Abstract

Examining individual households' load profiles and discovering contextual and temporal factors of energy usage (e.g., occupancy, time of use, and occupant activity) gain lots of popularity in recent studies. Given the proliferation of Home and Building Energy Management Systems (HEMS and BEMS) and the availability of high-resolution data of households’ energy usage, it is possible to gain a deeper understanding of temporal factors of load profiles and take advantage of the services offered by these systems. The incorporation of smart meters in the grid has several economic and environmental benefits. These technologies (1) provide the opportunity for appliance scheduling, which can reduce electricity costs on the customer-side, and (2) optimize the integration of intermittent renewable energy sources to the electricity grid. Despite the importance of temporal determinants, previous works mainly focused on determinants of the annual end-use load. Additionally, studies on residential energy mainly address district and city scales, while small-scale analyses are highly overlooked. Based on the identified limitations, in this study, two time-series analysis methods (k-shape clustering and change point detection) are implemented on historical, sensor-collected data of three residential units to discover the frequent occupancy schedule patterns of each household and identify the high- and low- consumption periods within each occupancy pattern. Then LASSO regression is utilized to find the comparative contribution of various activity factors on households’ energy usage (e.g., kitchen-, living room-, bathroom-, or bedroom-related activities indicated by plug loads recorded in specific rooms of the apartments) during the identified energy consumption periods. The results suggest that occupancy patterns are able to explain temporal variations in daily load profiles, and the shape of daily load profiles can be characterized by the occupancy schedule pattern of a day. Furthermore, the analysis of this study can make households aware of the most influencing activities during high-consumption periods. And as a result, households can reduce their energy bills by shifting the energy-consuming activities from high-consumption periods to off-peak periods.