Energy-based stochastic MPC for integrated electricity-hydrogen VPP in real-time markets

Publikationstyp
Journal Article
Date Issued
2021-06
Sprache
English
Author(s)
Appino, Riccardo R.
Wang, Han
González Ordiano, Jorge Ángel
Faulwasser, Timm  
Mikut, Ralf
Hagenmeyer, Veit
Mancarella, Pierluigi
TORE-URI
https://hdl.handle.net/11420/45677
Journal
Electric power systems research  
Volume
195
Article Number
106738
Citation
Electric Power Systems Research 195: 106738 (2021-06)
Publisher DOI
10.1016/j.epsr.2020.106738
Scopus ID
2-s2.0-85102870159
Publisher
Elsevier
Virtual Power Plants (VPPs) comprising renewables and hydrogen production through power-to-gas technologies can help to increase renewable penetration and to improve operational flexibility and economic performance. However, the uncertainty inherent to forecasts of renewable generation and energy prices renders cost effective operation difficult. The present paper approaches the issue by means of receding-horizon stochastic optimization (i.e. by stochastic Model Predictive Control (MPC)). Differently from previous works, we do not tackle computational tractability with a sampling-based approach, but by mapping quantile forecasts of virtual energy profiles to the mode of operation that has the highest probability of being optimal. This way, we reduce the computational load and the forecasting burden. Furthermore, simulation studies show that the proposed algorithm can attain a significant percentage of the revenue of optimal control with perfect forecasts.
Subjects
Power to gas
Probabilistic forecast
Real-time markets
Sector integration
Stochastic model predictive control
DDC Class
621: Applied Physics