Artificial neural networks applied to the analysis of performance and wear resistance of binary coatings Cr3C237WC18M and WC20Cr3C27Ni

The wear process, of hydroelectric turbine blades, is a complex and multifactorial phenomenon where coatings applied by thermal spray are often used. In this work, artificial neural networks were used to model the resistance against cavitation and slurry erosion wear of binary carbide coatings Cr3C237WC18M and WC20Cr3C27Ni, sprayed by HVOF processes. The influence of fuel (kerosene and hydrogen), fuel flow and stoichiometric ratios were evaluated. The speed and temperature of the powder particles were measured. The mechanical properties of the coatings: microhardness and fracture toughness, determined by Vickers indentation method, were considered. The thickness and porosity of the coatings were also evaluated as well as cavitation and slurry erosion tests (90 degrees erodent impact angle). Mass loss and wear rates were determined for each binary coating under the experimental conditions. Regardless of the type of coatings, the significant influence of the type of fuel and stoichiometry ratio on the cavitation and erosion wear rates was demonstrated. With the design of the Artificial Neural Network (ANN), it was possible to analyse 10 input variables, and the interaction between them. The resulting eight outputs produced a robust model. This result allows the prediction of thickness, porosity, microhardness, fracture toughness, and wear resistance of Cr3C237WC18M and WC20Cr3C27Ni binary coatings sprayed by the HVOF process.

Automated summary

This study utilized artificial neural networks to model the wear resistance of coatings, showing significant influence of fuel type and stoichiometry ratio on wear rates. The model allows for prediction of coating thickness, porosity, microhardness, fracture toughness, and wear resistance.