Surface roughness R prediction in Selective Laser Melting of 316L stainless steel by means of artificial intelligence inference

Abstract

Selective Laser Melting (SLM) is a widely used metal additive manufacturing process due to the possibility of elaborating complicated and customized tridimensional parts or components. This paper presents research on predicting surface roughness of 316L stainless steel manufactured SLM parts using the well-known multilayer perceptron (MLP) and an adaptive neuro-fuzzy inference system (ANFIS). Two models were adjusted to predict the top surface quality for different values of laser power, scanning speed, and hatch distance. The obtained results were evaluated and compared in order to ensure the goodness of fit of both techniques. The multilayer perceptron-based model has proved, to possess better predictive capability of the non-linear relationships of the SLM process. However, adequate results were also obtained with the adjusted ANFIS. The consistency of the presented models is also compared with previously published empirical formulations and discussed. As a final result, has been demonstrated that both fitted models outperform the previously published statistic-based approaches.