Hadi Parvaz - Ph.D. student in manufacturing engineering, Tarbiat Modares University
Mohammad Javad Nategh - Associate professor in manufacturing engineering, Tarbiat Modares University

Automatic recognition of machining features is one of the key modules in automation procedure of computer aided applications. In this study, a hybrid graph and hint based approach has been used for recognition of interacting and advanced features. A novel concept of convexity degree has been introduced adding capabilities of recognizing features beyond the usual FR systems. The introduced system is impregnated with neutral format read/write capability enabling its independence from design software. It incorporates a multi-TAD tree-view representation mechanism capable of presenting the recognized features in TAD-sorted format. Using incorporated innovations, the developed FR software is capable of recognizing diversities of isolated, interacting (of nested, volumetric and adjacent types) and advanced (of fillet, open pocket, 2.5D passages and protrusion types) features. Also, the developed system is capable of recognizing common features in cylindrical (lathe features) models, too. Finally, diversities of 3D models having complex features have been used for testing system recognition capabilities.

CAD/CAPP/CAM, Feature recognition, Hybrid G&H approach, Multi-TAD representation