Engineering sensor networks for energy studies of the built environment

This study provides insight into the deployment of a scaled up wireless sensor network (WSN) from residential settings to medium sized apartment complex in the Western United States with a dry climate, in order to provide for environmental and inhabitant information that could be used in energy studies. The WSN is comprised of Epic core modules fit with the following sensors; Honeywell HIH‐5030 (humidity), Panasonic AMS302 (photodiode – light intensity), Panasonic AMN41122 (PIR motion), Microchip TC1046 (temperature), and MEDER MK06 (door switch). The deployed WSN makes use of a split low‐power sensor/core network architecture to reduce energy consumption of the sensors. When active, sensor nodes utilize 2.2 mA of current and ∼660 μJ of power is required to read all sensors (excluding the PIR) at a single point in time. A successful transmission used 22.5 mA for 78 ms, consuming 5 mJ and a failed transmission used 22.5 mA for 534 ms, consuming 36 mJ. Such low‐power operation enables sensor nodes to operate for over one year from a single coin‐cell battery. The observed packet delivery was best for nodes located closest to a base station and decreased with distance from the base station. However, slightly less than half of the expected number of packets were received overall. Finally, base energy analysis of the building under study was cast as three hypotheses in regards to the system being deployed and available metrics with the data collected. Results indicate inhabited units with individual swamp coolers, on average, utilize more energy (12.4 kWh/d, n = 3) vs. units that make use of a central evaporative cooling system (5.4 kWh/d, n = 3). Further, increased relative humidity in this study correlates with increased energy use. Our study demonstrates a realistic rapid deployment of WSN used to characterize energy use in a medium sized apartment complex in a dry climate in the Western U.S. It is proposed that such studies provide information for more rigorous energy studies that could be used in retro‐fit decision frameworks, energy use behavior studies, or via integration with additional internet of things technologies. © 2017 American Institute of Chemical Engineers Environ Prog, 36: 539–547, 2017