Deep Reinforced Learning for the Governance of a Sample Microgrid - Information Technology in Disaster Risk Reduction Access content directly
Conference Papers Year : 2022

Deep Reinforced Learning for the Governance of a Sample Microgrid

Berkay Gür
  • Function : Author
  • PersonId : 1260138
Gulgun Kayakutlu
  • Function : Author
  • PersonId : 974826

Abstract

A proximal policy optimization reinforcement learning system is proposed to handle the energy dispatch management of a sample microgrid. The microgrid in question has 3 participants of different classifications, signifying their relative importance and how sensitive they are to energy shortages. The energy within the microgrid is generated by these participants, which are individually equipped with a solar panel and a wind turbine for energy generation, and an energy storage system to store this energy. The environmental conditions, i.e. temperature, wind velocity and irradiation figures of Istanbul are considered to obtain accurate energy generation figures. The microgrid is designed to be grid connected in order to compensate for the uncertainties caused by the weather changes, hence service of the utility is accessed when energy produced & stored cannot respond to the demand. Information security of the participants is respected and to that end, direct energy generation, consumption and storage figures are not supplied to the agent, instead only supply and demand figures are transferred. The agent, using this information, after a period of training, optimizes the system for a reward scheme that rewards energy exports and punishes energy deficits and imports. The results verify the feasibility of proximal policy optimization in managing microgrid energy dispatch.
Embargoed file
Embargoed file
0 7 28
Year Month Jours
Avant la publication
Wednesday, January 1, 2025
Embargoed file
Wednesday, January 1, 2025
Please log in to request access to the document

Dates and versions

hal-04120812 , version 1 (07-06-2023)

Licence

Attribution

Identifiers

Cite

Berkay Gür, Gulgun Kayakutlu. Deep Reinforced Learning for the Governance of a Sample Microgrid. IFIP International Workshop on Artificial Intelligence for Knowledge Management (AI4KMES), Aug 2021, Montreal, QC, Canada. pp.169-183, ⟨10.1007/978-3-030-96592-1_13⟩. ⟨hal-04120812⟩
14 View
0 Download

Altmetric

Share

Gmail Facebook X LinkedIn More