Computational Fluid Dynamics (CFD) analysis of thoracic aorta hemodynamics: a comparative study of spatial discretization methods

Abstract

[EN] Cardiovascular diseases, recognized as the leading global cause of death by the World Health Organization (WHO), emphasize the need for a comprehensive understanding of human hemodynamics. This collaboration between medical professionals and engineers has led to the advancement of numerical modeling techniques, aiming to provide a full description of patient-specific blood flow patterns. Such models prove valuable in the early diagnosis of pathologies, the identification of cardiovascular risk factors, and the monitoring of patient evolution both before and after surgical interventions. Computational Fluid Dynamics (CFD) holds significant potential in this regard, providing an accurate three-dimensional representation of patient hemodynamics. However, its effectiveness often relies on screening a large patient cohort to extract meaningful statistical population data, posing a challenge due to the associated high computational costs. Addressing this challenge involves drastic model optimization, where spatial discretization plays a crucial role in significantly enhancing accuracy and computational efficiency. This is particularly relevant in challenging environments, such as the cardiovascular apparatus, characterized by arbitrary geometries, intricate boundary conditions, and substantial fluid-structure interaction. In response to these complexities, this study aims to establish a CFD atlas for thoracic aorta hemodynamics within a small cohort of 15 patients, employing and comparing various spatial discretization approaches, all supported by a robust grid independence study based on the Richardson extrapolation method. One meshing approaches to compare entails the creation of an in-house, straightforward, and fully automated methodology based on a vessel coordinate system (VCS). This approach will enable the generation of high-quality adaptive structured meshes with complete control over quality parameters, including the identification of case-specific grid resolution requirements. Furthermore, the use of a customized VCS streamlines systematic comparisons of identical coordinate points among different patients, irrespective of their aortic geometry, thus facilitating inter-patient statistical analysis.