Search

Achieving further reductions in building electricity usage requires a detailed characterization of electricity consumption in homes. Understanding drivers of consumption can inform strategies for promoting conservation and efficiency. While there exist numerous approaches for modeling building energy demand, the use of regularization methods in statistical models can address challenges inherent to building energy modeling while also enabling more accurate predictions and better identification of variables that influence consumption. This paper applies five regularization techniques to regression models of original survey and electricity consumption data for more than one thousand households in California. It finds that of these, elastic net and two extensions of the lasso—group lasso and adaptive lasso—outperform other approaches in terms of prediction accuracy and model interpretability. These findings contribute to methodological approaches for modeling energy consumption in buildings as well as to our understanding of key drivers of consumption. The paper shows that while structural factors predominate in explaining annual electricity consumption patterns, habitual actions taken to save energy in the home are important for reducing consumption while pro-environmental attitudes and energy literacy are not. Implications for improving building energy modeling and for informing demand reduction strategies are discussed in the context of the low-carbon transition.

Demand-side flexibility has been suggested as a tool for peak demand reduction and large-scale integration of low-carbon electricity sources. Deeper insight into the activities and energy services performed in households could help to understand the scope and limitations of demand-side flexibility. Measuring and Evaluating Time- and Energy-use Relationships (METER) is a 5-year, UK-based research project and the first study to collect activity data and electricity use in parallel at this scale. We present statistical analyses of these new data, including more than 6250 activities reported by 450 individuals in 173 households, and their relationship to electricity use patterns. We use a regularization technique to select influential variables in regression models of average electricity use over a day and of discretionary use across 4-h time periods to compare intra-day variations. We find that dwelling and appliance variables show the strongest associations to average electricity consumption and can explain 49% of the variance in mean daily usage. For models of 4-h average “de-minned” consumption, we find that activity variables are consistently influential, both in terms of coefficient magnitudes and contributions to increased model explanatory power. Activities relating to food preparation and eating, household chores, and recreation show the strongest associations. We conclude that occupant activity data can advance our understanding of the temporal characteristics of electricity demand and inform approaches to shift or reduce it. We stress the importance of considering consumption as a function of time of day, and we use our findings to argue that a more nuanced understanding of this relationship can yield useful insights for residential demand flexibility.

Researching the dynamics of residential electricity consumption at finely-resolved timescales is increasingly practical with the growing availability of high-resolution data and analytical methods to characterize them. One methodological approach that is popular for exploring consumption dynamics is load profile clustering. Despite an abundance of available algorithmic techniques, clustering load profiles is challenging because clustering methods do not always capture the temporal aspects of electricity consumption and because clusters are difficult to explain without additional descriptive household data. These challenges limit the use of cluster analysis to better understand behavioral and other drivers of electricity usage patterns. We address these challenges by applying a novel clustering approach to a unique data set of high-resolution electricity and occupant time-use data from UK households. We cluster cumulative rather than raw load profiles to capture their full shape. Our clustering approach identifies two distinct patterns of electricity consumption during evening weekdays (5–9 p.m.), which are primarily differentiated by the timing of their peak demand. Next, we apply several classification algorithms to assess the potential for using time-use activity data to predict membership in these distinct usage clusters. The methods we use are suited to this predictive modeling context and are able to identify key activities driving patterns of electricity demand. We discuss how such an approach can inform more targeted strategies for residential peak demand reduction and response interventions as well as improve our understanding of constraints and opportunities for demand-side flexibility in the residential sector.