Process Parameter Optimization of Plasma Sprayed Nanostructured Al2O3-13 %TiO2 Coating Based on Genetic Algorithm

摘要

The process parameters of plasma sprayed nanostructured Al2O3-13 % TiO2 (mass fraction) coatings were optimized based on a genetic algorithm. A BP neural network was applied to compute the suitability of the genetic algorithm and a model was established. Four process parameters inputs were spraying distance, spraying electric current, primary gas pressure and secondary gas pressure. The bonding strength of the coating was the output. The network was trained by orthogonal test data. The BP neural network was tested by design of the orthogonal optimization and the margin of error was less than 5 %. It proved that the BP neural network could be applied to predict the bonding strength of plasma-sprayed nanostructured Al2O3-I3 % TiO2 (mass fraction) coatings. Process parameters of the coatings were optimized based on the genetic algorithm. The results showed that the maximal bonding strength of the coatings was 32.7 MPa. The process parameters obtained were: spraying distance of 107.51 mm, spraying electric current of 854.07A, primary gas pressure of 0.24 MPa and secondary gas pressure of 1.03 MPa. The results were superior to the design of the orthogonal optimization and provided a definite reference for selecting the best process parameters for plasma sprayed nanostructured Al2O3-I3 % TiCh (mass fraction) coatings.