TABLE OF CONTENTS

Background: Code Requirements: Technique: Mechanics and Software Display: Qualification and Trials: Results: Findings and Conclusions: Acknowledgements:

Background:

Major offshore petroleum projects are being carried out world wide. Drilling platforms used in the exploration are subjected to extensive inspection at the time of construction to ensure the integrity of components. The main load bearing components of such a platform are the "legs" and "piles". These are usually tubular sections welded together with the lower sections being of larger diameter and thicker wall than the upper sections. These tubular sections are constructed from plate rolled into cylindrical shapes with a submerged arc weld forming the long-seam. The "can" that is formed can then be joined to other "cans" again by submerged arc welding of circumferential seams. See Figure 1 and 2.

Traditionally both these seams, long and circumferential, have been tested using radiography. Results of radiography using Cobalt 60 through up to 100mm (4") of steel are dubious at best but not many facilities have the benefit of Linatrons for the better quality X-ray images. Under these conditions using Co-60, the standard 2-2T sensitivity is fuzzy at best. Normal production rates would see about 1 or 2 long seams radiographed in an eight or twelve hour shift.

Figure 1 Platform schematic Figure 2 Illustration of Chevron Offshore Platform Structure in Daewoo Heavy Industries shipyard

A large manufacturing company was awarded the contract to build two such structures in an 18 month period. This would be a significant burden on the radiography department but more important was the logistics of timing and handling. Radiography requires roped off areas to meet radiation regulatory requirements. This means that the very large "cans" would need to be lifted away to a dedicated radiography area or radiography would need to be carried out in the shop during night shift when the access restrictions imposed by roping areas off would have the smallest effect on the production activities. To address this potential "bottle-neck" situation the manufacturer requested that the customer consider mechanised ultrasonic testing (UT) as an option to radiography. The customer agreed to consider the option providing the technique was proven to be as reliable as radiography.
This paper describes the technique adapted and the experiences gathered as ultrasonics eventually replaced radiography and allowed the project to be completed in a timely fashion.

Code Requirements:

NDT on the project was regulated by the dictates of API Offshore Structural Fabrication as interpreted by the customer. In order for ultrasonics to be applied the procedures would need to meet the needs of "Recommended Practice for Ultrasonic Examination of Offshore Structural Fabrication and Guidelines for Qualification of Ultrasonic Technicians" API Recommended Practice 2X (RP2X).

API RP2X required inspection of 10% of all longitudinal seams and 100% of circumferential weld seams. Although API RP2X goes through extensive descriptions of general advantages and limitations of ultrasonics, its main coverage is devoted to Technical Recommendations for manual applications of ultrasonic testing and methods to qualify ultrasonic technicians. This concern for technique and operator ability is based primarily on the traditional application of ultrasonics to the very difficult T,K and Y joint configurations. Application of UT to the simple symmetric butt weld configurations used for this project would have less demand on operator ability to plot indications than in the in complex T,K and Y geometries. However, as with most similar such documents, API RP2X has not made special considerations to the specific aspects relating to mechanised ultrasonic testing. To address the absence of specific guidance in API RP2X both manufacturer and the customer developed a qualification programme to allow the introduction of mechanised ultrasonic testing. Since the effective application of mechanised UT was a first for both parties, considerable debate and co-operation was required in this new venture.

This became a multi-stage programme.

• Develop a procedure and techniques to inspect the range of butt welds proposed for inspection
• Train operators on the use of the equipment and application of the techniques
• Construct a mock-up of a weld with typical defects
• Qualify the equipment and technique on the mock-up
• Move to the field and perform trials of the equipment, techniques and operators to "qualify" the entire concept including operators.
• Correlate ultrasonic and radiographic results to establish a confidence in the system with decreasing dependence on radiography. This would mean moving from 100% radiography and 100% ultrasonic testing to 0% radiography and 100% ultrasonic testing in several stages of reduced radiography.

All efforts were made by both parties to preserve the intent of API RP2X while recognising that strict adherence to the document was not possible due to its lack of information on mechanised UT information. As a result of the extensive efforts to ensure that the results of radiography were the minimum expectations of the system, a significant advance in technology was possible.

To ensure maximum integrity the manufacturer agreed to use the Level A acceptance criteria in the Appendix D of API RP2X. This is based on workmanship quality and is not related to the component's fitness-for-purpose. This criteria would more closely relate to the radiography acceptance length criteria. However, this acceptance criteria bases evaluation response on a 1.5mm diameter side drilled hole at the depth of concern. This applies to a 1 inch thick section the same as a 4 inch thick section. Effectively a 1" (25.4mm) thick section is evaluated to a hole 6% of wall whereas a 4" (102mm) section is evaluated to a hole 1.5% of wall thickness.

This is in contrast to radiography where the IQI varies with the thickness trying to maintain a similar sensitivity with respect to wall thickness. In a 1" (25.4mm) section requiring a #25 penetrameter (ASME source side) the essential hole is the 2T and is 0.050" diameter (1.27mm). This represents about 5% with respect to the wall thickness. For a 2.5" (64mm) wall the source side IQI is the #40. The #40 2T hole is 0.08" diameter (2mm) or about 3% with respect to the wall thickness.

Technique:

Figure 3 Probe Configuration for 67mm Wall Butt Weld Figure 4 Traker being mounted on a tubular girth-weld for scanning Figure 5 Sketch of 40 mm plate Mock-up Figure 6 Scan Report Top/side/end view Figure 7 TOFD Scan Report

The ultrasonic technique developed was a combination of rastered pulse-echo inspection coupled with Time of Flight Diffraction. This ensured that the entire weld volume was inspected with the pulse-echo probes paying special attention to the weld bevel fusion faces and TOFD providing extra coverage for the volumetric defects and "off-angle" planar and volumetric defects not optimally orientated for the pulse-echo probes.

Figure 3 shows a typical set-up. This uses two pair of 45° probes for the full skip, a pair of 45° probes for the half skip and a TOFD pair set for nominal 60° refraction in compression mode. The data for the TOFD pair is collected with the probes symmetric on either side of the weld centreline. For the pulse-echo probes a raster scan is used whereby the probes are moved about 10mm perpendicular to the weld centreline after each linear scan pass and five such scans are made to cover the inner and outer third of the fusion faces.

In addition to the full volume being ultrasonically tested, further NDT using magnetic particle inspection was carried out on both the inside and outside surfaces. This ensured that fine surface defects that might be masked in the unground weld cap could be detected. As well, a manual UT inspection using 0° compression mode was used to ensure no laminations were present in the plate where the mechanised inspection was carried out.

Mechanics and Software Display:

The technique provided the required coverage using three to seven raster scans depending on the wall thickness of the work piece. Scanning was facilitated by placing a magnetic guide band at a fixed distance from the weld centreline and mounting the probe array so that a Hall sensor could follow the band. By means of two sets of motor driven magnetic wheels the "Traker" could be easily located over the magnetic guide strip in position for the scanning. Motorisation of the array allowed probe motion parallel to the weld centreline and a second axis of motion permitted the raster step perpendicular to the weld centreline. In both axes of travel, position was encoded and the ultrasonic output was co-ordinated with the position so as to permit A, B, and C-scan and volumetric presentation of the information.

After the operators had become familiar with the equipment a 4m long long-seam on a 70mm thick section could be inspected and evaluated within 1 hour. This was a significant improvement over radiography.

Qualification and Trials:

Calibration blocks using Side Drilled Holes and slots were made by the manufacturer and used to develop and optimise the techniques. Mock-ups were made by the manufacturer and used to qualify the techniques and operators. The mock-ups were long-seam welds made with imbedded welding defects at various depths and locations. The techniques were confirmed and qualified on the mock-ups. Figure 5 shows a sketch of the mock-up weld used for qualification with the positions of the imbedded defects. With the exception of the porosity, defects were made using ceramic inserts approximately 4mm high by 40-50 mm long. Figure 6 illustrates the processed data from the pulse-echo probes. Customised software allowed the data collected to be presented in a typical "Top/Side/End view" format.

Results:

After the technique had proven it could detect the defects in the mock-up there was a transition from radiography to mechanised UT. This involved inspection of actual production welds with both methods to start. During this transition only all significant indications found by RT were also detected by mechanised UT. As the transition progressed whereby less radiography was being performed, radiography was carried out where indications had been located by mechanised UT. All indications located by mechanised UT were confirmed by manual UT but in many cases they could not even be detected by RT. This did not negate the validity of either NDT method but merely reflects the differences in the methods to detect different types of defects.

Findings and Conclusions:

Laboratory techniques verified on customer mock-ups were used to phase out use of radiography. Further requirements to satisfy the customer's concerns involved site trials. Therefore, several hundred welds tested by mechanised UT were compared with radiography. It was seen that all significant defects found using radiography were also detected using mechanised UT.

It was concluded that mechanised ultrasonics was an effective replacement for radiography and offered a comparable confidence level in probability of detection of all significant defects.

Acknowledgements:

The authors would like to thank Chevron Oil for their help in adapting this new technology and for the image of the Lomba Project Platform in Figure 1 and 2. We would also like to thank RD-Tech and Daewoo Heavy Industries for the images of equipment and the report results.

Authors:

E. Ginzel is a consultant with Materials Research Institute, 368 Lexington Rd., Waterloo, Ontario, Canada N2K 2K2, tel. (519) 886-5071, fax (519)886-8363, E-mail: eginzel@mri.on.ca, Homepage http://www.mri.on.ca

G.Legault is an electrical design engineer with RD Tech, 1200 boul. St-Jean-Baptisite, #120, Québec, Québec, Canada G2E 5E8, tel. (418) 872-1155, fax. (418) 872-5231, E-mail: glegault@rd-tech.com, RDTech on NDTnet.

Y.M. Kim is manager at Dae Dong Engineering Co. Ltd, , 1 Ajoo-Dong, Kojesi, Kyungnam, Korea, tel. (0558) 681-7358-9 fax: (0558) 681-7360, E-mail: xsaint@nownuri.net