mSANN model benchmarks
Description
This data set contains scripts and dataset needed to reproduce the results in the following paper: Mohamed Amine Bouhlel, Sicheng He, and Joaquim R. R. A. Martins. Scalable gradient-enhanced artificial neural networks forairfoil shape design in the subsonic and transonic regimes. In this paper, we mainly produced three studies: * An analytical test case stored in the repository "Rosenbrock": This repository contains the dataset for running the training and validation of the models. It also contains three repositories ANN, SANN, and mSANN that contains the scripts needed to rerun the models, respectively. * The airfoil shape design analysis test case in both subsonic and transonic regimes in the repository "Analysis/mSANN": This repository contains three repositories "cd", "cl", and "cd" for training the mSANN model on the aerodynamic coefficients. Each sub-repository contains the dataset (for training, validation, and testing the model) and two script files run.py and prediction.py for training the neural network and make predictions. * The airfoil shape design optimization test case in both subsonic and transonic regimes in the repossitory "Optimization": This repository contains two repositories "subsonic" and "transonic" for running an optimization either using the mSANN model or CFD.
Files
Steps to reproduce
1. Download the repositories containing the dataset and the scripts 2. Modify the dataset paths in each script file, as described in the head of each script file 3. Run the desired test using the command python name_file.py