Seamless, Not Simple: The Case for Actual Wall Thickness in Metal Loss Depth Sizing

Seamless, Not Simple: The Case for Actual Wall Thickness in Metal Loss Depth Sizing

Pipeline operators know maintaining pipeline integrity is a complex and ongoing process, and nowhere is this more obvious than in this example from seamless pipe. Keep reading to learn why utilizing nominal wall thickness as the baseline measurement can be misleading when profiling metal loss, and just how important a nuanced data set with the right DA expertise is when it comes to seamless pipe.

It’s first important to know that seamless pipe is unique in that it is truly variable from joint to joint, and the detection and sizing tolerance of your integrity data may not be as tight as in seam pipe. Here’s a scene-setting image from an UHR MFL ILI data sample:

Scene setting image:

In this image, you can see adjacent segments of SMLS pipe between circumferential girth weld signatures in the data. If you look closely at the middle pipe joint, on the upstream end of the middle joint you will see the data is more blue, while the data on the downstream end of the joint is more yellow. The darker blue colors indicate that there is more metal present in that area. That means with this example you can tell that the wall thickness of the SMLS pipe on this joint is thicker on the upstream end then on the downstream end. We can confirm this with the Caliper data (shown beneath the zoomed in MFL example) as well, which also changes color from one side to the other, indicating a change in pipe wall thickness (within the same joint of pipe). The very bright yellow in the caliper data shows an ID gain – meaning a thinner wall on the pipe. The lined pattern of ribbing in the middle of the joint is something that can happen on SMLS pipe during the manufacturing process. We know that pattern is present in SMLS pipe, which shows that the pattern is from manufacturing and does not indicate new metal loss.

In the figure below, we have an AUT Scan of the pipe’s surface. The pipe was manufactured to be 0.312 inches nominal, which you can see in the reference strip is a royal blue color on the right of this image. Though the pipe was manufactured to be a standard thickness throughout, the data pictured shows multiple areas of darker blue, indicating a thicker wall at those spots. In some areas the wall gets as thick as 0.350 WT. Conversely, in the areas where the data is yellow and green, we know that those areas have a thinner WT, reading around the 0.280 inch level.

These two examples show us that there can be many variations in wall thickness of SMLS pipe in a relatively small area. This knowledge becomes extremely important to factor in when you have an incident of metal loss within these wall thickness variations.

In the example below, you can see the AUT data on the left compared to the MFL data on the right, taken of SMLS pipe with variable wall thickness. More of the diagonal ribs present in SMLS are pictured, this time with three metal loss features labeled.

Further comparing the two, we note that the “Target” feature on the left in the AUT data was deemed the deepest area of metal loss and therefore labelled as the primary feature. Yet in the MFL data on the right, it was determined that “A” was the deepest area of metal loss. Why is there a discrepancy? Let’s look closer.

Utilizing AWT vs. NWT in seamless pipe

In this illustration of the cross section of pipe, you can see what is happening. When looking at the Target feature, consider that AUT measures remaining wall from the outside of the pipe. Here, that is .228 in remaining at the Target point, with the UT showing 0.244 in remaining at Defect A. When comparing those measurements to the nominal wall thickness, you will get depth percentages of 22% and 27%, respectively, which makes Defect A appear shallower than the Target Feature from the AUT report.

However, if you look at the actual variation of the wall in the data from MFL, indicated by the dotted lines in the below image, you’ll see that the starting AWT at Defect A was 0.350in while the AWT at the Target Feature was only 0.277in.

This tells us that more metal was actually removed from Target A. In fact, the depth of Defect A from the AWT was double that of the Target Feature.  You can see the resulting percentage of metal loss was 30% for Defect A and only 18% for the Target Feature, as indicated by MFL data.

How does this happen? We know that when calculating WT from the in-the-ditch NDE equipment, often average values of pipe thickness or nominal wall thickness are used instead of calculating the precise actual wall thickness at many different points. Using this average is common practice, and is not wrong, but is also not accurate when dealing with seamless pipe and trying to achieve a percent depth.

Context is key.

What is the takeaway? When evaluating SMLS pipe and correlating ILI reports to in-the-ditch measurements, it is important to use the actual wall thickness for better correlation of features and to calculate percentage of depth, and to always make sure to understand the parameters being used in the pipeline assessment. This will require maintaining a wider sizing tolerance when validating the results. Following this process can prevent unnecessary reruns.