Development of an occupancy learning algorithm for terminal heating and cooling units

A significant portion of the North American workforce reports having the ability to alter their daily arrival and departure times for work. As a result, personal preferences translate into individual occupancy profiles. To accommodate these diverse personal schedules, building operators tend to use conservatively short temperature setback periods. In this paper, a year's worth of data gathered by motion sensors placed in private offices in an academic building were analyzed. The predictability of the recurring occupancy patterns was assessed. Drawing upon this, an adaptive occupancy-learning control algorithm which learns the arrival and departure times recursively and adapts the temperature setback schedules accordingly, was developed. Later, the algorithm was implemented in the Energy Management System (EMS) application of the building performance simulation (BPS) tool EnergyPlus. Simulations conducted with this tool indicate that a 10-15% reduction in the space heating and cooling loads can be achieved by applying individual and dynamically evolving temperature setback periods.