HPC for Cardiovascular System Simulations

The modeling of the cardiovascular system is receiving increasing attention from both the medical and mathematical environments on the one hand because of the great influence of haemodynamics on cardiovascular diseases (\cite{formaggia09:_cardiov_mathem} chap 1), and, on the other hand, due to its challenging complexity that keeps open the debate about the setting up of appropriate models and algorithms. A wide variety of approaches can be found in literature, dealing with different formulations of the problem and solution strategies.

This research project aims at furtherly developing and porting our research library LifeV to High Performant Computing infrastructure. LifeV is a parallel finite element library which has been jointly developed by the CMCS group at the EPFL, the MOX Laboratory at the Politecnico di Milano, and INRIA (REO project) since 1999/2000 (see www.lifev.org). It provides implementations of state--of--the--art of mathematical and numerical methods in multi-scale and multi-physics. It serves both as research and production library.

Our goal is to simulate the anatomic structure and the physiological response of the human cardiovascular system in healthy or diseased states. This demands to address many fundamental issues. Blood flow interacts both mechanically and chemically with the vessel walls and tissues, giving rise to complex fluid-structure interaction problems. The mathematical analysis of these problems and the related numerical analysis are complicated. In ongoing projects we are extending the recently achieved results on blood flow simulations by directing our analysis in several new directions, which encompass aspects of metabolic regulation, micro-circulation, the electrical and mechanical activity of the heart, and their interactions.

The emphasis of this project will be put on algorithm development and implementation, computational efficiency, and patient specific physiological simulations.  Our purpose is to set up a mathematical simulation platform eventually leading to the improvement of vascular diseases diagnosis, setting up of surgical planning, and cure of inflammatory processes in the circulatory system. This platform might also help physicians to construct and evaluate combined anatomic/physiological models to predict the outcome of alternative treatment plans for individual patients.

Because of the very high computational cost of our simulations (presently we can hardly run simulations for more than one heart beat and/or with particularly fine meshes), we have to develop algorithms and code suited for the coming HPC hardware. In collaboration with our running projects, we also want to develop domain decomposition methods for the solution of the fluid-structure interaction problem which enjoy good scalability properties for parallel simulations.

Our belief is that this project would enable us to run on HPC platforms both complex and accurate (enough) simulations of the cardiovascular system, in a sufficiently short computational time to make the solution for patient specific vascular geometries possible.

• Prof. Alfio Quarteroni, EPF Lausanne
• Dr. Simone Deparis, EPF Lausanne

• Paolo Crosetto, EPF Lausanne
  
• Dr. Cristiano Malossi, EPF Lausanne
  
• Gwenol Granperrin, EPF Lausanne
  
• Radu Popescu, EPF Lausanne