PulseExploreJournal ClubResearchersJournals
Instagram
HomeJournal ClubExplore
Synapse
⌘+K
Synapse
August 1, 2026Open Access

AI-Driven Energy Management in Smart Microgrids with Renewable Integration Using Deep Reinforcement Learning

View Full Paper
Ask AI
Bookmark
Share

Authors

DNDr. Praful NandankarNSNihar Shelke

Discussion

Loading...

Member takes

Overview

Randomized trial shows reduced operational costs and emissions in smart microgrids, suggesting a shift towards AI-driven solutions.

Key Points

  • This research aims to optimize energy management in microgrids utilizing renewable energy sources with a focus on mitigating costs and emissions.
  • Developed a Deep Q-Network (DQN) based energy management system for an islanded microgrid.
  • Trained the DQN agent in a Simulink environment using real data from Nagpur, India.
  • Compared DQN-EMS performance against traditional rule-based control and model predictive control.
  • DQN-EMS reduced daily operational costs by 52.2% compared to rule-based control and 35.7% compared to MPC.
  • CO2 emissions decreased by 56.8% compared to rule-based control and 41.9% compared to MPC.
  • Maintained battery state-of-charge (SOC) within safe limits for 98.4% of operating hours.

Cite This Study

Nandankar et al. (2026) studied this question.

synapsesocial.com/papers/6a6d98aae258b358b3c6c5c5https://doi.org/10.5281/zenodo.21704644
View Full Paper
Ask AI
Bookmark
Share