Real-Time Scheduling for Peak Load Reduction in a Large Set of HVAC Loads

This paper presents a technique to predictably coordinate the activation of Heating, Ventilation and Air Conditioning systems (HVACs) in order to limit the overall peak load of power consumption (peak shaving). The proposed solution represents a viable approach to the Demand-Side Management in the context of a smart grid for this type of loads. The coordination method performs a load shifting based on the discipline of real-time scheduling traditionally studied in the field of computing systems. With this approach, individual constraints on the temperature associated with the activation of each HVAC can be satisfied. The main advantage of the proposed technique is its low computational complexity, which allows to manage large sets of loads. A specific approach is proposed and evaluated to deal with large sets of loads by properly partitioning the load set into sub-sets (scheduling groups) that are scheduled independently from each other. Simulation results based on realistic parameters show that the peak load can be reduced by 35% in normal working conditions, and up to 60% with respect to worst case situations, without affecting the comfort achieved by each HVAC.