Emergent Structure Detection for Multi-Axis Machining

Abstract

This paper examines the phenomenon of emergent structures that occur in the transient stock material during multi-axis rough machining from a plurality of fixed orientations. Taking the form of thin webs and strings, emergent structures are stock material conditions that can lead to catastrophic failure during machining, even when tool path verification is successful. We begin by discussing the motivation for use of fixed orientations in multi-axis machining using multiple automated setups via rotary axes, which enables fast processing and ‘first part correct’ machining. Next, we demonstrate how unintended emergent structures occur in this paradigm of machining and can lead to catastrophic failure of the tool or work piece. Our original work focuses on the problem of geometric detection of these structures during process planning and prior to tool path planning, to the end of altogether avoiding emergent structure formation. To quickly simulate the machining process, we present an object-space method for determining the transient state of stock material based on the inverse tool offset. To identify emergent structures within this transient stock state, we propose a metric based on the medial axis transformation. Finally, we present our implementation of these methods and demonstrate realtime computation appropriate for an optimization scheme to eliminate emergent structures. Our methods provide consistent and logical results, as demonstrated with several freeform component examples. This work enables the development of robust algorithms for autonomous tool path planning and machining in multi-axis environments.