Evaluating Fracture Behaviour of Pipeline Steel in a Sour Environment at Ambient Temperature
By M Ali and T London
Industrial Need
Fracture mechanics based engineering critical analysis (ECA) is often used in the engineering industry to evaluate fitness for service through different routes, e.g. evaluating the stability of given flaws discovered during service, or defining flaw acceptance criteria during design and fabrication. A fracture mechanics based approach requires materials properties (tensile properties and fracture toughness) to be precisely defined to carry out reliable integrity assessments. The oil and gas industry has been exploring deeper and more difficult territories to discover new oil and gas reservoirs where the likelihood of the presence of hydrogen sulphide (sour gas) is high. It is a well-established fact that a sour service environment can adversely affect fracture resistance of carbon manganese steels to varying extents depending upon the severity of the environment. In order to carry out ECA for pipelines or any other pressurised components operating under sour service, it is required that the fracture toughness data for the given material be generated under conditions reflecting as closely as possible the service environment.
It is challenging to accurately quantify fracture behaviour of materials exposed to sour environments using conventional fracture mechanics testing techniques. Currently, there is no adequate guidance available on sour fracture toughness testing methods in existing standards. This report presents the outcome of the experimental and analytical work associated with an ambient temperature test programme. It includes a brief overview of existing guidance on sour fracture toughness testing in current standards and recommended practices as well as results from testing API 5L X65 pipeline steel in a specific sour environment at ambient temperature and pressure. Subsequently, numerical modelling was used to elaborate the effect of loading (applied during an unloading compliance fracture toughness test) on crack tip hydrogen concentration in order to understand its link with experimental observations. The report also incorporates the results from assessments carried out for some pre-defined service loading for a hypothetical pipeline system to evaluate the sensitivity of flaw tolerance to the various approaches used to define fracture resistance in a sour environment.
Key findings
- Although existing standards provide some guidance on relevant fracture toughness testing, significant work is required to produce more practical and generally acceptable guidelines due to the complexity associated with environmental effects.
- For fracture resistance or R curve testing, the crack extension can be measured to an acceptable accuracy using both unloading compliance (UC) and direct current potential drop (DCPD) techniques. The UC method is easier (requiring less equipment) but could result in lower R curves due to the potential effect of cyclic loading.
- In terms of convenience and practicality, it was found that a K rate within the range 3x10‑4 to 5x10‑4 MPam0.5s-1 can provide fracture toughness reasonably close to the expected lower bound fracture toughness.
- Partial R curve tests, where the specimens were held under a fixed displacement for a specified period, showed some crack growth during the displacement hold period, possibly due to sulphide stress corrosion cracking (SSC). Based upon TWI’s experience, it is likely that the point of stability might exist between 0.2mm and the crack extension associated with the maximum force for the specified environment/material combination.
- A multi-physics model of hydrogen diffusion and ductile crack growth has been developed. The results demonstrate that the hydrogen concentration ahead of the crack tip can reach approximately twice the initial hydrogen concentration in the specimen.
- ECAs based upon single point fracture toughness values and lower bound partial R curves have shown that the latter can yield higher flaw tolerance levels than the former. The extent of the benefit accrued by the use of the R-curve depends on the applied stress and the maximum crack extension allowed in the R curve.
Comparison of finite element model prediction with test data for clip displacement
