Search for: All records

Abstract The quantification of residential water end uses is an important component of improving the sustainability of urban water infrastructure. Disaggregation and classification methods based on statistical learning are used in research and practice to extract meaningful insights from smart water meter data. These insights can also reflect individual behaviors within the built environment, enabling end-user activity detection from water consumption patterns. In this study, we present an initial framework for classifying residential water end uses and assisting with discerning between perceived typical and atypical water-use behavior in a permanent supportive housing context. Classification schemes, based on fine-resolution temporal flow data, incorporated baseline activity to inform what typical water use was for individuals while also considering general trends in specific end uses such as showers, toilet flushes, and leaks. We found that while atypical activity based on end-use duration and frequency might fall outside the normally-distributed expected value for a period of interest, it need not be the case for all atypical activity. Defining atypical activity based on prescriptive guidelines might not align with normative behavior for an occupant transitioning into housing. Additionally, exogenous variables can affect occupant behavior regarding water end uses and this impact should be accounted for in analytical frameworks. Our findings can specifically inform supportive services provided by stakeholders responsible for the well-being of individuals in their care via non-intrusive, privacy-respecting insights on occupant behavior. 

Abstract Understanding shower habits is critical for developing effective residential water conservation and efficiency strategies. Previous research has focused on single-family homes, but less is known about shower behavior among college-aged individuals in university student housing. This study examines the shower habits of students at the University of Illinois Urbana-Champaign, comparing them with U.S. single-family residential households regarding shower duration, time-of-day, and day-of-week. Using Conditional Tabular Generative Adversarial Networks to generate synthetic data, we address sample size limitations and confirm the validity of our results. Our findings reveal that student housing showers tend to be longer in duration and more variable compared to showers in single-family residences. Unlike the predictable routines seen in single-family homes, student housing inhabitants display less consistent showering habits, with different time-of-day patterns that challenge typical conservation incentives. Major shower events also occur more frequently before weekends in student housing. These insights emphasize the need for tailored water conservation strategies in semi-permanent residential settings. We recommend further exploration of targeted interventions, including educational campaigns, real-time feedback mechanisms, and gamification, to foster sustainable shower habits among university students. This study contributes to sustainable water management by providing actionable strategies within a sociotechnical systems lens for enhancing water conservation in semi-permanent residential contexts. 

Abstract Water sustainability in the built environment requires an accurate estimation of residential water end uses (e.g., showers, toilets, faucets, etc.). In this study, we evaluate the performance of four models (Random Forest, RF; Support Vector Machines, SVM; Logistic Regression, Log‐reg; and Neural Networks, NN) for residential water end‐use classification using actual (measured) and synthetic labeled data sets. We generated synthetic labeled data using Conditional Tabular Generative Adversarial Networks. We then utilized grid search to train each model on their respective optimized hyperparameters. The RF model exhibited the best model performance overall, while the Log‐reg model had the shortest execution times under different balanced and imbalanced (based on number of events per class) synthetic data scenarios, demonstrating a computationally efficient alternative for RF for specific end uses. The NN model exhibited high performance with the tradeoff of longer execution times compared to the other classification models. In the balanced data set scenario, all models achieved closely aligned F1‐scores, ranging from 0.83 to 0.90. However, when faced with imbalanced data reflective of actual conditions, both the SVM and Log‐reg models showed inferior performance compared to the RF and NN models. Overall, we concluded that decision tree‐based models emerge as the optimal choice for classification tasks in the context of water end‐use data. Our study advances residential smart water metering systems through creating synthetic labeled end‐use data and providing insight into the strengths and weaknesses of various supervised machine learning classifiers for end‐use identification.