Pylsewave: A python toolkit to solve pulse wave dynamics

Georgios E. Ragkousis (georgios.ragkousis at port.co.uk)

Cardiovascular engineering research lab, Bioneer group, University of Portsmouth.






Summary. This is the documentation of the "pylsewave toolkit". The purpose of this open source software is to distribute an efficient objected-oriented tool to solve pulse wave dynamics and/or any other physical problem which can be modelled by a hyperbolic system of partial differential equations (PDEs). The core development language is cython, python and C++. The idea is that the most (computationally) expensive parts are developed in lower level languages (e.g C/C++) and; subsequently, called in high level interpreted (python) language. Moreover, the user can easily extend, modify or redistribute any part of this software with high flexibility. The presented library contains options of "elements" (e.g. for each vessel), meshes (arterial networks), different solvers (currently FD, later FE), optimisers and post-processing tools (matplotlib, VTK files, etc.).

Parallel CPU and GPU. Expensive computational parts can be run in parallel with OpenMP (via Cython.parallel module). In future, parallel GPU processing will be added; as well, via CUDA JIT (numba).

Table of contents

Mathematical problem formulation
      Derivation of 1D hyperbolic system
      Velocity profile
      State equation
      Conservational form and coupling of PDEs
      Characteristic equations
      Wave decomposition
      Intensity analysis
Numerical methods
      Finite Differences
      Boundary conditions
Case studies
      Upper extremity model
      A detailed arterial network
Bibliography
Appendix A: Non-linear system at conjuctions
      Bifurcation case
Appendix B: Wrapping C++ classes in pylseWave via Cython
      ctypes
      cython

© 2019, Georgios E. Ragkousis. Released under CC Attribution 4.0 license