law. It is thought that a virtual source may be responsible for measurements being generally higher than those predicted by the model.