VIVSIM

VIVSIM is a Python library for accelerated fluid-structure interaction (FSI) simulations based on the immersed boundary -lattice Boltzmann method (IB-LBM). It was originated from a research project requiring efficient simulation codes for studying vortex-induced vibration (VIV) of underwater structures.

Similar to projects like JAX-CFD and XLB, VIVSIM utilizes JAX as the backend to harness the power of hardware accelerators, achieving massive parallelism on GPU/GPUs.

Examples

Lid-driven cavity at $Re = 2\times 10^4$ on a 1000x1000 lattice grid.	Flow passes through some texts on a 1000x1000 lattice grid.
VIV of a cylinder with $U_r = 5$ and $Re = 100$. The simulation was conducted on a 2000x3000 lattice grid with 400 immersed boundary markers, taking about 8 min on a Nvidia A100.	VIV of a cylinder with $U_r = 5$ and $Re = 1\times 10^4$. The simulation was conducted on a 8000x4000 lattice grid with 1200 immersed boundary markers, taking about 30 min on 8 Nvidia A800.

Capabilities

Lattice Models

• D2Q9

Collision Models

• Single Relaxation Time (SRT) model
• Multiple Relaxation Time (MRT) model

Boundary Conditions:

• Velocity boundary using Non-Equilibrium Bounce Back (NEBB) method
• No-slip boundary using Halfway Bounce-Back method
• Outflow boundary simply by copying the second last column/row
• Periodic boundary

Fluid-Structure Interaction

• Multi Direct-Forcing Immersed Boundary method.

Getting Started

To locally install VIVSIM for development:

git clone https://github.com/haimingz/vivsim.git
pip install -e vivsim

This package is based on JAX, whose installation may depend on the OS and hardware. If the above command does not work well, please refer to the JAX Documentation for the latest installation guidance.

Alternatively, you can run the following command in a cell on Google Colab to install VIVSIM and run simulations using free/paid GPU on the cloud.

!pip install git+https://github.com/haimingz/vivsim

You can also create a Singularity image from the provided definition file vivsim.def and execute your code on High Performance Computing (HPC) clusters.

More detailed instructions can be found in our Documentation.

Cite VIVSIM

If you find this repo useful, please cite our paper:

@inproceedings{zhu2024computational,
  title={Computational Performance of IB-LBM Based VIV Simulation Using Python With JAX},
  author={Zhu, Haiming and Du, Zunfeng and Yang, Yuan and Han, Muxuan},
  booktitle={International Conference on Offshore Mechanics and Arctic Engineering},
  volume={87844},
  pages={V006T08A020},
  year={2024},
  organization={American Society of Mechanical Engineers}
}