Temporal Segmentation of Urban Water Consumption Patterns Based on Non-Parametric Density Clustering
Open Access
- 3 October 2025
- journal article
- research article
- Published by MDPI AG in Technologies
- Vol. 13 (10) , 449
Abstract
The management of modern water supply systems requires a detailed analysis of consumption patterns in order to optimize pump operation schedules, reduce energy costs, and support the development of intelligent management systems. Traditional clustering algorithms are applied for these tasks; however, their limitation lies in the need to predefine the number of clusters. The aim of this study was to develop and validate a non-parametric method for clustering daily water consumption profiles based on a modified DBSCAN algorithm. The proposed approach includes the automatic optimization of neighborhood radius and the minimum number of points required to form a cluster. The input data consisted of half-hourly water supply and electricity consumption values for the water supply system of Gomel (Republic of Belarus), supplemented with the time-of-day factor. As a result of the multidimensional clustering, two stable regimes were identified: a high-demand regime (6:30–22:30), covering about 46% of the data and accounting for more than half of the total water supply and electricity consumption, and a low-demand regime (0:30–6:00), representing about 21% of the data and forming around 15% of the resources. The remaining regimes reflect transitional states in morning and evening periods. The obtained results make it possible to define the temporal boundaries of the regimes and to use them for data labeling in the development of predictive water consumption models.Keywords
This publication has 31 references indexed in Scilit:
- Identification of Easily Accessible Urban Water Consumption Factors for Energy-Efficient Management of Pumping StationsWater Conservation Science and Engineering, 2025
- A SMOTE PCA HDBSCAN approach for enhancing water quality classification in imbalanced datasetsScientific Reports, 2025
- Short‐Term Memory and Regional Climate Drive City‐Scale Water Demand in the Contiguous USEarth's Future, 2025
- Integrated STL-DBSCAN algorithm for online hydrological and water quality monitoring data cleaningEnvironmental Modelling & Software, 2024
- Gypsum Binder With Increased Water Resistance Derived From Membrane Water Desalination WasteEngineering Reports, 2024
- K-DBSCAN: An improved DBSCAN algorithm for big dataThe Journal of Supercomputing, 2020
- Smart meters data for modeling and forecasting water demand at the user-levelEnvironmental Modelling & Software, 2020
- Implications of data sampling resolution on water use simulation, end-use disaggregation, and demand managementEnvironmental Modelling & Software, 2018
- ST-DBSCAN: An algorithm for clustering spatial–temporal dataData & Knowledge Engineering, 2007
- Clustering of time series data—a surveyPattern Recognition, 2005