A hybrid attention‐based long short‐term memory fast model for thermal regulation of smart residential buildings

Abstract

AbstractAn attention‐based long short‐term memory (ALSTM)‐fast model predictive control (MPC) thermal regulation system for buildings is presented. The proposed system is developed to address the challenges associated with traditional heating, ventilation, and cooling (HVAC) control systems, often designed with fixed setpoints and static control strategies, leading to poor performance and suboptimal energy efficiency. The ALSTM‐Fast MPC system, on the other hand, performs the integration of deep learning and optimisation algorithms to predict the thermal behaviour of buildings and optimise the HVAC system control for thermal comfort and energy efficiency. The ALSTM‐Fast MPC system was implemented and evaluated on a real‐world data collected from a building automation system. Additionally, extensive experiments were conducted to analyse the system's performance. The results demonstrated the system's adaptability to changing thermal dynamics and occupancy patterns and its ability to achieve robust and efficient thermal regulation. As a result, a solution for optimising HVAC control in buildings is provided by the proposed ALSTM‐Fast MPC system.