Abstract

Much discussion has been given the pros and cons of TOFD in weld inspection. The paper gives a background to the application of mechanised UT to pipeline construction and goes on to demonstrate how TOFD is not an effective stand-alone UT technique for the needs of pipeline inspections where ECA criteria are used but acts as an effective safety net for some critical defects.

Introduction

Both mechanised and manual welding methods are now used on the girth welds in pipeline construction. Ultrasonic techniques are rapidly taking the place of the traditional radiographic techniques that have long dominated but not without some debate. Although TransCanada Pipelines Ltd. has had by far the largest projects and the greatest number of welds tested using mechanised UT, others have also experimented with this technology. In Europe, Gasunie (Netherlands) carried out trials from around 1990 where UT and RT were done on 100% of the welds and in late 1993 they dropped RT and used mechanised UT on 100% of their mainline welds. Now Gasunie uses UT for all welds except for where geometry prohibits and they allow TOFD with additional pulse-echo in the root and cap areas. In Germany, Ruhrgas uses manual UT on manual welding and mechanised UT on mechanised welding. Wintershall on its projects uses RT on 30% of the welds and UT on the remaining 70% reserving the option to use 100% UT when deemed necessary. Similarly the technology is now being applied to off-shore projects. In 1994 Statoil and McDermott both had projects that started off using 100% RT and 100% mechanised UT and eventually both companies dropped RT and maintained 100% mechanised UT.

The methods used have been described by the authors and others (1,2,3,4,5,6). In a recent paper (7) the concept of adding TOFD as an enhancement to the standard pulse-echo techniques was discussed and was deemed a necessary prerequisite for mechanised inspections of manual welding. Field results of the technique incorporating TOFD on manual welds provided excellent correlation between UT and RT, even finding non-fusion defects in the manual welding that were undetected by radiography.

Limitations of TOFD have been covered elsewhere (7,8,9). These include:

1. Poor detection/sizing near the entry surface
2. Requires a second axis of motion to define the side of a weld a defect occurs
3. Optimum Probe Centre Spacing can result in probe interference with weld cap.
4. Mismatch conditions can mask root defects in back wall signal

However, when used with the standard pulse-echo techniques these limitations are minimised and in fact the combined techniques complement each other effectively.

This paper will illustrate examples of how the standard pulse-echo techniques and TOFD compliment one another.

Codes

Radiography has always had acceptance criteria based on workmanship. This is a very conservative assessment of the severity of any imperfection in the weld; however, since RT cannot determine the vertical extent of the defect the conservative treatment is necessary . Basic to the Canadian application of mechanised UT to girth weld inspection is the concept of Engineering Critical Assessment (ECA). This is merely the application of fracture mechanics equations to the weld metal to assess the effect of wall loss on the weld's fitness for purpose. Numerous codes exist that allow this type of calculation and if the inspection method can provide a reasonable assessment of the vertical extent of a flaw then more liberal treatment of the defect can be rationalised. Herein lies one of the greatest advantages of the application of mechanised UT used in combination with an ECA.

BS 4515 Welding of Steel Pipelines on Land and Off Shore references an outdated British Standard BS 3923 "Methods for ultrasonic examination of welds, Part 2: Automatic examination of fusion weld butt joints in ferritic steels, 1972. BS 3923 is out of date and quite general but BS 4515 allows the contracting parties to develop an acceptance criteria based on ECA.

Two of the most commonly referenced Codes in North America are the American Petroleum Institute's API 1104, and the Canadian Standards Association CSA Z-662. In API 1104 provision is made to calculate acceptance criteria using equations in the Appendix Section 7.2. Similarly Appendix K of CSA Z-662 contains a section called Determination of Maximum Acceptable Sizes of Imperfections.

API 1104 makes provision to use Ultrasonic inspection but then requires that the inspection procedure conform to ASTM E 164 which is a Standard Practise for Contact Ultrasonic Testing using manual techniques. Trying to apply API 1104 inspection requirements makes ultrasonics a poor production inspection method unless the provisions of the API 1104 Appendix - "Alternative Acceptance Standards for Girth Welds" are used, but even this requires approval to deviate from the ASTM-E-164 requirement.

CSA Z-662 has incorporated the lessons of a decade of experience in mechanised ultrasonic inspections. When mechanised UT is opted by the contracting parties, Z-662 incorporates specific requirements that must be incorporated by the inspection systems. This simplifies the application of the ECA as the inspection system requirements and capabilities are considered in the development of the acceptance criteria.

Apart from CSA Z-662 there exist few Codes that assist in developing a system that encompasses the needs of production welding and the special considerations imposed by an ECA.

A similar lack of guidance exists when we look for specifics on the application of TOFD. Despite the fact that the technique has been around since the 1970's, code agencies have been slow to incorporate useful guidelines for its application. Examples of codes where the TOFD technique is described include:

1. BS 7706: Guide to Calibration and setting-up of the Ultrasonic Time of Flight diffraction (TOFD) technique for defect detection, location and sizing of flaws. 1993
2. prEN 583 Ultrasonic Examination - Part 6: Time of Flight Diffraction Technique as a Method for Defect Detection and Sizing

But these are guides to the general application of TOFD and provide no acceptance criteria rules. For acceptance criteria it would seem that company specifications are all that can be looked to. When an ECA based acceptance criteria is available TOFD can be used as an aid to characterisation and allow more precise sizing as well as a safety net to detect flaws that are not well orientated to be detected by the pulse-echo probes yet seen to be of a serious nature.

Note: There is a draft for Acceptance Criteria for TOFD based on workmanship and is mentioned in Insight (April 1997 vol. 39 #4) by F.Dijkstra et al.

TOFD Principles

Figure 1: TOFD Theory

The TOFD technique uses a single probe pair in a transmitter-receiver arrangement (Fig. 1). Usually compression probes are used with a refracted angle of between 45° to 70° for the main compression mode wave. The diffracted signals are received via the receiver probe and evaluated using an ultrasonic imaging system to create B-scan images.
When ultrasound is incident on a linear discontinuity, diffraction takes place at its extremities in addition to the normal reflected wave. This diffracted energy is emitted over a wide angular range and is assumed to originate at the extremities of the flaw. This is significantly different from conventional ultrasonics, which relies on the amount of energy reflected by discontinuities.

In addition to energies diffracted by defects, the TOFD method will also detect a surface (lateral) wave traveling directly between the probes and a back wall echo from energies that reach the back of the test piece without interference from defects. The study of this phenomenon has led to the use of the time of flight diffraction method which is applied for flaw detection and flaw sizing.

Presentation of data is usually by means of a B-Scan. B-scans used in this discussion are in the traditional sense of the term whereby a two dimensional view of a cross section of the test piece is presented with one axis representing probe motion and the other axis representing depth showing front and back surfaces and flaws in between. Some consider two kinds of B-scan images based on the scan direction of probes with respect to a weld axis (but this would apply only to weld inspections);

1. Normal to the direction of the beam along a weld or flaw. That B-scan result is used to locate and size flaws. It is also known as linear, non-parallel or longitudinal scan. In literature this scan is often called D-scan but this is not consistent with the EU standard (8) prEN 1330-4 "Terms used in ultrasonic testing" wherein D-scan has another definition. Hence the EU standard draft for TOFD does not use the term "D-scan" in respect to TOFD.

2. In the direction of the beam transversely across a weld or flaw. That B-scan result is used to size flaws. It is also known as transverse, parallel or lateral scan.

Normally reference is made using the lateral wave response where depth of indications is calculated from the time of flight difference between the lateral wave and the diffracted pulse. The assumption that the flaw is positioned symmetrically between the probes introduces an error but this usually has little effect on the accuracy of the estimated flaw depth.

For more information see: TOFD in UT Online Journal 09/97 (NDTnet)

Pipeline Welding Defects

Mechanised and manual weld process defect differences are not too numerous for the most common defects. In Mechanised welding the main defect of concern is lack of fusion (LoF) on the bevel sidewall. Although LoF can occur in manual welding the occurrence of volumetric defects is more common, these would include slag, hollow bead and porosity. In both processes the ability to discriminate between true defects and geometries is crucial. The difference between a lack of fusion on the root bevel and the root surface geometry in a mismatch geometry might be only 1-2 mm. Any mechanised UT system must be able to detect the difference between these two situations. Centreline cracking is a trait associated with mechanised welding whereas under-bead cracking and root cracks are more likely to be seen in manual welding.

The table below indicates some of the more common indications and the weld process they are associated with. Also indicated is a very approximate probability of detection (POD) with the various NDT methods used in pipeline weld inspections.

Imperfection	Mechanised Welding	Manual Welding	POD Xray %	POD Pulse Echo (map) %	POD Strip-chart (TOF & amplitude) %	POD TOFD %
Lack of Fusion Surface	Yes	Yes	90	95	95	OD 0 ID 60
Lack of Fusion Subsurface	Yes	Yes	50	100	100	100
Lack of Fusion interpass	Yes	Yes	0	50	30	100
Slag	No	Yes	100	90	90	100
Porosity (>5%)	Yes	Yes	100	95	50	95
Undercut	Yes	Yes	100	90	90	ID 50 OD 0
Misfire	Yes	No	100	100	100	55
Lack of Cross Penetration	Yes	Yes	90	75	75	100
Incomplete Penetration	No	Yes	100	100	100	75
Hollow Bead	No	Yes	100	25	25	0
Centreline crack	Yes	No	75	75	75	100
Under-bead crack	No	Yes	60	100	100	100
Transverse crack	No	Yes	75	75	75	0
Mismatch (high-low)	Yes	Yes	60	10	10	100
Root Bead misalignment	Yes	No	50	100	100	50
Burn through	Yes	Yes	100	50	30	100

Notes:

1. Defects detected with X-ray do not reveal depth position or vertical extent.
2. Defects detected with UT are compared to calibration reflectors and therefore through wall thickness dimensions and position.
3. The percentages given are not intended to be exact values but merely estimates to show the POD due to the physical limitations of the various methods.

TOFD vs. Amplitude & Time-of-Flight Data

To illustrate the ease with which these defects can be identified the following examples are provided. To explain the stripchart presentation we have provided the graphic display of how the gated output is represented as a point on the chart.

The Chart

Figure 2: Data Displays

Figure 3: Gate Positioning for Pulse-Echo

Operators can quickly review the stripcharts for all channels (zones) and assess the signals for nature and acceptability. A typical series of steps assesses the following:

For example:


Nonfusion Root: Several variations can occur. Misfire: The internal welding head did not fire or sputtered. No metal is deposited. Ideally this presents 2 smooth root faces, however, welders have been known to see this from the outside and the Hot Pass bug can be run over the area twice. This can cause some metal to penetrate and reduce the surface area of nonfused root face.	Operator Evaluation Checklist Characteristic ------------------- Comment
	Indication over threshold Channel(s) affected Symmetry (US & DS) Transit Time Length	Yes Root & LCP Yes At calibration target distance and smooth Acceptable/Rejectable
Nonfusion Root: Missed edge: Due to misalignment of the internal head or high low conditions, one side of the root bevel may not get metal deposited on it. This was called LFS on RT reports.	Operator Evaluation Checklist Characteristic ------------------- Comment
	Indication over threshold Channel(s) affected Symmetry (US & DS) Transit Time Length	Yes Root only No At calibration target distance and smooth Acceptable/Rejectable
Nonfusion Fill 1: The sources of this defect are the same as for any nonfusion defect in the fill passes. Fill 1 nonfusion is often associated with the corner where the hot pass bevel and the fill bevel meets. This would be called LFSS on the radiographic report.	Operator Evaluation Checklist Characteristic ------------------- Comment
	Indication over threshold Channel(s) affected Symmetry (US & DS) Transit Time Length	Yes Fill 1 No At calibration target distance Acceptable/Rejectable

Illustrations of defects

TOFD, B-scan/Mapped data and Amplitude -ToF Strip charts
The following Figures illustrate how each of the presentation formats has a place in the evaluation of indications for rapid and accurate assessment of welds.

Figure 4: Burn Through prior to MisFire
Burn Through prior to MisFire Misfire seen well on Pulse-Echo and TOFD but B-scan/mapping shows the subtle travel time arc back thereby providing the best evidence for the Burn Through.

Figure 5: Manual weld containing heavy porosity in cap.
Manual weld containing heavy porosity in cap. Seen easily on TOFD but not notable on Pulse-Echo strip chart display.

Figure 6: A 16mm wall Manual weld with Slag on the upstream side.
A 16mm wall Manual weld with Slag on the upstream side. Also, seen is Mismatch: US (upstream) short travel time (low side) and DS (downstream) longer time (high side) seen in root time gate. TOFD can see slag but cannot be sure how close to OD (Pulse-Echo indicates it is within 1mm of OD)

Figure 7: Crack seen with both TOFD and Pulse-Echo
Crack seen with both TOFD and Pulse-Echo Irregular travel time on Pulse-Echo and loss of Back wall (and subsequent loss of shear/comp.) signal in TOFD indicates it is ID surface breaking. Based on amplitude exceeding a threshold over a minimum length this defect could be acceptable without TOFD information to aid in evaluation.

Summary

All techniques presently used provide a relatively high P.O.D When used in combination the techniques complement each other very well. Generally;
• X-radiography scores low on planar defects not well aligned with the beam (such as for nonfusion and cracks).
• Strip-chart (peak amplitude & TOF) scores low on hollow bead, small porosity, unacceptable mismatch and burnthrough.
• Raw data (B-scans or mapping) ) scores low on hollow bead, small porosity, unacceptable mismatch and burnthrough.
• TOFD scores low on transverse cracks and defects near the OD surface

  A combination of all NDT methods would provide the highest P.O.D. but this is a costly option as two separate systems must be used. The best compromise is obtained from mechanised ultrasonics where pulse-echo and TOFD results can be combined in a single system.

Acknowledgements:

The authors would like to thank Weldsonix International Inc. (Canada) for the images provided.
The authors would also like to thank TransCanada Pipelines Ltd. for their ongoing efforts to promote ultrasonic inspection in the pipeline industry.

References

1. E.Ginzel & R.Ginzel, B.Gross, M.Hoff, P.Manuel, Developments in Ultrasonic Inspection for Total Inspection of Pipeline Girth Welds, 8th Symposium on Pipeline Research, Houston, Texas, August 1993
2. Glover et al, Inspection and Assessment of Mechanized Pipeline Girth Welds, Proceedings: Weldtech 88, London, UK, 1988
3. E.Ginzel, Further Developments In Ultrasonic Inspection of Pipeline Girth Welds, UT Online Journal, http://www.ultrasonic.de, 1996
4. B.Gross et al, Pipeline Technology Conference, Ostende, Belgium, 1990
5. J.A.de Raad, R. van Agthoven, Mechanical Ultrasonic Test Systems for Pipelines Welds, Proceedings: International Conference on Pipeline Inspection, Edmonton, Canada, 1983
6. J.A.de Raad, High Speed Ultrasonic Inspection of Field Girth Welds During Pipeline Construction, Pipeline Technology Conference, Ostende, Belgium, 1990
7. E.Ginzel, P. den Boer, Merv Hoff, Application of Mechanised Ultrasonic Inspection to Manually Welded Pipeline Girth Welds, UT Online Journal, http://www.ndt.net, Vol. 2 No. 5 (May 1997)
8. The British TOFD standard BS 7706: Guide to Calibration and setting-up of the Ultrasonic Time of Flight diffraction (TOFD) technique for defect detection, location and sizing of flaws. Review NDTnet Vol. 2 No. 9 (Sep 1997)

Authors

Merv Hoff is the senior NDT coordinator for TransCanada Pipelines. He can be contacted at TransCanada Pipelines Ltd. 111 5th Avenue SE Calgary, Alberta Canada T2P 4K5 tel. (403) 267-6401, fax (403) 267-6242 email merv_hoff@tcpl.ca

Henk van Dijk is manager of Automated UT Inspections for WeldSonix International Inc. He can be contacted at WeldSonix International Inc. 2507 84th Avenue Edmonton, Alberta Canada T6P 1K1 tel. (403) 417-3114, fax (403) 417-1185 email Henk@Weldsonix.com

Ed Ginzel is an independent consultant with the Materials Research Institute. He can be contacted at Materials Research Institute 368 Lexington Road Waterloo, Ontario Canada N2K 2K2 tel. (519) 886-5071, fax (519) 886-8363 email mailto:eginzel@mri.on.ca Homepage http://www.mri.on.ca Materials Research Institute on NDTnet