Recognition of Machining Features Based on HSPCE Decomposition, Feature Composition, and Process Centered Classification

Abstract

A comprehensive set of computational procedures have been developed for transforming a geometric model based on design features to alternative sets of machining features. In the paper we only discuss a sub-set of these procedures; those related to machining feature recognition. There are three phases in the recognition process: 1) orthonormal decomposition of the removal volumes based on half space partitioning at concave edges (HSPCE) to determine regions of interaction and to recover portions of features lost by interaction 2) concatenation of decomposed cells into candidate machining features based on cycles in cell adjacency graphs, 3) classification of these volumes with respect to accessibility and tool motion for subsequent feasible process selection. This strategy is process centered—it recognizes all features that can be produced by common machining operations in a uniform way. It is not restricted to a few predefined features that each require specific, predefined rules. It deals with many types of feature interactions with one general purpose algorithm and produces alternative sets of machining feature sequences. Several measures have been incorporated to reduce computational complexity.