Data Efficiency Assessment of Generative Adversarial Networks in Energy Applications

---

📄 Paper

Nabila, U. M., Lin, L., Zhao, X., Gurecky, W. L., Ramuhalli, P., Radaideh, M. I. (2025). Data efficiency assessment of generative adversarial networks in energy applications. Energy and AI, 20, 100501. https://doi.org/10.1016/j.egyai.2025.100501

⚙️Environment Installation

This project uses PyTorch for most models (WGAN, FNN, GRU, etc.) and TensorFlow for the cGAN CHF model and FNN model. Please set up the environments accordingly:

🔵 PyTorch Environment (for cGAN, FNN, GRU, etc.)

# 1. Create a new conda environment with Python 3.11
conda create -n torchgpu python=3.11

# 2. Activate the environment
conda activate torchgpu

# 3. Install PyTorch and related libraries
pip install torch torchvision torchaudio

# 4. Install other relevant packages
pip install pandas matplotlib scikit-learn seaborn
pip install optuna

🟠 TensorFlow Environment (only for cGAN CHF model and FNN model)

# 1. Create a new conda environment with Python 3.11
conda create -n tfgpu python=3.11

# 2. Activate the environment
conda activate tfgpu

# 3. Install TensorFlow with CUDA support
pip install tensorflow[cuda]

# 4. Install other relevant packages
pip install pandas matplotlib scikit-learn seaborn

📂 Dataset Access

Due to GitHub's file size limit, the dataset CAISO_zone_1_.csv (187 MB) is hosted externally.

👉 [Download CAISO_zone_1_.csv from this link] (https://drive.google.com/file/d/1coOdL7Lq1hBkMSt8t9sRT3f5M7pPv7Jb/view?usp=sharing)

After downloading, place the file in the data/ folder of this repository.

📊 How to generate the results

• The folder data contains the CHF test data file (with all the CHF data points) and PSML dataset.
• Datasets for different case scenarios are organized into appropriate case-specific folders under the model directory.
• Navigate to the desired case folder inside models and run the appropriate script to start training or evaluation (e.g., cGan_all.py for case1/all):

cd model/cGAN/CHF/case1/all/
python cGan_all.py

Results will be saved automatically in the working folder.

Note: The seasonal variation experiment (PSML case 3) is not included in a separate folder in this repository. This is because the results were briefly discussed in the paper without being presented in a dedicated table or figure.