SUPR
Integrating reinforcement learning and predictive control for smart home energy management
Dnr:

NAISS 2024/22-551

Type:

NAISS Small Compute

Principal Investigator:

Meng Yuan

Affiliation:

Chalmers tekniska högskola

Start Date:

2024-04-15

End Date:

2025-05-01

Primary Classification:

20202: Control Engineering

Webpage:

Allocation

Abstract

The usage of this HPC platform is based on a project related to smart home energy management funded by the European Commission. This project aims to use reinforcement learning with model-based methods such as model predictive control to develop an energy management system to improve energy efficiency at home. At the starting phase of this project, we aim to use the HPC with Python to simulate the charging and discharging performance of the battery. The simulation involves many factors such as temperature, and state of charge, and is very time-consuming. Moreover, it can take hundreds and thousands of cycles to make the battery degrade. Using this platform accelerates the simulation time.