GPU-Accelerated Multiphase Eulerian-Lagrangian Solver with Adaptive Mesh Refinement

Award Information
Department of Energy
Award Year:
Phase I
Agency Tracking Number:
Solicitation Year:
Solicitation Topic Code:
Solicitation Number:
Small Business Information
CFD Research Corp.
215 Wynn Drive, NW, Flr 5, Huntsville, AL, 35805-1926
Hubzone Owned:
Minority Owned:
Woman Owned:
Principal Investigator:
Vladimir Kolobov
(256) 726-4800
Business Contact:
Deborah Phipps
(256) 726-4884
Research Institution:

Computer simulations of multi-phase flows are important for many applications. In the proposed work, we plan to advance Eulerian-Lagrangian methods by using Adaptive Mesh Refinement for multi-scale resolutions, and CPU-GPU parallelization to drastically reduce computing time. Eulerian-Lagrangian methods treat the gas/liquid phase as a continuum (Eulerian) fluid, and solid particulates as discrete Lagrangian particles. Graphics Processing Units (GPU) have demonstrated the ability to accelerate Lagrangian models by two orders of magnitude compared to the traditional Central Processing Unit (CPU) computations. The goal of this project is to develop a modern computational tool for parallel simulations of multiphase flows on heterogeneous multi-core CPU-GPU systems. In Phase I, we plan to demonstrate GPU-accelerated computations of gas flows with Lagrangian particles on multi-GPU systems. The computational algorithms will be developed to expose fine-grained parallelism of Lagrangian methods and octree data structures. During Phase II, we will add dense spray models, particle- particle interactions, and gasification and combustion reactions. The capabilities of the new tool will be demonstrated for practical simulations of reacting fluid-solid multiphase flows and devices used in fossil energy applications. Develop parallel Eulerian-Lagrangian models for simulations of reacting gas-solid flows on multiple graphical processor units (GPUs) to obtain orders of magnitude speed-up compared to the current state-of-the-art Eulerian-Eulerian parallel models without loss of fidelity. Combine computing power of modern hardware with state-of-the-art computational algorithms and data structures that expose the fine-grained parallelism inherent in Largangian methods to achieve maximal utilization of CPU-GPU processing. Utilize adaptive Cartesian mesh methodology to address the challenge of multiple spatial and temporal scales. The proposed advances in the Eulerian-Lagrangian modeling will have great impact throughout science and engineering, by enabling the solution of previously unsolvable problems. Commercial industrial applications include chemical engineering, material processing, combustion, and power generation. The DoE will be able to use the new capabilities for the exploration, analysis, design, virtual prototyping, and troubleshooting of multiphase flow devices and dense-phase systems such as fluidized beds used in fossil energy applications.

* information listed above is at the time of submission.

Agency Micro-sites

SBA logo

Department of Agriculture logo

Department of Commerce logo

Department of Defense logo

Department of Education logo

Department of Energy logo

Department of Health and Human Services logo

Department of Homeland Security logo

Department of Transportation logo

Enviromental Protection Agency logo

National Aeronautics and Space Administration logo

National Science Foundation logo
US Flag An Official Website of the United States Government