A reduced-order model of a patient-specific cerebral aneurysm for rapid evaluation and treatment planning

Abstract

We discuss a computationally-efficient numerical Reduced Order Model (ROM) for patient-specific cerebral aneurysms, which can be used to examine the rupture-related hemodynamic parameters over a range of relevant physiological flow parameters, rapidly. The ROMs were derived using a Proper Orthogonal Decomposition (POD) technique, which also took advantage of the method of Snapshot POD. Initially a series of Computational Fluid Dynamics(CFD) training runs were performed, which were subsequently improved using a QR-decomposition technique to satisfy the various boundary conditions in physiological flow problems. We used the obtained ROMs to study the influence of Pulsatility Index (PI) on a patient-specific aneurysm’s Wall Shear Stress (WSS) and Oscillatory Shear Index (OSI). In addition, we discuss how each of the obtained high-energy POD modes represents a separate significant flow pattern that is believed to influence the aneurysm’s WSS and OSI.