This article was originally published in two parts in the April and July 2019 issues of ASNT NDT Technician Newsletter. It appears here with minor edits.

.

by Kervan Govender

Although it may seem simple at first, detecting and correctly classifying internal corrosion has proven to be quite a complex skill to acquire and master, even to the most experienced of technicians. Time and resources are used to re-inspect false calls. The problem lies when the NDT technician detects a reduction in wall thickness on the vessel or piping and is unable to supplement the finding with an alternative method or technique such as radiographic or visual testing. Using two or more methods/techniques to confirm an indication is good engineering practice. Since most false calls are attributed to laminations and inclusions, we will look at ways of differentiating laminations and inclusions from internal corrosion. These laminations and inclusions existed in the component from the initial manufacturing stage so are not necessarily detrimental to the component’s service life.

Part 1

In the first part of this article, we will look at the basic outline of corrosion detection theory and what practical options are available to an ultrasonic technician working in the field.

Overview of Corrosion Detection

The inherent difficulties of extracting oil and gas from subsea reserves are numerous. One of the most notable is the vast amount of seawater that is pumped up together with the crude oil. The seawater acts as the perfect electrolyte to initiate and sustain internal corrosion in associated piping and vessels. As stated in the text Corrosion for Science and Engineering, “sodium chloride (NaCl) in water is an extremely aggressive corrosion medium.

Corrosion is the degradation of a metal by an electrochemical reaction with its environment” (Trethewey and Chamberlain 1995, pp. 28 and 30).

There are many mechanisms of corrosion that may cause wall loss: for example, galvanic, crevice, pitting, or intergranular corrosion. Flow-induced erosion corrosion is expected in oil extraction systems, due to high flow rates pumping out of the subsea oil reserves together with solids such as sand. Knowing which type of corrosion to expect may help the NDT technician in determining which areas are of more concern to test, meaning when flow-induced erosion corrosion is suspected, then bends and reducers are typical areas of interest due to impingement of particles on the inner wall.

As stated in the text Corrosion for Science and Engineering, “a sudden change in inner diameter or a change in direction of pipe will result in turbulence and therefore an increase in likelihood of flow-induced erosion corrosion” (Trethewey and Chamberlain 1995, p.192).

Compression (longitudinal) waves are used extensively in the oil and gas industry for the detection of internal corrosion in process piping and pressure vessels. It has proven to be a successful technique when performed by competent technicians.

Ultrasonic waves are high-frequency mechanical vibrations transmitted by the probe, which travel through the test part, bounce back off the opposite surface or anything in between (discontinuities), and are then detected again by the probe. It is similar to striking a ball off a flat surface and then catching the ball on the rebound.

Corrosion detection, on the other hand, may be compared to bouncing a ball off a bumpy surface; the ball (here, the signal) could possibly rebound in any direction. Low corrosion signal or even no corrosion detection at all would result.

The ideal reflector (discontinuity) would be parallel to the outside scanning surface and have a smooth contour. But internal corrosion is much more difficult to detect due to its irregular profile (Drury 1997).

.

.

Note in Figure 1 the corrosion echo is lower in amplitude and more “rounded” compared to the normal first backwall echo (BWE), meaning the first BWE is of higher amplitude and has a sharp peak. Also note that the rounded corrosion echo is more to the left of the screen when compared to the first BWE, meaning it’s shorter on the time base, indicating a reduction in measured thickness.

Another telltale sign of corrosion is the loss of the second BWE. As stated in the Nondestructive Testing Handbook, Vol. 7: Ultrasonic Testing, third edition, “corrosion testing ideally uses the same principle of operation as thickness testing,” but corrosion testing is more complicated because with corroded materials, the series of backwall echoes and the corrosion signal itself rarely shows above the system’s noise level (ASNT 2007, p. 445). According to the same text, “Ultrasonic scattering from inside surface pits produces a noisy reflected signal” (p. 467).

Careful consideration should be given by the technician during corrosion testing due the fact that the only indication of internal corrosion is a loss of repeat backwall echoes. Suspected areas must be interrogated and other possible causes for the loss of signal eliminated (for example, insufficient couplant, loose cable connections, or an irregular or rough outer surface). Codes that regulate the industry, such as API 510 (American Petroleum Institute), require vessels to be visually inspected internally within specified time frames (for example, 5 or 10 years).

Under certain conditions the internal inspection may be substituted with an alternate on-stream inspection, which uses ultrasonic or radiographic testing to determine the remaining thickness of the vessel. Ultrasonic testing (UT) and radiographic testing (RT) are considered volumetric NDT methods, meaning the full through thickness of the test part is tested as compared to magnetic particle testing, where basically only the surface is inspected. When an indication that may hinder the service life of the vessel is detected with UT or RT, then the responsible inspector may shorten the inspection time periods (for example, six months).

In-service inspection is much favored by the client compared to internal visual inspection since the latter would require stopping the production process to allow entry into the vessel. The costs and labor hours can raise quite considerably—for example, to allow for labor to open up flanges, high-pressure cleaning for internal inspection, and pressure tests upon completion—not to mention any delays in production will result in a loss of revenue for the oil producer.

Most training schools and international certification bodies, such as ASNT and BINDT, have set guidelines on the training and certification of UT technicians. Certification is further subdivided into smaller categories for which the technician may seek competence (for example, butt welds in plate, pipe or nozzle welds, and the like). The training and examination samples have discontinuities similar to those found in new fabrications, such as lack of fusion, lack of penetration, and slag inclusions. But rarely do we find training or certification samples that are specific to in-service discontinuities such as internal corrosion. This poses its challenges as competency in corrosion testing falls solely on the employer’s in-house training program.

Some shortfalls that may be encountered with regard to this include the following scenarios:

• The Level III technologist might be based in a different area/country other than where testing is carried out by Level I and II technicians
• No clear transfer of knowledge from mentor (Level II technician) to trainees or Level I technicians, meaning trainees have to rely solely on company procedures to acquire skills for corrosion detection
• Trainees not taking the initiative to seek knowledge themselves or ask questions related to their field
• In-house or client competency checks may not be adequate enough to simulate real corrosion conditions found in the field (meaning the examination sample is fabricated from scratch and not an actual in-service corrosion failure sample)

Due to the nature of internal corrosion, it is well suited for detection with UT and RT from the outside of the vessel, but the basic forms of liquid penetrant, magnetic particle, magnetic flux leakage, laser profilometry, and eddy current techniques cannot detect this in-service degradation from the outside surface (ASNT 2007, p. 437). Some possible alternatives to complement corrosion testing are:

• Phased array – zero-degree
• Phased array – dual array
• Tangential RT (digital or conventional film)
• Visual testing (a last resort since this requires equipment to be shut down to allow for internal inspection)

“Phased array dual transducers are an improvement in ultrasonic detection and characterization of internal corrosion compared to conventional transducers” (Pellegrino and Nugent 2015). Phased array is possibly one of the best alternatives to confirm corrosion detected with conventional UT, since access to only the external side of the vessel is needed (unlike visual testing), and it does not matter if the vessel is full of seawater or oil, which may hinder RT. The drawback is the client is going to have to cover the additional costs involved with phased array, which may require more specialized technicians, purchase of new equipment, and probes.

Another option is profile radiography. Tangential radiography is popular for corrosion detection (ASNT 2002, p. 519). Due to the large source-to-film distances needed, this technique is not practical for large-diameter vessels; however, tangential radiography is practical for small- to medium-diameter piping.

As stated in the Nondestructive Testing Handbook, Vol. 4: Radiographic Testing, third edition, radiography does not work very well for in-service examinations as the vessel is likely to contain water or product of some type, which would attenuate the radiation beam and cause an unsatisfactory radiographic image.

Random spot thickness readings on a vessel are not enough to detect internal corrosion on a vessel. The vessel should be scanned in 300 mm × 300 mm (12 in. × 12 in.) grids at areas where internal corrosion is suspected. Also, each grid should not be scanned slowly but rather with a fast back-and-forth motion (while ensuring probe overlap), since we humans are much better at detecting a change in pattern when it happens sharply. According to research by Drury, this fast “windshield wiper action” is quite efficient at detecting small-diameter pitting due to a sudden drop in the echo pattern (Drury 1997). Ultrasonic angle beam examinations are more suitable in most instances because they need access to one side only, and the water in the vessel would not pose a problem (ASNT 2002, pp. 526–527).

.

.

Angle beam inspection would be suitable to detect an in-service discontinuity such as fatigue cracks but as illustrated in Figure 2, angle beam inspection is not a very reliable technique for corrosion testing (ASNT 2007, p. 220). The corrosion indication gives a very low-signal response or even no echo at all (ball bouncing off bumpy surface).

Part 2

In Part 2 of this article, we look at classifying indications seen on the A-scan display of the UT machine using a zero degree twin crystal probe due to its widespread availability and cost effectiveness. We will also look at two cases (as shown in Figures 1 and 2) where wall loss was reported initially as internal corrosion but was later reviewed as either inclusions or laminations after numerous follow-up inspections.

.

.

Figure 1 is a glycol suction scrubber where internal corrosion was reported next to the inlet nozzle on the shell, and Figure 2 is a firewater pump start air receiver where internal corrosion was reported on the blind flange.

Methodology

Below is a step-by-step description on how to analyze the signal:

• The UT machine must have an A-scan display.
• Connect a 3/8 in. (10 mm) diameter twin crystal 5 MHz delay line probe.
• Calibrate the timebase on the UT machine according to the company’s procedure.
• Apply couplant and place the probe on a clear area of the test piece, meaning an area with no internal corrosion.
• Set the leading edge of the first BWE to 40% of the timebase by adjusting the range setting, meaning place the echo on graticule 2 of 5 horizontally.
• Set the peak of the first BWE to 80% of the full screen height (FSH) by adjusting the gain setting, meaning place the echo height at graticule 4 of 5 vertically.
• The leading edge of the second BWE will automatically fall on 80% of the timebase, since the second BWE is double the time of flight of the first BWE.
• Scan quickly in a back-and-forth motion, ensuring you cover the whole test area, and observe any changes in the echo pattern. One of five patterns may be observed, as summarized in Figure 3 with an accompanying sequence flowchart illustrated in Figure 4.

.

.

Report Analysis

This section covers equipment and methods and findings.

Equipment

The following is a list of equipment used in these examples:

• UT machine with an A-scan display
• 5 MHz 3/8 in. (10 mm) diameter twin crystal delay line probe
• 4 MHz 3/8 in. (10 mm) diameter single crystal angle probes (45°, 60°, and 70°)
• Wallpaper paste as couplant
• IIW-type V2 calibration block
• Step wedge

Methods and Findings

Figure 5 provides a summary of inspection reports for the glycol suction scrubber. In the original inspection of the glycol suction scrubber (Visit 1), Technician A is the one who initially classifies the discontinuity as internal corrosion and does the same again in Visit 3.

.

.

But in Visit 5, Technician A now classifies it as entrapped slag in the nozzle-to-shell weld. Note that in Figure 5b, although the signal moves to the left on the timebase and the second BWE has disappeared, the indication does not have the pattern of a typical corrosion echo (low amplitude and rounded). Please refer to Figure 3 (summary of scanning patterns) and Figure 4 (discontinuity classification flowchart) for further illustration of the scan patterns.

Figure 6 provides a summary of indication reports for the firewater pump start air receiver. For the firewater pump start air receiver, Technician D initially classifies the indication as internal corrosion in 2011. In 2014, Technician A also classifies the same indication as internal corrosion, but later revises this decision upon subsequent visits.

.

.

The indication is typical of a lamination. Time and resources could have been spared if it was classified correctly initially. In Figure 6b, it can be seen that the indication comes up on the left of the timebase, on its own, and the first BWE is still present at 40% of the timebase. Also, in Figure 6c, we can see the multiple repeat echo pattern. Both of these key factors are indicative that the discontinuity is in fact a lamination and not internal corrosion. Please refer to Figure 3 (summary of scanning patterns) and Figure 4 (discontinuity classification flowchart) for further illustration of the scan patterns.

During the plant turnarounds, the opportunity to inspect the two vessels in this study with internal visual inspection was seized. The findings are illustrated in Figures 7 and 8. As can be seen, there was no internal corrosion present in the glycol suction scrubber (Figure 7). The areas marked in red in Figure 7 were incorrectly classified as internal corrosion and later revised to inclusions in the weld. This nozzle is quite unusual since it has a nominal thickness much larger than the vessel shell thickness (the nozzle wall thickness was 50 mm and the vessel shell thickness was 12.5 mm). The technicians should not have made a final call on this finding without having the vessel weld details and should have consulted with other technicians.

.

.

.

As can be seen, there was no internal corrosion present in the firewater pump start air receiver (Figure 8). This blind flange had been incorrectly classified as internal corrosion, and then later on subsequent visits as containing laminations. The laminations finding is correct.

Conclusions

False calls of internal corrosion when there actually is none present in a vessel or pipe are a waste of valuable resources. Inclusions are present in steel from the original ingot up until the final product. Inclusions that are flattened out in steel shaping processes such as “rolling” to form plate will become laminations. Laminations and inclusions are not necessarily detrimental to the service life or strength of a part. The fewer inclusions/laminations the client wishes to have in steel, the more expensive the steel is to produce (such as vacuum degassed steel).

If numerous technicians are based in an area, then the cost-efficient option would be to have all the technicians qualified to perform conventional UT, and any areas suspected of having internal corrosion to be reevaluated by a suitably trained and certified phased array technician using phased array zero degree or phased array dual array transducers.

One way of standardizing readings among a team of technicians would be to have a master calibration block with the supervisor in addition to the calibration blocks held by each technician. Once a month, the technicians would measure the supervisor’s block to see whether they are within the company’s tolerance (for example, ±0.1 mm). This could help tremendously when multiple technicians test a component over a period of years to determine the actual corrosion rate for the condition monitoring location trending.

Client competency checks could be upgraded to include 10 samples from actual vessels that are no longer in service that contain a mixture of either laminations, inclusions, and/or internal corrosion, and some with no indications at all. The samples would be cut into 8 in. ´ 8 in. (200 mm ´ 200 mm) pieces and the wall loss side covered with a thin sheet of steel attached with screws so it can be removed later for training. The supervisor should prepare the master answer sheets by using ultrasonic and visual methods with the aid of pipe pit gauges. A phased array technique should also be used to record a permanent image of the deepest pits; this would add more reliability to the master answer sheets and makes good engineering sense. The technician would need to correctly measure the minimum remaining thickness for 8 out of the 10 samples, but if the technician classifies a lamination as internal corrosion or classifies internal corrosion as a lamination on any one of the samples, then this is grounds for further training.

As it can be seen, being able to differentiate between laminations, inclusions, and internal corrosion requires some practice, but with the right guidelines and by always referring your findings to a more experienced technician, fewer false calls will be experienced in the field.

.

_______

.

Author

Kervan Govender: Oceaneering Mobile Workforce, Durban, Kwa-Zulu Natal, 4068, South Africa, govenderkervan@hotmail.com.

References

• ASNT, 2002, Nondestructive Testing Handbook, Vol. 4: Radiographic Testing, third edition, American Society for Nondestructive Testing, Columbus, OH
• ASNT, 2007, Nondestructive Testing Handbook, Vol. 7: Ultrasonic Testing, third edition, American Society for Nondestructive Testing, Columbus, OH
• Drury, J.C., 1997, “Corrosion Monitoring and Thickness Measurement – What Are We Doing Wrong?” (Link no longer available.)
• Pellegrino, Bruce A., and Michael J. Nugent, 2015 (access date), “Nondestructive Testing Technologies and Applications for Detecting, Sizing and Monitoring Corrosion/Erosion Damage in Oil & Gas Assets,” accessed on 15 December 2015 (Link no longer available.)
• Trethewey, K., and J. Chamberlain, 1995, Corrosion for Science and Engineering, 2^nd edition, Longman Scientific & Technical, London, England

.

Acknowledgments

• Thank you to George Niday for allowing me to use the company’s computers and internet access for my research, and Pierre Hershensohn for his invaluable suggestions and help at the right time.
• A special thank-you to Logan Govindasamy, who is one of the great mentors in the corrosion industry.
• And to everyone else who imparted their knowledge to me over the years (you know who you are), there are too many names to mention here.