I-Corps: Algorithms and software for resolving intersection problems in computer aided design

University Of Texas At Austin, Austin TX

Investigators

Abstract

The broader impact/commercial potential of this I-Corps project is that it allows users of Computer-Aided Design (CAD) systems to create product models that can be used directly in downstream applications. Current workflows require extensive manual repair of geometric models before they can be used for Computer-Aided Engineering/Manufacturing. Within the US automotive industry alone nearly $600M/year is spent in direct cost and lost productivity on mitigating CAD problems. Entire business units or job shops are employed to deal with this expensive problem by repairing or reworking gap-ridden geometric models. Commercializing the algorithm developed will result in improved productivity in the engineering design process, resulting in better products that are designed more quickly than is possible with current CAD technology. This I-Corps project focuses on development and commercialization of a novel software algorithm that resolves a ubiquitous problem in the Computer-Aided Design (CAD) industry, the gaps that arise at the intersections of surfaces in CAD models. The majority of modern geometric modeling relies on parametric freeform surfaces built as complex combinations of much simpler geometric shapes. However, the intersection of freeform surfaces in CAD currently results in highly approximated, incomplete solutions. The approach developed here is based on a thorough understanding of the parametric space underlying the geometry of most modern CAD models. The algorithm classifies and segments the intersection curve in parametric space, then reconstructs the intersecting surfaces such that their boundaries are guaranteed to lie on the intersection curve to the tolerance set by the user of the CAD system. The approach solves the problem in a way that produces objects in the same form as that of the original models, making them available in the modeler's native software package.