Prediction models for non-revenue water ratio

Year 2021, Volume: 10 Issue: 1, 276 - 283, 15.01.2021

Burak Kızılöz , Eyüp Şişman

https://doi.org/10.28948/ngumuh.789694

Abstract

In this study, Non-Revenue Water Rate (NRWR) predictions have been made by taking into account two-year data (2018 and 2019) of Kocaeli and using the main parameters of consumed water amount, network length, service connection length, total network length, network age and network pressure in only six districts with the highest water loss. Model predictions have been made by both Artificial Neural Network (ANN) models with two inputs one output and Kriging method. In this study, the total network length for all model combinations and the service connection length for ANN combinations with two inputs have been used as a model input for the first time. The model output performances of the above-mentioned methods have been evaluated in accordance with R2 and HKOK criteria. In conclusion, it is obvious according to the results that the model prediction performances made by Kriging method are better than the other one (ANN method). On the other hand, while evaluating the NRWR prediction model outputs established by Kriging method is easier through the prediction charts, this requires a specialized skill set for evaluating the ANN results

Keywords

Kriging, Artificial neural network, Non revenue water ratio, Water distribution network

Gelir getirmeyen su oranı tahmin modelleri

Abstract

Bu araştırmada Gelir Getirmeyen Su Oranı (GGSO) tahminleri, Kocaeli’nin en fazla su kaybı yaşanan altı ilçesinin 2018 ve2019 yıllarına ait iki yıllık verisi dikkate alınarak ve tüketilen su miktarı, şebeke uzunluğu, servis bağlantı uzunluğu, toplam şebeke uzunluğu, şebeke yaşı ve şebeke basıncı ana parametreleri kullanılarak gerçekleştirilmiştir. Model tahminleri iki girdi ve tek çıktılı Yapay Sinir Ağı (YSA) modelleri ve Kriging yöntemi ile gerçekleştirilmiştir. Modellerde toplam şebeke uzunluğu ve iki girdili YSA model kombinasyonlarında ise, servis bağlantı uzunluğu ilk kez bu araştırmada model girdisi olarak kullanılmıştır. Yöntemlerin model çıktı performansları R2 ve HKOK performans ölçütleri üzerinden değerlendirilmiştir. Sonuç olarak; Kriging yöntemi ile gerçekleştirilen modellerin tahmin performansları YSA yöntemine göre oldukça iyidir. Kriging tekniği ile oluşturulan GGSO tahmin model çıktılarının değerlendirilmesi ve yorumlanması elde edilen tahmin haritaları sayesinde daha kolay yapılabilirken, kapalı model yapısına sahip olan YSA model sonuçlarında bu durum nitelikli uzmanlık gerektirmektedir.

Keywords

Kriging, Yapay Sinir Ağı, Gelir Getirmeyen Su Oranı, Su Dağıtım Şebekesi, Kocaeli

References

• [1] R. Liemberger and A. Wyatt, Quantifying the global non-revenue water problem. Water Science and Technology: Water Supply, 19 (10), 831–837, 2019. https://doi.org/10.2166/ws.2018.129.
• [2] IBNET, The international benchmarking networks. https://database.ibnet.org/Reports/Indicators/HeatMap?itemId=27, Accessed 06 November 2020.
• [3] AWWA, Best practice in water loss control: improved concepts for 21st century water management, American Water Works Association. https://www. awwa.org/Portals/0/AWWA/ETS/Resources/WLCFlyerFinal.pdf?ver=2015-02-10-083650-287.
• [4] IWA, Assessing Non-Revenue Water and its Components: A Practical Approach. The IWA water loss task force, Water 21, 2003.
• [5] V. Kanakoudis and H. Muhammetoglu, Urban water pipe networks management towards non-revenue water reduction: Two case studies from Greece and Turkey. Clean-Soil, Air, Water 42 (7), 880-892, 2014. https://doi.org/10.1002/clen.201300138.
• [6] F. Boztaş, Ö Özdemir, F. M. Durmuşçelebi and M. Firat, Analyzing the effect of the unreported leakages in service connections of water distribution networks on non-revenue water. International Journal of Environmental Science and Technology, 16 (8), 4393–4406, 2018. https://doi.org/10.1007/s13762-018-2085-0.
• [7] M. Tabesh, A. Asadiyami Yekta, R. Burrows, An integrated model to evaluate losses in water distribution systems, Water Resources Management. 23 (3), 477–492, 2009. https://doi.org/10.1007/s112 69- 008-9284-2.
• [8] H. E. Mutikanga, S. K. Sharma, K. Vairavamoorthy, Multi-criteria decision analysis: A strategic planning tool for water loss management, Water Resources Management. 25, 3947-3969, 2011. https://doi.org/ 10.1007/s11269-011-9896-9.
• [9] H. E. Mutikanga, S. K. Sharma, K. Vairavamoorthy, Assessment of apparent losses in urban water systems. Water and Environment Journal, 25 (3), 327–335, 2011. https://doi.org/10.1111/j.1747-6593.2010 00 225.x.
• [10] C. van den Berg, Drivers of non-revenue water: A cross-national analysis. Utilities Policy, 36, 71–78, 2015. https://doi.org/ 10.1016/j.jup.2015.07.005.
• [11] S. H. Zyoud, L. G. Kaufmann, H. Shaheen, S. Samhan and D. Fuchs- Hanusch, A framework for water loss management in developing countries under fuzzy environment: Integration of Fuzzy AHP with Fuzzy TOPSIS. Expert Systems with Applications, 61, 86–105, 2016. https://doi.org/10.1016/j.eswa.2016.05.016.
• [12] M. Tabesh, A. Roozbahani, B. Roghani, N. R. Faghihi and R. Heydarzadeh, Risk Assessment of factors influencing non-revenue water using bayesian networks and fuzzy logic. Water Resources Management, 32 (11), 3647–3670, 2018. https:// doi.org/10.1007/s11269- 018-2011-8.
• [13] K. Gonelas and V. Kanakoudis, Reaching economic leakage level through pressure management. Water Science and Technology: Water Supply, (16), 3, 756–765, 2016. https://doi.org/10.2166/ws.2015.181.
• [14] G. Güngör-Demirci, J. Lee, J. Keck, R. Guzzetta and P. Yang, Determinants of non-revenue water for a water utility in California. Journal of Water Supply: Research and Technology – AQUA, vol. 67 (3), 270–278, 2018. https://doi.org/10.2166/aqua.2018.152.
• [15] D. Jang and G. Choi, Estimation of non-revenue water ratio using MRA and ANN in water distribution networks. Water, (10), 1, 1-13, 2018. https://doi.org/ 10.3390/w10010002.
• [16] D. Jang and G. Choi, Estimation of non-revenue water ratio for sustainable management using artificial neural network and Z-score in Incheon, Republic of Korea. Sustainability, (9), 11, 1933, 2017. https://doi.org/ 10.3390/su9111933.
• [17] E. Şişman and B. Kizilöz, Artificial neural network system analysis and Kriging methodology for estimation of non-revenue water ratio. Water Supply, (20), 5, 1871-1883, 2020. https://doi.org/ 10.2166/ ws.2020.095.
• [18] İSU, Kocaeli Su ve Kanalizasyon İdaresi Genel Müdürlüğü, Yıllık faaliyet raporları, 2019.
• [19] B. Kizilöz, E. Çevik ve B. Aydoğan, Estimation of scour around submarine pipelines with Artificial Neural Network. Applied Ocean Research, (51), 241–251, 2015. https://doi.org/10.1016/j.apor.2015.04.006.
• [20] C. M. Yeşilkanat, Y. Kobya, H. Taşkin and U. Çevik, Dose rate estimates and spatial interpolation maps of outdoor gamma dose rate with geostatistical methods; A case study from Artvin, Turkey. Journal of Environmental Radioactivity, (150),132–144, 2015. https://doi.org/10.1016/j.jenvrad.2015.08.011.