Energy simulation of the fused deposition modeling process using machine learning approach

Abstract

Fused deposition modeling (FDM) is an additive manufacturing process using a nozzle to squeeze filaments of thermoplastic materials to create layers. At current research state, the energy performance evaluation of FDM processes is mainly realized by means of experimental methods, and a simulation-based approach is still absent and called for. As a promising tool for analyzing and constructing data, machine learning approaches have been widely implemented in energy performance analysis of different engineering areas but not FDM. Hence, this paper aims at this research gap and introduces the application of random forest algorithm, which is a typical inductive machine learning approach, for predicting the energy consumption as well as the power curve imitation of a desktop FDM system. In addition, an experimental validation is carried out.