In our new series “Managing Pipeline Threats – The Way Forward,” we explore the approach necessary to achieve a holistic maintenance mindset when it comes to the integrity management of oil and gas assets.
After focusing on diagnostics and having discussed threats and the data needed to assess condition in the first installment, our experts Michael Beller and Roland Palmer Jones will now consider how this data is turned into useful information on pipeline condition today and what advances we can expect to see in the future. It is this process of taking what we know (ideally everything for all locations at any time) and turning it into useful information to help us make decisions and move us from watching pipelines fail to actively managing threats.
Pipelines are a key element in global energy transportation and constitute the safest, most environmentally friendly way to transport large quantities of oil and gas. They are found in all segments of the oil and gas infrastructure in the up-, mid- and downstream sectors and will retain their important role in the future.
High-pressure pipelines are also important in other industries, including water, mining and chemicals. Within the energy industry, their use may be extended – for instance to the future transportation of hydrogen.
Apart from being extremely useful from a logistics perspective, pipelines are highly valuable assets, in a strategic as well as an economic sense, since they are needed to keep homes warm (or cool), industry running and transport moving. Therefore, they need to be protected. Their safe and economical operation must be ensured at all times.
This is the focus of an effective maintenance process, which includes the need to identify potential threats, assess their impact on mechanical integrity and derive useful information regarding actions to be taken.
In this three-part series that started in the last issue, we are exploring the approach necessary to achieve a holistic maintenance mindset.
In Part 1, we will address diagnostics, the discipline of identifying and collecting the data needed to define a given status.
In Part 2, we will see how this data can be used to provide information on the present and likely future mechanical integrity of a pipeline.
In Part 3, we will investigate how this condition information can assist in the development of a “predictive maintenance mindset,” address some of the new requirements raised by the Mega Rule for gas pipelines in the US and review some of those requirements from a global perspective.
In this series of articles, we will focus on the actual line pipe and not consider other important parts of the pipeline infrastructure, such as pumps, compressors, valves, etc.
Having discussed threats and the data needed to assess condition in Part 1, we will now consider how this data is turned into useful information on pipeline condition and what advances we can expect to see in the future.
It is this process of taking what we know (ideally everything for all locations at any time) and turning it into useful information to help us make decisions that move us from watching pipelines fail to actively managing threats.
PART 2: DATA TO INFORMATION
Good decision-making requires robust, reliable and accurate information – useful information derived from data. In the previous article, we discussed the collection of data required for pipeline integrity assessment purposes and used the analogy of the three-legged stool, as shown in Figure 2 below.
Figure 2 – The integrity “stool” using a data collection perspective
We can already collect data on the presence and size of defects such as dents, cracks and corrosion metal loss in pipelines using a variety of in-line inspection technologies.
DEFECT ASSESSMENT METHODS
Traditionally, these data have been used in semi-empirical formulae to evaluate the significance or acceptability of the defects. An example is the widely used Modified B31.G method for assessing the significance of a corrosion metal loss defect, shown here calculating the hoop stress in the pipe to cause failure:
d = depth of corrosion defect
t = thickness of pipe
l = length of corrosion defect
D = pipe diameter
SMYS = specified minimum yield strength
Graphical representations of these equations can be used to aid understanding. An example of a Modified B31.G feature sentencing plot is shown in Figure 3.
Figure 3 – Feature sentencing plot (MB31.G)
For the example shown in Figure 3, we can see that the feature is acceptable, but that it would only need to grow in depth by another 15 percent through the pipe wall to become unacceptable, whereas it could more than double in length and still be acceptable.
Other formulae and associated plots help us to assess different types of defects. For example, cracks are typically plotted on a Failure Assessment Diagram (FAD). An example of a FAD is shown in Figure 4.
Figure 4 – Example Failure Assessment Diagram as typically used in the sentencing of crack-like features
These semi-empirical formulae include inherent conservatisms and assumptions. In the case of Modified B31.G some examples are:
- The formula was adjusted using empirical (test) data to ensure lower-bound (conservative) predictions of failure load.
- The shape of the corrosion defect is taken as two-dimensional (length and depth, but no width), and the profile is approximated by the 0.85 factor to account for the fact that corrosion tends to result in uneven depth of metal loss. Historically, measuring the actual shape of the defect in detail has been impractical, and in many cases only a small benefit can be gained with more detailed measurements given other approximations.
- The material failure stress is based on the specified minimum yield strength, despite the fact that in many cases the steel will actually be stronger than specified. In the past, of course, data on the actual strength of each pipe was not readily available. Even the best manufacturing records only confirm that certain samples from a batch of pipe meet the minimum requirements of the specification.
- Even though many pipelines rarely operate at their maximum allowable levels, acceptance criteria for defects are based on the maximum expected load condition, namely the Maximum Allowable Operating Pressure (MAOP).
- It is generally assumed that the only significant cause of loading is internal pressure, and only hoop stress is considered. There may be many other sources of loading, such as thermal expansion, overburden, ground movement and construction alignment. All of these loads may affect the resistance to the presence of metal loss.
Approaches that treat the inputs in a statistical or probabilistic manner as distributions have been developed. These can help us to understand the influence of different uncertainties, but they are also ultimately constrained by the availability of data to define the input distributions.
For features such as cracks, the influence of assumptions regarding materials and loading is even more significant than for metal loss. The basic crack assessment formula is:
We can thus see that crack size is less significant than either applied stress level or the material properties when it comes to predicting crack failure, since we take the square root of the size. This may be particularly important for the future of the pipeline industry, as there is some evidence that H2 service can result in changes in material properties and to a reduction in fracture toughness.
LIMITATIONS IN THE ACCURACY OF PREDICTIONS
It is important to be aware that although the formulae given above will give answers to as many decimal places as one wants, there are inherent limitations in the accuracy of predictions. Even with perfect data on the material properties, loading condition and defect dimensions, we cannot expect them to deliver exact predictions of failure stress or critical feature depth. This is a function of the semi-empirical nature of the methods; they are not a true representation of the physics of the situation. Even with perfect input data, the scatter in the predictions of most methods is around +/-10 percent. Consequently, ever more accurate feature sizing data will have diminishing returns in terms of threat management, unless the assessment methods are also improved and we get better data on material properties and local loading conditions.
A further challenge is combined or interacting features. As discussed above, our assessment methods for specific features such as cracks or corrosion contain inherent uncertainties or errors, and they are constrained by uncertainties in the inputs. When we get two or more features combined, such as when we get a dent, gouge and cracking, these uncertainties multiply. Assessment becomes extremely challenging, even with perfect input data, and the collection of data also becomes more difficult because the coincidence of the different features can influence the performance of the inspection systems.
A good example is fatigue cracking in the long seam. Roof topping is a geometric anomaly that affects some pipes more than others and accelerates the development of fatigue cracks due to stress concentration effects. Unfortunately, roof topping can also affect the positioning of the sensors of an in-line inspection tool, reducing the system’s capability to detect and size features. Consequently, an awareness of the potential for multiple features to be present and the deployment of systems to identify all that could be significant aids threat management. Returning to the roof topping example: if the presence of roof topping is not known, then sufficient attention may not be given to the possibility of hard-to-find cracks in these locations.
ADVANCEMENT OF DEFECT ASSESSMENT METHODS
Technologies that allow us to understand the stress state of pipelines with a certain degree of confidence have been available for some time. By considering the geometry and profile of a pipeline based on caliper and XYZ data, together with measurements and models of temperature, reasonably accurate estimates of stress can be derived. These methods are already deployed in the management of threats such as ground movement loading of onshore pipelines and thermal expansion of offshore pipelines. Sensor systems that give indications of axial stress levels are now coming to market, and more such systems can be expected in the future. Most defect assessment methods, however, do not consider the true loading condition experienced by a defect. Methods such as Modified B31.G tend to assume that the hoop stress is dominant and that other loads can be neglected. The reality is that a better understanding of biaxial load state will improve our predictions. Some work has already been done in this area, and the DNV-recommended practice for assessing corrosion defects (DNV RP F-101), for example, does allow for the effect of a compressive axial load to be taken into account (above a certain level, it reduces the failure pressure, so not including it can be non-conservative). The assessment of defects using 3D non-linear Finite Element Analysis (FEA), as illustrated in Figure 5, does of course allow for complex loading conditions to be taken into account.
Figure 5 – Examples of 3D FEA of dent, gouge and crack features
Historically, the combination of relatively coarse information, computing power restrictions, and the time and expertise required to build the models meant that this approach was limited to certain critical cases. Recent advances in cloud computing, automation and data granularity mean that using 3D FEA methods may become normal practice, even for relatively simple defects.
Inspection systems that can identify variations in material strength are available. These are increasingly accepted for demonstrating that a material meets certain strength requirements where records have been lost. They may also be used to identify “rogue” pipes that may have been used for undocumented repairs many years ago. As experience with these systems increases and the technology is improved, accuracy will increase, and the ability to identify variations in fracture toughness will no doubt be developed.
Defect assessment methods have traditionally been based on specified minimum material properties – on the basis that material properties will vary – and we have not known the specific properties at the location of any particular defect. With technology that allows us to measure material strength and identify local variations such as hard spots, we now have the prospect of using defect-specific material properties in our assessments.
A DIGITAL TWIN
We can thus anticipate that in the near future integrity assessments will be based not only on ever more accurate measurements of defects and alignment of coincident defects but also on measured material properties and loading conditions for the specific location. The assessment is likely to be completed using a detailed 3D FEA model of the pipe that closely represents true behavior – or, in other words, a digital twin rather than a semi-empirical model. These assessments provide us with a more precise understanding of the condition and the available margin of safety for defect growth or changes in load or materials. It is important to remember that if it is possible to run an in-line inspection, the pipeline must be operating and unlikely to be at the point of failure.
Having baselined our pipeline condition with a comprehensive digital twin (3D FEA), we now want to focus on the management of threats. For threat management, it is instructive to think about failures. Failures generally occur for one of the following reasons:
- A defect is introduced suddenly that is critical given the load and material – a dent and gouge introduced by impact from a backhoe, for example.
- A defect during pipeline operation grows to a critical size given the load and material. Active stress corrosion cracking caused by interaction of high stress and a corrosive environment and fatigue cracking driven by stress cycles are both examples of this phenomenon.
- The load on the pipeline increases such that the load on an existing defect becomes critical, such as when a sudden pressure increase due to a control system malfunction causes a corrosion defect to fail.
- The load on the pipeline increases such that undamaged material is overloaded, which may happen in the event of a landslide.
MANAGING THREATS INTO THE FUTURE
In the future, we may also see failures that are the result of material changes. There is some concern that the introduction of hydrogen into gas transmission systems could result in some materials becoming embrittled. If these material changes are coincident with existing defects, the combination could be critical and result in a failure.
So, once we can define condition with more confidence, how do we manage threats into the future? If we consider the ways in which failure can occur, it is clear that ideally we want to be able to do three things:
First, we want to predict accurately how defects grow, loads vary and materials change so that we can estimate when the pipeline condition may become unsafe and take action before that happens.
Second, we want to be able to inspect for changes in any of these key properties so that we can be confident that our predictions are valid.
Finally, we want to monitor and inspect for predictive indicators or changes that may lead to defect development, load increases or material property degradation.
In practice, this means collecting and aligning data on the things that could affect pipeline integrity. Using internal inspection, for example, we could monitor depth of cover, coating condition and cathodic potential. We can also collect and store data on pressure variations, product composition, soils, rainfall, geology, ground profile, land use, etc. – all of which may be useful in helping to predict degradation.
REPEATABILITY OF INSPECTIONS – INCREASINGLY IMPORTANT
Returning to the second point of inspecting for changes in the defects, loads and materials, it is worthwhile to think about what is actually being measured and, therefore, what changes we should look for. When baselining condition using in-line inspection data, we interpret what is recorded from sensors to estimate geometry or material properties. A certain level and pattern of magnetic flux leakage is linked to a certain metal loss defect size (length, width and depth); a particular time for a sound wave to be reflected is used as a measure of sensor standoff or remaining wall thickness. Each of these conversions has an associated error or uncertainty. If we take the resulting estimates of defect size and look for changes, we have to deal with that initial measurement uncertainty, with the result that small changes may be neglected because they cannot be distinguished from measurement conversion error.
However, if we directly compare the original measurements of flux leakage or sound wave reflection time, our confidence that there has, or has not, been a change is much higher. We can then look to understand any change and the cause of that change. In some cases, changes in measurements may be due to spurious effects, such as tool speed differences; in others, they will be caused by actual changes in defects. If we have evidence that defects, loads or materials are changing, then we can compare those with our predictions and take action accordingly. Consequently, the repeatability of inspections is likely to become increasingly important for threat management as we get better at accurately baselining condition with digital twins and at predicting deterioration taking account of monitoring information.