The 2nd International Conference on Technical Inspection and NDT (TINDT2008)- October 2008 - Tehran, Iran
1
Approaches for Developing Solutions for Specific NDT Problems
Casper Wassink1, Frits Dijkstra2
1- Applus RTD Technological Center, Rotterdam, The Netherlands AND Delft University of Technology, department
of Technology, Strategy and Entrepreneurship; 2- Applus RTD Technological Center, Rotterdam, The Netherlands
casper.wassink@applusrtd.com
Abstract
As oilfields are being explored in deeper and deeper water, the inspection of welds in deepwater Steel
Catenary Risers (SCR's) is a growing issue in the oil and gas industry. These structures are subject to fatigue,
and because of this the requirements for the welds are very demanding. In construction of the risers, basic
inspection is a quality assurance process. Apart from concerns about welding quality, inspection is
increasingly tailored to find flaws that might become a crack initiation later on. During the lifetime of the
installation the concern then shifts to monitoring the welds for the presence and growth of potential cracks.
Two approaches for developing NDT solutions are presented in this paper. The first one is starting from the
problem analysis, the other is starting from the practical toolbox of the NDT technician. This paper will argue
that both are in essence right in their own way, but that both perspectives are needed to develop breakthrough
NDT technology. We will look at the problem of weld inspection in two distinct situations: new construction
of the weld and inspection during the use of the equipment. These situations give rise to distinct failure
modes, and associated flaws. We will then look at the ultrasonic toolbox, from a fundamental point of view
and try to match fundamental characteristics of ultrasonic techniques to the flaws present in both situations.
The research presented in this paper has been conducted by Applus RTD and Delft University of Technology
and aims to better understand the innovation process in NDT and other areas where new technology and
safety interact. The input of Delft University is in the innovation system model presented. The research is
motivated by Applus RTD trying to improve facility in employing new technologies. The intent is to
understand how to maximize the value created for customers while retaining high quality and consistently
providing customer confidence.
Keywords: Deepwater risers, ultrasonic crack monitoring, problem solving approach
Introduction
Education teaches us that solving a problem is best started by analysing the problem, and then testing possible
solutions, before choosing the best one. The craftsman however will often try to apply his craft, and modify his
solutions to fit the immediate situation. NDT is mostly a craftsman industry, leading to a discourse that focuses
mostly on comparing available techniques. A clear example of this is e.g. Dubé [1]. In this article the suitability
of ultrasonic inspection for girth welds is related as a replacement of radiography, instead of looking at the
suitability of ultrasonic inspection to find the types of flaw that might give rise to failures. As Scruby noted in
ENDE2007 [2], the state of the industry is now a disconnect between the NDT academia and NDT application
field. In practical life this mostly shows up as a very long period needed to get a new NDT technology accepted,
and during this implementation period as a feeling of disillusionment around new NDT technology.
Technological thinking in the NDT industry is dominated by single techniques. Many times clients of Applus
RTD will request for installations to be simple Radiographed, Guided Waved, or Phased Arrayed, without the
problem or objective of inspection being specified. Even when discussing single NDT techniques it is possible to
distinguish two levels of technology. There is the inspection tool, and the application of this tool. Codes and
1- Manager NDTI Technological Center
2 - Senior NDT Consultant
The 2nd International Conference on Technical Inspection and NDT (TINDT2008)- October 2008 - Tehran, Iran
2
standards clearly make this distinction, in having a separate code for the inspection equipment, and for the
inspection process. For the purpose of illustration, one could say that building a race car is different from driving
it, and that different skills are needed for both. However, to know something about the car you drive certainly
helps to drive it faster, and to know about driving when building one helps to make a faster car.
Knowing how to drive a car however still won't let you drive from A to B without additional knowledge about
the road systems, traffic rules and where B actually is. In addition a car is not the only way to go from A to B,
there are also planes and trains. Translated back to NDT, designing an inspection solution needs considerable
knowledge about the actual problem, including metallurgy, corrosion processes, mechanical loading of the
component, and what process the component is being used for. Additionally, Ultrasonics is not the only NDT
technology, there is Eddy Current and X-ray as well.
Finally, design of NDT equipment cannot be done without underlying fundamental knowledge about
measurement physics, and other fundamental scientific knowledge about electronics, mechanics, industrial
design and a large number of other technical disciplines. After Heron [3], this can be presented as an Up-
hierarchy or as a Cyclic model. The Cyclic model will be left for future papers. In this paper, the structure of
NDT technology is presented as an Up-hierarchy as shown in fig. 1.
Solution
Technology Application
Engineering
Fundamental Scientific Knowledge
NDT solution
NDT procedure
NDT equipment
NDT
fundamentals
Going from A to B
Driving a Car
Building a Car
Car fundamental
knowledge
Figure 1: Up-hierarchy of NDT technology
Looking at the problem of riser inspection we can now approach this in two ways.
1. Fundamental knowledge up
2. Integrity problem down.
To further illustrate the inspection situation and the difference between manufacturing and in-service
inspection we will start from the Integrity Problem.
The integrity problem of riser welds at construction and during their lifetime
As described by Bouma [4], the inspection process during the construction of a pipeline is mainly a quality
assurance process. A weld is being produced and the welders need feedback about the quality they produce to
keep making quality welds. Already during the welding procedure qualification, NDT is used to fine tune the
welding. The main kind of flaws that need to be found are flaws of the welding process. Knowing that this is the
problem to be solved, it is logical to find out what kind of flaws are actually produced, and which NDT
technology can detect those. Most girth weld AUT procedures are designed to detect lack of fusion type defects
in a GMAW welding process. Such a procedure would not be appropriate if the dominant flaw would be e.g.
solidification cracking.
In new construction NDT, the link between the NDT technique used and the quality level required by the QA
process is in the acceptance criteria for weld imperfections. These are, for critical components, often specific for
a specific project and for the NDT technique used.
The 2nd International Conference on Technical Inspection and NDT (TINDT2008)- October 2008 - Tehran, Iran
3
In in-service NDT the focus is not the weld quality, but the detection and quantification of possible service-
induced damage such as fatigue cracks. In this case the approach is completely different: we are not talking
about a quality level, but about the possible occurrence of an individual defect which is possibly growing and, if
it does, might one day become the cause of a catastrophic failure. No wonder that one wants to know the growth
rate (if any), the critical size, and the remaining lifetime as a result of that. Inspection intervals will have a direct
relationship with inspection coverage, detection probability and sizing accuracy. Together with fracture
mechanics considerations, NDT plays a crucial role in monitoring such critical components.
Fatigue cracks in welds as a potential integrity problem
For detection and sizing of imperfections in e.g. welds, a number of NDT solutions exist. Many NDT experts
are familiar with ultrasonic techniques such as Pulse-Echo, Time of Flight Diffraction (ToFD), Phased Arrays,
Tandem Technique etc. But it is not commonly realized that these techniques are fundamentally different in their
principles for defect detection and sizing and, on top of that, can be applied in different ways. Of course, they all
use ultrasonic shear waves and / or compression waves, and they all use reflection and/or diffraction on the
defect as detection and sizing mechanisms. But the nature of a technique and the way it is used can make it either
very suitable or completely unsuitable to solve a specific type of problem.
When looking at the Pulse Echo technique, most commonly used in ultrasonic testing of new construction, we
see that application on welds does not guarantee for instance perpendicular incidence on a lack of fusion (LOF)
defect. Although a reflected signal often has a large amplitude, most of the reflected signal will not necessarily
return to the probe, or does not reach the probe under the angle the probe is designed for (fig 2).
Luckily, in addition to reflection on the defect's face (blue), also diffraction at its edges will occur (green) .
Some of these (weak) signals can return to the probe, enabling a certain likelihood of detection. If the defect is
rough, also reflections from its facets can be present. These are often the primary detection mechanisms that play
a part in weld inspection according to a major code or standard. Although there is a fair chance of detection,
there is no way to size the defect.
In new construction NDT, these limitations have been proven to be not a problem at all. History shows that
NDT according to a well-established code or standard provides a sufficiently high safety and integrity level.
NDT using new construction is more intended to check the performance of the welder than the weld itself, it is
primarily intended to check weld quality.
Figure 2: Pulse Echo technique on a LOF defect
Fig. 3: detection of an (almost) perpendicular defect
In monitoring existing structures for the presence of fatigue cracks, fundamentally different questions should
be asked. In this case, it is crucial to be able to tell how small or large the crack is, how fast it grows, how often
the structure has to be inspected and what the remaining lifetime can be expected to be. This means that not only
extremely reliable detection is required, but also accurate through-thickness sizing.
A high probability of detection can be achieved in various ways. It would help to adapt the probe angle in such
a way that the expected defect is hit perpendicularly. This often works for defects oriented along a V-bevel, but
not for (almost) perpendicular defects oriented along narrow gap bevels or perpendicular fatigue cracks. Fig. 3
shows that the specular reflection of a perpendicular defect will not return to the probe. The obvious solution
here would be to place a separate receiver behind the transmitter probe; this is the idea behind Tandem
Technique.
Whatever measures we take to ascertain optimum incidence on the expected defect, they have one thing in
common: they enable using amplitude for defect sizing, because they create a situation in which there is a
relationship between flaw size and amplitude. This principle is used in e.g. pipeline AUT, where even the angles
of transmitter and receiver are adapted in such a way that an LOF defect in a narrow gap weld is hit in an
The 2nd International Conference on Technical Inspection and NDT (TINDT2008)- October 2008 - Tehran, Iran
4
optimum way (fig. 4). In this way, sizing algorithms can be used to size defects on the basis of their amplitude
[5, 6]. Different zones can be addressed by different probe pairs, or, alternatively, by a phased array with
sufficient length (fig. 5).
Figure 4: Tandem technique with adapted angles
Figure 5: pipeline AUT using a phased array
Whilst amplitude-based detection and sizing require optimum angles for reflection, the use of another well-
known phenomenon, tip diffraction, requires the opposite: the defect shall NOT be hit perpendicularly. By not
obeying this "law", the strong reflection signal will completely mask the weak diffraction signal. This is the
principle used by techniques such as ToFD and Phased Array Sector scan. Both techniques are designed for
receiving and viewing reflected and/or diffracted signals in such a way that time differences can provide accurate
information on flaw through-thickness size.
Figure 6: ToFD technique heavily relies on the presence of diffraction signals
In ToFD, the above is evident. Fig. 6 shows how diffraction signals from flaw tips can reach the receiver,
whilst the specular reflection cannot. The presence of tip diffractions, together with complementing observations
such as interruption of lateral wave and/or bottom reflection, is used for detection. Timing of the tip diffractions
is used for sizing, using lateral wave and bottom reflection as references [7]. Amplitude is not used.
Something similar, but perhaps less evident, is happening in Phased Array sector scan [8]. A sector scan
consists of a large number of separate pulse-echo shots. If, for example, the sweeping range is between 40 and
70º with increments of 1º, 31 separate pulse-echo measurements are performed, as if many different angle probes
were used successively. This looks at the defect in many different ways. A few of the situations are depicted in
fig. 7. A surface breaking crack is taken as an example.
At 42º: The corner trap effect will give a strong signal. This is a reliable means of detection.
At 46º: The reflection does not reach the probe; signals from the crack face will be received, but only if
the crack is sufficiently rough.
At 50º: Only a diffraction signal from the crack tip is received. Sizing is possible by using its timing.
At 54º: No signal.
At 62º: Only a diffraction signal from the crack tip is received. Sizing is possible by using its timing.
At 70º: The reflection does not reach the probe; signals from the crack face will be received, but only if
the crack is sufficiently rough.
The 2nd International Conference on Technical Inspection and NDT (TINDT2008)- October 2008 - Tehran, Iran
5
Figure 7: Different situations during a Phased Array sector scan
So, sizing is done by looking at the timing of the diffraction signal, using e.g. the corner trap signal as a
reference. It is evident that this only works if the corner trap signal and the tip diffraction signal are sufficiently
far apart in time; this means that too shallow cracks cannot be sized.
Figure 8: Sector scan of a rough fatigue crack
This principle of height sizing by means of tip diffraction signals was already designed in 1978, before phased
arrays had been introduced in NDT [9].
Fig. 8 shows an example of a sector scan of a fatigue crack. Note that even the roughness of the crack is
visible (the arc between corner effect and tip diffraction). So, sizing is done by looking at the timing of the
diffraction signal, using e.g. the corner trap signal as a reference.
Linking NDT problem to fundamentals
So, which technique is to be used when? The more important it is to achieve reliable detection and sizing, the
more essential it will be to analyse the problem first, and then carefully select the matching NDT solution. It is
clear that an amplitude-based sizing technique is only possible if prior exact knowledge of the defect's location
and orientation is present. If this knowledge is not present, diffraction-based techniques are probably more
suitable. Furthermore, other arguments can play a role: Is the crack expected to be very tight (possible
The 2nd International Conference on Technical Inspection and NDT (TINDT2008)- October 2008 - Tehran, Iran
6
transparency)? Is its orientation known? Is it large enough to be sizeable within the technique's resolution? Is it
rough? For maximum detection reliability and sizing accuracy, it can be advantageous to use redundant
techniques, using different fundamentals. An approach could be to do the detection with a reflection-based
technique (pulse-echo, tandem, phased array sector scan directed to detection of corner effect, creeping waves
etc). For sizing, additional amplitude-based techniques (e. g. tandem technique) and/or or diffraction-based
techniques can be used (phased array sector scan, ToFD, focusing techniques). This illustrates the importance of
seeing ultrasonic techniques as a toolbox and not more than that. Which tool(s) to use and in what way has to be
dictated by the nature of the expected defect.
Conclusions
- It was shown that NDT technology can be evaluated at several levels of abstraction
- It was shown that different objectives for NDT, require different solutions
- It was shown that several tools in the ultrasonic tool box can be tailored for the problem of sizing fatigue
cracks in SCR risers. In case of highly critical components it may be wise to use multiple, complementing
NDT techniques.
References
[1] Dubé, N., Ginzel, E. A. and Moles, M. D. C., "Mechanized Ultrasonic Inspection of Large Diameter Gas
Pipeline Girth Welds", PACNDT 98, Toronto 1998.
[2] Scruby, C. B., "NDE Research makes a difference", ENDE 2007, Cardiff, 2007.
[3] Heron, J., "Co-operative inquiry", Sage publications, London 1996.
[4] Bouma, T., Pörtzgen, N. and Dijkstra, F. H., "Advances in Non-destructive Testing (NDT) of Pipeline Girth
Welds", Proceedings of the Conference Application and Evaluation of High-Grade Line Pipes in Hostile
Environments, November 2002, Yokohama.
[5] De Sterke, A., "Flaw tip reflection as a help in ultrasonic flaw size estimation", paper to the 1976 Annual
Conference on NDT, organised by the British Institute of NDT et al, London, April 26-28, 1976.
[6] De Raad, J. A., "Inspection of Girth Welds by AUT", Applus RTD, 2007.
[7] Charlesworth, J. P. and Temple, J. A. G., "Engineering applications of ultrasonic time of flight diffraction,"
Research Studies Press Ltd., Second Edition, 2001.
[8] Moles, M. D. C. et al, "Introduction to phased array ultrasonic technology applications", Olympus NDT,
2004 / 2007.
[9] Dijkstra, F. H., "Flaw evaluation with particular reference to flaw size estimation", paper to the IIW
Colloquium on nondestructive determination of type, position, orientation and size of weld defects,
1978.
pdf