6/04/02 -- NAS Researcher Awarded for New Mesh Refinement Strategy
NAS Division researcher Michael Aftosmis and colleague Marsha Berger, professor and deputy director of Courant Institute, New York University, have received the American Institute of Aeronautics and Astronautics (AIAA) best paper award for 2002. Their paper, “Multilevel Error Estimation and Adaptive h-Refinement for Cartesian Meshes with Embedded Boundaries” describes the team’s development of new techniques for error estimation and adaptive refinement for computational fluid dynamics (CFD) solutions.

The award by the AIAA Fluid Dynamics Technical Committee will be presented at a special luncheon held in conjunction with the 32nd AIAA Fluid Dynamics Conference in St. Louis, Missouri on June 25.

Aftosmis and Berger’s new adaptive mesh refinement, or h-refinement technique, provides an optimal mesh refinement strategy for flow field adaptive CFD solutions. Meshes, or grids are used to calculate flow fields surrounding vehicles, such as the Space Shuttle. “The novelty in our work is that it removes the ambiguity of mesh generation, and identifies the single best solution,” says Aftosmis.

Several features put this new h-refinement strategy a cut above previous adaptive meshing techniques: “It is general enough that it can be applied to any type of unstructured or hierarchical mesh, and, since it is parameter-free, it can be fully automated,” explains Aftosmis. The new meshing strategy was also designated to remove the dependence on user skill, making it quick and easy to generate meshes.

CFD groups at Department of Energy laboratories, Michigan State University, and Massachusetts Institute of Technology have been experimenting with the new error estimator and refinement criteria. Aftosmis and Berger plan to integrate these error estimation and refinement techniques into their grid generation and solution software package, Cart3D this fall.

For more information, contact Aftosmis at aftosmis@nas.nasa.gov, (650) 604-4499.

5/14/02 -- Growing the Next Generation of Computers
NAS researchers Deepak Srivastava and Chris Henze have been awarded the Ames Research Center Director’s Discretionary grant for an innovative proposal encompassing nano-, bio-, and information technology. Their proposal for the “biomimetic simulation of signal transmission in nanotube-based dendritic networks” represents an innovative step forward in the division’s work on nanotechnology.

Inspired by the structure and operations of biological neural systems, the scientists plan to investigate the properties of a 3-D, tree-like architecture of carbon nanotubes, and prove its viability.

The field of molecular electronics using carbon nanotubes has progressed rapidly in recent years, resulting in several newly patented technologies, which experts believe will ultimately lead to the next generation of electronic computing devices. The National Science Foundation estimates that by 2015, the market for products manufactured using nanotechnology will reach $1 trillion, and in his latest budget proposal, President Bush included a 17 percent increase in spending on nanotechnology research.

Henze and Srivastava’s proposal of a 3-D, tree-like circuitry takes a unique approach to current research by private industry and academia, which has so far focused on 2-D models.
“We recently showed that three-terminal nanotubes grown as a y-junction serve as rectifier switches and have analog logic capacity,” says Srivastava, senior scientist and lead of computational nanotechnology investigations at NAS. “These properties inspired us to consider the concept of a system architecture similar to the biological neural system, but made up of synthetic material.”

Taking data derived from activating the nerve cells of crickets, Henze, a senior research scientist, can simulate their electronic behavior. Such simulations are a major tool in determining exactly how signals are transformed as they pass through the branch structures of neurons. “What’s becoming clear is that there’s a lot of non-linear processing going on in the complex branching structures,” says Henze, “and it has a lot to do with the geometry of the branching.” This observation, combined with Srivastava’s simulations in building branched nanotube systems, led to a confluence of ideas for the proposal that the scientists hope will ultimately demonstrate how signals could be transmitted and processed as they pass through a branched-tree system made of carbon nanotubes.

The Director’s Discretionary Fund is a well-established grant that provides seed money for innovative and high-risk research that, if successful, could lead to fundamental scientific breakthroughs. The NAS scientists hope that showing a high level of functionality will inspire other groups to try growing the tree and eventually lead to very complex sensing and computing applications.

For more information, contact Deepak Srivastava at deepak@nas.nasa.gov, 650-604-3486.

4/03/02 -- Barth Appointed to Springer-Verlag Editorial Board
Timothy Barth, a computer scientist in the NASA Advanced Supercomputing (NAS) Division’s Physics and Simulation Modeling Office, has been appointed to the editorial board of a book series published by prestigious international scientific publisher Springer-Verlag.

Established in 1997, The series, “Lecture Notes in Computational Science and Engineering,” is a part of the publisher’s mathematics division. Barth has been a contributing author to two editions of the lecture series. The most recent, released in January, is devoted to the topic of multi-scale and multi-resolution methods. All editions are in some form related to computational science and engineering (CSE), which is the application of applied mathematics and computations.

“Springer is really into CSE — they are embracing this as a new subject area, even though it’s not a very focused subject, it’s all applied mathematics with an emphasis on computations,” Barth explained.

His duties on the editorial board include approving new books for the series and providing input to the senior editor about potential authors and new topics. For additional information about the computational science and engineering series, contact Tim Barth at barth@nas.nasa.gov.