AVLWrapper

Python interface for MIT AVL (Athena Vortex Lattice)

Description

Currently implemented:

• Geometry definition
• Case definition
• Mass distribution definition
• Running operating-point run cases
• Eigen-mode analysis
• Results parsing

Installation

AVLWrapper can be installed from PyPI:

$ pip install avlwrapper

Or can be installed from Git:

$ pip install git+https://github.com/renoelmendorp/AVLWrapper.git@master

Requirements

AVL 3.40 or later (link) should be installed. Older versions don't write the machine-readable output the wrapper reads; the wrapper checks the version before it uses AVL and raises an AvlVersionError for an older version. If installed on a location in $PATH or in the module directory, the wrapper will locate it with the default configuration. See Changing settings how to change the executable path to a custom location.

(optional) Ghostscript is required to convert and save plots as pdf, jpeg, or png. Ghostscript can be installed on Linux/MacOS with a package manager:

Linux:

$ apt-get install ghostscript

MacOS:

$ brew install ghostscript

For Windows, Ghostscript can be found on the website.

Usage

For usage examples, see the example.ipynb notebook.

Changing settings

To change settings, make a local copy of the settings file:

from avlwrapper import default_config
default_config.local_copy()

By default the wrapper will look for a configuration file in the working directory and module directory. If you would like to use a different configuration file, you need to give the path to the session:

from avlwrapper import Configuration
my_config = Configuration(path_to_file)
session = Session(..., config=my_config)

AVL options

AVL settings which change the meaning of the results can be set per session:

options = avl.Options(
    profile_drag=False,          # leave out the profile drag of CDCL polars
    stability_axis_rates=False,  # rates of the cases are about the body axes
    standard_axes=True,          # X forward, Z down
    trailing_leg_forces=True,    # include forces on the trailing legs
)
session = avl.Session(geometry=aircraft, cases=cases, options=options)

Options which are not set keep the AVL default. The defaults differ between AVL versions, e.g. trailing-leg forces are included by default in AVL 3.40 but not in 3.52, so set the options your results depend on.

Trimmed flight

AVL can set up a case as level or banked horizontal flight, or as steady looping flight, from the states of the case:

case = avl.Case(
    name="cruise",
    trim=avl.Trim.level_flight,
    velocity=20.0,
    mass=10.0,
    density=1.225,
    bank=30.0,
)
results = avl.Session(geometry=aircraft, cases=[case]).run_all_cases()
results[1]["States"]  # CL, turn radius, load factor, ... as solved

Eigenmodes and mass properties

session = avl.Session(geometry=aircraft, cases=cases, mass_dist=mass)
modes = session.run_mode_analysis(cases=[1, 3])  # defaults to all cases
modes["EigenModes"][1]   # eigenvalues with their eigenvectors
modes["SystemMatrix"][1]
session.mass_properties()  # mass, CG, inertia and apparent mass

Off-body flow, surface pressures and design changes

session = avl.Session(
    geometry=aircraft,
    cases=cases,
    survey_points=[avl.Point(10.0, 0.0, 1.0)],  # results in "OffBodyFlow"
    design_changes={"twist": 1.5},  # DESIGN variables of the geometry
)

Surface pressures on the outer mould lines are computed when SurfacePressures = yes in the configuration. AVL computes them only for surfaces with airfoils over their full chord, and needs a separate run per case.

Output precision

The wrapper reads AVL's machine-readable output, which has full precision. AVL has no machine-readable format for the strip forces in body axes (5 to 6 significant digits), the off-body flow (6 decimals), the eigenvectors (4 decimals) and the mass properties (4 significant digits).

Error handling

All errors raised by the wrapper derive from avlwrapper.AvlError:

• InputError: invalid geometry, case or mass input
• AvlExecutionError: AVL did not produce the expected results. The AVL output is available in the output attribute
• AvlVersionError (an AvlExecutionError): the AVL version is older than 3.40, or could not be determined
• OutputError: an AVL output file could not be parsed

Development

Tests

To run tests in development, first install the development requirement into your environment:

pip install -r requirements-dev.txt

Then, from the source directory of the repo:

pytest -vv tests

Tests which run AVL are skipped when the AVL executable is not found. To build AVL from source (Linux, macOS), use ci/build-avl.sh, which is also used by the CI workflow:

ci/build-avl.sh 3.52 ~/.local