/DiffPLF

Primary LanguagePython

DiffPLF: A Conditional Diffusion Model for Probabilistic Forecasting of EV Charging Load

Authors: Siyang Li, Hui Xiong, Yize Chen (HKUST-GZ)

Paper: https://arxiv.org/abs/2402.13548, accepted to PSCC 2024

Several representative forecasts generated by our model:
Charging load prediction intervals

Datasets are available at:

  1. EV Charging Station Usage Open Data: https://www.kaggle.com/datasets/venkatsairo4899/ev-charging-station-usage-of-california-city
  2. Meteostat weather forecasts: https://dev.meteostat.net/

Implementation of the whole framework by running main.py with the following mode configuration:

  1. pretraining stage: isTrain=True;
  2. collecting profiles for subsequent refinement: isCollect=True;
  3. fine-tuning stage: isRefine=True;

Evaluation on experimental results: eval.py, figure.py