SBIR/STTR Research Award
Hybrid CPU-GPU Parallel Development of the Eulerian-Langrangian Barracuda Multiphase Program.
Award amount
$99,542Program and phase
SBIR · Phase IAward year
2010Award details
- Awarding agency
- Department of Energy
- Branch
- Not reported
- Recipient company
- Cpfd Software, Llc
- Proposal award date
- Not reported
- Contract end date
- 2011-03-18
- UEI
- Not reported
- Solicitation number
- DE-FOA-0000161
- Topic code
- 23 d
Research abstract
Computer-based simulations of fossil fuel processing plants can help make significant improvements to plant efficiency, as well as with procedures to reduce emissions, e.g. of CO2. The Barracuda engineering software is in wide spread usage for such applications at DOE Labs and with industry because of its demonstrated success for high-fidelity predictions of the multiphase flows. Simulation of these complex flows is very computer intensive and can require days or even weeks of computer time; however, use of new GPU parallel methodology holds the potential to reduce the compute time by several orders of magnitude. CPFD Software, LLC is the developer of the commercial Barracuda multiphase code, with its underlying data structure and numerical algorithms. Knowledge of the code structure, working in close concert with a recognized expert company in GPU penalization, will result in a successful development effort. It is emphasized that this combined approach of a specific Eulerian-Lagrangian model (Barracuda) and demonstrated expertise on GPU parallization (e.g. EMPhotonics or Tech-X) is required to achieve success. Phase 1 will demonstrate significant GPU speedup on a key subroutine(s) of Barracuda, and then Phase 2 will extend this proven methodology to the entire code. Commercial Applications and Benefits: The high-fidelity Eulerian-Lagrangian model for multiphase flow will be a seamless part of the commercial Barracuda engineering software. This software is used by DOE Laboratory researchers, and industrial companies including ExxonMobil, Alstom, Dow Chemicals and many others. The GPU parallization effort will enable extremely fast computer simulations that can help engineer an optimized process for fossil power production, gasification of coal and biomass, plus reduce greenhouse gas releases.
Technology keywords
Not reported
Awards with the same topic code
Verified by an exact topic code match. This is not a text-similarity match.
- Measurement of the Radial Electric Field in the Plasma Edge of the Lithium Tokamak Experiment
- A Novel Sorbent to Reduce CO2 Emissions from Existing Coal-Fired Power Plants
- A Novel Sorbent to Reduce CO2 Emissions from Existing Coal-Fired Power Plants
- Backside Contact Multijunction Solar Cells for Concentrator Applications
- Backside Contact Multijunction Solar Cells for Concentrator Applications