The first phase of the SIM Home project discussed the potential effects of variations in user behavior on estimates of energy consumption for plug load devices. Three types of variations were explored: (1) the duration or number of times per day the device is used, (2) the frequency or periodicity of that use (e.g., five hours all at once, or in five separate sessions), and (3) use or disabling of power management features. The goal of this project is to further explore the range of behavioral device use tests proposed in the initial SIM Home final report. The main research question driving the testing would be how much each of the three aspects contributes to higher energy consumption, for each type of device tested.

California’s ongoing push to Zero Net Energy (ZNE) for residential and commercial buildings encourages reducing energy consumption at all levels, including miscellaneous electric loads and other plug load devices. Understanding this problem from both a top-down and bottomup approach is required to produce deep, forward-facing and sustainable technological solutions. In an earlier project for SCE, “Technology Roadmap towards 2030 and Beyond” (Klopfer, Rapier, Luo, Pixley, & Li, 2017), CalPlug presented a top-down view of plug loads and the impact of these on ZNE efforts. In another related project for SCE, CalPlug presented a bottom-up view of plug load energy use in the first SIM Home (Simulation, Integration, and Management Home) report (Xia, Pixley, & Gago-Masague, 2017). The current project is an extension of the analyses and methodologies outlined in the original SIM Home project, assessing the role of behavior and device configuration on energy use. To this end, CalPlug developed new open-source tools – the Plug Load Simulator Suite 1.2 (PLSim)1 and the Marginal Intervention Savings of Energy Reporter (MISER)2 – as well as a database for device state-wise energy usage. With these tools, CalPlug calculated energy usage across a defined range of usage conditions for household consumer electronics and other plug load devices, then analyzed each set of outcomes to determine points of substantial energy use related to usage behavior. In the first SIM Home report, CalPlug proposed testing a range of device use profiles to supplement insights gained through standard energy testing. Standardized test protocols such as ENERGY STAR® assess and compare the energy consumption of all devices of a certain type using a single (presumably average) set of parameters. In most cases, this is based on a controlled and repeatable testing scenario rather than field observations of actual usage. The device use profile method supplements this approach by asking how much consumption could vary depending on how devices are used across a wide range of households. CalPlug discussed three important aspects of use that could affect energy consumption. The first aspect is the amount of active use, such as the number of hours the users watch TV, or how many cups of coffee they make each day. The second aspect is the pattern of use over time, which may have an impact, particularly if gaps between periods of active use require additional warm-up stages or more periods of idle prior to automatically transitioning to standby or sleep mode. The third aspect is the extent to which Power Management (PM) options are used, such as disabling default sleep or auto-off settings (on-board, or automatic PM), or turning off the device when finished (manual PM). In the original SIM Home project, CalPlug presented an automated testing and display system to evaluate the energy impact of these behavioral aspects of usage on a set of common household plug load devices. The scope of the work limited the original tests to three profiles for each device: high, low, and moderate usage. The results from these simulations indicated that even if households contained the same plug load devices, actual household energy consumption could be substantially higher or lower than standard estimates, depending on how devices are used. The current project expands on the original project, and describes how profile-based evaluation can be expanded as a core component to assessing plug load device energy consumption. Compared to the previous study, a larger set of profiles was constructed for each device to fully address the roles of the three aspects of behavior (active use, pattern of use, and PM). Eleven example devices were selected and tested: two TVs (HD and 4K), a sound bar, a satellite set-top box, a streaming device, a video game console, desktop and laptop computers, two pod coffee maker models, and a rice cooker. For each device, a minimum of three levels (low, moderate, and high) was defined for each aspect (active use, pattern, and PM). Each device's profile set included all logical combinations of levels of the three aspects. The levels of each aspect were based on ENERGY STAR protocols (for "moderate" active usage) and, whenever possible, on prior research, self-report survey data, and the devices’ PM options and factory default settings. For any devices that did not have sufficient field data, educated assumptions were required to create the profiles. The plug load devices were tested using high-resolution equipment to capture the states of operation and the power consumption at each state. CalPlug then created schedules that were entered into the energy modeling software as parameters to calculate energy consumption over time, given the usage pattern described by the device use profile. Lastly, the energy modeling software was used to output the annual energy consumption of the plug load devices. This investigation evaluated two primary points: the use of the specific methodology and tools developed here, based on granular usage evaluation across multiple residential devices, and the demonstration of a generalized approach for incorporating behavior into device energy use evaluation. This report details the estimated effects of active use, pattern of use, and PM on energy usage for selected devices. This analysis first identifies which devices exhibit large variations in energy use across profiles. It then quantitatively assesses whether the variation for each device is more strongly driven by the amount of active use, pattern of use, or PM behaviors, each suggesting different remediation strategies.