Perdikaris Paris, Grinberg Leopold, Karniadakis George Em
Department of Mechanical Engineering, Massachusetts Institute of Technology , Cambridge, Massachusetts 02139, USA.
IBM T.J Watson Research Center , 1 Rogers St, Cambridge, Massachusetts 02142, USA.
Phys Fluids (1994). 2016 Feb;28(2):021304. doi: 10.1063/1.4941315. Epub 2016 Feb 8.
The aim of this work is to present an overview of recent advances in multi-scale modeling of brain blood flow. In particular, we present some approaches that enable the study of multi-scale and multi-physics phenomena in the cerebral vasculature. We discuss the formulation of continuum and atomistic modeling approaches, present a consistent framework for their concurrent coupling, and list some of the challenges that one needs to overcome in achieving a seamless and scalable integration of heterogeneous numerical solvers. The effectiveness of the proposed framework is demonstrated in a realistic case involving modeling the thrombus formation process taking place on the wall of a patient-specific cerebral aneurysm. This highlights the ability of multi-scale algorithms to resolve important biophysical processes that span several spatial and temporal scales, potentially yielding new insight into the key aspects of brain blood flow in health and disease. Finally, we discuss open questions in multi-scale modeling and emerging topics of future research.
这项工作的目的是概述脑血流多尺度建模的最新进展。特别是,我们介绍了一些能够研究脑血管系统中多尺度和多物理现象的方法。我们讨论了连续介质和原子尺度建模方法的公式化,提出了它们并行耦合的一致框架,并列出了在实现异构数值求解器的无缝和可扩展集成时需要克服的一些挑战。所提出框架的有效性在一个实际案例中得到了证明,该案例涉及对患者特异性脑动脉瘤壁上血栓形成过程进行建模。这突出了多尺度算法解析跨越多个空间和时间尺度的重要生物物理过程的能力,有可能为健康和疾病状态下脑血流的关键方面带来新的见解。最后,我们讨论了多尺度建模中的开放性问题以及未来研究的新兴主题。