An accurate estimation of residential water consumption at the end-use level is helpful to implement strategies aimed at developing efficient water systems, demand models, water saving technologies, or providing feedback to users. If not obtainable through direct metering, this information can be gathered by disaggregating and classifying water consumption data collected at the household inlet point. However, most of the automated techniques for water end-use disaggregation and classification require high-resolution (e.g. 1-s) data, whereas the validation of methods exploiting data at lower resolutions (i.e., 1-min) was generally performed only with synthetic or limited baseline data collected in field. This study aimed to test the robustness of an automated method for end-use disaggregation and classification, applicable on water consumption data at 1-min temporal resolution and originally validated only with a limited water use dataset collected in Italy. The method was tested with new data observed at nine households in the Netherlands. These data – collected at the inlet point of the domestic plumbing systems with 1-s resolution – were automatically pre-processed by means of a new, rule-based, filtering algorithm for the segmentation of the combined water uses into individual end-use events and, in turn, manually labelled by expert analysists. The end-use dataset obtained was then aggregated at the 1-min temporal resolution and input in the model to test its robustness. The results obtained confirmed the potential of the method in disaggregating and classifying water end-use events efficiently and proved that end-use disaggregation and classification is possible even with data whose resolution is closer to that of most commercial water meters.
Application of water consumption smart metering for water loss assessment: a case study
Mazzoni F.;Alvisi S.;Franchini M.
2022
Abstract
An accurate estimation of residential water consumption at the end-use level is helpful to implement strategies aimed at developing efficient water systems, demand models, water saving technologies, or providing feedback to users. If not obtainable through direct metering, this information can be gathered by disaggregating and classifying water consumption data collected at the household inlet point. However, most of the automated techniques for water end-use disaggregation and classification require high-resolution (e.g. 1-s) data, whereas the validation of methods exploiting data at lower resolutions (i.e., 1-min) was generally performed only with synthetic or limited baseline data collected in field. This study aimed to test the robustness of an automated method for end-use disaggregation and classification, applicable on water consumption data at 1-min temporal resolution and originally validated only with a limited water use dataset collected in Italy. The method was tested with new data observed at nine households in the Netherlands. These data – collected at the inlet point of the domestic plumbing systems with 1-s resolution – were automatically pre-processed by means of a new, rule-based, filtering algorithm for the segmentation of the combined water uses into individual end-use events and, in turn, manually labelled by expert analysists. The end-use dataset obtained was then aggregated at the 1-min temporal resolution and input in the model to test its robustness. The results obtained confirmed the potential of the method in disaggregating and classifying water end-use events efficiently and proved that end-use disaggregation and classification is possible even with data whose resolution is closer to that of most commercial water meters.I documenti in SFERA sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
