An Immersive Environment for Exploration of CUBIT Meshes

Proceedings, 6th International Meshing Roundtable, Sandia National Laboratories, pp.47-48, October 1997

MESHING
RESEARCH
CORNER

Abstract
This work explores the use of immersive technologies, such as those used in synthetic virtual environments (commonly referred to as virtual reality, or VR), in enhancing the mesh-generation process for 3-dimensional (3D) models. The objective is to enable interaction with meshes, either complete or under construction, in a highly visual and intuitive manner, allowing a much greater understanding of the mesh, as well as allowing interactive feedback to the mesh generation process. This capability is particularly useful for examining mesh quality and mesh topology in certain spatial localities where the automated mesher may have produced certain complexities or anomalies.

This work partners with Sandia's CUBIT mesh generation research, and explores the use of capabilities developed for Sandia's Multi-dimensional User-oriented Synthetic Environment (MUSE). We have successfully implemented a prototype system for viewing and understanding CUBIT meshes which led one CUBIT developer to comment: "A capability like this on the desktop would increase our productivity by a factor of 4 or 5 for looking at meshes." The system has been linked to CUBIT (using sockets), so, in addition to being able to import meshes from a file, it is possible to import meshes directly from CUBIT, while the meshing system is actively generating a mesh. This further enables mesh editing and/or other feedback while the meshing algorithm is running - to date, we have only begun to explore this aspect by demonstrating the ability to move nodes spatially in the mesh. While the prototype system has been demonstrated with a high-end, equipment-rich VR system, the software is also running on a lower-end desktop system for which we are currently working to enable certain VR features, such as stereo, head-tracking, and use of a space-ball for input. Other current work includes an effort to enhance the system for use with relatively large meshes - the initial prototype worked very well for very small meshes, but experimentation with certain real application meshes has mandated a need to investigate other approaches, particularly of a hierarchical or selective nature, for allowing high-performance manipulation and exploration of such meshes.

Functionality of the prototype system includes: the ability to differentiate between nodes in the mesh which belong to different numbers of elements; the ability to turn on/off mesh edges (i.e. to see only nodes, or nodes and edges); the ability to display node identifiers with the nodes; the ability to highlight specific elements in the mesh for visual scrutiny; the ability to "tether" to (i.e. visually focus attention to) a specific node, which also highlights the edges of all elements which contain the node; the ability to grab and move nodes in the mesh; and the ability to interact with all of these features in a fully stereoscopic environment, supported by advanced human-computer interface capabilities. We are also currently implementing a capability which would enable CUBIT researchers working on the "WhiskerWeaving" algorithm to explore a resulting mesh together with associated sheet diagrams, highlighting certain primaldual relationships. To accommodate larger meshes, we provide a couple-of simplified ways to look at a global mesh space (colored bounding boxes for each element block or display of exterior faces only) together with mechanisms for pickin- -a certain locality for detailed display and scrutiny - for example, one approach uses a movable transparent sphere to select a spherical portion of the global mesh space for detailed display.

Our presentation will include a description of the prototype system, a discussion of lessons learned, and a video segment showing a live session of the prototype in use.

Download Full Paper (Postscript Format)