Application of Artificial Neural Networks for water demand forecasting to optimize water network operations
DOI:
https://doi.org/10.17512/INSTAL.2026.07.02Keywords:
artificial neutral networks, multilayer perception, water demand forecasting, time series analysis, smart water supply systemsAbstract
The article presents an analysis of the feasibility of using artificial neural networks to forecast daily water demand in systems with a complex consumer structure. The objective of the study was to investigate whether the implementation of neural models allows for precise consumption prediction, which is crucial for optimizing network operation, including the reduction of water losses. The research material consisted of measurement data from the years 2020–2025, obtained from a water supply system characterized by a high industrial share, reaching 40%. Within this period, data from 2020–2024 were isolated to construct the model, while data from 2025 were used for its verification. A multilayer perceptron (MLP) was applied to predict daily water demand; the process was divided into two stages: the first utilized uncategorized data, while in the second stage, a "day of the week" type was additionally assigned to the data. The conducted analyses indicate that the introduction of a variable describing the day of the week significantly improves prediction quality. The best generalization capability was demonstrated by the MLP 112-13-1 model, for which the relative mean squared error tested on an independent dataset was 8.40%, and the root mean squared error was 257.3 m³/d. The obtained results confirm the deterministic-stochastic nature of water demand and the high effectiveness of artificial neural networks in mapping non-linear relationships occurring in dynamic water supply systems.
Downloads
References
1. Adamowski, J. F. (2008). Peak daily water demand forecast modeling using artificial neural networks. Journal of Water Resources Planning and Management, 134(2), 119–128. https://doi.org/10.1061/(ASCE)0733-9496(2008)134:2(119)
2. Alvisi, S., & Franchini, M. (2014). Assessment of the predictive uncertainty within the framework of water demand forecasting by using the model conditional processor. Procedia Engineering, 89, 893–900. https://doi.org/10.1016/j.proeng.2014.11.522
3. Bishop, C. M. (2006). Pattern recognition and machine learning. Springer.
4. Candelieri, A., Soldi, D., & Archetti, F. (2015). Layered machine learning for short-term water demand forecasting. Environmental Engineering and Management Journal, 14(9), 2061–2072. https://doi.org/10.30638/eemj.2015.221
5. Cieżak, W., Łomotowski, J., Siwoń, Z., Licznar, P., & Cieżak, J. (2011). Modele neuronowe w analizie i prognozowaniu rozbiorów wody. Instal, (7/8), 61–63.
6. Hotloś, H., Głowacka, J., & Kołodziej, A. (2012). Zmienność poboru wody w systemie wodociągowym Wrocławia. Ochrona Środowiska, 34(4), 23–28.
7. Hutton, C. J., & Kapelan, Z. (2015). A probabilistic methodology for quantifying, diagnosing and reducing model structural and predictive errors in short term water demand forecasting. Environmental Modelling & Software, 66, 87–97. https://doi.org/10.1016/j.envsoft.2014.12.021
8. Kowal, A. L., Świderska-Bróż, M., & Wolska, M. (2022). Oczyszczanie wody (T. 1–2). Wydawnictwo Naukowe PWN.
9. Licznar, P., & Łomotowski, J. (2006). Zastosowanie sztucznych sieci neuronowych Kohonena do prognozowania dobowego poboru wody. Ochrona Środowiska, 28(1), 45–48.
10. Osowski, S. (2013). Sieci neuronowe do przetwarzania informacji. Oficyna Wydawnicza Politechniki Warszawskiej.
11. Piasecki, A., Jurasz, J., & Marszelewski, W. (2016). Wykorzystanie wielowarstwowych sztucznych sieci neuronowych do średnioterminowego prognozowania poboru wody – studium przypadku. Ochrona Środowiska, 38(2).
12. Siwoń, Z., Cieżak, W., & Cieżak, J. (2006). Analiza i prognozowanie szeregów czasowych krótkotrwałego poboru wody. Instal, (2), 44–49.
13. Siwoń, Z., Cieżak, W., & Cieżak, J. (2011). Modele neuronowe szeregów czasowych godzinowego poboru wody w osiedlach mieszkaniowych. Ochrona Środowiska, 33(2), 23–26.
14. Siwoń, Z., Łomotowski, J., Cieżak, W., Licznar, P., & Cieżak, J. (2008). Analizy i prognozowanie rozbiorów wody w systemach wodociągowych. Polska Akademia Nauk.
15. Stańczyk, J., Kajewska-Szkudlarek, J., Łomotowski, J., Lipiński, P., Rychlikowski, P., & Konieczny, T. (2018). Prognozowanie zapotrzebowania na wodę z wykorzystaniem uczenia maszynowego. Gaz, Woda i Technika Sanitarna, 92(10), 372–377. https://doi.org/10.15199/17.2018.10.5
16. Śmigiel, S., Stańczyk, J., Dzimińska, P., Ledziński, D., Andrysiak, T., & Licznar, P. (2023). Predykcja krótkoterminowa zużycia wody w budynkach wielorodzinnych z wykorzystaniem technik uczenia maszynowego. Instal, (12), 72–78. https://doi.org/10.36119/15.2023.12.11
17. Tadeusiewicz, R. (1993). Sieci neuronowe. Akademicka Oficyna Wydawnicza RM.
18. Tiwari, M. K., & Adamowski, J. F. (2015). Medium-term urban water demand forecasting with limited data using an ensemble wavelet-bootstrap machine-learning approach. Journal of Water Resources Planning and Management, 141(2), 1–12. https://doi.org/10.1061/(ASCE)WR.1943-5452.0000454
Downloads
Published
Issue
Section
Categories
License
Copyright (c) 2026 Maciej Bełcik, Monika Nowakowska (Autor)
This work is licensed under a Creative Commons Attribution 4.0 International License.
Authors retain copyright and grant the journal the right of first publication, with the work simultaneously licensed under a Creative Commons Attribution 4.0 International License, which permits others to share the work with appropriate acknowledgment of the authorship and initial publication in this journal.
