Fluid-structure interaction analysis of pulsatile flow in arterial aneurysms with physics-informed neural networks and computational fluid dynamics


Rehman M. A. U., EKİCİ Ö., Farooq M. A., Butt K., Ajao-Olarinoye M., Wang Z., ...More

PHYSICS OF FLUIDS, no.3, 2025 (SCI-Expanded) identifier

  • Publication Type: Article / Article
  • Publication Date: 2025
  • Doi Number: 10.1063/5.0259296
  • Journal Name: PHYSICS OF FLUIDS
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Academic Search Premier, Aerospace Database, Aquatic Science & Fisheries Abstracts (ASFA), Chemical Abstracts Core, Chimica, Compendex, INSPEC, zbMATH
  • Hacettepe University Affiliated: Yes

Abstract

Marfan syndrome (MS) is a genetic disorder often associated with the development of aortic aneurysms, leading to severe vascular complications. The progression of this condition is intricately linked to hemodynamic factors such as wall shear stress (WSS) and von Mises stress, as abnormal distributions can contribute to thrombus formation, endothelial damage, and the worsening of aneurysmal conditions. In this study, six vascular models were analyzed: four representing diseased aortas with Marfan syndrome aneurysms and two healthy aortic models for comparison. The models were sourced from Vascular Model Repository, and computational fluid dynamics (CFD) simulations were conducted using a Newtonian fluid model and the shear stress transport (SST) k- omega turbulent transitional model to evaluate WSS and von Mises stress. Fluid-structure interaction was employed to incorporate vessel wall interaction, and pulsatile inlet velocity profiles were used to simulate physiological blood flow, capturing time-dependent hemodynamic variations. The results revealed significant differences between healthy and diseased aortic models. In healthy models, WSS was uniformly distributed, with values consistently below 40 Pa, reflecting stable vascular conditions. Conversely, the diseased models exhibited highly non-uniform WSS distributions, with notably lower values in aneurysmal regions, contributing to thrombus formation, with elevated WSS in areas like the carotid and subclavian arteries due to geometric and hemodynamic complexities. The von Mises stress analysis identified regions of heightened rupture risk, particularly on the superior side of case MS1, where both von Mises stress and WSS reached their highest values among all cases. Physics-informed neural networks demonstrated strong agreement with CFD results while significantly reducing computational cost, highlighting their potential for real-time clinical applications. These findings underscore the critical role of hemodynamic factors in aneurysm progression and rupture risk, offering valuable insights for optimizing diagnostic and therapeutic strategies in vascular diseases.