Next Generation Blast Simulation
Small Business Information
77 West 200 South, Suite 210, Salt Lake City, UT, 84101
Mgr. Computational Technologies
Mgr. Computational Technologies
AbstractOBJECTIVE: Develop a fast running Computational Fluid Dynamics (CFD) Finite Element code taking advantage of hardware and software technologies developed in the gaming industry. The CFD code will be used to model highly non-linear and dynamic external blast events in urban settings and internal blasts inside structures. DESCRIPTION: The Defense Threat Reduction Agency (DTRA) seeks proposals for development of the next generation high fidelity CFD codes for faster simulation of the blast environment. Existing codes used for this purpose are primarily written for computations on Central Processing Units (CPUs). The calculations are performed either in a serial or parallel fashion on single or parallel CPUs. At present a typical blast calculation for an urban environment or calculations for building collapse takes weeks to months of CPU run time. With ever increasing popularity of computer and TV games and animated movies, the gaming industry and chip manufacturers are investing heavily on improving the processing speed of the hardware and improving the software for life-like visualizations. For example, Graphic Processing Units (GPUs) in today's computers and gaming consoles are used for 50-70 times faster computations than CPUs. Other hardware improvements include the so called desktop massively parallel supercomputers with hundreds of energy efficient CPUs bundled into a desktop computer. New hardware consoles combined with innovative software allow users to interact with games in more natural and life-like fashion using handheld devices with motion sensors and touch-screen displays. To take advantage of these technologies, researchers in the academia have recently started experimenting with high fidelity finite element models written for processing on GPUs that have led to impressive speed up in computations. This solicitation is for development of a commercial fast running CFD code on GPU processors or desktop supercomputers. In the first two phases of the program a clear demonstration of the viability of the proposed method is required including mitigations of risks associated with limitations of gaming technologies in general purpose computing. Faster running codes with user friendly interfaces are important requirements in order to assess damage to Weapons of Mass Destruction (WMD) facilities or to plan for protection of personnel and mission critical equipment in a protected facility. PHASE I: The successful Phase I project should develop the proposed methodology in sufficient mathematical detail to show technical competency. At a minimum the Phase I work should clearly demonstrate the expected speed up or ease of use of the new methodology by working out an example problem. PHASE II: The successful Phase II project should develop a prototype CFD code and compare the results with legacy codes and experimental data. An important aspect of this phase of the program is to develop a detailed plan for commercializing the prototype for use by the government and the private sector. PHASE III DUAL USE APPLICATIONS: Potential dual use, non-defense applications of fast running CFDs are in weather modeling, modeling of wind loading on structures and more realistic and physics based computer and TV games. A Phase III project would develop a commercial version of the Phase II prototype CFD code to include user friendly Graphic User Interface with licensing for software releases, user manuals and training materials and worked out examples. REFERENCES: 1. D. Goddeke, R. Strzodka, and S. Turek, "Accelerating Double Precision FEM Simulations with GPUs," In Proceedings of ASIM 2005 - 18th Symposium on Simulation Technique, Sept. 2005. 2. GPGPU -- CFD-Wiki, the free CFD reference, http://www.cfd-online.com/Wiki/GPGPU. 3. Harris M.J., Fast Fluid Dynamics Simulation on the GPU, Chapter 38, Excerpted from GPU Gems: Programming Techniques, Tips, and Tricks for Real-Time Graphics. 4. Z. A. Taylor, M. Cheng, S. and Ourselin, "High Speed Non-linear Finite Element Analysis for Surgical Simulation Using Graphics Processing Units," IEEE Transactions on Medical Imaging, , May 2008, Vol. 27, Issue 5, pp 650-663.
* information listed above is at the time of submission.