Monitoring Home Electrical Power Consumption Patterns Using Secondary Wireless Meters

This report presents analysis of high-temporal-resolution whole-house electrical consumption data obtained from 17 residential buildings located in the Denver metro and Aspen regions of Colorado as well as survey data collected from the participants related to building characteristics and equipment types. The electrical consumption (load) data were obtained using wireless devices manufactured by Copper Labs, Inc. that collected data being broadcast from advanced metering infrastructure (AMI) or automated meter reading (AMR) -type whole-house electrical meters and relayed the data to a central server. Traditional metering systems, which typically collect data at hourly or daily intervals, lack the granularity required to capture nuanced energy behaviors, such as the power cycles of individual appliances or brief but intense usage spikes. By using minute-resolution data collection, this study uncovers valuable insights into household electricity use, identifying peak load times and patterns that contribute to unnecessary energy use. NREL worked with Copper Labs, a Colorado utility (Holy Cross Energy) and other local agencies (BrightVault, LLC., City of Boulder Climate Initiatives, Rocky Mountain Institute, Xcel Energy, and CET and Associates) to identify houses with suitable AMI/AMR meters. The participants from the identified houses were recruited through NREL internally (NREL employees' homes). The anonymized data generated by this study were shared with the participants and will be made publicly available for future researchers. In this study, we first evaluate the general characteristics of raw time series data and 24-hour average load profile for specific houses that participated in the study. We also present the aggregate data analysis after down-selecting houses with overlapping data collection periods. We then perform categorical comparisons of aggregated data that highlight features of the electrical power load being influenced by factors such as the presence of rooftop solar photovoltaics (PV), electrical versus gas appliances for space and water heating, number of occupants, and occupants working from home. Although the analysis presented here is performed in a small sample size after down-selecting (n=17), we present the data collected from all the houses that participated in the study in the appendices of this report. From our data analysis, we first note that houses with rooftop solar PV systems have drastically different load profiles compared to houses without rooftop PV systems. Instead of the morning and evening peaks expected in a load profile, houses with rooftop solar PV tend to have a dip toward the middle of the day and have little or no significant peak during the morning and evening hours. We noted that majority of the time, we can expect houses to experience roughly half a kilowatt of average load and peak loads of roughly 12 kW. Single-family detached houses have slightly greater (roughly 200 W more) average load, but a much wider load range (the difference between maximum and minimum) of power loads compared to multifamily apartment building units. There is also an increasing trend for the power load ranges with increasing square footage. Houses with less than 1,000 sq. ft. area can be expected to have significantly less power load than bigger houses. Houses with electric vehicles (EV) have roughly the same median load, but a much wider range of loads. Rooftop solar PV systems significantly reduce the median load by up to half a kilowatt. In terms of load categorization by space heating type, the difference in overall loads are very small across houses heated with gas, heat pumps, and propane in a colder day. Since our data mostly stretch across the cooling season, we also analyzed the load patterns for cooling system types categorized as heat pumps and traditional central air conditioning types. During a heat wave event, houses with heat pumps can experience lower median and peak loads compared to houses with traditional central air conditioning systems. Houses with electric water heaters have a lower median load than those with electric heat pump water heater systems but almost double the load range. Houses with electric induction stoves have higher loads than those with electric resistance stoves. But houses with electric resistance stoves have a much wider range of load. In terms of the number of occupants, the median load is only slightly different between houses with 2, 3, or 4 occupants. Houses with maximum work-from-home fraction also have the highest variability in the power load. And finally, newer houses do not necessarily have lower load, but a much wider range of load. Ultimately, this report emphasizes the importance of high temporal resolution in advancing both individual energy efficiency and larger sustainability goals.