Propeller.jl

Julia implementation of a hybrid blade element momentum (BEM) propeller model based on Davoudi [1].

The model supports:

• Linear inflow coupled with BEM, using Pitt-Peters by default
• Multi-threaded sweeps when Julia is started with multiple threads
• Prandtl hub/tip-loss correction [2]
• Glauert Mach correction for polar data
• Offline and online rotational 3D polar correction
• Export helpers for DUST and FLOWUnsteady workflows

Repository Layout

• src/: package implementation
• config/rotor/: rotor YAML files and source geometry data
• data/airfoils/: airfoil coordinate data
• data/polars/: polar CSV data
• example.jl: small RPM sweep example
• rotor_calculations.jl: pre-calculation helper, for example Reynolds number estimates
• validation/: validation scripts and reference data
• misc/: export and data-generation utilities
• test/: package tests

Installation

Install Julia, Git, and Conda or Miniconda first. Run all project commands from the repository root unless noted otherwise.

On Windows PowerShell:

git clone https://github.com/christianhauschel/Propeller.jl C:\path\to\Propeller.jl
Set-Location C:\path\to\Propeller.jl
julia --project=.

If the repository is already on a mapped drive, use the normal Windows path:

Set-Location P:\code\09_aero\Propeller.jl
julia --project=.

On Linux/macOS:

git clone https://github.com/christianhauschel/Propeller.jl <mypath>
cd <mypath>
julia --project=.

In the Julia REPL, instantiate the project:

import Pkg
Pkg.instantiate()

This repository contains a Manifest.toml, so Pkg.instantiate() is the preferred path. If Julia cannot resolve the private/unregistered dependencies, add them by URL:

import Pkg

url = "https://github.com/christianhauschel/"
packages = [
    "AirfoilPolars.jl",
    "AirfoilFast.jl",
    "PrettySections.jl",
    "Gradient.jl",
    "FlightConditions.jl",
    "GeometricTools.jl",
    "GridSpacing.jl",
]

Pkg.add([Pkg.PackageSpec(url=url * p) for p in packages])
Pkg.instantiate()

AirfoilPolars.jl must be available before AirfoilFast.jl, because AirfoilFast.jl depends on it.

Python Plotting Setup

Some examples and validation scripts use Python plotting through PyCall and ultraplot.

Create a Python environment.

On Windows PowerShell:

conda create -n julia python=3.10
conda activate julia
python -m pip install ultraplot
where python

Use the Python path inside the julia environment. It usually looks like:

C:\Users\<user>\AppData\Local\miniconda3\envs\julia\python.exe

If conda activate julia is not recognized in PowerShell, run once:

conda init powershell

Then close and reopen PowerShell.

On Linux/macOS:

conda create -n julia python=3.10
conda activate julia
python -m pip install ultraplot
which python

Configure PyCall from Julia:

import Pkg
Pkg.add("PyCall")
ENV["PYTHON"] = raw"C:\Users\<user>\AppData\Local\miniconda3\envs\julia\python.exe"
Pkg.build("PyCall")

On Linux/macOS, use the path returned by which python:

import Pkg
Pkg.add("PyCall")
ENV["PYTHON"] = "/home/<user>/miniconda3/envs/julia/bin/python"
Pkg.build("PyCall")

Restart Julia, then test:

using PyCall
pyversion
uplt = pyimport("ultraplot")

Running The Example

From the repository root:

julia --project=. example.jl

Or from the Julia REPL:

include("example.jl")

For threaded runs:

julia --threads=4 --project=. example.jl

The default example uses:

name_rotor = "apc_11x47SF"

which loads:

config/rotor/apc_11x47SF.yaml

Plots and generated files are written below out/.

Rotor Configurations

Available rotor YAML configs include:

• apc_11x47SF.yaml
• apc_845MR.yaml
• apc_845MR_model.yaml
• davoudi.yaml
• dji9443.yaml
• djimatrice300rtk.yaml
• rotor_anopp.yaml
• rotor_simple.yaml
• test.yaml

When changing scripts, set name_rotor to the YAML filename without .yaml. For example:

name_rotor = "dji9443"

not dji_9443, because config/rotor/dji_9443/ is a data directory, not a YAML config file.

Useful Scripts

• example.jl: runs an RPM sweep and optionally plots thrust, torque, C_T, and C_Q
• rotor_calculations.jl: estimates Reynolds number, Mach number, and velocity ranges for a rotor
• validation/apc11x47.jl: APC 11x4.7 static validation sweep
• validation/davoudi.jl: comparison against Davoudi reference data
• validation/loading.jl: normal and tangential loading comparison
• validation/polarmodel.jl: compares model polars against polar data
• misc/export_dust.jl: exports a rotor blade geometry for DUST
• misc/export_flowunsteady.jl: exports rotor data for FLOWUnsteady
• misc/generate_training.jl: generates training samples and writes trainingdata.csv
• misc/airfoil_csv2dat.jl: converts airfoil CSV data to DAT-style files

Most scripts have settings near the top. Edit name_rotor, RPM ranges, plotting flags, and output folders there.

Tests

Run the package tests with:

julia --project=. -e "import Pkg; Pkg.test()"

Troubleshooting

If Julia cannot find a config file, check that you are in the repo root:

pwd()
readdir("config/rotor")

If build("PyCall") fails with build not defined, use:

import Pkg
Pkg.build("PyCall")

If pyversion is not defined, load PyCall first:

using PyCall
pyversion

If using PyCall fails because Python cannot import ctypes, verify the Python environment outside Julia:

conda activate julia
python -c "import ctypes; print('ctypes ok')"
conda install -y libffi

Then rebuild PyCall and restart Julia.

Validation

APC 11x4.7 Static Case

Simple comparison with uniform inflow:

Davoudi Measurements

Data taken from Davoudi [1]:

Normal And Tangential Loads

Comparison to FLOWUnsteady BEM code:

Conventions

References

1. B. Davoudi, "A Hybrid Blade Element Momentum Model for Flight Simulation of Rotary Wing Unmanned Aerial Vehicles," AIAA Paper, 2019.
2. https://flow.byu.edu/FLOWUnsteady/examples/rotorhover-aero/