Multi-objective optimization of HVAF-sprayed Fe-based amorphous alloy coatings via machine learning for superior corrosion resistance

Here we studied the effect of high-velocity air fuel (HVAF) spraying parameters, i.e., air pressure, traverse velocity, power delivery rate, and spray distance, on the porosity and crystallization fraction of Fe₄₀Ni₇Cr₁₇Mo₁₂C₁₃B₉Y₂ (at%) amorphous alloy coatings. An inherent conflict between the porosity and crystallization fraction of amorphous alloy coatings was identified, making it challenging to minimize both simultaneously. However, electrochemical measurements indicate that achieving superior corrosion resistance in the coatings requires both low porosity and crystallization fraction. Traditional orthogonal experiments, which focus on individual parameters, are insufficient to achieve the goal. To address the issue, hybrid models integrating backpropagation neural network (BPNN) with intelligent optimization algorithms such as particle swarm optimization (PSO), sparrow search algorithm (SSA), and genetic algorithm (GA) were employed. Among these models, GA-BPNN demonstrated the excellent predictive performance for porosity and crystallization fraction. By coupling the GA-BPNN model with the Non-dominated Sorting Genetic Algorithm II (NSGA-II) multi-objective optimization algorithm, a set of Pareto-optimal solutions was obtained. The optimal HVAF spraying parameters were determined as follows: air pressure of 88.5 psi, traverse velocity of 956.8 mm/s, power delivery rate of 3.8 rpm, and spray distance of 383.7 mm. Coatings prepared under these optimized conditions exhibited low porosity and crystallization fraction values of 0.66 % and 1.00 %, respectively, with prediction errors within 5 %. Electrochemical tests demonstrated that the optimized Fe-based amorphous alloy coating exhibited remarkable corrosion resistance compared with conventional alloys such as GH3536 and 316 L.