RESUMO
An intelligent cooling management system with a smart home application is proposed to evaluate optimal target temperatures and air conditioner fan modes, thereby maximizing energy efficiency while ensuring residents' comfort. The proposed system integrates a home energy management system with a sophisticated backend infrastructure designed to enable seamless hardware connectivity for real-time data acquisition from various sensors, a gateway, internet of things (IoT) devices, and servers. Furthermore, it serves as a platform for implementing a software data analytics model, structured upon a microservice architecture, aimed at providing optimal feedback control. The data analytics platform utilized in this research integrates two sets of artificial neural networks (ANNs) and a particle swarm optimization (PSO) algorithm for computation and control. This platform is designed not only to gather real-time ambient data and air conditioner usage records but also to regulate the air conditioner's operation autonomously. By considering aprevailing ambient air condition, the ANNs accurately predict power consumption, indoor temperature, and indoor humidity following adjustments in target temperature and fan mode. The PSO-based data analytics model efficiently selects the most suitable target temperature and fan mode, thereby achieving a dual purpose of enhancing energy conservation while minimizing potential occupant discomfort. This optimization is driven by utilizing the predicted mean vote (PMV) calculated through the analysis performed by the ANNs. Validation of the developed intelligent cooling management system was conducted in a real smart home environment inside a single detached two-story house, using an 8,000 BTU air conditioner as the testbed within an 8 × 5 m2 space accommodating four occupants. The implementation results indicate that the proposed intelligent cooling management system can reliably predict the behavior and ambient data of the air conditioner and give the best-operating settings in any different environment scenarios and therefore shows potential for energy savings in smart home applications.
