Leveraging feature generalization and decomposition to compute a well-connected midsurface. (January 2017)

• Record Type: Journal Article
• Authors: Kulkarni, Yogesh; Sahasrabudhe, Anil; Kale, Mukund
• Is Part Of: Engineering with computers. Volume 33:Number 1(2017)
• Page Start: 159
• Page End: 170
• Publication Date: 2017-01
• Subjects: CAD -- CAE -- Features -- Midsurface -- Sheet metal features -- Cellular decomposition; Engineering design -- Data processing -- Periodicals; Computer-aided design -- Periodicals; Conception technique -- Informatique -- Périodiques; Conception assistée par ordinateur -- Périodiques; Electronic journals; 620.00285
• DOI: 10.1007/s00366-016-0466-x
• Languages: English
• ISSNs: 0177-0667

Abstract

Abstract Computer-aided design (CAD) models of thin-walled parts, such as sheet metal or plastic parts, are often represented by their corresponding midsurfaces for computer-aided engineering (CAE) analysis. The reason being, 2D surface elements, such as "shell" elements, which need to be placed on the midsurface, provide fairly accurate results, while requiring far lesser computational resources time compared to the analysis using 3D solid elements. Existing approachesof midsurface computation are not reliable and robust. They result in ill-connected midsurfaces having missing patches, gaps, overlaps, etc. These errors need to be corrected, mostly by a manual and time-consuming process, requiring from hours to even days. Thus, an automatic and robust technique for computation of a well-connected midsurface is the need of the hour. This paper proposes an approachwhich, instead of working on the complex final solid shape, typically represented by B-rep (boundary representation), leverages feature information available in the modern CAD models for techniques such as defeaturing, generalization, and decomposition. Here, first, the model is defeaturedby removing irrelevant features, generating a simplified shape called "gross shape". The remaining features are then generalizedto their corresponding generic loft-feature equivalents. The model is then decomposed into sub-volumes, called "cells" having respective owner loft features. A graph is populated, with the cells at the graph nodes. The nodes are classified into midsurface patch-generating nodes (called 'solid cells' orsCell s) and interaction-resolving nodes (called 'interface cells' oriCell s). Using owner loft feature's parameters, sCell s compute their own midsurface patches. Using a generic logic, the patches then get connected appropriately in theiCell s, resulting in a well-connected midsurface. The efficacy of the approach is demonstrated by computing well-connected midsurfaces of various real-life sheet metal parts.