Keywords

crash, isogeometric analysis, CAD, Ricci flow, quadrilateral layout

Abstract

Isogeometric analysis (IGA) has attracted attention from academia and industry because of its high-fidelity results, ability to represent geometry exactly, and potential to streamline the engineering design-through-analysis process. However, one of the greatest challenges limiting the scope of IGA is the ability to rapidly convert CAD geometry into a set of splines suitable for engineering analysis¾particularly for a wide set of shapes of industrial relevance. In this presentation, we describe a new, mathematically rigorous, potentially automatable framework using Ricci flow and subsequent metric optimization through which surface geometries can be rebuilt as sets of watertight, analysis-suitable, boundary-conforming semistructured NURBS patches. Reconstruction of models using this framework leads to geometries that are immediately suitable for both isogeometric analysis and any additional design operations. We demonstrate the utility of this method by rebuilding a trimmed CAD model of the US Army’s DEVCOM Generic Hull vehicle as a set of watertight splines. We also reconstruct the chassis of a 1996 Dodge Neon from a finite element model as a set of analysis-suitable splines. Finally, we perform crash analysis of the Dodge Neon vehicle using the reconstructed spline spaces. It is shown that the Bezier elements of the isogeometric splines can be much coarser than the quadrilateral and triangular elements of the original finite element model while still producing analysis results with much higher accuracy than those of the original model. (Here, accuracy is determined by comparing analysis results with those on refined finite element meshes.) Ultimately, this work both showcases a high-fidelity isogeometric model reconstruction paradigm and also affirms the merit of isogeometric analysis in automotive crash simulations.

Original Publication Citation

Isogeometric Reconstruction and Crash Analysis of a 1996 Body-in-White Dodge Neon. 15th World Congress on Computational Mechanics 2022

Document Type

Peer-Reviewed Article

Publication Date

2022-07-31

Publisher

15th World Congress on Computational Mechanics

Language

English

College

Ira A. Fulton College of Engineering and Technology

Department

Civil and Environmental Engineering

University Standing at Time of Publication

Assistant Professor

Share

COinS