What is the CHeart Project?
The CHeart Project is a coding initiative started at Kings College London to support translational modeling projects. Specifically, the CHeart Project centered initial efforts around cardiac modeling challenges and has since branched out into other areas of biomedical simulation. The focus of the team is to support these efforts and bring advanced simulations from theory to the clinic.
CHeart started as a for purpose simulation program developed at the University of Oxford by Dr David Nordsletten as part of his DPhil focused on fluid-structure interaction in the heart. The initial code provided flexible implementation of different finite element schemes as well as multi-physics coupling, but focused predominantly on fluid-only, solid-only or fluid-solid interaction systems.
Joining King’s College London in 2010, Dr Nordsletten joined together with Prof Nicolas Smith, Dr Jack Lee and others to start the CHeart Project. The project aimed to expand the initial platform to enable the simulation of a more diverse array of systems, with a core focus around physical phenomena present in the heart. Initial work focused on new physical systems — monodomain, scalar advection-diffusion-reaction, eikonal equations, Darcy flow — all important for simulating cardiovascular biomechanics in the heart. Extension to wave equations, helmoltz systems and 1D blood flow formulations were also devised.
The continuing effort of the CHeart Project is provide an environment enabling computational techniques for translational goals, bringing simulation technology closer to the heart of the clinic. This effort requires continual development from the foundational theory through to tools for real-world applications.
To date, CHeart has been responsible for over 50+ scientific papers and studies.
ABOUT THE CODE
The core of the CHeart Project is its software platform. CHeart software is designed to be flexible and enable linking of multi-physics systems for a diverse range of applications. The code itself is a modest 40K lines written in the Fortran 2008. CHeart exploits multi-core platforms using MPI and uses a number of libraries for efficient parallel computing, including PETSc, MUMPs, SUPERLU, PARMETIS and others. For more on the computational engine driving the CHeart Project, see our paper.
There are a number of strong coding efforts from research groups worldwide, with some focused on functionality, efficient HPC implementation, modeling phenomena, multi-physics. So why CHeart?
The reason is simple: necessity is the mother of invention. To support the translational research of the group, a malleable simulation environment was necessary. One that would accommodate integration of new techniques, new simulation ideas and new sources of medical data. That could be adapted at short notice to integrate the newest ideas coming from our research team. One that could seamlessly integrate different physical systems to address questions of multiphysics phenomena. One that would efficiently exploit available computational resource.