Managing the integrity of a pipeline is a challenging task, particularly when your pipeline is susceptible to cracking. For Stress Corrosion Cracking (SCC), the materials, stress state and environment all synergize to contribute to the development and growth of cracks. With fatigue cracking, factors such as manufacturing flaws, stress cycles and environment can all also have a major impact. Managing that threat and demonstrating fitness for service represent significant challenges.
In these situations, you can never have too much information. Anything that contributes to understanding the cause and severity of cracking helps to define appropriate methods for safe and efficient threat management.
In-line inspection, of course, is a great source of information. The advent of Ultrasonic Testing (UT) and Electromagnetic Acoustic Transducer (EMAT) crack-detection tools has revolutionized the management of cracking threats in recent years by enabling operators to identify and repair significant cracks. The contribution of other inspection technologies, such as material property inspections, geometry inspections and metal loss inspections, can also provide substantial additional valuable information. However, all inspection systems have limitations; they cannot provide direct information on the history of operation (pressures, temperatures, products, etc.), they cannot give information on manufacturing and construction history and they cannot detect the local environment. Plus, it goes without saying, there are limitations on what can be detected and on accuracy.
Therefore, collecting additional data is a vital part of a comprehensive integrity management program. Cutting out samples of pipe, particularly with examples of defects present, can provide an enormous amount of valuable additional information of the kind that is extremely hard to generate from internal or in-the-ditch inspections or historical records.
Let us consider the example of an operator that experienced a failure due to a pipe body axially orientated crack. The cracking mechanism at the failure location was confirmed to be SCC. Managing the threat safely, efficiently and effectively became their top priority.
Following two extensive inspection campaigns using multiple internal inspection technologies, including Electromagnetic Acoustic Transducer (EMAT), Magnetic Flux Leakage (MFL) and geometry mapping tools, many features were reported, verified in-ditch and remediated.
Figure 1: Crack-like feature found in-ditch on a 16" natural gas pipeline
For one site along a 16-inch section, an axially oriented SCC colony was reported and confirmed in-field by Phased Array (PAUT) to be more than 50% wt deep and greater than 4 in (100 mm) in length.
So, while the operator already had a robust cracking threat management process in place, they were aware of the need to gain a deeper understanding of the contributing factors and demonstrate to stakeholders that the approach they were taking was conservative. Therefore, rather than install a structural repair, the operator decided to remove the pipe joint for further analysis, shipping it to ROSEN’s testing facility in Newcastle upon Tyne with the following objectives:
- Confirm feature and pipe dimensions
- Validate the material properties
- Burst the feature and confirm failure pressure
- Evaluate crack morphology
- Validate the predictive assessment model
A condition assessment, material testing program and burst test followed by an in-depth post-test analysis in a laboratory environment was proposed to provide confidence in in the physical and causal attributes of the feature.
The deepest crack-like feature was confirmed by magnetic particle inspection (MPI) and noted as a colony 4.7 in (120 mm) in length by 1.5 in (40 mm) in width.
Figure 2: Post-MPI crack measurements
The peak depth was measured at 0.2 in (5.4 mm) using PAUT and found to be coincident with deformation bands present due to cold field bending.
A 3D laser scan was performed to generate a representative model of the actual spool geometry for finite element analysis simulations. This was for the purpose of having a full set of geometric data for the pipe, recognizing that very few pipes are actually straight and round. In addition, there was the possibility of identifying minor local stress variations or any other geometric features that may have contributed to the formation of SCC in particular locations.
Figure 4: 3D laser scan
A series of mechanical tests was performed to validate the grade and quality of the material in the parent material, long seam and girth weld. The material exhibited ductile behavior during Charpy and fracture toughness testing with adequate tensile properties, and was found to meet the grade requirements of API 5L. Hardness measurements were taken across the long seam and confirmed to meet the requirements of API 1104.
Figure 5: Material testing
A vessel was fabricated from the removed pipe spool section. A number of strain gauges were installed on the external surface of the spool in locations around the field bends.
Figure 6: Strain gauge installation in close proximity to SCC colony
A thermocouple was installed on the surface of the spool to monitor temperature throughout the test, and pressure transducers were installed at the pressure inlet to the spool and at the pump to monitor real-time pressure. The vessel was situated in the burst chamber and pressure introduced in stages, each with a target pressure and 5-minute hold points to observe stabilization.
Figure 7: Burst test
The spool burst at approximately 290 psi (20 bar) higher than the pipeline MAOP. The failure location was consistent with the position of the peak crack depth reported from PAUT.
By performing post-test metallography and fractography in the laboratory, critical information was obtained regarding crack morphology, crack size and profile, initiation point, and whether any additional features were present.
Figure 8: Sectioned colony for profile measurements
During this post-test assessment, two axial indications were selected for further examination. The first indication revealed a deep pre-existing through-wall elliptical crack that was more than 99% deep. Thick corrosion product was evident on the fracture surface; it was measured to be approximately 1.4 in (35 mm) in length, with the near through-wall section being just 0.3 in (8 mm) in length.
Figure 9: Section of deepest crack
The second indication exhibited a longer pre-existing almost through-wall elliptical crack that also exhibited thick corrosion product and was measured to give a 0.04 in (1 mm) remaining ligament, with the length of the deepest section being just 0.4 in (10 mm).
A detailed crack profile was observed and measured along the length of both leak locations and compared to the measurements observed by PAUT.
Figure 10: Composite image showing crack profile measurements
Focused analysis of the fracture faces using Scanning Electron Microscopy (SEM) and optical microscopy provided a positive identification: the morphology was principally trans-granular crack propagation, typical of near-neutral pH SCC.
Figure 11: High magnification SEM image showing secondary trans-granular crack propagation
Figure 12: Optical microscope image section showing principally trans-granular crack propagation
This provided ROSEN’s integrity engineers with all of the key data to consider at the assessment stage.
FINITE ELEMENT ANALYSIS
A relationship was observed between the geometry data and the SCC, showing a peak hoop stress at the surface in the locations where circumferential deformation bands were evident. In this instance, the field bending process has likely contributed to the formation of local minor stress variations, promoting the development of cracks in this pipeline with a possibility of low-level cracking in other locations with similar patterns.
The through-wall stress distribution at the crack location indicated a bending stress, with a decrease in hoop stress from the external wall to the internal wall, noting that the small remaining ligament was of course at the internal wall and, due to the short-length bulging, was minimal. This may have contributed to the ability to retain pressure above MAOP, as demonstrated during the burst test.
PREDICTED BURST PRESSURE
With all of the data gathered from the test and post-test analysis, the intent was to predict the burst pressure of the spool based on the size of the feature to determine whether the most applicable assessment methodology would provide a result close to the actual burst pressure.
Using actual tensile, Charpy and fracture toughness data from the material testing tasks, the integrity engineers were able to conclude that the material was fully ductile at the test temperature. Therefore, a plastic collapse failure model could be expected to give a reliable result.
The post-test assessment conclusion on the physical attributes of the failure at the crack locations suggested that a flow-stress-dependent methodology is likely to be appropriate for future assessment, considering the calculation of the predicted burst pressure vs. the actual burst pressure was a variation of < 2%.
A follow-up test on the remaining material containing a shallower feature was recommended and is being considered.
This work showcases how comprehensive testing and subsequent detailed analysis together can really help us to understand more completely the many complex contributors to cracking threats while adding significant value to pipeline operators globally. This particular example provided extremely valuable data to support:
- An understanding of crack morphology and critical dimensions
- Validation of material properties
- Refinement and selection of the most appropriate assessment model for the pipeline