The complexity of many components being coated in the aircraft industry today makes the traditional trial and error approach to obtain uniform coatings inadequate. To reduce programming time and further increase process accuracy a more systematic approach to develop robot trajectories is needed. In earlier work, a mathematical model was developed to predict coating thickness for thermal spray deposition on rotating objects with rotationally invariant surfaces. The model allows for varying spray distance and spray direction but is simple enough to give very short simulation times. An iterative method for robot feed optimization to obtain uniform coatings was also proposed. Currently, the use of the model in engineering practice is being evaluated. A MATLAB implementation of the model has been integrated with a commercial off-line programming system, giving a powerful and efficient tool to predict and optimize coating thickness. Simulations and experimental verifications are presented for two zirconia plasma sprayed parts.

The industrial flame spraying process has been analyzed by three-dimensional Computational Fluid Dynamics (CFD) simulation. The actual process is employed at the Volvo Aero Corporation for coating of fan and compressor housings. It involves the Metco 6P gun where the fuel, a mixture of acetylene and oxygen, flows through a ring of 16 orifices, while the coating material, a powder of nickel-covered bentonite, is sprayed through the flame with a stream of argon as a carrier gas by a central orifice. The gas flow was simulated as a multi-component chemically reacting incompressible flow. The standard, two equations, k-e turbulence model was employed for the turbulent flow field. The reaction rates appeared as source terms in the species transport equations. They were computed from the contributions of the Arrhenius rate expressions and the Magnussen and Hjertager eddy dissipation model. The particles were modeled using a Lagrangian particle spray model. In spite of the complexity of the system, the complex geometry and the numerous chemical reactions, the simulations produced fairly good agreement with experimental measurements. The powder size distribution was found to play a critical role in the amount of unmelted fraction of particles. The modeling approach seems to give a realistic description of the physical phenomena involved in flame spraying, albeit some model refinement is needed.