DGAC / Acropole

This repository contains the Acropole model for aircraft fuel flow prediction and Python packages for aircraft trajectory processing and fuel flow enhancement.

Acropole

This repository contains the Acropole model for aircraft fuel flow prediction and Python packages for aircraft trajectory processing and fuel flow enhancement.

Easy Install

To easy install, creating dedicated anaconda environment is recommended :

conda create -n acropole python=3.8 -c conda-forge

Activate conda environment :

conda activate acropole

Install dependencies :

conda install numpy tensorflow scipy joblib scikit-learn pandas  

If you want to add Jupyter notebooks and matplotlib :

conda install matplotlib jupyter jinja2==3.0.3 

Clone repository :

git clone https://github.com/DGAC/Acropole.git

Finally, install lib :

cd acropole
pip install .

Python packages

The Acropole Python library includes the following packages :

• columns: that contains default data column names and list of columns

• utils: that contains shared functions

• predictor: that contains functions to load and apply Acropole model

• trajectory: that contains trajectory processes and pipelines

Data and models

Available data

Aircraft parameters from open data to feed the model are available in https://github.com/DGAC/Acropole/blob/main/acropole/data/acft_params.csv and loaded by the packages.

Model training and evaluation

The Acropole model is a neural network built using data from Quick Access Recorder (QAR) from different aircraft types. Evaluation of the model and list of aircraft is available in https://github.com/DGAC/Acropole/tree/main/evaluation/Dense_Acropole_FuelFlow_Scaling.

Example of use

For example of use please refer to https://github.com/DGAC/Acropole/blob/main/examples/examples.ipynb