Manufacturability evaluation of parts using descriptor-based machining feature recognition

Abstract

Manufacturability validation refers to activities to evaluate the manufacturing cost of parts, manufacturing methods, and manufacturing difficulty, and to determine whether parts can be manufactured from three-dimensional (3D) computer-aided design (CAD) models. To automate the manufacturability evaluation process, it is important to first recognize the machining features of an input 3D CAD model. Thereafter, various pieces of information required for manufacturability evaluation are extracted from the recognized features, and whether manufacturing can be realized according to the predefined manufacturability criteria is determined. This study proposes a method to evaluate manufacturability according to 19 criteria, after automatic recognition of the machining features of turning, milling, and drilling from boundary representation (B-rep) 3D CAD models, which are widely used in work sites. In addition, this study implements a prototype system, and discusses the results of an experiment involving recognition of the features of two test cases, and their manufacturability evaluation.