Disaggregation of Household Water Use by Means of a Rule-based, Automated Methodology

An improved management of water resources is required to face the demand increase in urban areas and improve water infrastructure efficiency. The application of smart-metering technology to the residential sector allows to understand where and when the water is used supporting both customers and utilities. Moreover, additional information about water use at the domestic individual level can be obtained by coupling smart-metered data with a methodology for water disaggregation into end-uses. Recently, several techniques for automated water disaggregation have been introduced. Most of the developed disaggregation methods include the installation of a single smart meter (typically at the household water inlet point) and require high temporal resolution data, e.g. 1-10 second [1]. On the other hand, methods relying on medium resolution data e.g. minute [2], have not been tested in the water field yet, while lower temporal resolutions are not sufficient to perform a complete water use disaggregation at end-use level. In this work, a new, rule-based and automated disaggregation methodology is presented. The approach relies on volume information recorded with minute readings by a single smart meter at the household inlet point. The developed algorithm makes use of deterministic rules based on physical water use parameters (i.e. duration, flow rate, consumed volume) to classify water events into end-uses. Unlike similar approaches, the disaggregation can be achieved with medium frequency (i.e. minute resolution), which is comparable to the frequencies characterizing the most widespread commercial smart meters.