Optimizing Air-Conditioner Target Temperature and Fan Mode for Energy Conservation Based on Long-Short Term Memory and Particle Swarm Optimization

A data analytics model for a cooling management system is proposed to find the optimal adjustment of target temperatures and air-conditioner fan mode to maximize energy efficiency while maintaining residents’ comfort. The ambient scenarios and usage of air conditioner data can be collected from sensors and Internet of Things (IoT) devices installed in an occupied home. Long-short-term memory (LSTM) algorithms have been developed to predict the power consumption of the air conditioner and the indoor temperature and humidity from ambient scenario data and adjustment data of target temperatures and air-conditioner fan mode. A particle swarm optimization (PSO) algorithm has been developed to be capable of selecting the target temperatures and the air-conditioner fan mode that are most appropriate for energy savings while controlling comfort for the occupants by using a predicted mean vote (PMV) as a criterion. The implementation results indicate that the proposed data analytics model can effectively predict the power consumption of the air conditioner and the indoor ambient conditions and succeed in finding the best adjustment case for the air conditioner in any different ambient scenarios, thereby increasing the potential for home energy savings.