Long-term oxidation of newly developed HIPIMS and PVD coatings with neural network prediction modelling

Three protective coatings (TiAlCr coating, TiAlN + Al2O3 and TiAlYN/CrN + Al2O3) were deposited on a Ti–45Al–8Nb alloy and exposed at 750 °C for oxidation test for 5000 h. The materials with Al2O3 top coat showed much better resistance than TiAlCr coating where mainly TiO2 and Al2O3 phases developed. The corrosion kinetics were determined by means of discontinuous gravimetry. The exposed samples were characterised using SEM, EDS and XRD. Artificial neural network was used to predict the kinetic behaviour of the exposed alloys after 2000–4000 h of exposure. The predicted kinetic after 5000 h shows good agreement with the experimental data.

► TiAlYN/CrN + Al2O3 had a better oxidation resistance than the TiAlCr coating.
► Interlayer thermal expansion mismatch is paramount in oxidation behaviour.
► Dynamic 3rd order LR NNs give good prediction results for the kinetics.
► The middle of nano-structured TiAlYN/CrN + Al2O3 coating was covered by Al2O3–Cr2O3.
► TiAlN + Al2O3 developed Al2O3 scale and underneath a layer of Al2O3/TiO2 formed.