In a Nutshell:

With time comes experience, and success comes from experience. In the case of tank bottom inspection by means of Magnetic Flux Leakage (MFL), ROSEN can look back at more than 25 years of experience. The equipment and technologies used for such inspections have, of course, been improved and updated frequently. As of last year, the experts from our Non-Destructive Testing business line launched the new Tank Bottom Inspection Technology (TBIT) Ultra service. With its advanced technology, TBIT Ultra is pushing the detectability and accuracy of tank bottom inspection data to a new, never-before-reached level.


TBIT Ultra

The storage of hydrocarbon or chemical liquids in above-ground storage tanks is an important element in the energy supply chain. A major concern for the integrity of a storage tank is the condition of the tank bottom, with corrosion being one of the most common threats.

Early detection of the presence or absence of any small defect indications in tank bottoms enables operators to obtain more-precise corrosion-rate determinations. Precise data allows for extended inspection intervals because it enables the repair of smaller defects. Reliable detection and prompt assessment of features of time-dependent damage mechanisms, including corrosion, are essential for confidence in inspection findings and the determination of the next inspection date (NID).

CAPACITIES OF TBIT ULTRA

Through sensitive detection and diagnosis, our TBIT Ultra service is capable of accurately sizing even the smallest metal loss defects, such as pinholes or signs of microbiologically induced corrosion, helping to optimize storage tank inspection programs.

Key advantages include a high sensitivity, facilitating the early detection of small features at a 10% metal loss threshold, as well as automatic defect sizing through signal-based analysis, delivering faster and more reliable results than manual verifications. The accuracy of the automatic sizing capabilities is so high, it exceeds the real-world accuracy of Ultrasonic Testing (UT) validation in the tank and includes automated internal-external discrimination. The defect dimensions are calculated instantly. During the scan (the data is stored online), the software automatically detects and classifies the feature and calculates the metal loss right away. Fast calculation makes it possible for the field technician to see the feature on the laptop screen – which is mounted on the scanner – in real-time in order to take immediate action if needed.

Additionally, the tool is equipped with 360-degree steering so it can inspect annular plates, sketch plates and areas close to the plate welds to maximize coverage.

It is not necessary to remove any coating for the bottom inspection and verification of inspection results, as the instrument can inspect right through it (for coatings 0.24 inches/6 millimeters). Additionally, we do not need to follow up on all indications. That said, we still need to perform a limited number of feature verifications to ensure that our inspection results are within specifications for the given tank. The coating thickness is measured and collected on a 1-millimeter (0.039-inch) sample and can be displayed in the report. This enables the customer to visualize the coating thickness distributed over the entire floor scan. The technology is suitable for use in restricted areas, under piping and heating coils, and close to plate weld and shell and annular plates (up to .63 inches/16 millimeters). A single unit can inspect up to 328 square feet/100 square meters of tank bottom per hour. For larger tank floors, multiple inspection tools can be utilized, and the data can be merged into one report, significantly increasing the inspection speed.

TBIT ULTRA FIELD EXPERIENCE

A few months after launching our new TBIT Ultra service, we would like to present some field experiences we have collected up to this point and share our most recent results. The chosen inspection results from a bottom scan using TBIT Ultra represent a typical tank bottom, making the results comparable to a large number of tank bottoms.

In this case, with a diameter of nearly 15 meters/50 feet, the inspected tank bottom was equipped with none-lined bottom plates 8 millimeters (0.314 inches) thick. The previously performed cleaning of the bottom was acceptable according to the ROSEN document “Cleaning requirements for MFL Inspection by means for the ROSEN TBIT,” which is similar to the document SSPC-SP 12/NACE 5 for a visual surface preparation of WJ-1, WJ-2 or WJ-3.

The surface was very representative for a non-lined bottom with a certain level of internal roughness and internal pitting present (see Figure 1).

Figure 1 – A tank bottom with internal roughness and pitting

Figure 1 – A tank bottom with internal roughness and pitting

DELIVERING THE DATA

This tank bottom was scanned within one day, and ROSEN experts crosschecked the proper functioning of the algorithm specific to the conditions of the storage tank. The customer had instant access to integrity-relevant data and clear recommendations on mitigation and repair right after the inspection.

The results were handed over to the client by means of the free-of-charge software ROSOFT for Tanks. In this software, each individual feature can be looked at individually, even years after the inspection, thanks to the pre-determined coordinate system.

The ROSOFT for Tanks data management and planning software also lets the user define required repair patches, easily manipulating their placement and location using drag-and-drop functionality, and to create multiple repair scenarios and strategies. Additionally, the software allows the display of the estimated weld length, the patch area and coating/liftoff, and it includes continuous updating of anomaly statistics during the “virtual repair” process.

Figure 2 – Overview of the floor map

Figure 2 – Overview of the floor map

THE FINDINGS

Each colored dot in the general overview of the floor map represents a single metal loss feature (see Figure 2). In total, our TBIT Ultra tool detected 2,219 features during the inspection, which was in line with expectations based on the condition of the asset and the inspection interval. No hole was found – just one single feature with 80% metal loss.

In general, bigger indications are quite easy to detect during a tank bottom inspection, because the signal-to-noise level is typically high, meaning the signal clearly “sticks out” from the noise level. With smaller indications, we are not talking about smaller indications in depth but in length and width. With the new TBIT Ultra technology, ROSEN is able to reliably find even difficult-to-detect indications, so-called pinholes.

Another equally, or even more, important question we have to ask ourselves in this context is: How accurate are the reported feature depths? In order to find an answer to this question, we performed quite a large number of feature verifications (in this case around 60). On a tank bottom like this, the standard procedure only requires 40 feature verifications. Thus, 20 more verifications have been performed to confidently determine the accuracy of the metal loss values in this case as well as the tool’s performance.

VARIOUS DEPTH RANGES

Since metal loss sizing is not linear over the depth-diameter ratio, we grouped feature verification into three depth ranges:

  1. 10% to < 20% metal loss features
  2. 20% to < 50% metal loss features
  3. 50% to 100% metal loss features

Figure 3 – Details of the feature verification. a) Metal loss range 10% to < 20%; b) Metal loss range 20% to < 50%; c) Metal loss range 50% to 100%

Figure 3 – Details of the feature verification. a) Metal loss range 10% to < 20%; b) Metal loss range 20% to < 50%; c) Metal loss range 50% to 100%

Above, you can see the details of the feature verification. With a confidence factor of 85%, we reached the following accuracy (vs. UT measurements) in each of the metal loss ranges:

  1. 10% to < 20% an accuracy of ± 5%
  2. 20% to < 50% an accuracy of ± 6%
  3. 50% to 100% an accuracy of ± 6%

The aforementioned accuracy values tie into the repair strategy for the asset. The higher the accuracy frame, the more conservative the repair threshold needs to be. A smaller accuracy value allows for an efficient repair strategy and may result in avoiding unnecessary repairs and waste of resources.

CONCLUSION

In the case presented, the customer was able to directly benefit from the accuracy of the collected data since the inspection went much faster than anticipated, along with the proof that we were able to deliver the promised highly accurate results. With the help of our state-of-the-art TBIT Ultra technology, we size anomalies automatically with the same or better accuracy as using UT or pit measurements, but we do it rapidly and in an automated manner. In this specific case, around 40 feature verifications were supposed to be performed. Calculating eight minutes per feature verification, that would have added up to 5.5 hours spent for feature verifications if the calculation had been done manually. The tank in this case contains more than 2,200 features; thus, verifying them all manually would have taken our experts more than 293 hours, and the manual verifications would still have involved the human factor. Completely removing the human factor leads to fewer mistakes and a smoother performance overall.

With this inspection data and the comparison with the feature verifications, we can conclude that the TBIT Ultra service meets performance specifications not only on machined reference/calibration plates but also on corroded bottom plates.