Forschungsprojekt AI4Grids

Abstract

In this paper, we propose a novel method for real-time control of electric distribution grids with a limited number of measurements. The method copes with the changing grid behaviour caused by the increasing number of renewable energies and electric vehicles. Three AI based models are used. Firstly, a probabilistic forecasting estimates possible scenarios at unobserved grid nodes. Secondly, a state estimation is used to detect grid congestion. Finally, a grid control suggests multiple possible solutions for the detected problem. The best countermeasures are then detected by evaluating the systems stability for the next time-step.

Contact

Marcel Arpogaus - marcel.arpogaus@htwg-konstanz.de

BibLaTex

@INPROCEEDINGS{9841826,
  author={Arpogaus, M. and Montalbano, J. and Linke, M. and Schubert, G.},
  booktitle={CIRED Porto Workshop 2022: E-mobility and power distribution systems}, 
  title={Probabilistic real-time grid operation management of future distribution grids with high penetration of renewable generators and electrical vehicles based on artificial intelligence}, 
  year={2022},
  volume={2022},
  pages={147-151},
  doi={10.1049/icp.2022.0681}}

Preprint on arXiv

Abstract

The transition to a fully renewable energy grid requires better forecasting of demand at the low-voltage level to increase efficiency and ensure reliable control. However, high fluctuations and increasing electrification cause huge forecast variability, not reflected in traditional point estimates. Probabilistic load forecasts take future uncertainties into account and thus allow more informed decision-making for the planning and operation of low-carbon energy systems. We propose an approach for flexible conditional density forecasting of short-term load based on Bernstein polynomial normalizing flows, where a neural network controls the parameters of the flow. In an empirical study with 363 smart meter customers, our density predictions compare favorably against Gaussian and Gaussian mixture densities. Also, they outperform a non-parametric approach based on the pinball loss for 24h-ahead load forecasting for two different neural network architectures.

Contact

Marcel Arpogaus - marcel.arpogaus@htwg-konstanz.de

BibLaTex

@unpublished{Arpogaus2022a,
  title = {Short-{{Term Density Forecasting}} of {{Low-Voltage Load}} Using {{Bernstein-Polynomial Normalizing Flows}}},
  author = {Arpogaus, Marcel and Voss, Marcus and Sick, Beate and Nigge-Uricher, Mark and Dürr, Oliver},
  date = {2022-04-29},
  eprint = {2204.13939},
  eprinttype = {arxiv},
  primaryclass = {cs, stat},
  archiveprefix = {arXiv}
}

Open Access

Abstract

The transition to a fully renewable energy grid requires better forecasting of demand at the low-voltage level. However, high fluctuations and increasing electrification cause huge forecast errors with traditional point estimates. Probabilistic load forecasts take future uncertainties into account and thus enables various applications in low-carbon energy systems. We propose an approach for flexible conditional density forecasting of short-term load based on Bernstein-Polynomial Normalizing Flows where a neural network controls the parameters of the flow. In an empirical study with 363 smart meter customers, our density predictions compare favorably against Gaussian and Gaussian mixture densities and also outperform a non-parametric approach based on the pinball loss for 24h-ahead load forecasting for two different neural network architectures.

Contact

Marcel Arpogaus - marcel.arpogaus@htwg-konstanz.de

BibLaTex

@inproceedings{Arpogaus2021,
  title = {Probabilistic {{Short-Term Low-Voltage Load Forecasting}} Using {{Bernstein-Polynomial Normalizing Flows}}},
  booktitle = {{{ICML}} 2021 Workshop on Tackling Climate Change with Machine Learning},
  author = {Arpogaus, Marcel and Voss, Marcus and Sick, Beate and Nigge-Uricher, Mark and Dürr, Oliver},
  date = {2021-06-23},
  url = {https://www.climatechange.ai/papers/icml2021/20}
}