Quality evaluation method by simple 3D measurement of machine parts made from FDM 3D printers for educational use

Abstract

In recent years, commercial mass production of mechanical parts has been realized using powder 3D printers. However, the FDM (Fused Deposition Modeling) type 3D printers used in industrial education are often used only to roughly confirm the actual shape. This is due to suffering from decrease in molding accuracy owing to such as distortion caused by the thermal expansion characteristics of the material and the effects of the relative humidity. In other words, the problem is that machine parts designed with 3D CAD often do not faithfully reproduce the actual shape. Therefore, evaluation of molding accuracy is necessary to ensure the function of the machine itself in case that the machine parts created by the FDM type 3D printer are combined to produce a machine. In this paper, we propose a method to evaluate the quality of mechanical parts by means of a machine learning that uses the results of simple shape measurements and strength tests.