Neural Computing of Erosion Assessment in Al-20TiO2 HVOF Thermal Spray Coating

Abstract

Abstract Stainless steel (SS) 316L is widely used for hydraulic machinery of ash disposal slurry pump. In this work, the Al-20TiO2 coating powders were sprayed on SS316L materials using HVOF technique. The various properties such as erosion resistance, microhardness, microstructure, roughness, etc. were tested during the experiments. A pot tester was used to examine the rate of erosion. At an impact angle of 60 degrees, Al-20TiO2 coatings were found to erode the most. The neural computing was performed by using the artificial neural network model (ANN). Present ANN model produced the Pearson coefficient (R) of 0.99903, 0.99301, and 0.99194 respectively for training, validation and testing. The overall R value of model was found as 0.99686. Microscopically, Al-20TiO2 demonstrated semi-brittle erosion behavior. Craters and ductile fractures were the most common erosion wear mechanisms detected on the Al-20TiO2 coating, indicating that this material had semi-ductile properties.