• Land and Housing Review
• /
• v.10 no.3
• /
• pp.9-16
• /
• 2019

This study aimed at developing control algorithms for operating a variable refrigerant flow (VRF) heating and cooling system with optimal system parameter set-points. Two artificial neural network (ANN) models, which were respectively designed to predict the heating energy cost and cooling energy amount for upcoming next control cycle, was developed and embedded into the control algorithms. Performance of the algorithms were tested using the computer simulation programs - EnergyPlus, BCVTB, MATLAB in an incorporative manner. The results revealed that the proposed control algorithms remarkably saved the heating energy cost by as much as 7.93% and cooling energy consumption by as much as 28.44%, compared to a conventional control strategy. These findings support that the ANN-based predictive control algorithms showed potential for cost- and energy-effectiveness of VRF heating and cooling systems.