Robust Learning Model Predictive Control for Periodically Correlated Building Control

TL;DR

This paper introduces a robust learning model predictive control framework that leverages periodic correlations in uncertainties like weather and occupant behavior to optimize building energy management.

Contribution

It presents a novel control scheme that explicitly incorporates periodic correlation structures into the learning model predictive control for building operations.

Findings

Enhanced robustness to weather and occupant variability

Improved energy efficiency in building control

Effective handling of periodic uncertainty patterns

Abstract

Accounting for more than 40% of global energy consumption, residential and commercial buildings will be key players in any future green energy systems. To fully exploit their potential while ensuring occupant comfort, a robust control scheme is required to handle various uncertainties, such as external weather and occupant behaviour. However, prominent patterns, especially periodicity, are widely seen in most sources of uncertainty. This paper incorporates this correlated structure into the learning model predictive control framework, in order to learn a global optimal robust control scheme for building operations.