A multiagent saver for the automatic management of HVAC systems

The high penetration of renewable energy generation has strongly modified the behavior of those users, or prosumers, that after shifting from passive receivers to active contributors, often aim to reach a zero net energy condition through a smart management of production, consumption and storage systems. A method for reducing energy consumptions of a Heating, Ventilation and Air Conditioning (HVAC) system of a prosumer node is herein illustrated and verified by means of experimental field tests. Based on information from internal and external sensors, the proposed predictive controller can effectively minimize HVAC energy use by activating, for each working condition, the best agent, which is to be chosen, time by time, among three different, predefined energy savers. In all operating situations comfort levels in the conditioned zone were maintained within predefined limits.