Automatic Feature Recognition System for Prismatic Parts Using STEP-Based Feature Extraction and ANN-Based Pattern Recognition

Abstract

Feature technology is considered an essential tool for integrating design and manufacturing. Automatic feature recognition (AFR) has provided the greatest contribution to fully automated CAPP system development. After a brief review of approaches based on application of artificial neural networks (ANN) for solving major AFR problems, this paper presents an AFR system for prismatic parts with external extraction of geometric information out of a STEP CAD model, part representation based on B-rep coding and ANN-based pattern recognition.