Pipe Life Prognosis in Water Distribution Networks using Reliable Data-based Approaches*

The assessment and prognosis of pipe life in water distribution networks has great potential in optimizing asset investment and protecting water resources. In the state-of-the-art, most of the research work about pipe life assessment focuses on revealing associated variables and regulations for the occurrence of pipe failures, which has scientific value but still far from assisting water industry directly in real operation. In order to provide a pipe life assessment and prognosis approach with practical significance, this paper presents: 1) a comparable approach to quantify impact of different factors (mainly age, material and diameter) on the occurrence of pipe failures using statistical reliability model based on cumulative Weibull distribution, survival model based on neural networks and evolutionary polynomial regression model for pipe deterioration; 2) a prognosis method for the remaining useful life of pipes using previous algorithms; 3) a maintenance and renewal plan of the network to assist daily operation of water operators by means of a checklist including risk levels (low, medium, high) under different factor ranges. The Barcelona water distribution network is used as a real life case study, demonstrating how the proposed approaches can be used.

[1]  R. Bharathi,et al.  Weibull Cumulative Distribution based real-time response and performance capacity modeling of Cyber-Physical Systems through software defined networking , 2020, Comput. Commun..

[2]  Vicenç Puig,et al.  Leak Localization in Water Distribution Networks Using Pressure and Data-Driven Classifier Approach , 2019, Water.

[3]  Ida Scheel,et al.  Time-to-Event Prediction with Neural Networks and Cox Regression , 2019, J. Mach. Learn. Res..

[4]  T. Awolusi,et al.  Performance comparison of neural network training algorithms in the modeling properties of steel fiber reinforced concrete , 2019, Heliyon.

[5]  Markus Haltmeier,et al.  Pipe failure modelling for water distribution networks using boosted decision trees , 2018 .

[6]  Uri Shaham,et al.  DeepSurv: personalized treatment recommender system using a Cox proportional hazards deep neural network , 2016, BMC Medical Research Methodology.

[7]  K. Pietrucha-Urbanik,et al.  Analysis and assessment of water distribution subsystem failure , 2016 .

[8]  Lisa Scholten,et al.  Statistical failure models for water distribution pipes - A review from a unified perspective. , 2015, Water research.

[9]  Dan D. Koo,et al.  A Comparison Study of Water Pipe Failure Prediction Models Using Weibull Distribution and Binary Logistic Regression , 2015 .

[10]  Luigi Berardi,et al.  Development of pipe deterioration models for water distribution systems using EPR , 2008 .

[11]  D. Savić,et al.  A symbolic data-driven technique based on evolutionary polynomial regression , 2006 .

[12]  Balvant Rajani,et al.  Comprehensive review of structural deterioration of water mains: statistical models , 2001 .

[13]  Okitsugu Fujiwara,et al.  Reliability improvement for water distribution networks through increasing pipe size , 1991 .

[14]  K. Lansey,et al.  Reliability‐Based Optimization Model for Water Distribution Systems , 1987 .

[15]  Symeon E. Christodoulou,et al.  A Neurofuzzy Decision Framework for the Management of Water Distribution Networks , 2010 .

[16]  Dragan Savic,et al.  Assessing pipe failure rate and mechanical reliability of water distribution networks using data-driven modeling , 2009 .

[17]  D. Ouazar,et al.  Reliability tester for water-distribution networks , 1996 .