Grant: Discovery Grants Program - Group
Université de Montréal
$29,000.00
GPU implementation of lattice Boltzmann method for time-resolved three-dimensional flow with complex and moving boundary: in-vitro and in-vivo experimental validation
EF 2012-13
Résumé
Context: Finite element method (FEM) is widely applied to manifold areas of computational fluid dynamics (CFD). FEM, however, may be limited in the presence of time-dependent fluid geometries. Dealing with large deformations requires particular procedures like remeshing that may increase computation time considerably and introduce additional inaccuracy into the solution. The so-called Lattice Boltzmann method (LBM) can provide accurate and stable simulations even with extremely deformable geometries.
Objective: The first objective is to develop the LBM for simulating blood flow and vortex dynamics inside the left heart. The simulations will be validated with volumetric three-dimensional measurements issued from particle image velocimetry (PIV) and phase-contrast magnetic resonance imaging (PC-MRI). The second objective is to accelerate the LBM by using Graphics Processing Units (GPUs). The goal is to obtain very fast simulations by building up algorithmic strategies to implement the LBM on GPUs efficiently.
Methods: The coding and validation phases of the LBM algorithm will be first achieved on a CPU-based framework. In vitro measurements of intracardiac velocities will be performed in a pulse duplicator system (ViVitro Labs Inc) using volumetric three-component particle velocimetry (V3V system, TSI Inc). In vivo blood velocity acquisitions will be also completed in volunteers using PC-MRI with a 1.5 Tesla scanner. The LBM-derived velocity fields will be compared with those measured in vitro and in vivo. After the completion of the validation step, the LBM algorithm will be parallelized on GPUs in order to minimize the computation time and open the door to real-time simulations.
Discussion: Implementation and validation of CFD in highly deformable flow geometries is still a persistent challenge. Such complex flows can be simulated using the LBM method. Our group possesses all the knowledge, skill and state-of-the-art tools for carrying the present research program through to a successful conclusion.
Programme
Discovery Grants Program - Group
Bénéficiaire
Emplacement
Quebec, Canada
Type de bénéficiaire
Academia
Type
Grant
Source
NSERC's Awards Data
Tel que publié
Application id
386636-2010
Name nom
Kauffmann, Claude
Department département
Radiologie, radio-oncologie et médecine nucléaire
Organization id
63
Institution établissement
Université de Montréal
Province en
Québec
Province fr
Québec
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2012
Competition year année de concours
2010
Award amount
29000
Program id
RGPGP
Program name en
Discovery Grants Program - Group
Program name fr
Programme de subventions à la découverte - équipe
Group en
Discovery grants programs
Group fr
Programmes de subventions à la découverte
Committee code
1512
Committee name en
Mechanical Engineering
Committee name fr
Génie mécanique
Area of application code
801
Area of application group en
Information and communication services
Area of application group fr
Services d'information et de télécommunications
Area of application en
Computer software
Area of application fr
Logiciel
Research subject code
2200
Research subject group en
Fluid mechanics
Research subject group fr
Mécanique des fluides
Research subject en
Fluid mechanics
Research subject fr
Mécanique des fluides
Partie
3
Nb partie
5
Application title
GPU implementation of lattice Boltzmann method for time-resolved three-dimensional flow with complex and moving boundary: in-vitro and in-vivo experimental validation
Keyword
NSERC
Application summary
Context: Finite element method (FEM) is widely applied to manifold areas of computational fluid dynamics (CFD). FEM, however, may be limited in the presence of time-dependent fluid geometries. Dealing with large deformations requires particular procedures like remeshing that may increase computation time considerably and introduce additional inaccuracy into the solution. The so-called Lattice Boltzmann method (LBM) can provide accurate and stable simulations even with extremely deformable geometries.
Objective: The first objective is to develop the LBM for simulating blood flow and vortex dynamics inside the left heart. The simulations will be validated with volumetric three-dimensional measurements issued from particle image velocimetry (PIV) and phase-contrast magnetic resonance imaging (PC-MRI). The second objective is to accelerate the LBM by using Graphics Processing Units (GPUs). The goal is to obtain very fast simulations by building up algorithmic strategies to implement the LBM on GPUs efficiently.
Methods: The coding and validation phases of the LBM algorithm will be first achieved on a CPU-based framework. In vitro measurements of intracardiac velocities will be performed in a pulse duplicator system (ViVitro Labs Inc) using volumetric three-component particle velocimetry (V3V system, TSI Inc). In vivo blood velocity acquisitions will be also completed in volunteers using PC-MRI with a 1.5 Tesla scanner. The LBM-derived velocity fields will be compared with those measured in vitro and in vivo. After the completion of the validation step, the LBM algorithm will be parallelized on GPUs in order to minimize the computation time and open the door to real-time simulations.
Discussion: Implementation and validation of CFD in highly deformable flow geometries is still a persistent challenge. Such complex flows can be simulated using the LBM method. Our group possesses all the knowledge, skill and state-of-the-art tools for carrying the present research program through to a successful conclusion.