A Cautionary Note on Using Smart Plugs for Research Data Acquisition

The availability of low-cost smart plugs that measure and control loads has encouraged their use as a research tool to disaggregate end-user loads. While there has been extensive research on the cybersecurity implications, and demand response capabilities, of smart plugs, there has been little investigation of their primary function: Measuring loads. In this study, we analyze the accuracy of power and energy measurements reported by 5 smart plugs from different manufacturers, with a specific focus on the accuracy of the meters when load levels change rapidly, as is often the case with power electronic loads. The study included Belkin, TP-Link, Etekcity, Emporia, and Sonoff smart plugs, with loads switching at frequencies of 0.01Hz to 2Hz, in addition to the steady-state loads often used for testing these devices. Data indicate significant errors in both energy and power measurement when used to monitor highly variable loads. While maximum energy measurement error for the Belkin product never exceeded 1.7%, maximum energy measurement errors were 54–100% for the other 4 brands. Power measurement accuracy was highly dependent on load variability, with all units exhibiting substantial errors when loads varied, even at rates as low as 0.01-0.05 Hz. All units also exhibited reporting delays of 3–6 seconds, with some delays as high as 20 seconds. The large, load-dependent, errors, call into question the use of these devices for research data acquisition, particularly for monitoring highly variable power electronic loads.