Predicting consumption events in a water monitoring system

Abstract

Monitoring water consumption has multiple benefits nowadays. Big data collected from the sensors provide a consistent basis for the decision-making processes in terms of establishing the indices and criteria needed to optimize the water demand. In this study, the data provided by four distinct water consumption outlets (hot/cold water sink, toilet, and shower) from multiple households were analyzed. A clustering analysis revealed a visual overview of the consumption events from each outlet. Then, classification methods were used to predict the source of water consumption events using four algorithms based on machine learning and deep learning. The proposed methods and results are promising towards the development of a decision support system for streamlining water consumption in urban water distribution systems.