Feature based methods for machining process planning of automotive powertrain components.

Abstract

Automatic NC code generation from manually created high-level process plan is at the core of computer-aided process planning systems for job shops. Methods for automatic creation of the high-level plan are recently being incorporated using the techniques for feature extraction and generative planning. In the future we see incorporation of part retrieval techniques to provide combined benefits of both generative and variant methods of planning. We also see incorporation of solutions for problems that are specific to planning for high volume production such as modeling of intermediate geometry for in-process part inspection. In this thesis, we present a method for modeling of intermediate geometry of automotive powertrain components and methods for feature extraction and part retrieval for non-standard parts of machining systems on which automotive components are produced. Interacting features pose an important challenge to feature extraction. Previous methods have addressed this by employing a two-step method of decomposition and composition. By contrast, we present a method for decomposition in a single-step using face visibility, face classification and machining relevant primitives defined for each face class. Results for decomposition for a selected set of parts are shown. For the application of part retrieval for machining process planning, it is important to compare parts using machining relevant shape characteristics. Previous methods have used feature intersection graphs and feature types. Equally important are the spatial and dimensional relationships amongst features that influence the setups and tools. We present shape characteristics for comparing these using sets, vectors and arrays and show results of sorting on a sample set of parts using different query parts. The current techniques for modeling intermediate parts use cutter swept volumes as volumetric features, which are subtracted sequentially from the initial workpiece. However, the intermediate geometry at the end of each operation is independent of the cutter geometry and path. We present a method for modeling volumetric machining features by selectively offsetting and extruding face boundary. Results for features on an automotive cylinder block are shown. In the future computer-aided process planning systems, we see incorporation of solutions for other important problems such as tolerance stack up analysis.