Risk and reliability analysis for oil and gas pipelines using data-driven methods

• Author / Creator

  Chen, Yinuo

• Oil and gas pipelines (OGP) play a crucial role in sustaining the economy. With the increasing speed of pipeline network building, there is a corresponding growth in energy supply and demand. Nevertheless, numerous pipeline network safety operating issues arise during oil and gas transportation, such as corrosion failure leading to leakage, often resulting in fatalities, significant environmental damage, and economic losses. Therefore, it is necessary to conduct risk and reliability analyses to identify accident precursors and prevent accidents before they happen. It can also predict the pipeline degradation process, avoid major failures, determine the priority of risk mitigation, and optimize resource allocation. Various categories of OGP systems generate a large amount of data during operation. However, current studies have encountered challenges using these data to model risk and reliability effectively. Existing studies have constructed generalized models that ignore specific characteristics of different OGP systems. Furthermore, some of these studies failed to use appropriate data sources and inadequately addressed the uncertainty associated with input data, resulting in inaccurate results. In addition, the current models exhibit computational inefficiency and operational complexity.
  Therefore, this thesis aims to utilize data sources from different OGP systems to develop more efficient and accurate data-driven models for risk and reliability assessment. This thesis fully considers the characteristics of different OGP systems and creates different risk and reliability analysis models in a targeted manner. The proposed models are more comprehensive. At the same time, the structure is simplified and the operation is more convenient, which can significantly improve the computational efficiency of the models while obtaining more accurate analysis results.
  For pipelines where in-line inspection (ILI) cannot be conducted, a novel method of cloud-variable weight function is proposed to analyze the pipeline’s risk level and critical risk factors by establishing a pipeline risk assessment index system. The proposed method fully considers the uncertainty in the evaluation process, resolves the contradiction of existing methods to model the fuzzy concepts accurately, optimizes the weight distribution, and obtains a more scientific and reasonable assessment result. For gas transmission systems (GTS), a structure mapping method based on failure modes and effects analysis (FMEA) is proposed to form the GTS’s object-oriented Bayesian network (OOBN) framework, making the model more user-friendly. An accident precursor identification approach is proposed based on the piecewise aggregate approximation-cumulative sum (PAA-CUSUM) algorithm, which can better discover the potential risks in real-time. The proposed method identifies process anomalies through monitoring data and analyzes the events and propagation patterns with the highest potential risk. For pipelines where ILI can be conducted, a finite element (FE) model is established. A reliability prediction method based on residual neural networks (ResNet) that can directly map the magnetic flux leakage (MFL) inspection data to the pipeline’s reliability is proposed. Pipeline defect-effective area models, rather than those based on just depth, are effectively integrated with deep learning models. Moreover, an innovative approach for reconstructing the defect profile using a novel hybrid neural network to accurately and efficiently map three-axial MFL signals to the defects’ 3-D profile is also proposed. It utilizes the neural ordinary differential equation (ODE) as a module within the neural network architecture, which can map the MFL signals to the spatial position of each point on the defective concave surface. Additionally, the proposed model incorporates the Fourier integration kernel to enhance computational efficiency.
  The contributions of this study lay the foundation for OGP potential risk discovery, risk control and rehabilitation, and pipeline digital integrity management. The proposed research can be extended to investigate the risk and reliability problems for OGP systems, considering more complex events and situations in future work.

• Subjects / Keywords

  • Risk
  • Reliability
  • Pipeline
  • Data-driven

• Graduation date

  Spring 2024

• Type of Item

  Thesis

• Degree

  Doctor of Philosophy

• DOI

  https://doi.org/10.7939/r3-yny8-3712

• License

  This thesis is made available by the University of Alberta Libraries with permission of the copyright owner solely for non-commercial purposes. This thesis, or any portion thereof, may not otherwise be copied or reproduced without the written consent of the copyright owner, except to the extent permitted by Canadian copyright law.