The present paper presents an intelligent automation of the electron beam physical vapour deposition (EBPVD) process to achieve high quality and cost efficient coatings for low volume part production, using realtime feedback control. A computational model of EBPVD for predicting coating thickness is used with an optimisation heuristic for reducing coating thickness variance and feedback control approaches for substrate temperature control and melt pool control. The computational model can be readily generated using a standard computer aided design (CAD) model of the workpiece, which makes the method applicable to workpieces with complex three-dimensional geometry. Based on this model, an optimisation heuristic for the EBPVD process is developed to control workpiece motion systematically with the objective of reducing coating thickness variance, i.e. providing a uniform coating. These computational developments are illustrated using a simulation of a turbine blade coating in which the coating thickness variance is reduced significantly. Process level intelligence is incorporated using realtime feedback control for substrate temperature and melt pool control using an open architecture control system. Results using thermocouple based temperature control and realtime vision for melt pool control are presented. Video images of the melt pool are analysed on a block by block basis, using a technique to identify critical regions of the melt pool. Simulation results demonstrate the feasibility of automating the electron gun beam steering sequence. The proposed methods offer the prospect of eliminating dedicated tooling/fixtures and improving the cost effectiveness of the process, especially for low volume production.
Read full abstract