SOLID MODELING FOR MOLD DESIGN AND MANUFACTURE (CADICAM, COMPUTER GRAPHICS)

This work is part of the Cornell Injection Molding Program which is aimed at the establishment of a scientific base for advancing the state-of-the-art of injection molding. The goal is to develop and implement an integrated CAD/CAM (Computer-Aided Design and Manufacture) system for mold design and manufacture. Two CAD/CAM applications of solid geometric modeling to molding are addressed in this Thesis.; For mold design, a hierarchical constructive solid geometry system called the TIPS-1 Parts Operation System (TIPS-POS) has been developed. TIPS-POS is based on the TIPS-1 geometric database. It extends the current modeling capability of TIPS-1 to include operations at the "Part level". Three types of set operators are provided in TIPS-POS to define either a more complex object based on existing TIPS-1 Parts or an assembly which could be an aggregate of these Parts. The regularized Boolean operators Union, Difference, and Intersection are available for constructing objects such as core- and cavity-plates "derived" from the geometric shape of the molded part. To facilitate the design of multiple-cavity molds, Translation and Rotation operators are included for duplication of geometry at different positions. In addition, an Assemble operator can be used to construct mold assemblies.; For mold machining, a system has been developed to simulate the NC (numerical control) milling of molds. The swept volume of the cutting tool is continuously subtracted from the workpiece to yield the "as machined" part. The resulting part-in-progress is dynamically displayed on a color raster terminal in near real-time for visual review. This is made possible by the development of a fast shading algorithm for solid objects. With such a system, potential errors to NC programming as well as collisions between the cutter and the NC machine can be detected. The mathematical basis for the volume swept by the motion of a solid is presented. The modeling of swept volumes can be useful for prediction of interference in other mechanisms as well.