Neural computation analysis of alumina-titania wear resistance coating
Résumé
Pin-on-disc tests were performed on alumina–13 wt.% titania coatings obtained under several APS conditions. Friction coefficient data were analysed using artificial neural network. This permitted to predict parameter ranges for which good wear resistance is possible when varying each of the process parameters individually with respect to a reference condition. In this case, results suggest that large parameter ranges did not permit to obtain a significant friction coefficient variation which was mainly between 0.51 and 0.61. In addition, injection parameters and total plasma gas flow rate were the control factors.