One-Year-Ahead Neural Network-Based HVAC Electricity Consumption Optimization: The Influence of Occupancy Schedules

Although knowing the occupancy schedule of a building can save significant energy, ensuring the heating, ventilation, and air-conditioning (HVAC) system does not run needlessly, its uncertain nature has long challenged the development of accurate long-term non-Boolean occupancy-based HVAC management systems. In this paper, we propose an occupancy-based one-year-ahead HVAC electricity consumption optimization approach using feedforward neural networks. The results confirm that including the number of occupants improves the prediction accuracy and provides an optimized profile that allows for a 33.56% of annual electricity saving, a 3.8% more than in the case where neither occupancy-based prediction nor optimization is performed.