Microstructure evolution of Al 1050 in cold spray process using dislocation-based model and machine learning algorithms

Predicting microstructure evolution during the cold spray process remains a major challenge, particularly in relation to dislocation density and grain refinement under varying process parameters. In this study, the dislocation density and grain size parameters and microstructure evolution during the cold spray process of Aluminium 1050 were investigated using a numerical approach – dislocation-based model. To this aim, a total of 80 simulations were conducted, varying particle size, velocity, and particle count to explore their effects on the final microstructure. Subsequently, a Multi-Layer Perceptron (MLP) with the Levenberg-Marquardt algorithm were utilised to further analyse the simulation data. According to the results, it can be concluded that an increase in critical parameters, including particle size, velocity, and particle count, results in a corresponding increase in dislocation density also smaller grain sizes are correlated with larger particle sizes. Due to high dislocation density at the interface between the particle and the substrate, the grain size is 141 nm, which is the minimum value. Furthermore, under conditions with more particles, larger particle sizes, and higher particle velocity, the dislocation density reaches zero at a depth of 80 µm from the surface. Dislocation density has been obtained maximum value at 15 particles and 850 m/s, 7.32 × 108 mm−2.