Modelling Physical Water Losses in the Urban Water Supply Networks of Balıkesir Province Using Artificial Neural Networks

Öz

Successive sluice gates are designed to maintain a stable discharge rate in open channels despite variations in upstream water levels. This study numerically investigates the hydraulic performance of successive sluice gates under different operational conditions. Four gate configurations and five upstream flow depths were simulated to analyze their flow regulation efficiency. The numerical results were validated against experimental data, revealing differences in outlet discharge of less than 10%, indicating that the simulation model provides reliable accuracy. Following model validation, the computed outflow rates for each configuration were compared with the corresponding design discharge, showing an average deviation of approximately 8% in all tested cases. These results confirm that successive sluice gates are capable of maintaining a nearly constant flow rate, even under fluctuating upstream water levels. Moreover, energy dissipation analysis showed that losses were negligible when only the first gate was active. However, as the upstream depth increased and additional gates were engaged, energy dissipation rose to about 15–22%. Overall, the findings demonstrate that successive sluice gate systems can effectively combine precise flow regulation with adequate energy dissipation, making them suitable for hydraulic structures requiring stable discharge control under variable inflow conditions.

Anahtar Kelimeler

• Energy dissipation
• FLOW-3D software
• Successive sluice gates
• Free surface flow.

Balıkesir ili Şehir Şebeke Hatlarındaki Fiziki Su Kayıplarının Yapay Sinir Ağları ile Modellenmesi

Öz

Dünyanın üçte ikisi sularla kaplı olmasına rağmen bu kısmın büyük bir bölümü (%97,5) tatlı olmayan sulardan oluşmaktadır. Tatlı olan suyun ise üçte ikisi buz halinde olduğu için her an kullanıma hazır değildir ve tatlı suya ihtiyaç dünya nüfusunun hızla artmasına paralel olarak her yıl (12.500 km3) artmaktadır. Tatlı suların yetersiz olduğu toplumlarda deniz suyu gibi uygunsuz kaynaklara yönelimi artırmıştır. Dünya genelinde içilebilir su miktarının azalması son zamanlarda su kayıplarını önleme çalışmalarına önemi artırmıştır. Türkiye’de de bu çalışmaların önemi anlaşılmakta olup, su kayıplarının kontrolüne yönelik 8 Mayıs 2014 tarihinde 28994 sayılı yönetmelik yayınlanmıştır. Su kayıpları, fiziki ve idari kayıp olmak üzere iki sınıftan oluşur. Toplam su kayıplar içinde fiziki ve idari kayıpların payları ülkeden ülkeye farklılık gösterse de genel olarak büyük kısmı (yaklaşık %60’ı) fiziksel kayıplardan oluşmaktadır. Türkiye’de de genellikle kayıpların büyük bir kısmı fiziki su kayıplarından oluşmaktadır. Fiziki su kaybı Su Dağıtım Şebekeleri’nde (SDŞ) kaybolan ve kullanılmayan su olarak tanımlanır ve SDŞ’lerdeki ve servis bağlantılarındaki kayıp-kaçaklar ve depolarda oluşan taşmalar ve sızıntılardan oluşur. Fiziki su kaybı ile ilgili önceki çalışmalar ağırlıklı olarak sızıntı tespiti ve sızıntı yeri belirleme üzerine yapılmıştır. Bu durum, SDŞ’lerde gerçek kayıpların toplam sisteme giren su hacmine oranını ifade eden Fiziki Su Kayıp Oranı’nı (FSKO) etkileyen parametreler ve tahmini üzerine çalışmalara olan ihtiyacı artırmıştır. Bu çalışmada, FSKO tahminleri, Balıkesir ilinin yirmi ilçesinin 2018-2022 yıllarına ait nüfus, sisteme giren su, tüketilen su miktarı, şebeke yaşı, ortalama şebeke basıncı ve ortalama boru çapı parametreleri kullanılarak gerçekleştirilmiştir. Model tahminleri MLFFNN (LM) yöntemi ile gerçekleştirilmiştir. Modellerin tahmin doğruluğu, MAPE, MAD, RMSE, R² ve NSE performans ölçütlerine göre yapılmıştır. FSKO tahmin modellerinde en iyi sonuç, performans ölçütleri (n: 71 Test MAPE: 13.41, MAD: 2.96, RMSE: 3.79, R²: 0.66, NSE: 0.59) ile M20’dir.

Anahtar Kelimeler

• Balıkesir
• Makine Öğrenmesi
• Fiziki Su Kayıp Oranı
• Su Dağıtım Şebekesi.

Kaynakça

1. [1] Songur, M., Dabanlı, A., Yılmazel, B. and Kürkçüoğlu, M. A. Ş., 2021. The importance of SCADA in preventing physical losses in water distribution networks: The case of ASKİ, Afyon Kocatepe University Journal of Science and Engineering, 21, 6, 1424-1433.
2. [2] Kurban, R., 2022. Industry 4.0, internet of things, cloud computing, machine learning and optimization applications in the automation of water supply and distribution systems, Research and Evaluations II, 49-70.
3. [3] Liemberger, M., 2024. Reducing non-revenue water, Asterra.
4. [4] Fırat, M. and Güleç, A., 2021. Control and management of water losses and leakages in drinking water distribution systems, in: Sarıkaya M.S. (Editor), Turkish Water Institute (SUEN), Istanbul. URL Address: https://www.suen.gov.tr/uploads/icmesuyu-dagitim-sistemlerinde-su-kayip-ve-kacaklarinin-kontrolu-ve-yonetimi.pdf Access date: 20.04.2026.
5. [5] Ministry, K., 2013. Tenth Development Plan 2014-2018, Water resources management and security, Specialization commission report, Ankara. URL Address: https://www.hidropolitikakademi.org/uploads/wp/2017/02/Su_Kaynaklari_Yonetimi_ve_Guvenligi_oik.pdf Access date: 14.05.2026.
6. [6] Ramana, G. V., 2013. Efficient water distribution system using supervisory control and data acquisition (SCADA): A case study of Vijayawada Municipal Corporation, International Journal of Engineering Sciences and Research Technology, 2, 5, 1055-1061.
7. [7] Kanakoudis, V., Papadopoulou, A., Tsitsifli, S., Curk, B. C., Karleusa, B., Matic, B., ... and Banovec, P., 2017. Policy recommendation for drinking water supply cross-border networking in the Adriatic region, Journal of Water Supply: Research and Technology—AQUA, 66, 7, 489-508.
8. [8] URL , Access date: 08.05.2026.

9. [9] Muhammetoğlu, H. and Muhammetoğlu, A., 2017. Handbook for the control of water losses in drinking water supply and distribution systems, Republic of Türkiye Ministry of Forestry and Water Affairs, General Directorate of Water Management.
10. [10] WWF-Türkiye, 2026. Current status of water resources in Türkiye. URL Address: https://www.wwf.org.tr/kesfet/tatli_su/turkiyede_su_kaynaklarinin_guncel_durumu/ Access date: 13.05.2026.
11. [11] Özkan, M., 2023. National water plan prepared, Agriculture and Forestry Magazine. URL Address: https://www.turktarim.gov.tr/Haber/869/ulusal-su-plani-hazirlandi Access date: 27.04.2026.
12. [12] URL , Access date: 20.04.2026.
13. [13] Farley, M., Wyeth, G., Ghazali, Z. B. M., Istandar, A. and Singh, S., 2008. The manager’s non-revenue water handbook: A guide to understanding water losses, in: Dijk, N., Raksakulthai, V. and Kirkwood, E. (Editors), Ranhill Utilities Berhad and the United States Agency for International Development, Bangkok, Thailand. URL Address: https://bear.warrington.ufl.edu/centers/purc/DOCS/PAPERS/other/BERG/SandysSelections/1302_The_Managers_NonRevenue.pdf Access date: 07.05.2026.
14. [14] Pearson, D., 2019. Standard definitions for water losses, in: Donnelly A. (Editor), UK: IWA Publishing, London. [15] Ríos, J. C., Santos-Tellez, R. U., Rodríguez, P. H., Leyva, E. A. and Martínez, V. N., 2014. Methodology for the identification of apparent losses in water distribution networks, Procedia Engineering, 70, 238-247.
15. [16] Alegre, H., Baptista, J. M., Cabrera Jr, E., Cubillo, F., Duarte, P., Hirner, W., ... and Parena, R., 2017. Performance indicators for water supply services, Third Edition, IWA Publishing, London.
16. [17] Kingdom, B., Liemberger, R. and Marin, P., 2006. The challenge of reducing non-revenue water in developing countries—How the private sector can help: A look at performance-based service contracting, The World Bank, Washington. URL Address: https://documents1.worldbank.org/curated/en/385761468330326484/pdf/394050Reducing1e0water0WSS81PUBLIC1.pdf Access date: 22.04.2026.
17. [18] Oertlé, E. and Knobloch, A., 2010. Guidelines for water loss reduction, Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ). URL Address: https://www.icafrica.org/fileadmin/documents/Knowledge/GIZ/Guidelines-water-loss-reduction.pdf Access date: 14.05.2026.
18. [19] Puust, R., Kapelan, Z., Savic, D. A. and Koppel, T., 2010. A review of methods for leakage management in pipe networks, Urban Water Journal, 7, 1, 25-45.
19. [20] Besner, M. C., Prévost, M. and Regli, S., 2011. Assessing the public health risk of microbial intrusion events in distribution systems: Conceptual model, available data, and challenges, Water Research, 45, 3, 961-979.
20. [21] Adedeji, K. B., Hamam, Y., Abe, B. T. and Abu-Mahfouz, A. M., 2017. Towards achieving a reliable leakage detection and localization algorithm for application in water piping networks: An overview, IEEE Access, 5, 20272-20285.
21. [22] Gerger, R. and Aslan, A., 2019. Detection of drinking water losses and leakages for Şanlıurfa province, Harran University Journal of Engineering, 4, 2, 26-35.
22. [23] Leu, S. S. and Bui, Q. N., 2016. Leak prediction model for water distribution networks created using a Bayesian network learning approach, Water Resources Management, 30, 8, 2719-2733.
23. [24] Jang, D., Park, H. and Choi, G., 2018. Estimation of leakage ratio using principal component analysis and artificial neural network in water distribution systems, Sustainability, 10, 3, 750.
24. [25] Kizilöz, B., 2021. Prediction model for the leakage rate in a water distribution system, Water Supply, 21, 8, 4481-4492.
25. [26] URL , Accessed on: 02 August 2025.
26. [27] URL , Accessed on: 01 August 2025.
27. [28] URL , Accessed on: 6 August 2025.
28. [29] URL , Accessed on: 6 August 2025.
29. [30] Box, G. E. and Cox, D. R., 1964. An analysis of transformations, Journal of the Royal Statistical Society Series B: Statistical Methodology, 26, 2, 211–243.
30. [31] Meloun, M., Sáňka, M., Němec, P., Křítková, S. and Kupka, K., 2005. The analysis of soil cores polluted with certain metals using the Box–Cox transformation, Environmental Pollution, 137, 2, 273–280.
31. [32] Sennaroglu, B. and Senvar, O., 2015. Performance comparison of the Box–Cox transformation and weighted variance methods with the Weibull distribution, Journal of Aeronautics and Space Technologies (Havacılık ve Uzay Teknolojileri Dergisi), 8, 2, 16.
32. [33] Şanlı, M. and Özerbaş, D. H. S., 2021. The effect of STEM activities on students’ attitudes towards STEM fields and their motivation towards science, Manisa Celal Bayar University Journal of Social Sciences, 19, 3, 139–154.
33. [34] Tadese, M. T., Kumar, L., Koech, R. and Zemadim, B., 2019. Hydro-climatic variability: a characterisation and trend study of the Awash River Basin, Ethiopia, Hydrology, 6, 2, 35.
34. [35] Erdem, F., 2022. Parameter estimation in crystal sugar production using MLR, ANN and ANFIS, Pamukkale University Journal of Engineering Sciences, 28, 7, 987–992.
35. [36] Yavuz, S. and Deveci, M., 2012. The effect of statistical normalisation techniques on the performance of artificial neural networks, Erciyes University Journal of Economics and Administrative Sciences, 40, 167–187.
36. [37] Moradi, G. R., Dehghani, S., Khosravian, F. and Arjmandzadeh, A., 2013. The optimised operational conditions for biodiesel production from soybean oil and the application of artificial neural networks for estimating biodiesel yield, Renewable Energy, 50, 915–920.
37. [38] Eğrioğlu, E., Aladağ, C. H., Yolcu, U., Uslu, V. R. and Başaran, M. A., 2009. A new approach based on artificial neural networks for high-order multivariate fuzzy time series, Expert Systems with Applications, 36, 7, 10589–10594.
38. [39] Koyuncu, I., Yılmaz, C., Alçın, M. and Tuna, M., 2019. Artificial neural network optimisation of a geothermal energy-assisted hydrogen generation system, In 4th International Conference on Engineering Technology and Applied Sciences, 360–364.
39. [40] İclal, G., 2016. Solving a system of linear differential equations using a multi-layer perceptron artificial neural network, 18th Academic Computing Conference, Aydın, 3–5.
40. [41] Atalay, M. and Çelik, E., 2017. Artificial intelligence and machine learning applications in big data analysis, Journal of the Institute of Social Sciences, Mehmet Akif Ersoy University, 9, 22, 155–172.
41. [42] Uçar, M. K., Bozkurt, M. R., Bilgin, C. and Polat, K., 2017. Retracted Article: Automatic detection of respiratory arrests in OSA patients using PPG and machine learning techniques, Neural Computing and Applications, 28, 10, 2931–2945.
42. [43] Karakurt, M., Saygılı, H. and Karcı, A., 2024. Comparison of activation functions in the KarcıFANN method, In 2024 8th International Artificial Intelligence and Data Processing Symposium (IDAP), IEEE, 1–6.
43. [44] URL , Accessed: 16 December 2025.
44. [45] Ghaedi, M., Ansari, A., Assefi Nejad, P., Ghaedi, A., Vafaei, A. and Habibi, M. H., 2015. Artificial neural network and bees algorithm for removal of Eosin B using cobalt oxide nanoparticle-activated carbon: Isotherm and Kinetics study. Environmental Progress and Sustainable Energy, 34, 1, 155–168.
45. [46] Bilgiç, H. H., Yağlı, H., Koç, A. and Yapıcı, A., 2016. Power estimation in an experimental organic Rankine cycle using artificial neural networks (ANNs), Selçuk University Journal of Engineering, Science and Technology, 4, 1, 7–17.
46. [47] Meng, X., Jia, M. and Wang, T., 2014. Neural network prediction of biodiesel kinematic viscosity at 313 K. Fuel, 121, 133–140.
47. [48] Sharma, M. and Sarma, K. K., 2017. Soft computation-based spectral and temporal models of linguistically motivated Assamese telephonic conversation recognition, CSI Transactions on ICT, 5, 2, 209–216.
48. [49] Akman, M., Uçar, M. K., Uçar, Z., Uçar, K., Baraklı, B. and Bozkurt, M. R., 2022. Determination of body fat percentage by gender using photoplethysmography signals with a machine learning algorithm. IRBM, 43, 3, 169–186.
49. [50] Özkan, Ö., Yıldız, M. and Köklükaya, E., 2011. Improving diagnostic accuracy by supplementing laboratory tests used in the diagnosis of fibromyalgia syndrome with sympathetic skin response parameters, Sakarya University Journal of Science, 15, 1, 1–7.
50. [51] Keleşoğlu, Ö., 2008. Determination of annual heat loss and energy requirements in a reinforced concrete building using artificial neural networks. Engineering Sciences, 3, 2, 381–390.
51. [52] Charfi, O. M., Charfi, C. M. and Louhichi, N., 2018. Intelligent prediction of daily irrigation water requirements for citrus orchards using a feedforward neural network with the Levenberg-Marquardt algorithm, International Journal of Engineering and Technical Research (IJETR), ISSN: 2321-0869, 8, 12, 2454–4698.
52. [53] Doğan, E., Ateş, A., Yılmaz, E. C. and Eren, B., 2008. Application of artificial neural networks to estimate wastewater treatment plant inlet biochemical oxygen demand, Environmental Progress, 27, 4, 439–446.
53. [54] Sönmez, O. and Zengin, K., 2019. Demand forecasting in food and beverage establishments: a comparison of artificial neural networks and regression methods, European Journal of Science and Technology, 302–308.
54. [55] Doğru, F., 2015. Parameter estimation from residual gravity anomalies using modern optimisation methods, Earth Sciences, 36, 1, 31–44.
55. [56] Moriasi, D. N., Arnold, J. G., Van Liew, M. W., Bingner, R. L., Harmel, R. D., and Veith, T. L., 2007. Model evaluation guidelines for systematic quantification of accuracy in watershed simulations, Transactions of the ASABE, 50, 3, 885–900.
56. [57] Oraee, K. and Asi, B., 2006. Prediction of rock fragmentation in open-pit mines using neural network analysis, In Fifteenth International Symposium on Mine Planning and Equipment Selection (MPES 2006).
57. [58] Izzudin, M. S. ve Novandari, W., 2018. The Effect of perceived quality, brand image on customer satisfaction and brand awareness toward repurchase intention, Journal of Research in Management, 1, 3, 32-43.
58. [59] Taylor, K. E. (2001). Summarizing multiple aspects of model performance in a single diagram. Journal of geophysical research: atmospheres, 106(D7), 7183-7192.