Integração de imagens por ressonância magnética na dinâmica dos fluidos computacional: aplicações em petrofísica e neurociência

Abstract

Computational fluid dynamics (CFD) is widely used to study the flow of fluids in complex systems. Magnetic resonance imaging (MRI) is a consolidated technique in clinical routine and in applications in medicine. However, despite MRI being able to provide diverse information for simulations, combining CFD and MRI is still a challenge. The objective of this thesis is to develop and apply MRI and CFD in a petrophysics and a neuroscience problem. For the petrophysics study, we used MRI of wormholes in carbonatic rocks obtained in a pre-clinical magnet. For the neuroscience study, we used angiography and perfusion data from cerebral arteries of healthy subjects and subjects with moyamoya disease. Both data were acquired in clinical systems. The processing and simulation approaches were different for each problem. To study the wormholes we used image processing and three-dimensional reconstruction algorithms to generate models and then calculated the simulations in the OpenFOAM software. For the study of cerebral arteries, we used the software SimVascular to generate models and calculate the simulations. Specifically for subjects with moyamoya disease, we compared velocity data acquired by MRI with the simulations through images generated by the Paraview software with the resample filter and a python script. The results of the simulations in wormholes were compared with experimental pressure measurements and were within the experimental deviation and the mesh independence analysis, despite some exceptions. The results of the simulations in cerebral arteries were compared with the perfusion division. The perfusion division is better correlated with the healthy subject than with the subjects with moyamoya disease. In future studies with wormholes, transient and multiphase applications can be considered. For cerebral arteries, it is still necessary to develop new outlet boundary conditions and consider the use of selective perfusion acquisitions. In addition, the simulations translated into resample images used in the analysis of subjects with moyamoya disease can help in the development of MRI techniques. The contribution of this study was to establish the basis for more complex simulations and to define MRI as an essential complement for CFD.