| description abstract | Mechanical damage in transportation pipelines is a threat to their structural integrity. Failure in oil and gas pipelines is catastrophic as it leads to personal fatalities, injuries, property damage, loss of production, and environmental pollution. Therefore, this issue is of extreme importance to pipeline operators, government and regulatory agencies, and local communities. As mechanical damage can occur during the course of pipeline life due to many reasons, appropriate tools and procedures for assessment of severity is necessary. There are many parameters that affect the severity of the mechanical damage related to the pipe geometry and material properties, the defect geometry and boundary conditions, and the pipe state of strain and stress. The main objective of this paper is to investigate the effect of geometry, material, and pressure variability on strain and stress fields in dented pipelines under static and cyclic pressure loading using probabilistic analysis. Most of the published literature focuses on the strain at the maximum depth for evaluation, which is not always sufficient to evaluate the severity of a certain case. The validation and calibration of the base deterministic model was based on full-instrumented full-scale tests conducted by Pipeline Research Council International as part of their active program to fully characterize mechanical damage. A total of 100 cases randomly generated using Monte Carlo simulation are analyzed in the probabilistic model. The statistical distribution of output parameters and correlation between output and input variables is presented. Moreover, regression analysis is conducted to derive mathematical formulas of the output variables in terms of practically measured variables. The results can be used directly into strain based assessment. Moreover, they can be coupled with fracture mechanics to assess cracks for which the state of stress must be known in the location of crack tip, not necessarily found in the dent peak. Furthermore, probabilities derived from the statistical distribution can be used in risk assessment. | |
