CLED: Computer Lab Energy Dataset

The increasing availability of sensors to track device activities enables smart environment applications, for example, occupancy pattern identification in building offices. However, the lack of open datasets for analyzing and testing such applications delays the deployment and complicates adaptation to unknown scenarios. Despite the existence of a varied catalog of energy datasets available at the residential level, there are few at the industrial and university levels. In this work, we present a Computer Lab Energy Dataset (CLED) of electric energy obtained in a controlled environment with devices commonly found in an engineering classroom laboratory. The baseline of devices we will present is a good representation of typically used devices we may find in such a variety of situations, especially in South America. We begin by describing the scenario and the devices under study, then explain the telemetry infrastructure and protocols used to obtain the data. Finally, we analyze the characteristics of the data obtained at the electrical level and demonstrate various applications, such as load characterization and non-intrusive disaggregation, which may help organizations looking for energy sustainability to develop electricity policies based on this data analysis.