FLAPJAX — FLexible Adjoint-enabled Panel code in JAX

nonlinear · differentiable · adjoint-enabled

FLAPJAX is a differentiable nonlinear aeroelastic analysis framework which couples a nonlinear structural model with unsteady vortex lattice method (UVLM) aerodynamics. This allows for a range of structural, aerodynamic and coupled analyses, with gradients available using the adjoint method. The full codebase is implemented in JAX, which allows for efficient automatic differentiation, GPU acceleration and multi-case parallelisation.

Full documentation, including tutorials, API reference and theory, is available at ben-l-p.github.io/flapjax.

Installation

Github

For the latest version from git, clone the repository and install with pip:

git clone https://github.com/ben-l-p/flapjax/
cd flapjax
pip install -e .  # uses the -e flag to install in editable mode

PyPI

Two PyPI distributions are provided:

Install one distribution — they provide the same flapjax import path.

pip install flapjax        # minimal
pip install flapjax-full   # includes tests and tutorial notebooks

Testing

An extensive suite of over 300 tests is included to verify the correctness of the code. This checks the numerics, and takes approximately 40 minutes to run on an M2 MacBook Air. Tests can be run with pytest, either against a flapjax-full install (pytest --pyargs flapjax) or from a repository clone (uv run pytest).