Published July 22, 2026 | Version v1
Journal article Open

Unsupervised Machine Learning for Risk-Based Integrity Assessment of Underground Steel Pipelines: Bulk Water Distribution Utilities

Authors/Creators

  • 1. School of Mechanical, Industrial and Aeronautical Engineering University of Witwatersrand Johannesburg, South Africa.

Description

ABSTRACT:

Aging big-diameter underground steel pipelines pose significant sustainability and operational challenges for bulk water distribution utilities. These challenges include, but are not limited to, an increased risk of underground big-diameter pipeline failure and rising costs for asset condition assessments. We developed and tested a unsupervised machine-learning framework to improve pipeline condition assessment, predictive maintenance, and inspection prioritisation using real-world secondary data. We combined mixed-data clustering, non-linear dimensionality reduction, anomaly detection, and association rule mining to identify complex patterns in the condition of underground steel pipelines without excavation. Our results indicate that mixed-type clustering methods produce stable, well-separated condition groups and outperform numeric-only methods. Non-linear embeddings show clear separability, and anomaly detection reliably pinpoints high-risk pipeline segments. Association rules reveal hidden connections between pipeline attributes, enhancing clarity and engineering relevance. This framework enables data-driven decision-making, reduces unplanned maintenance, and supports efficient resource use by extending the lifespans of large-diameter underground pipeline assets and boosting operational reliability. This study supports sustainable asset management and improved operations in bulk water distribution pipeline systems through practical, scalable unsupervised analytics.

Keywords: Unsupervised Machine Learning, Big-diameter steel pipeline, Water Utilities.

Files

MSIJAT512026 GS.pdf

Files (1.2 MB)

Name Size Download all
md5:eaf045c29b541545e5e5e71188a5d358
1.2 MB Preview Download

Additional details

Software

Repository URL
https://msipublishers.com/wp-content/uploads/2026/07/MSIJAT512026-GS.pdf

References

  • Agbenowosi, N., Loganathan, G. V., Deb, A. K., Grablutz, F., Hasit, Y., & Snyder, J., Methods of analysis for pipeline replacement, Proceedings of the American Society of Civil Engineers, pp. 1–10, 2003.
  • Amaya-Gómez, R., Vacca-Sánchez, F., & Cardona-Montoya, J., Machine learning techniques applied to pipeline corrosion assessment: A review, Journal of Loss Prevention in the Process Industries, vol. 72, pp. 1–15, 2021.
  • Atambo, D. O., Brdjanovic, M., & Wols, J. B., Development and comparison of prediction models for sanitary sewer pipes condition assessment using small datasets, Sustainability, vol. 14, no. 10, pp. 1–18, 2022.
  • Benslimane, A., Smilkovic, M., & Sonnemann, G., Risk-based asset management for pipeline integrity using data-driven approaches, Journal of Pipeline Engineering, vol. 21, no. 2, pp. 85–97, 2022.
  • Cabral, M., Gray, D., Brentan, B. M., & Covas, D., Assessing pipe condition in water distribution networks, Water, vol. 16, no. 3, pp. 1–22, 2024.
  • Chen, R., & Wang, T., Interpretable anomaly detection for water distribution networks using unsupervised learning, Water Research, vol. 198, pp. 1–12, 2024.
  • Chen, S., Li, X., & Zhang, Y., Unsupervised clustering for anomaly detection in pipeline systems, Engineering Applications of Artificial Intelligence, vol. 117, pp. 1– 11, 2023.
  • Concetti, G., Giunta, G., & Righetti, M., Visual analytics for condition assessment of buried pipelines using unsupervised learning, Automation in Construction, vol. 148, pp. 1–14, 2023.
  • Dia, A. K., Diallo, M., & Sene, S., Unsupervised neural network for data-driven corrosion detection of a mining pipeline, Proceedings of the FLAIRS Conference, pp. 234–239, 2022.
  • Dolire, P., Singh, R., & Kumar, A., Predictive maintenance of water pipelines using unsupervised machine learning, Journal of Infrastructure Systems, vol. 31, no. 1, pp. 1–13, 2025.
  • Doorsamy, W., & Bokoro, P. N., Explainability challenges in unsupervised learning for infrastructure condition monitoring, IEEE Access, vol. 12, pp. 45612–45628, 2024.
  • Fingerhut, N., Wörnle, C., & Reinhart, G., Industrial validation of online anomaly detection for large-scale infrastructure systems, Computers in Industry, vol. 151, pp. 1–12, 2024.
  • Giunta, G., Righetti, M., & Concetti, G., Early-warning systems for pipeline failures based on unsupervised learning, Journal of Hydroinformatics, vol. 22, no. 4, pp. 897– 912, 2020.
  • Gu, H., Li, Z., & Wang, Y., DBSCAN-based data fusion for defect matching in pipeline inspection data, Measurement, vol. 221, pp. 1–12, 2024.
  • Hao, Y., Intelligent early warning method for pipeline failure guided by pressure time series clustering, Proceedings of the IEEE International Conference on Industrial Engineering and Engineering Management, pp. 567–572, 2025.
  • Heng-yu, L., Zhang, X., & Chen, B., Laboratory-based validation of pipeline defect detection using machine learning, Journal of Pipeline Systems Engineering and Practice, vol. 15, no. 2, pp. 1–10, 2024.
  • Kleiner, Y., Adams, B. J., & Rogers, J. S., Water distribution network renewal planning, Journal of Computing in Civil Engineering, vol. 15, no. 1, pp. 15–26, 2001.
  • Liu, Y., & Tong, L., Noise-robust clustering for underground pipeline condition assessment, Structural Health Monitoring, vol. 24, no. 1, pp. 1–16, 2025.
  • Liu, Z., Wang, H., & Zhao, Y., Dimensionality reduction methods for pipeline inspection data interpretation, Advanced Engineering Informatics, vol. 59, pp. 1–14, 2024.
  • Manchanda, R., Data-driven predictive maintenance strategies for pipeline assets, Journal of Infrastructure Preservation, vol. 6, no. 2, pp. 45–58, 2022.
  • Manchanda, R., & Pervez, M., Label scarcity challenges in pipeline failure prediction models, Engineering Failure Analysis, vol. 137, pp. 1–10, 2022.
  • May, Z. B., Mohd Isa, M. H., & Ahmad, R., Machine-learning-based classification for pipeline corrosion with Monte Carlo probabilistic analysis, Energies, vol. 16, no. 8, pp. 3456–3478, 2023.
  • Mohammadagha, M., Khosravi, A., & Nahavandi, S., Hybrid machine learning metamodel for condition assessment of urban underground pipes, Journal of Infrastructure Systems, vol. 31, no. 2, pp. 1–15, 2025.
  • Ouadah, L., Risk-based maintenance planning for pipeline systems, International Journal of Pressure Vessels and Piping, vol. 164, pp. 12–21, 2018.
  • Park, S., Kim, S., & Kim, J., Unsupervised learning-based plant pipeline leak detection system using autoencoder and transfer learning, IEEE Access, vol. 12, pp. 45678–45692, 2024.
  • Ren, Q., Zhou, L., & Li, M., Environmental noise effects on pipeline anomaly detection models, Measurement, vol. 181, pp. 1–11, 2025.
  • Smith, J., & Johnson, L., Clustering-based inspection prioritisation for oil and gas transmission pipelines, Journal of Pipeline Engineering, vol. 21, no. 4, pp. 189–204, 2023.
  • Su, H., Zhang, Y., & Liu, J., Comparison of shallow clustering and deep learning methods for infrastructure monitoring, Neurocomputing, vol. 350, pp. 98–110, 2019.
  • Yeo, I., Lee, S., & Kim, D., Feature extraction using unsupervised learning for infrastructure anomaly detection, Sensors, vol. 19, no. 18, pp. 1–17, 2019.