Mut, Fernando2008-09-03NO_RESTRIC2008-09-032008-09-03https://hdl.handle.net/1920/3284Image-based patient-specific CFD modeling of blood flows is important for better understanding the hemodynamics in cerebral aneurysms and their treatment. Some limitations of current methodologies have been identified, including 1) model size, 2) endovascular device modeling, 3) missing information and 4) extraction of relevant data. This work addressed the first two of these limitations. First, a Deflated Preconditioned Conjugate Gradients (DPCG) algorithm was developed to accelerate the computation of incompressible flows in the elongated geometries typically encountered in vascular models. This technique has enabled the modeling of the blood flow in complex arterial networks in a timely manner making these models practical for clinical purposes. Second, a methodology to model stented aneurysms on a patient-specific basis has been developed. This methodology has allowed the computation of the blood flow in cerebral aneurysms after the treatment with stents or other flow diverters. These two developments have extended the range of applicability of image-based CFD techniques applied to cerebral hemodynamics.en-USComputational HemodynamicsCerebral AneurysmsEndovascular StentingDeflated Conjugate GradientsDissertation