报告人：FelipeGalarce(智利瓦尔帕莱索天主教大学)
报告时间：2026年9月30日 （周三）下午14:00-15:30
报告地点：6号楼C203
报告摘要：Cardiovascular flows provide a compelling setting in which mathematicalmodeling, numerical simulation, and data-drivenmethods must operate together.Although imaging techniques such as Doppler ultrasound and 4D-flow MRI provide increasingly rich information about blood flow,measurements remain sparse, noisy, and incomplete, while clinically relevant quantities such as pressure differences, wall shear stress, and rheological properties are often difficult to measure directly. In this talk, I will discuss mathematical and computational approaches for recovering such hidden information by combining physical models with available measurements. Starting from the Navier–Stokes equations and non-Newtonian descriptions of blood rheology, cardiovascular reconstructionwill be formulated as an inverse and data-assimilation problem, with particular emphasis on reduced-order modeling and scientific machine learning. Approaches based on low-dimensional representations, parametrizedmanifolds, and learned flow dynamics will illustrate how sparse or low-dimensional velocity measurements can be used to reconstruct complete three-dimensional, time-dependent flow fields and estimate quantities such as pressure, viscosity, and hemodynamic biomarkers. The broader objective is to show how physical models, numerical simulation, and machine learning can be integrated to transform incomplete cardiovascular measurements into physically consistent and quantitativelyuseful descriptions of the underlying flow.
专家简介：FelipeGalarce 博士现为智利瓦尔帕莱索天主教大学（PontificiaUniversidadCatólicadeValparaíso，PUCV）副教授。他曾于法国巴黎萨克雷大学获得硕士学位，并于索邦大学获得应用数学博士学位，博士期间主要研究血流动力学中的反问题，以及基于医学数据的血流参数估计。此后，他还曾在德国柏林魏尔斯特拉斯研究所从事博士后研究。其主要研究方向包括数学建模、计算模拟、流体力学、反问题和数据同化等。
