Yapay Zeka İle Atıksu Toplama Sistemlerinde Katı Madde Birikiminin ve Tıkanıklığın Tespit Edilmesi

Abstract

Atıksu toplama sistemlerinde katı madde birikimi, tıkanmalara, akış kapasitesinin azalmasına ve Ayrık veya Birleşik Kanal Taşmaları (SSO veya CSO) riskinin artmasına yol açarak sel ve patojen bulaşması gibi halk sağlığını tehdit eden ciddi sorunlara neden olmaktadır. Bu tür birikimlerin erken tespiti kritik öneme sahip olsa da mevcut otomatik yöntemler, yalnızca derinlik verilerine dayanan ve hız bilgisi eksikliği nedeniyle tutarsız sonuçlar veren akademik çalışmalarla sınırlıdır. Bu çalışmada, Yağmur Suyu Yönetim Modeli (SWMM) ile oluşturulan 1.482.000 simülasyon sonucunda derinlik-hız dağılım grafiklerinden 82.482 etiketli silt birikim seviyesi grafiği kullanılarak bir Derin Evrişimli Sinir Ağı (CNN) eğitilmiştir. CNN, test verilerinde (katı madde miktarı tahminlerinde %10’a kadar olan sapmalarla) %99 doğruluk oranına ulaşmış ve 20 cm çapındaki borularla yapılan üç gerçek deney ile doğrulanmıştır. Bu deneylerde, tıkanmalar gerçek koşullara sırasıyla 0 cm, 0,3 cm ve 1,5 cm yakınlıkta tahmin edilmiştir. Sonuçlar, CNN’in silt birikimini tespit etmede etkili bir araç olduğunu, geleneksel yöntemlere göre daha hızlı ve hassas bir çözüm sunduğunu, bakım süreçlerini iyileştirme ve sistem arıza risklerini azaltma potansiyeline sahip olduğunu göstermektedir. Ancak, modelin daha geniş kapsamlı gerçek hayattaki uygulamalar için optimize edilmesine ve ek araştırmalara ihtiyaç duyulmaktadır.

Keywords

• Atıksu ve Yağmur suyu Toplama Sistemleri
• Katı Madde Birikimi
• Yapay Zeka
• Evrişimli Sinir Ağı

LEVERAGING AI TO IDENTIFY SILT ACCUMULATION AND BLOCKAGE IN COLLECTION SYSTEMS

Abstract

Solid waste accumulation in wastewater collection systems poses significant challenges, leading to blockages, reduced flow capacity, and an increased risk of Separate or Combined Sewer Overflows (SSO or CSO), threatening public safety through flooding and pathogen contamination. Proactive detection of such buildups is critical, yet current automated methods are limited to academic studies that rely solely on depth data, lacking velocity information essential for determining silt depth accurately. This study addresses this gap by training a Deep Convolutional Neural Network (CNN) on depth-velocity scatter plots generated from 1,482,000 simulations using the Storm Water Management Model (SWMM), producing 82,482 labeled plots of varying silt accumulation levels. The CNN achieved 99% accuracy (for solid content predictions with deviations of up to 10%) on test data and was further validated with three real-world experiments using 20 cm pipes, where blockages were predicted within 0 cm, 0.3 cm, and 1.5 cm of actual conditions. These findings demonstrate CNNs as an effective tool for identifying silt accumulation, offering a faster and more precise alternative to traditional methods, with the potential to enhance maintenance and reduce system failure risks. Further research is needed to optimize the model for broader real-world applications.

Keywords

• Sanitary Sewer and Stormwater Collection Systems
• Silt Accumulation
• Artificial Intelligent
• Convolutional Neural Network

References

1. Berggren, M., Pettersson, T., and Eriksson, E. (2012). Modeling urban flooding under extreme conditions, Water Resources Management, 26(2), 431-442. doi:10.1007/s11269-012-0158-3.
2. Ebtehaj, I., Azimi, H., and Bonakdari, H. (2015). Numerical analysis of sediment transport in sewer pipe, International Journal of Engineering, 28(11), 1564-1570.
3. El-Zaemey, A. K. S. (1991). Sediment transport over deposited beds in sewers (Doctoral dissertation, Newcastle University).
4. Enfinger, K. L., and Kimbrough, H. R. (2004). Scattergraph principles and practice: A comparison of various applications of the Manning equation. Pipeline Engineering and Construction: What’s on the Horizon?, 1-13.
5. Faris, N., Zayed, T., Aghdam, E., Fares, A., and Alshami, A. (2024). Real-Time sanitary sewer blockage detection system using IoT, Measurement, 226, 114146.
6. Ghani, A. A. (1993). Sediment transport in sewers (Doctoral dissertation, Newcastle University).
7. Goodfellow, I., Bengio, Y., and Courville, A. (2016). Deep learning. MIT Press.
8. Guo, W., Soibelman, L., and Garrett Jr, J. H. (2009). Automated defect detection for sewer pipeline inspection and condition assessment, Automation in Construction, 18(5), 587-596.

9. Khan, A., Sohail, A., Zahoora, U., and Qureshi, A. S. (2020). A survey of the recent architectures of deep convolutional neural networks. Artificial intelligence review, 53, 5455-5516.
10. Krizhevsky, A., Sutskever, I., and Hinton, G. E. (2012). Imagenet classification with deep convolutional neural networks, Advances in Neural Information Processing Systems, 25.
11. LeCun, Y., Bengio, Y., and Hinton, G. (2015). Deep learning, Nature, 521(7553), 436-444.
12. Li, N., Wang, X., Li, Z., Zhao, F., Nair, A., Zhang, J., and Liu, C. (2023). Real-time identification and positioning of sewer blockage based on liquid level analysis in rural areas, Processes, 11(1), 161.
13. Li, T., Tan, Q., and Zhu, S. (2010). Characteristics of combined sewer overflows in Shanghai and selection of drainage systems, Water and Environment Journal, 24(1), 74-82.
14. Manning, R. (1891). On the flow of water in open channels and pipes, Transactions of the Institution of Civil Engineers of Ireland.
15. Montes, C., Vanegas, S., Kapelan, Z., Berardi, L., and Saldarriaga, J. (2020). Non-deposition self-cleansing models for large sewer pipes, Water Science and Technology, 81(3), 606-621.
16. Muttil, N., Nasrin, T., and Sharma, A. K. (2023). Impacts of extreme rainfalls on sewer overflows and WSUD-based mitigation strategies: A review, Water, 15(3), 429.
17. Perrusquia, G. S. (1992). An experimental study on the transport of sediment in sewer pipes with a permanent deposit, Water Science and Technology, 25(8), 115-122.
18. Raschka, S., and Mirjalili, V. (2019). Python machine learning: Machine learning and deep learning with Python, scikit-learn, and TensorFlow 2. Packt Publishing Ltd.
19. Rawat, W., and Wang, Z. (2017). Deep convolutional neural networks for image classification: A comprehensive review. Neural computation, 29(9), 2352-2449.
20. Regulation on Wastewater Collection and Disposal Systems. (2017) https://webdosya.csb.gov.tr/db/altyapi/icerikler/6-ocak-2017-cuma-20180215122614.pdf
21. Semadeni-Davies, A., Hernebring, C., Svensson, G., and Gustafsson, L. G. (2008). The impacts of climate change and urbanisation on drainage in Helsingborg, Sweden: Suburban stormwater, Journal of Hydrology, 350(1-2), 114-125.
22. Stevens, P. L., and Sands, H. M. (1995). Sanitary sewer overflows leave telltale signs in depth-velocity scattergraphs, Seminar Publication–National Conference on Sanitary Sewer Overflows.
23. Tan, Y., Cai, R., Li, J., Chen, P., and Wang, M. (2021). Automatic detection of sewer defects based on improved You Only Look Once algorithm, Automation in Construction, 131, 103912.
24. Türk, T., Kuşkonmaz, H., and Karadağlı, F. (2022). Atıksu toplama sistemlerindeki işletme problemlerinin tanımlanması, Uludağ Üniversitesi Mühendislik Fakültesi Dergisi, 27(1), 205-218.
25. Vitorino, D., Coelho, S. T., Santos, P., Sheets, S., Jurkovac, B., and Amado, C. (2014). A random forest algorithm applied to condition-based wastewater deterioration modeling and forecasting, Procedia Engineering, 89, 401-410.
26. Vongvisessomjai, N., Tingsanchali, T., and Babel, M. S. (2010). Non-deposition design criteria for sewers with part-full flow, Urban Water Journal, 7(1), 61-77.
27. Xie, Q., Li, D., Xu, J., Yu, Z., and Wang, J. (2019). Automatic detection and classification of sewer defects via hierarchical deep learning, IEEE Transactions on Automation Science and Engineering, 16(4), 1836-1847.