Prediction of Gloss in Plastic Injection Parts Based on 3D Surface Roughness from Virtual Machining with Artificial Neural Networks

Abstract

The gloss is one of the most important characteristics of plastic injection molding parts, whereby the molding process needs to consider the influence of the three main factors such as the surface roughness of the cavity, chemical properties of the plastic, and injection parameters. The surface roughness of the plastic injection mold was considered for the gloss that occurs with the parts of the plastic injection processes. Therefore, the objective of this research is to predict the gloss for plastic injection parts based on the artificial neural network method from the input parameters of 3D surface roughness from virtual machining on a 3-axis CNC machining center, and plastic injection parameters. The shape generation motions were mathematically described by combining 4 × 4 transformation matrices including the kinematic motion deviations, machining parameters, end milling ball nose geometries, and cutting force. The results of the research showed the prediction of gloss in plastic injection parts from a neural network model compared to the gloss value of the actually measured workpiece and response surface methodology combined with central composite design. It was found that the average error of the gloss was 2.36%. The proposed method provides us with a systematic method to estimate the gloss for plastic injection parts before producing the actual cavity mold, which leads to increased accuracy as well as efficiency in manufacturing plastic injection parts.