isat_ffd is a reduced order model trained by FFD simulations that can be used for fast predictions of airflow in the building. For the detailed performance evaluation of isat_ffd, one should refer to the paper titled Fast and Self-Learning Indoor Airflow Simulation Based on In Situ Adaptive Tabulation.
ffd, which is the abbreviation of Fast Fluid Dynamics, is parallelized in OpenCL to run full-scale simulations on multi-core devices, and generates training data for the reduced order model isat. For the performance of ffd parallelized in OpenCL, one can refer to the paper titled A Systematic Evaluation of Accelerating Indoor Airflow Simulations Using Cross Platform Parallel Computing
isat was originally developed by Professor Stephen B. Pope at the Cornell University to efficiently simulate the reacting flow with detailed chemistry. For more information regarding isat, one can go to the webpage
TBA
TBA
Wei Tian, [email protected]
This research was supported by the U.S. Department of Energy under Contract No. DE-EE0007688.