Procedural Framework for Modeling the Likelihood of Failure of Underground Pipeline Assets Academic Article uri icon

abstract

  • 2015 American Society of Civil Engineers. Reliable prediction of asset condition and its likelihood of failure is one of the core requirements for a utility to establish effective asset management strategies for optimized maintenance, rehabilitation, and replacement plans. Although there have been many research efforts in academia to predict the failure of pipe assets, many utilities across the United States still find it challenging to effectively predict the likelihood of failure (LOF) of their pipeline assets. Most of them still use subjective scales and rely on engineers' anecdotal experience and judgments. This study developed a holistic procedural framework that utilities can follow to develop a data driven LOF prediction model of their pipeline assets. The unique contribution of this paper is that the framework addresses issues that a utility will encounter from data collection and data organization to LOF prediction model development, and discusses possible solutions as well. Historical performance records of sewer pipes from a major city were used to demonstrate and validate the framework. The procedural framework developed in this study is anticipated to facilitate and accelerate the practical use of advanced data-driven methods for underground pipeline asset management, which will result in more reliable and high-quality investment decisions. As a derivative of the case study, the study also found that different lengths of sewer pipes actually do change the expected life of a sewer pipe, which indicates that most of the previous deterioration models for sewer pipes without consideration of pipe length may be seriously flawed.

published proceedings

  • JOURNAL OF PIPELINE SYSTEMS ENGINEERING AND PRACTICE

author list (cited authors)

  • Park, H., Ting, S. H., & Jeong, H. D.

citation count

  • 8

complete list of authors

  • Park, Heedae||Ting, See Hyiik||Jeong, H David

publication date

  • November 2015