Peter Mudge joined TWI's NDT Research department in 1976 and was manager of that department from 1985-97. His major interest has been the development of ultrasonic testing techniques, latterly concentrating on the use of guided waves for long-range inspection. Peter is a past President of the British Institute of Non-Destructive Testing and a Member of the Institute of Materials. In 1990 he was awarded the Leslie Lidstone/ESAB Gold Medal for his contribution to welding technology. As Operations director of P i Peter is responsible for managing both manufacture of Teletest ® equipment and the delivery of services and demonstrations.
Long range ultrasonic testing (LRUT) was introduced commercially by Plant Integrity Ltd in the form of the Teletest ® technique in early 1998 for the in-service monitoring of pipes and pipelines. As Peter Mudge reported in part one of this two part piece it is largely used by the oil and gas industry for detection of corrosion and other metal loss defects, and is becoming widely accepted as a valid means of assessing the condition of pipes and pipelines, particularly where they are difficult or expensive to access for inspection. In part two he addresses the applications of Teletest.
Confirmation of sensitivity achieved
Earlier work showed that the smallest area of metal loss which long range UT can detect is approximately 3% of the pipe wall cross-section. The reporting level which is normally used is a signal amplitude equivalent to 9% area. This is to ensure that false call rates are kept to an acceptable level. However, if clear and unambiguous indications are detected below the reporting level, they are identified as minor defects.
An opportunity arose to determine whether these thresholds were capable of being met through involvement in the joint European RACH project, which was managed by University College London. A major part of this was the gathering of NDT data from controlled corroded pipe specimens using eight different methods in order to determine their detection and evaluation performance. The trials were conducted 'blind' without knowledge of the defects present and the results were evaluated by an independent team from Bureau Veritas, Paris. Figure 12 shows the results from the Teletest ® technique on 36 individual defects. The plot is in terms of depth and circumferential extent of the defects and indicates whether each was detected or not. The lines representing 3% and 9% defect area for the 6" diameter pipes tested are also included.
The figure shows that under blind trials conditions the Teletest ® technique performs as expected from the development work. The limit of detection is clearly at the 3% level, with virtually no successful detections below this size. The data show the classic probability of detection characteristics, with an increasing likelihood of detection above the 3% level. All flaws examined which were around or greater than the 9% reporting level were detected. These results are important as they demonstrate that the performance of the technique, determined from 'open' tests on known specimens, could be reproduced when testing real corroded pipes with unknown (internal) flaws. Thus these tests provide valuable evidence of the performance level which may be expected in the field.
There is also evidence that the sensitivity to defects is greater in large diameter lines where the ratio of diameter to wall thickness is much greater than in small diameter pipes. Figure 13 shows the results from a total of 86 machined flaws in 24" diameter, 0.344" (8.74mm) wall pipe. In this case, flaws considerably smaller than the 9% area line were detected and there was a good detection rate for flaws smaller than 3%. These results clearly show that on this larger pipe size the minimum sensitivity of detection of loss of wall equivalent to 9% area is readily exceeded and that the practical target for detection of 3% area was likely to have been reached in the majority of cases. Of the five defects which were not detected (two being of 20% wall thickness and one inch wide), the two giving cause for some concern are those of 65% and 80% depth (see Fig.13), which would have reduced the remaining strength of the pipe to around 50% and 70% of its original strength respectively. These were not detected because they were narrow (1in. or 25.4mm) in circumferential extent, and gave a small response echo despite being relatively deep. The reduction in pipe strength owing to the presence of these defects was large in each case because they were long in the axial direction. The axial length has little effect on detection by the long range UT method because the waves hit the defects 'end on'. The three cases of non detection of flaws of 20% depth is not significant, as these would have reduced the strength of the pipe by less than 10%.
Case studies
Case Study 1 - 14" Ammonia line
This line was the feed to a reactor vessel in a chemical complex and was normally insulated. It emerged from the reactor around 2m above ground level, ran vertically for around 7m then horizontally for a further 10m. Visual inspection was not easily carried out owing to the insulation and access was not feasible to the elevated section without scaffolding. External corrosion under insulation (CUI) was suspected. The Teletest ® transducer was attached near the base of the vertical section, with the aim of examining the vertical and elevated horizontal legs. An example of the Teletest ® A-Scan output is shown in Fig.14a. The two lines are distance amplitude correction (DAC) curves, the upper representing the amplitude from butt welds in the pipe and the lower being the reporting level. The large peak at around 5m from the transducer is a weld at the elbow where the pipe turned to the horizontal. A number of defects were reported in the region 13 to 19m from the transducer ( marked '+' on the plot). On removal of the insulation for cleaning and visual inspection, these were confirmed as areas of CUI attack. Some of these are shown in Fig.14b.
Case Study 2 - 24" Slurry line
This line carried a water-based slurry. It was not insulated and was at ground level, so that access for inspection was not difficult. However, the main concern was local high levels of erosion internally where eddies in the flow caused turbulence and consequent high impact of particles in the slurry on the inside of the pipe wall. Since service history had shown that these occurrences were difficult to predict, spot thickness measurements were ineffective in detecting thinned areas before leaks had occurred. Application of Teletest ® overcame this problem as 100% of the pipe wall is examined. During initial trials, a test was carried out on a section where a small leak had occurred. The result is shown in Fig.15a. In this instance, the signal at approximately 12m from the datum is from a butt weld in the pipe. The very large signal which follows it at 14m (marked '+') coincided with the location of the leak. This suggested extensive metal loss and it was found by subsequent examination that there was a band of erosion almost through the wall for the majority of the pipe circumference. The pipe was therefore at the end of its service life. The Teletest ® equipment may be seen in Fig.15b.
Case Study 3 - 10" Buried line
This work was carried out on a network of water injection lines at an oil processing facility. These were partially above ground, but the majority were buried. The concerns were external corrosion around the soil to air interface and where the coating had been damaged in the underground sections. Tests were carried out either from above ground sections or from 'bell hole' excavations. An example of the Teletest ® equipment set up on an above ground section is shown in Fig.16a. Figure 16b shows the Teletest ® result from a buried section of the line, which was totally inaccessible at the time of inspection. The scale has been enlarged to show the feature of interest, which is at around 26m from the transducer. There are welds at 11.5 and 25.5m. The first is shown off-scale; the steepness of the DAC curves indicates the higher levels of attenuation which are generally observed on buried lines with protective coatings. It should also be noted that the scatter on the baseline (similar to ultrasonic 'grass') is greater for such lines.
Immediately beyond the second weld is an additional signal at 26m from the transducer, with associated mode converted components, which are plotted in colour (marked '+' on the plot). This region was reported as a moderate-to-severe defect and the area was excavated. The pipe was found to be heavily corroded at that point and a repair was put in place immediately. The section removed is shown in Fig.17. The weld can be clearly seen with the heavy corrosion adjacent to it. It is unlikely that this defect would have been detected by any other means before it had caused a failure of the line.
Factors influencing performance
The main factors which affect the sensitivity of the technique to corrosion defects are:
- The size of the corrosion as 'seen' by the wave propagating along the pipe. The detectability is related to the proportion of the pipe wall cross-section which is missing, ie it is a combination of the depth and the circumferential extent. Our experience is that the limit of detection is an area of 3% of the original pipe wall cross-section. The technique is equally sensitive to internal and external defects. Similarly, the sensitivity is uniform around the pipe, unless it is affected by pipe features (see below).
- Depth of corroded area. Although the primary parameter affecting detection is area, the technique is more sensitive to depth than circumferential extent, ie a deep, short region of corrosion will produce a stronger signal than a shallow, wide region of the same cross-sectional area.
- The axial extent of the corroded area. The technique is less sensitive to this dimension, as the waves approach the defect end-on. However, we do have some evidence that a long defect produces a stronger signal than a short one, provided that the circumferential extent is large enough for it to be detected.
- Pipe features. All discontinuities and changes of geometry affect the ultrasound signals and therefore give rise to responses. There are a number of pipe features which may affect the ability to detect defects. Some of these are listed below:
- Butt welds. These are widespread in any pipe system and it is essential that there is a means of distinguishing between these and other defective, regions. The Teletest ®
system achieves this by examining the different characteristics of the weld and defect signals during the data treatment and display process.
- Bends. Bends give a response owing to their geometry. When testing a pipe system containing bends it is necessary to determine that there is sufficient sensitivity in the region following the bend to allow testing (see below).
- Attachments. Items, such as pipe supports, welded to the pipe again produce signals. It is necessary to determine an interpretation procedure for such cases, depending on the geometry of the support brackets.
- Ts and branches. Small diameter branches, such as instrumentation lines, do not create too much of a problem. However, if there are Ts which attach branches of a diameter approaching that of the main pipe, the signals may become very complicated and this may prevent a satisfactory test being performed in that region.
- Coatings. Some types of coating affect the rate of attenuation of the ultrasound and therefore reduce the test range achievable. It is known that the polyurethane foam insulation does affect test range, but not to the extent that testing will not be feasible. Internal deposits may also cause some absorption of the ultrasound and may cause some reduction of test range.
- Test sensitivity. To perform an adequate test, a certain level of ultrasound has to be generated. Also, there needs to be a minimum signal-to- noise ratio in order to maintain the expected sensitivity to defects. These aspects are assessed at the time of testing, and they determine the effective test range at that point. The range may be influenced by features such as bends etc.
Conclusions
The growing body of evidence for the performance of long range ultrasonic testing in general and Teletest ®
in particular, as illustrated here, is supporting the wider application of this novel technology. LRUT has already crossed the technology transfer threshold from a curiosity to a usable and highly effective tool, and the range of applications continues to grow.
Proven applications are for:
- Painted pipework
- Mineral wool insulated lines
- Polyurethane foam insulated lines
- Bund wall penetrations
- Buried pipelines
- High temperature lines (<+125°C)
- Low temperature lines (>-25°C)
- Mixed phase lines
- Road crossings
- Sleeved sections
- Spirally welded pipe
- Stainless steel pipe
- Offshore riser inspection
