Prediction of Surface Roughness in Wire Arc Additive Manufactured Components Using Machine Learning

This chapter presents investigations on surface quality of beads made by super duplex stainless steels (SDSS ER2594) material by wire arc additive manufacturing (WAAM) considering the effects of input factors, i.e., wire feed rate (f), torch travel speed (s), voltage (V) and gas flow rate (g). The SDSS wire material was deposited using GMAW-based WAAM by varying the process parameters, and surface roughness values of the accumulated material were measured. Artificial neural network-based machine learning method has been employed to model the input parameters with corresponding surface roughness of the deposited beads. The built equation could estimate the surface roughness with mean absolute deviations of 22% when compared with the experimental data.