Probabilistic failure investigation of small diameter cast iron pipelines for water distribution

Jian Ji, Jia Hong Lai, Guoyang Fu, Chunshun Zhang, Jayantha Kodikara

Research output: Contribution to journalArticleResearchpeer-review

31 Citations (Scopus)

Abstract

At present, a significant proportion of Australia's ageing water infrastructure is composed of cast-iron pipes. The corrosion of cast-iron pipes is the main triggering factor for Australia's water industry. It has been found that the failure mode of small diameter water mains (<300 mm external diameters) from Yarra Valley Water is mainly the longitudinal failure (hoop-stressed) instead of broken back failure, which is unexpected. Therefore, the purpose of this paper is to examine the failure mechanism of longitudinal failure of small diameter cast-iron pipes. The observational failure database provided by Yarra Valley Water show limited data availability. This has stimulated the focus of this paper: to examine the corrosion mechanism by utilising probabilistic inverse analysis for unknown physical parameters. Focusing on the likelihood of failure, this paper investigates the performance of small diameter cast-iron pipes by using statistical analysis based on the observational failure lifetime data within pipe cohorts. Background studies on the deterioration of cast-iron pipes due to corrosion are made to help set up the probabilistic physical modelling (PPM). Using the updated corrosion parameters, the lifetime probabilities of failure and the hazard rate of the cast iron model are derived from PPM. Last, it is found that the modelling results agree reasonably well with prediction curves of statistical data, indicating that the proposed method for pipe lifetime prediction is promising.

Original languageEnglish
Article number104239
Number of pages14
JournalEngineering Failure Analysis
Volume108
DOIs
Publication statusPublished - Jan 2020

Keywords

  • Cast-iron pipes
  • Corrosion
  • Failure prediction
  • Probabilistic inverse analysis (PPM)
  • Probabilistic physical modelling

Cite this