Distributed reinforcement learning energy management approach in multiple residential energy hubs

Mehdi Ahrarinouri, Mohammad Rastegar, Kiana Karami, Ali Reza Seifi

Research output: Contribution to journalArticlepeer-review

19 Scopus citations

Abstract

Energy management optimization in residential buildings plays an essential role in addressing the problem of energy crisis in the world. This paper introduces a novel method to optimize the energy scheduling for multiple residential buildings in an interconnected framework through transferring energy concepts. To this end, a distributed reinforcement learning energy management (DRLEM) approach is proposed to manage the energy scheduling in multi-carrier energy buildings. In such facilities, equipped with the micro-combined heat and power (micro-CHP) and the gas boiler, the possibility of heat and electrical energy transfer among energy hubs is provided. The effectiveness of the proposed method is verified in a test residential interconnected energy hubs (EHs). Results show a noticeable improvement in energy costs while transferring energy concept is available. In a test frame consisting of three residential EHs, the proposed DRLEM approach in an interconnected mode leads to a daily cost reduction of 3.3% and wasted heat energy decrement of about 18.3% in a typical day compared to the independent mode. Furthermore, in peak tariff energy hours, EHs tend to share their produced excess energy about 23% more than low tariff energy hours to reduce overall energy price.

Original languageEnglish (US)
Article number100795
JournalSustainable Energy, Grids and Networks
Volume32
DOIs
StatePublished - Dec 2022

All Science Journal Classification (ASJC) codes

  • Control and Systems Engineering
  • Renewable Energy, Sustainability and the Environment
  • Energy Engineering and Power Technology
  • Electrical and Electronic Engineering

Fingerprint

Dive into the research topics of 'Distributed reinforcement learning energy management approach in multiple residential energy hubs'. Together they form a unique fingerprint.

Cite this