Classification of materials for selective laser melting by laser-induced breakdown spectroscopy

In the present work, we introduce a possibility to improve the rapid prototyping process of selective laser melting (SLM) using laser-induced breakdown spectroscopy (LIBS) which provides a material analysis. SLM uses many disparate materials for manufacturing of parts. The elemental composition of raw materials and constructed parts is obtained from a characteristic spectrum, which is a result of LIBS measurement. We compared a high-end LIBS instrumentation with a low-cost one; the latter could be easily implemented to a SLM device. The measured data were processed using multivariate data analysis algorithms. First, the principal component analysis was employed for a visualization and dimensionality reduction. Second, the reduced data set was classified using support vector machines. Moreover, we have suggested a procedure for an automatized classification of materials and parts during the SLM process without any supervision of a spectroscopy-specialist.