Skip to content

A multi-physics finite element solver for patient-specific blood flow simulation including fluid-structure interaction and cardiac electrophysiology

Notifications You must be signed in to change notification settings

zinanhu0810/svFSI

 
 

Repository files navigation

Introduction

svFSI is a multi-physics finite element solver designed for computational modeling of the cardiovascular system. Some of the unique capabilities of svFSI include modeling cardiac electrophysiology, biological tissue mechanics, blood flow, and large deformation fluid-structure interaction (FSI). svFSI also offers a wide choice of boundary conditions for performing patient-specific modeling of cardiovascular biomechanics. The code is parallelized using message-passing-interface (MPI) and offers multiple options to choose a linear solver and preconditioner. svFSI can be used as part of the SimVascular software or can be used as a stand-alone solver.

Dependence

The following packages are required to build and use svFSI.

  • cmake
  • cmake-curses-gui
  • cmake-gui
  • gcc (version>=4.8.5) with gfortran
  • openmpi or mpich
  • blas & lapack
  • trilinos (optional)

On Ubuntu, most of the dependencies can be installed using apt install. On macOS, the dependencies may be installed using brew. Apart from GNU compilers, svFSI can also be built with Intel oneAPI Toolkits. For more details, please refer to INSTALL.md and INSTALL-DEPS.md.

Quick Build

Precompiled binaries for Ubuntu and MacOS are available for download from SimTK.

Users are recommended to build from the source code to access the most recent features and bug fixes. Instructions for a quick build are provided here for a Linux/Mac OS system.

  1. Clone or download the current repository.

  2. Create a build directory

    cd svFSI && mkdir build && cd build
  3. Initiate the CMake terminal interface to generate makefiles.

    ccmake ..
  4. This will automatically search for compilers. Follow instructions if necessary. Press “c” to configure repeatedly until CMake parameters no longer change and CMake presents the option “g” to generate. Press “g” to create makefiles and exit. Run make in the build directory:

    make

    A successful build will generate a solver binary, called svFSI in the following directory build/svFSI-build/bin.

    For more advanced users, please refer INSTALL.md for detailed platform-specific instructions to install svFSI.

Build With Trilinos

svFSI also supports compilation with Trilinos. Users can build Trilinos locally following its online documentation.

The recommended Trilinos third-party libraries (TPLs) include Boost, BLAS, HDF5, HYPRE, LAPACK, MPI, and MUMPS. The required Trilinos packages are Amesos, AztecOO, Epetra, EpetraEXT, Ifpack, ML, MueLU, ROL, Sacado, Teuchos, and Zoltan.

To enable Trilinos in svFSI, users need to turn on the option SV_USE_TRILINOS located in the file Code/CMake/SimVascularOptions.cmake as,

option(SV_USE_TRILINOS "Use Trilinos Library with svFSI" ON)

In most cases, users can proceed to build svFSI following the Quick Build, and CMake should be able to locate Trilinos automatically through find_package. In case the automatic way fails, users can also specify the path to Trilinos through ccmake -DCMAKE_PREFIX_PATH:PATH="<Path_to_Trilinos>/lib/cmake/Trilinos;<Path_to_any_other_package>;".

For more detailed instructions, please refer INSTALL.md.

Run Simulation

svFSI requires a plain-text input file to specify simulation parameters. The syntax of the input file can be found here.

A master template is provided in the current repository, svFSI_master.inp. Users are also recommended to go through the input files in the examples and modify them for their needs.

An MPI-based run can be initiated through

mpiexec -np <number of MPI processes>  <Path to Build>/svFSI-build/bin/svFSI <Path to input file>

Features

svFSI provides the capability to model a variety of physics including unsteady diffusion, linear and nonlinear elastodynamics, convective heat transfer, fluid flows, fluid-structure-interaction (FSI), and cardiac electrophysiology. As the code is modular, the users are provided with a choice to couple these physics depending on their needs. We strongly recommend users to browse through the examples provided in the GitHub repository svFSI-Tests to get a detailed insight into the capability of the code.

Below, we provide a list of the available choice of constitutive models for different types of equations being solved. Users are also encouraged to implement new constitutive models. Users may use global search tools such as grep to locate the implementations of the available constitutive models in the code using the abbreviated names below.

Abbreviation refers to the variable name in the source code; Full name refers to the generic name of the model; Input keyword refers to the phrase in the input file that can invoke such model.

  1. Available isochoric constitutive models for the structure equation

    Abbreviation Full name Input keyword
    stIso_stVK Saint Venant-Kirchhoff "stVK", "stVenantKirchhoff"
    stIso_mStVK modified Saint Venant-Kirchhoff "m-stVK", "modified-stVK", "modified-stVenantKirchhoff"
    stIso_nHook Neo-Hookean model "nHK", "nHK91", "neoHookean", "neoHookeanSimo91"
    stIso_MR Mooney-Rivlin model "MR", "Mooney-Rivlin"
    stIso_HGO_d Holzapfel-Gasser-Ogden (decoupled) "HGO", "HGO-d", HGO-decoupled"
    stIso_HGO_ma HGO model (modified anisotropy) "HGO-ma", "HGO-modified"
    stIso_Gucci Guccione model "Guccione", "Gucci"
    stIso_HO_d Holzapfel-Ogden model (decoupled) "HO", "Holzapfel", "HO-decoupled", "HO-d"
    stIso_HO_ma HO model (modified anisotropy) "HO-ma", "HO-modified"
  2. Available volumetric constitutive models for the structure equation

    Abbreviation Full name Input keyword
    stVol_Quad Quadratic model "quad", "Quad", "quadratic", "Quadratic"
    stVol_ST91 Simo-Taylor91 model "ST91", "Simo-Taylor91"
    stVol_M94 Miehe94 model "M94", "Miehe94"
  3. Available constitutive models for the fluid equation

    Abbreviation Full name Input keyword
    viscType_Const Constant viscosity (Newtonian model) "Constant", "Const", "Newtonian"
    viscType_CY Carreau-Yasuda non-Newtonian model "Carreau-Yasuda", "CY"
    viscType_Cass Cassons non-Newtonian model "Cassons", "Cass"
  4. Available cardiac electrophysiology models

    Abbreviation Full name Input keyword
    cepModel_AP Aliev-Panfilov model "AP", "Aliev-Panfilov"
    cepModel_BO Bueno-Orovio-Cherry-Fenton model "BO", "Bueno-Orovio"
    cepModel_FN Fitzhugh-Nagumo model "FN", "Fitzhugh-Nagumo"
    cepModel_TTP tenTusscher-Panfilov model "TTP", "tenTusscher-Panfilov"

Additional Resource

More details can be found here:

Citation

In preparation.

About

A multi-physics finite element solver for patient-specific blood flow simulation including fluid-structure interaction and cardiac electrophysiology

Resources

Stars

Watchers

Forks

Packages

No packages published

Languages

  • Fortran 84.5%
  • Makefile 4.8%
  • CMake 4.8%
  • C 2.4%
  • C++ 2.0%
  • Shell 0.8%
  • Other 0.7%