It is widely accepted that economically viable long-distance onshore transportation of large quantities of gases and liquids in petrochemical industry and, recently, carbon dioxide (CO2) in Carbon Capture and Sequestration projects, can be achieved using pressurised steel pipelines. In cases where such pipelines are passing close to residential areas, their safety becomes of primary concern for the pipeline design, which should ensure minimal risks posed by the release of a toxic, as it would be, e.g. in the case of CO2, or flammable inventory, as in the case of hydrocarbons, in a hypothetical scenario of the pipeline accidental failure.
One of the most dangerous failure scenarios is associated with the pipeline running fractures, damaging long pipeline sections and resulting in rapid releases of large amounts of the inventory. For example, a 0.23 m diameter 250 m long pipeline operating at 80 bar pressure would typically contain ca 4 tonnes of dense-phase CO2 fluid, which in case of the pipeline full-bore rupture will be released in only ca 20 s. In order to reduce the risks of the pipeline running fractures, the pipeline material, diameter and wall thickness, as well as mitigation measures (e.g. placement of the pipeline crack arrestors and using emergency isolation valves) are carefully considered in the design relying on predictions using mathematical models of fracture propagation and arrest.
The occurrence of a longitudinal crack propagating along a pipeline is a catastrophic event, which involves both economic losses and environmental damage. Therefore, the fracture propagation estimation is an essential strategy to ensure pipeline integrity. Fracture prediction is a challenging task, since it requires knowledge of the interaction between the dynamic forces driving crack growth. Moreover, plenty of material properties should be taken into consideration. To this end, under the framework of a European CO2QUEST FP7 project, OCAS has developed a coupled numerical model of the pipe decompression and the ductile fracture, which is based on a feedback algorithm.
For modelling fluid-structure interaction, the first step involves the computation of the bulk fluid pressure and the corresponding crack tip pressure for an arbitrary small initial longitudinal crack opening along the major axis of the pipeline, formed for example, as a result of third party damage. The corresponding crack tip velocity is then calculated using the fracture model. A zero crack velocity meant no propagation and the calculations are terminated. For a positive value on the other hand, the new crack opening area is determined after an arbitrary small time increment. Based on the new crack opening area and time interval, the mass of fluid escaping and hence the new crack tip pressure is determined using the fluid flow model. This procedure is repeated for further time increments up to the point in time at which the crack velocity reached zero.
The picture above shows finite element simulation of running ductile fracture in a full-scale buried pipeline test performed by British Gas Company (BGC). This test involved rupture of an X70 grade steel pipeline, of 50 m in length, an outer diameter of ca. 1.2 m and 18 mm wall thickness. Prior to the rupture the pipeline was filled with rich natural gas (containing ca 89.55 % (v/v) of CH4) compressed to 116 bar (ac 11.6MPa) at –5 °C. To mimic the conditions of the test, the fracture propagation was initiated by a pre-existing through-wall crack. To account for the effect of backfill pressure from the soil an external pressure on the outer surface of the pipe was applied at 5 MPa.
The figure also illustrates the deformed shape of the steel pipeline as predicted by the numerical model at 0.2 s after the initiation of the crack propagation. It can be seen that, following the fracture propagation, the wall of the unzipped section of the pipe becomes corrugated, which can be attributed to the plastic deformations of the pipe during the fracture. Remarkably, the shape of the fractured pipe predicted by the model is in a qualitative agreement with the shape observed in a real burst test.
- R. H. Talemi, S. Brown, S. Martynov & H. Mahgerefteh (2016). “Hybrid fluid–structure interaction modelling of dynamic brittle fracture in steel pipelines transporting CO2 streams.” International Journal of Greenhouse Gas Control, 54: 702-715
- S. Martynov, R. H. Talemi, S. Brown, H. Mahgerefteh (2016). “Assessment of fracture propagation in pipelines transporting impure CO2 streams.”, 13th Conference on Greenhouse Gas Control Technologies (GHGT-13), Ecublens, Switzerland
“Although experimental campaigns are essential for material characterization, it is not an easy task to evaluate fracture behaviour of large-scale components such as steel pipelines experimentally. In such a situation, finite element analysis can be a competitive alternative to characterize the fracture behaviour for such large components.”