Integrity Assessment of Dents in Pipelines using Finite Element Analysis and Artificial Neural Networks

• Author / Creator

  Woo, Janine B.

• Dents are common occurrences along oil and gas pipelines and can be formed due to pipe contact with external forces such as rocks or construction equipment. There are many factors that contribute to the level of integrity concern of a dent including its shape, size, location on the pipe, and proximity to other features, which has made it difficult for industry agreement on a single, consistent, accurate, and efficient dent assessment method. Current regulations for dents, which are based on depth and interaction with stress risers, have been proven to have limitations by other researchers and have been shown to be too conservative in some cases while failing to identify concern in others. A method is proposed to assess dents herein, which uses the results from finite element analysis (FEA) to train artificial neural networks (ANNs) and the trained ANNs can then be used to assess real-life dent features reported by in-line inspection (ILI) tools.

  First, a methodology to use FEA to model dents in pipelines was developed and validated against full-scale tests published by another researcher. The effect of using different element properties was investigated and it was concluded that the use of S4R elements (shell, linear elements with reduced integration) and five integration points in the thickness direction was most suitable for this problem. Automation techniques were developed to make the FEA model generation and extraction of results more efficient and ensure the methodology was performed consistently.

  Precision of the dent profiles achieved by FEA compared to the profiles reported by ILI tools was found to be important to attain accurate stress and strain results within a dented region. In order to achieve an accurate profile match, a trial and error method was proposed but was found to be time-consuming. It was determined that a ring torus is an appropriate shape to use as an indenter in the FEA to model smooth, symmetric dent profiles and that two-dimensional profiles aligned with the longitudinal and circumferential axes through the most significant, or deepest, point of the dent, can be used as a simplified representation of the full three-dimensional shape of the dent.

  Lastly, ANNs were used to produce results as accurate as could be found using FEA but in significantly less time. Given only the dent profile information and basic pipe properties, the ANNs were able to accurately output stress and strain results. This thesis demonstrates the feasibility of using FEA in conjunction with ANNs to provide an accurate and efficient method for assessing dents.

• Subjects / Keywords

  • Abaqus
  • Artificial Neural Networks
  • Pipelines
  • Pipeline Integrity
  • Finite Element Analysis
  • Dents
  • In-Line Inspection

• Graduation date

  Fall 2019

• Type of Item

  Thesis

• Degree

  Master of Science

• DOI

  https://doi.org/10.7939/r3-05w4-0h45

• License

  Permission is hereby granted to the University of Alberta Libraries to reproduce single copies of this thesis and to lend or sell such copies for private, scholarly or scientific research purposes only. Where the thesis is converted to, or otherwise made available in digital form, the University of Alberta will advise potential users of the thesis of these terms. The author reserves all other publication and other rights in association with the copyright in the thesis and, except as herein before provided, neither the thesis nor any substantial portion thereof may be printed or otherwise reproduced in any material form whatsoever without the author's prior written permission.