Smart Solar Community

Table of contents

• General info
• Data Sources
• Files
• Sources

General

Reinforcement learning algorithm that optimizes the distribution of solar power within a community that shares solar so that the total amount of electricity drawn from the grid is minimized.

Data Sources

• Install the PVlib library to generate solar data
• Existing electricity usage data taken from the COMED region

Files

• Comed data: Gets usage data
• Solar data: Generates solar data from 2011 to 2018 from PVlib
• RL agent: Class containing Q_learning agent
• Environment: Class that sets up solar environment and functions to get state and reward

Sources

This approach is inspired by a paper by R. Leo, R. S. Milton and A. Kaviya, "Multi agent reinforcement learning based distributed optimization of solar microgrid"