Application description

openBF is an open-source 1D blood flow solver based on MUSCL finite-volume numerical scheme, written in Julia and released under Apache 2.0 free software license. The software is developed by Alessandro Melis and Alberto Marzo at the Insigneo Institute at the university of Sheffield (UK). 

openBF has been used for modelling pathologies of the cardiovascular system such as vasospasm and ischaemic stroke. It has also been used for modelling mechanical thrombectomy procedures and support the development of next-generation thrombectomy devices together with Anaconda, a Spanish SME.

Available for academic and clinical researchers, the code is Distributed open-source through Github. Information and documentation on the code are available on Github. Users can download the code from the Gihub repository and install it on their local machines/HPC clusters. The code is deployed on ShARC, USFD’s HPC cluster, but external users need a guest account to access it.

Technical specifications 

OpenBF is written in Julia and has a typical memory requirement of 500 MB/job. The output size of a typical simulation is below 100 MB. The code is not parallel at the moment and has been deployed and tested on Intel CPUs. It is distributed with in house developed pre-processing and post-processing Python scripts (NumPy, GPy, SALib). Work is in progress to incorporate openBF with the VECMA toolkit to support sampling in the pre-processing stage and management of simulations


HPC usage and parallel performance

openBF is used for performing biomarker identification and sensitivity analysis of 1D cardiovascular models defined by hundreds of input parameters. The sensitivity analysis is performed through a statistical emulator that is trained on datasets whose training requires around 8000 simulator runs, each of which requires 15 minutes to complete on a single CPU. This would result in a total computing time of 2000 hours. Concurrent simulations on 12 full ShARC nodes bring the needed time to 15 hours. Current developments are focusing on the improvement of the criteria for evaluating the convergence of openBF models and defining criteria for pathology-driven exploration of the input space.

Clinical Use:

In Silico Trials

License type:

Open source (Apache 2.0) on GitHub, free

User Resources

Related articles

  • PhD Thesis: Gaussian process emulators for 1D vascular models. Link
  • Melis A. et al. 2017, Bayesian sensitivity analysis of a 1D vascular model with Gaussian process emulators. DOI
  • Melis A. et al. 2019, Improved biomechanical metrics of cerebral vasospasm identified via sensitivity analysis of a 1D cerebral circulation model. DOI
For more information about the applications supported in CompBioMed, you can contact us at "software at".