ACEEE 2010: How Much Does Household Behavior Count

Jamie Woods, Loren Lutzenhiser, Mithra Moezzi

Portland State University

  Building simulation models are not necessarily intended to provide accurate estimates of energy use for any individual dwelling. They do present an exceptional opportunity to examine the impact of variations in household behavior – such as occupancy patterns and heating and cooling control practices – on expected energy use. This move, in turn, sets the stage to ask how well current assumptions about occupant behavior represent the range of actual behavior in households, and accordingly to ask about the consequences of inaccuracy and averaging, whether on the scale of individual buildings, such as building design or retrofit recommendations, or on larger scales, such as building codes and policy strategies.

  Our analysis addresses the question of the importance of capturing variability by applying observed data on thermostat temperature setting to residential building energy simulation. We used EnergyPlus and data from 78 California households with programmable thermostats for this analysis. The thermostat-setting data shows that many households manage household temperature considerably more actively than implied in default assumptions employed for demonstrating Title 24 compliance. Temperature-setting practices vary substantially among households, and often from day to day for a given household. Furthermore, this sample revealed that the air conditioning was often off even in the summer, and quite rarely in “auto” programmed mode.

  Results from the simulation experiments suggest that failing to account for such variability in thermostat management, and instead modeling with highly averaged temperature settings, may lead to sizable overestimates of cooling energy use. Furthermore observed differences in thermostat temperature management from household to household accounted for a factor of two difference in estimated average household electricity, under otherwise identical assumptions about the physical house. 

  Working from these results, and drawing on further sources of behavioral data (e.g., the American Time Use Survey), the team has begun to test the impacts of other aspects of occupant behavior, such as window-opening and occupancy patterns, on model estimates of energy use. Simulation model architecture, however, may not always readily allow expression of things that people really do in their homes. Based on results from these explorations, we will continue to assess, for a variety of different applications of building simulation, the implications of differences between real household behavior and the standardized behaviors assumed in much simulation work.

Videos and slides from a poster session at the ACEEE Summer Session on Buildings held in Asilomar, CA.

This is a series of four youtube videos in a playlist.

This is a prezi presentation. Prezi moves away from the concept of a slide and seems to be a big improvement over the Powerpoint, "slide of bullet points" style of presention.