AUT Handbook Extracts.en

Automated Ultrasonic Testing for Pipeline Girth Welds A Handbook

E. A. Ginzel

Automated Ultrasonic Testing for Pipeline Girth Welds: A Handbook Series coordinator: Noël Dubé Technical reviewer and adviser: Dr. Michael D. C. Moles (Olympus NDT) Layout, graphics, editing, proofreading, and indexing: Technical Communications Service, Olympus NDT Published by: Olympus NDT, 48 Woerd Avenue, Waltham, MA 02453, USA Marketing and distribution: Olympus NDT This guideline and the products and programs it describes are protected by the Copyright Act of Canada, by laws of other countries, and by international treaties, and therefore may not be reproduced in whole or in part, whether for sale or not, without the prior written consent from Material Research Institute. Under copyright law, copying includes translation into another language or format. The information contained in this document is subject to change or revision without notice. R/D Tech part number: DUMG070A © 2006 by Olympus NDT All rights reserved. Published 2006. Printed in Canada ISBN 0-9735933-2-6 Notice To the best of our knowledge, the information in this publication is accurate; however, the Publisher does not assume any responsibility or liability for the accuracy or completeness of, or consequences arising from, such information. This book is intended for informational purposes only. Final determination of the suitability of any information or product for use contemplated by any user, and the manner of that use, is the sole responsibility of the user. The Publisher recommends that anyone intending to rely on any recommendation of materials or procedures mentioned in this publication should satisfy himself or herself as to such suitability, and that he or she can meet all applicable safety and health standards. Trademarks Olympus and the Olympus logo are registered trademarks of Olympus Corporation. R/D Tech, the R/D Tech logo, OmniScan, and PipeWIZARD are registered trademarks, and “Innovation in NDT” and Tomoscan are trademarks of Olympus NDT Corporation in Canada, the United States, and/or other countries. CANDU (CANada Deuterium Uranium) is a registered trademark of Atomic Energy of Canada Limited (AECL). Cycolac is a trademark of Marbon Chemical Corp. Ethernet is a trademark of Xerox Corporation. Lexan is a registered trademark of General Electric Company. Lucite is a registered trademark of E.I. DuPont Co. Microsoft, MS-DOS, Windows, and the Windows logo are registered trademarks of Microsoft Corporation in the United States and/or other countries. Polysulfone is a trademark of Union Carbide Corp. Profax is a trademark of Hercules, Inc. Rexolite is a registered trademark of C-Lec Plastics Company. All other product names mentioned in this book may be trademarks or registered trademarks of their respective owners and are hereby acknowledged.

Table of Contents

Preface .................................................................................................... xi Introduction ............................................................................................ 1 1. History — Early Days ...................................................................... 3 1.1 Early Ideas ......................................................................................................... 3 1.2 The Introduction of Mechanised Welding .................................................... 5 1.3 UT Adaptations to the Mechanised Welding ............................................... 6 1.4 Probe Design Changes .................................................................................... 8 1.5 Mechanised UT Enters the Computer Age ................................................ 10 1.6 Standardising Concepts ................................................................................ 16 References to Chapter 1 ........................................................................................ 17

2. Zones—How to Keep Them Apart ............................................. 19 2.1 Principles of the Zonal Technique ............................................................... 2.2 Beam Control .................................................................................................. 2.2.1 Beam Spot Size ..................................................................................... 2.2.2 Beam Angle .......................................................................................... References to Chapter 2 ........................................................................................

20 22 23 30 48

3. Beam Angles—Designing a Technique ..................................... 49 3.1 Defining Fusion Zones and Volumetric Zones .......................................... 3.2 Volumetric Detection ..................................................................................... 3.3 Some Transverse Ideas .................................................................................. 3.4 Adding TOFD ................................................................................................. 3.5 SMAW and SAW Needing Vertical Targets ............................................... References to Chapter 3 ........................................................................................

50 66 69 73 74 77

4. Calibration—How to Set Up the Targets ................................... 79 4.1 Block Drawings and What the Machinist Needs to Know ...................... 4.2 Acoustic Quality Check ................................................................................. 4.3 Surface Condition .......................................................................................... References to Chapter 4 ........................................................................................

80 87 89 96

Table of Contents

v

5. System Requirements—Equipment and Displays .................. 97 5.1 Scanner Head .................................................................................................. 98 5.2 Data Acquisition Units ................................................................................ 101 5.3 Displays ......................................................................................................... 108 5.4 Records .......................................................................................................... 116 References to Chapter 5 ...................................................................................... 118

6. Making a Good Calibration ....................................................... 119 6.1 A Review of Targets ..................................................................................... 6.1.1 Zonal Targets ..................................................................................... 6.1.2 Volumetric Targets ............................................................................ 6.1.3 Transverse and TOFD Notches ....................................................... 6.1.4 Various Calibration Aspects ............................................................ 6.2 Limitations on Calibration Quality ........................................................... 6.3 Characteristics of a Good Calibration ....................................................... References to Chapter 6 ......................................................................................

120 120 123 127 127 137 139 140

7. Project Preparation and Pre-Project Planning ........................ 141 References to Chapter 7 ...................................................................................... 148

8. Qualification—Systems, Operators, and Techniques ........... 149 8.1 What Is Required and Is It Reasonable? ................................................... 8.1.1 Testing at Elevated Temperature ..................................................... 8.1.2 Repeatability with Guide Band Movement ................................... 8.1.3 Accuracy in Circumferential Positioning/Location ...................... 8.1.4 Repeatability with Scanner Orientation (Upside Down) ............ 8.1.5 Detection Level on an Unwelded Bevel ......................................... 8.1.6 Coupling Check—Alarm Levels Verified (TR and PE) ................ 8.1.7 Repeatability with Scan Direction .................................................. 8.1.8 Clearance Test .................................................................................... 8.1.9 Probability of Detection and Sizing ................................................ 8.1.10 Accuracy in Height and Length Sizing .......................................... 8.1.11 Other Equipment Tests ..................................................................... 8.2 Operator Qualifications ............................................................................... 8.3 In Summary .................................................................................................. References to Chapter 8 ......................................................................................

153 153 155 156 157 157 158 159 161 161 163 164 165 166 168

9. Interpretation ................................................................................ 169 9.1 What Is Evaluated? ...................................................................................... 9.2 Evaluation Thresholds ................................................................................. 9.3 Flaws .............................................................................................................. 9.3.1 Misfire ................................................................................................. 9.3.2 Undercut ............................................................................................. 9.3.3 Shrinkage Cracks ............................................................................... vi

Table of Contents

169 173 174 175 175 176

9.3.4 Porosity ............................................................................................... 9.3.5 Cold Lap ............................................................................................. 9.4 Geometry ....................................................................................................... 9.4.1 Root ..................................................................................................... 9.4.2 Cap Geometry .................................................................................... 9.5 False or Overtrace Signals. What are they? .............................................. 9.6 Examples of AUT Indications .................................................................... 9.7 Flaws .............................................................................................................. 9.7.1 Lack of Fusion Root .......................................................................... 9.7.2 Lack of Fusion Midwall .................................................................... 9.7.3 Lack of Fusion Cap ........................................................................... 9.7.4 Centreline Crack ................................................................................ 9.7.5 Root Porosity ...................................................................................... 9.7.6 Fill Porosity ........................................................................................ 9.7.7 Burn Through .................................................................................... 9.7.8 Geometry: High-Low ........................................................................ 9.7.9 Cap Geometry .................................................................................... 9.7.10 False Signals ....................................................................................... 9.8 Special Flaws ................................................................................................. 9.8.1 Transverse Cracks ............................................................................. 9.8.2 Copper ................................................................................................ 9.8.3 General ................................................................................................ 9.9 What to Do with Volumetric and TOFD Data ......................................... 9.10 Computer-Assisted Positioning ................................................................ References to Chapter 9 ......................................................................................

176 177 178 179 182 184 186 187 187 188 189 190 191 192 194 195 198 199 201 201 202 202 203 204 206

10. Acceptance Criteria ...................................................................... 207 10.1 Workmanship ............................................................................................... 207 10.1.1 API 1104 .............................................................................................. 209 10.1.2 AS 2885.2 ............................................................................................ 210 10.1.3 CSA Z662 ............................................................................................ 212 10.1.4 ISO 13847 ............................................................................................ 213 10.1.5 DNV OS F101 ..................................................................................... 214 10.1.6 General Considerations in Workmanship Acceptance Criteria for AUT ..................................................................................................... 217 10.2 ECA-Based Acceptance Criteria ................................................................ 218 10.2.1 The Principles of Engineering Critical Assessment ..................... 219 10.2.2 API and BS Rules .............................................................................. 226 10.2.3 Generic Acceptance Criteria ............................................................ 228 References to Chapter 10 .................................................................................... 231

11. AUT Sizing and PODs ................................................................ 233 11.1 Sizing “Techniques” .................................................................................... 235 11.2 Amplitude Techniques ............................................................................... 235

Table of Contents

vii

11.2.1 Simple Zone Sizing ........................................................................... 11.2.2 Amplitude Corrected Zone Sizing ................................................. 11.2.3 Amplitude Corrected Zone Sizing with Overtrace Allowance .. 11.2.4 Amplitude Comparisons with Zone Characterisation ................ 11.2.5 Beam Boundary by dB Drop ........................................................... 11.3 Temporal Techniques .................................................................................. 11.3.1 TOFD ................................................................................................... 11.3.2 Backscatter ......................................................................................... 11.4 Amplitude Sizing Comparison ................................................................. 11.5 POD and POS ............................................................................................... 11.6 Statistical Techniques .................................................................................. 11.7 Precision and Accuracy .............................................................................. 11.8 Summary ...................................................................................................... References to Chapter 11 ....................................................................................

235 236 237 238 240 243 244 245 246 249 250 255 262 263

12. Great Expectations and Mythologies ....................................... 265 12.1 Procedure Policies ....................................................................................... 12.2 Audit Guidelines ......................................................................................... 12.3 The Operator’s Log ..................................................................................... 12.4 Wall Thickness Variation Effects ............................................................... 12.5 AUT Myths ................................................................................................... References to Chapter 12 ....................................................................................

265 268 271 272 274 284

13. Applications .................................................................................. 285 13.1 Pipeline Land Lay ....................................................................................... 13.2 Offshore Projects ......................................................................................... 13.3 Offshore Spoolbase ..................................................................................... 13.4 Dual-Product Longitudinal Seam ............................................................. 13.5 Non-Pipeline Applications of the Zonal Technique ............................... 13.6 Future of AUT .............................................................................................. 13.7 Phased Arrays .............................................................................................. 13.7.1 Seamless Pipe ..................................................................................... 13.7.2 Inspections on Thick Section Risers ............................................... 13.7.3 1.5D Arrays for Improved Sizing in Risers and Tendons ........... 13.7.4 Focusing for Small Diameter Pipes ................................................ 13.7.5 Short Cutback .................................................................................... 13.7.6 Clad Pipe ............................................................................................ 13.7.7 Portable Phased Array Units ........................................................... 13.8 Conclusions .................................................................................................. References to Chapter 13 ....................................................................................

285 289 292 294 295 296 297 297 298 300 301 302 303 304 305 307

Appendix A: Beam Size Tables ....................................................... 309

viii

Table of Contents

Appendix B: Beam Size Approximations from Calibration  Scans  ............................................................................................... 313 Appendix C: Suggested Techniques ............................................... 317 C.1 CRC Bevel ......................................................................................................  317 C.2 CRC Techniques  ...........................................................................................  318 C.3 SMAW Single Vee Bevel of 60° (Manual Welding)  .................................  324

Appendix D: Statistical Concepts  ................................................... 329 D.1 Comparing Samples  ....................................................................................  335 D.2 The Significance of Significance .................................................................  336 D.3 A Tail of a Curve  ..........................................................................................  337 D.4 Probability .....................................................................................................  340 Bibliography of Appendix D: “Statistical Concepts” .....................................  342

Glossary of Terms and Abbreviations  ........................................... 343 List of Figures  ..................................................................................... 349 List of Tables  ....................................................................................... 355 Index  ..................................................................................................... 357

Table of Contents    ix

Preface

In 1977, the company in which I was a partner, was approached by a company from the Netherlands to see if a Canadian company would be interested in setting up with them to address a potential nondestructive inspection market being created by a new high-impact welding process. The welding process was to use a belled flange arrangement that would be fused using an explosive charge. The resultant weld would be a laminar fusion face for which radiography would not be able to detect any flaws. The inspection system looked like it had a great potential for several reasons not the least of which was the fact that it would have a niche market irreplaceable by radiography. However, the welding process proved to have an unacceptable side effect. It was prone to knocking dishes off shelves and cracking windows of nearby houses (within a kilometre or two). The original plans we had to get into the pipeline girth weld inspection and replace radiography did not mature at that time. But this did not stop progress in welding development and eventually a very different process was developed that was again found to be more effectively tested using ultrasonics. But by that time I had moved on and my company was no longer involved in pursuing potential pipeline inspection services. In 1989 I had my next exposure to the infant automated ultrasonic testing (AUT) of pipeline industry when asked to look at an application of mechanized inspection that was designed for ultrasonic inspection of pipeline girth welds. This came as a request from TransCanada Pipelines Ltd. (TCPL) to look into concerns for disagreement between radiography and ultrasonic results. My function was, as an auditor and researcher, to identify problems and make recommendations that would assist the engineering department to assess the efficacy of the ultrasonic testing being provided. This analysis was allowed one year to identify and correct any problems or the research to replace radiography would be terminated. Over the next 12 months, I learned a lot about pipeline AUT as it has become known. During that time, in addition to identifying problems with the AUT process, I also identified problems with the assessment of results. Like most others in this, and other industries, there Preface

xi

was a pre-conceived notion that ultrasonics was supposed to provide identical results to radiography in order to replace it. Thanks to the understanding of two people in particular in TCPL, we managed to move forward and build the AUT into the worldwide success it is today. I would like to take this opportunity to thank Mr. Merv Hoff and Mr. Ben Gross for the confidence they showed in my decisions and the support they provided. Many people had great ideas and innovations before and since that time, but had the decision been made to stop AUT in 1990, it is doubtful that the progress we have made since then would have been possible. E. A. Ginzel December 2005

xii

Preface

Introduction

Since 1993, mechanised ultrasonic examination of girth welds using zonal discrimination with focused probes has replaced traditional radiography on all major pipeline construction projects in Canada. Origins of the concepts date back to 1972. At that time, the Alberta Gas Trunk Line (later called NOVA Corporation and subsequently TransCanada Pipelines) began to investigate the feasibility of using mechanised welding techniques to replace the traditional manual shielded metal arc welding process. The increasing rate of demand for natural gas has meant an ever increasing pipeline expansion programme. The mechanised welding technique evaluated by NOVA involved complex weld bevel geometry. An internal welding machine was used to deposit a root pass, and then welding “bugs” moving along guide tracks clamped to the outside of the pipe were used to deposit the hot pass and fill passes. When evaluating the welding process NOVA’s engineers determined that the 37.5° root bevel and the 45° hot pass bevel were poorly inspected by the traditional X-ray radiography, which relies on defect orientation parallel to the beam to permit detection. To address this shortcoming, NOVA initiated a programme to develop a mechanised ultrasonic inspection system. RTD of Rotterdam was invited to carry out the first field tests in 1977 and 1978. Results were encouraging and the NOVA recommendations indicated that the complimentary nature of ultrasonic testing and radiography would make it beneficial to use ultrasonics to supplement radiography. In the 1978 report, it was predicted that, “Ultimately, with adequate development and proven performance, ultrasonic testing might someday replace radiography as the standard inspection technique.” In 1982, TransCanada Pipelines Limited, one of the world’s largest gas transmission companies, joined the development programme. They experimented in both summer and winter conditions with mechanised ultrasonic testing to inspect over 3000 welds made using the mechanised welding process in their initial evaluation. This allowed for a much-needed, large-scale statistical evaluation of the technology. By the mid 1980s, ultrasonic beam spot sizes were consistently held to around 2 mm in diameter at the area of interest. False repair calls that were caused by beam edges Introduction

1

interacting with weld surface geometries were virtually eliminated. This allowed a new philosophy to be considered: engineering critical assessment (ECA). The ECA concept uses the principles of fracture mechanics to assess the severity of a defect based on its vertical extent. The small spot sizes now achieved allowed the weld to be divided into several discrete zones. Over the five-year period between 1988 and 1993, more than 100,000 welds were tested by both ultrasonics and radiography on NOVA and TCPL projects. During this period, a better understanding of the parameters causing variability was gained. By 1993, the standards were set sufficiently high to ensure that the mechanised ultrasonic inspection technique would provide reliable probability of detection of all significant defects. TransCanada Pipelines Limited applied to the National Energy Board and received approval to replace radiography with mechanised ultrasonics on mainline production projects. This book provides an overview of the principles involved in automated ultrasonic testing (AUT) of girth welds as well as presents some of the origins and the many parameters that influence the results of these inspections. Also considered will be some of the more controversial aspects of the process, including sizing and acceptance criteria. In addition, we will look at the similarities of the basic concepts that are applied by the major players in the industry. Finally, we will look at some of the features used and the limitations of the technique and speculate on potential future enhancements.

2

Introduction

1. History — Early Days

A new scientific truth does not triumph by convincing its opponents and making them see the light, but rather because its opponents eventually die out, and a new generation grows up that is familiar with it. —Max Planck (1858–1947) Use of mechanised ultrasonic inspection systems has now become a common occurrence in nondestructive testing (NDT). Any review of an NDT journal or NDT related Web page will quickly reveal a wide variety of scanners and data acquisition software systems available. Goldman1 and McMaster2 describe mechanised systems using motorised carriages that were used in the 1940s and 1950s for aircraft structures, plate scanning, and for rail testing on trainmounted systems. Data display and recording in early UT was first just the RF traces on a scope. A bit later, there were B-scan and C-scan scope displays using phosphorescent screen persistence and later still pen recorders using ink or hot wire elements on heat- or light-sensitive paper. Today, we seem to take for granted the speed and accuracy associated with mechanised UT inspections. One of the main beneficiaries of mechanisation enhancements in ultrasonic testing is the pipeline industry. It has taken the best part of half a century but the traditional radiographic inspection of pipeline girth welds is now being replaced, in many locations, by mechanised UT. In this chapter, we will identify the highlights of the progression of events that led from an idea to a major industry.

1.1

Early Ideas In any discussion of pipeline girth weld inspection by mechanised ultrasonics, one must look to RTD b.v. in the Netherlands as a pioneer in the industry. Since 1959, they have been working on options for the pipeline industry. An example of one of the early efforts made by RTD is shown in Figure 1-1. This early version of the Rotoscan shows a split ring support on History — Early Days

3

which the probe holder moves. A single probe was used with a single channel ultrasonic instrument. Three separate UT instruments were used for each of the three probes: two probes set opposite each other to detect longitudinal flaws and a third probe to detect transverse defects. The instruments were not multiplexed and so would have been subjected to the possibility of cross talk.

Photo courtesy of RTD b.v.

Figure 1-1 RTD three-probe Rotoscan (single-channel unit shown, around 1959)

Simple applications of ultrasonic testing in pipe mills on longitudinal or helical seams were about the extent of pipeline UT during the 1960s. During the 1970s there again seemed to be an increased interest in a faster, radiationfree option to inspect girth welds. One of the early efforts in the 1970s came out of Japan. M. Nakayama et al.3 at Nippon Steel Product Research and Development Laboratories, Sagamihara City described a prototype with two probes calibrated on through holes 3.2 mm in diameter (similar to the calibration technique for the submerged arc longitudinal seam welds). Scanning could be done from 100 mm/min to 1000 mm/min and coupling checks could be made using the opposing probe configuration, which used a pair of probes of 5 MHz, 10 × 10 mm, 70°. Recorded outputs were made on a polar graph plot that indicated amplitude with angular position around the circumference (see Figure 1-2). 4

Chapter 1

Nakayama et al. noted that detection of flaws was closely related to test sensitivity, but if sensitivity was too great, the echoes from the weld bead geometry lowered the signal-to-noise ratio.

Figure 1-2 Polar plot from the Nippon Steel prototype system

On the left is a polar graph made with the eight through holes 3.2 mm in diameter used for calibration. On the right is a weld scan showing: slag (1 o’clock), crack (3 o’clock), and lack of fusion (7 o’clock).

1.2

The Introduction of Mechanised Welding One of the reasons for the slow advances in the application of ultrasonics to girth weld inspection was the variability in the weld cap and root bead geometry. As noted by the Japanese authors in the 1974 paper, the geometry signals were a cause of poor signal-to-noise ratio. Around this time, work was being done to mechanise the welding process. Some early efforts had been made during the 1940s using an oxyacetylene heated upset-butt welding process. In 1958, Esso Research and Engineering Company funded a project at Battelle in Columbus, Ohio, USA. This used a then new process called gas metal arc welding (GMAW) with a CO2 cover gas. Problems with the process were mostly due to fit-up as it had used the standard 60° included bevel angle. In the mid 1960s Crutcher-Rolf-Cummings merged with M. J. Crose Manufacturing and operated as CRC-Crose International until sometime in History — Early Days

5

the 1970s when they became Crutcher Resources Corp. In the mid 1960s Jerry Nelson (who had been the engineer working on the Battelle project) approached CRC with an idea for a new and improved system. He purposed to make a system that would be composed of a bevel facing machine, an internal line-up clamp/welding machine to deposit the root pass and external welders to run on bands clamped to the pipe outside surface. This was the foundation of the CRC Automatic Welding System and it remains essentially unchanged today. The CRC welding system had some growing pains. Welding parameters were then poorly understood, the operators were not welders and needed to have extensive training and like most new technology, it was not readily accepted in the industry. Finally, some trial runs were made, welding parameters were improved, and the operators became proficient. In about 1983 CRC Pipeline International broke away from Crutcher Resources Corp. and became a privately held operation. In 1985, CRC Pipeline International merged with Evans Pipeline Equipment to form CRC-Evans Pipeline International. Mechanised GMAW is now the preferred method of producing large diameter pipeline girth welds. CRC Automatic Welding System was one of the first systems developed but now others are also proving effective. Vermaat from the Netherlands, SERIMAR from France, and RMS from Canada, are but a few of the systems used today. Some use internal clamp and welding machines while others use just the internal clamp with a copper backing ring. When well tuned, all produce a narrow uniform cap and a smooth, nearly flat root profile.

1.3

UT Adaptations to the Mechanised Welding In 1971, just as the CRC Automatic Welding System was beginning to show promise, Vetco Offshore Inspection saw the CRC system in Houston, Texas and sent a welding “bug” and band along with some sample welds to their Canadian office. There, a young engineer named Tony Richardson designed the first UT inspection system for the CRC welding process (see Figure 1-3 and Figure 1-4). This consisted of a Branson SonoRay UT instrument, a multiplexer, a Clevite-Brush four-channel light beam chart recorder, and four immersion probes. The immersion probes had one pair directed at the root (one each side) and the other pair was directed at the fill. This system also included an odometer (encoder). The problems associated with the annoying signals from the weld cap and root were addressed by setting gate lengths on a special calibration block that limited the data collected to the weld volume.

6

History — Early Days

Figure 1-3 Tony Richardson with the first multiprobe scanner mounted on a CRC welding band

Figure 1-4 Tony Richardson’s scanning head using four immersion probes (1972) History — Early Days

7

This system was ahead of its time. Attempts to get TransCanada Pipelines Ltd. (TCPL) interested failed as they saw no application for it in 1972. It was sent to the Vetco European office and had an evaluation by British Gas but nothing further developed with Vetco with that specific design. However, in the same year, NOVA (then Alberta Gas Trunk Ltd.) began development of the GMAW process using the CRC Automatic Welding System and by 1977 decided that UT was the best option for the 37.5° root bevel and 45° hot pass bevel angles. Not being aware of the Canadian development already established five years previously using the CRC equipment, NOVA opted to have RTD develop a system based on the CRC Automatic Welding System. RTD adapted their existing yoke version Rotoscan used on offshore applications and put it on a CRC welding band. This land-based version was dubbed the Bandscan to differentiate it from the offshore version that was called Rotoscan. Soon the name Bandscan was dropped and reverted to “Rotoscan.” In the late 1970s there were several other companies developing mechanised systems but when demonstrating their abilities on the girth welds, all wanted to show how much information could be collected and how it duplicated manual style scanning. To do this, they used a raster scan. Scanning with a raster movement can provide useful information and the relationship of signal-to-position (echo dynamics) using raster motion could be helpful in sorting out defects from geometries; but the scan process is extremely slow. When demonstrated on typical welds 42 in. (1067 mm) in diameter the scanning time was on the order of 10–15 minutes and interpretations would require even more time. The linear scan showed the best hope for production rates typically achieved by welding (80–100 welds per day at that time). The Rotoscan used standard contact probes (just as the Nippon Steel system had) but used several probes on each side trying to optimise for the weld bevel (just as the Vetco scanner had). With the combined interest of a potential pipeline user (NOVA) and the manufacturer/service company (RTD), the development moved faster. However, standard contact probes and simple amplitude gating still resulted in many false calls. It became apparent that probe design had to be reconsidered.

1.4

Probe Design Changes The obvious solution to annoying signals caused by off-axis components of the beam was to limit the divergence of the beam. This led to the development of focused transducers designed for the application. The well-known pseudofocus effect using a transmit-receive pair of elements did not provide an adequately small focal area and suffered from skew effects on longer path

8

Chapter 1

lengths. Focused transducers, even contact focused transducers, are not new. Back to the earliest editions in Ultrasonic Testing of Materials by the Krautkramer brothers, they described the principles involved in this application. The options in the 1970s were based on one of three principles: •

a flat element with a curved lens,

•

a curved element with a matching lens insert,

•

a mosaic element array on a curved wedge surface.

Although the principles were well documented, the previous uses for this technology had not demanded the sort of precision required by this application. Improved techniques for the relatively thin pipeline girth weld applications were possible only with focused probes. The first of these applied to the girth weld inspections used internal lenses and was developed by BAM in Germany. Later, others experimented with shaped elements. By the mid 1980s spot sizes were consistently held to around 2 mm in diameter at the area of interest and false calls that were caused by beam edges interacting with weld surface geometries were virtually eliminated. This improved signal-to-noise ratio allowed a new philosophy to be considered: engineering critical assessment (ECA). The ECA concept uses the principles of fracture mechanics to assess the severity of a defect based on its vertical extent. The small spot sizes now achieved allowed the weld to be divided into several zones. This linking of ultrasonic results to fracture mechanics was probably the single most important aspect in the development of mechanised UT on girth welds. Although the concept of dividing the weld into zones was popularised in the pipeline industry by Glover et al.4 in the late 1980s, the concept was used in a much earlier report (work done in 1981/82) by Moles and Allen5. Not only did they utilise the concept of zones to size flaws but they used a computer display to indicate the zones. They also used the tandem pitch-catch arrangement of elements, which was later re-developed by Canmet in the study reported by Glover as the recommended option for the inspection of the 5° vertical bevel of GMAW. The application of the tandem probe arrangement has long been popular in Europe for heavy-wall vessel inspections. Krautkramer references A. de Sterke6 regarding this technique, who described the practice that had been used in Europe for some time by then.

History — Early Days

9

1.5

Mechanised UT Enters the Computer Age By persistence and support from a few believers like NOVA, TransCanada Pipelines and GasUnie, RTD managed to push on with the development of the concept of mechanised UT on pipelines. In about 1991, Guardian-Hyalog (now Shaw Pipeline Services Ltd., or SPSL), after encouragement and specific direction from NOVA, undertook to construct their own girth weld inspection system. But RTD had set the standards for what was expected, even down to the presentation format using multichannel strip charts and symmetry of output display to simulate the weld opened along the centreline. SPSL’s first unit consisted of a bank of portable UT instruments, one for each probe. This bank of units was multiplexed and the outputs from the various UT instruments were fed to an analogue-to-digital conversion (ADC) card where the signals were processed and stored to a 286-PC style computer. The chart that was printed from the computer was made to look almost exactly like the strip chart output that RTD had been using since the mid 1980s. The SPSL system utilised the system encoder to space information equally on the strip chart hard copy, thereby making the information more precisely located. Prior to this the position was marked on the charts by a timing marker that estimated speed. Scanner speed variations were shown on the chart recording as different length intervals between the nominal 10-mm markers. Now the speed of development was accelerated even more. RTD developed a MS-DOS®-based computerised system and developed a mapping display in 1993 to improve the discrimination between flaws and surface geometry, and it had the serendipitous advantage of characterising porosity. Mapping was just the modern equivalent of the B-scan. RTD collected stacked A-scans and assigned colours to the various signal amplitudes thereby making porosity appear like porosity on the computer monitor. Examples of the earliest versions of the strip charts used in Canada are illustrated in Figure 1-5 and Figure 1-6.

10

Chapter 1

Figure 1-5 Original RTD strip chart format printed on a RMS recorder and light-sensitive paper

History — Early Days

11

Figure 1-6 SPSL computerised strip chart format printed on a laser printer

In the early 1990s parallel development was being done by two other companies. SGS Gottfeld in Germany had designed its MIPA system and R/D Tech in Canada had developed its initial system subsequently used by AIB Vinçotte in Belgium. The SGS Gottfeld system seemed to be a flashback to the Vetco system of the 1970s but now modernised with computer technology. It too used immersion probes and a skirt to hold the water. The monitor display consisted of a series of bands for each channel with amplitude represented as a colour (a single line C-scan). Figure 1-7 and Figure 1-8 are courtesy of Rolf Diederichs and www.NDT.net, which can be found at http://www.ndt.net/article/schulz/.

12

Chapter 1

Courtesy of Rolf Diederichs and www.NDT.net

Figure 1-7 SGS scanner

Courtesy of Rolf Diederichs and www.NDT.net

Figure 1-8 SGS chart display

The R/D Tech® system used a probe array with contact probes similar to that used in Canadian projects but their system also had a single line colour per History — Early Days

13

channel C-scan display as the hard copy Figure 1-10.

Figure 1-9 AIB scanner

Figure 1-10 R/D Tech chart display used by AIB around 1992

Both the R/D Tech and SGS systems were used on a project in North Africa but failed to get accepted in Canada. By then the ease with which the RTD strip chart format using time and amplitude information combined on the same chart could be read by the operators made it the preferred presentation method.

14

Chapter 1

An interesting side note: in 1992 a Canadian company, Canspec, used a prototype version of the AIB system and experimented with the B-scan display available in the R/D Tech software package. They found that root geometry and porosity were more easily identified using the B-scan information. This was over a year before RTD officially came out with their Roto-map which was designed for the same purpose (another example of parallel thinking). In 1996 WeldSonix introduced their system. This system came up with a smaller scanning head (Figure 1-11) and full waveform data collection for all channels but also used the industry-accepted strip chart format.

Figure 1-11 WeldSonix scan head

In 1997 R/D Tech decided to conform to the data presentation initiated by RTD nearly 20 years earlier and developed a newer version of both hardware and software based on their Tomoscan™ technology. In 1998/99 R/D Tech used this display (see Figure 1-12) for the data collected by the new phased array system they had developed.

History — Early Days

15

Figure 1-12 R/D Tech display 1997

1.6

Standardising Concepts Until 1998 all mechanised UT on pipeline girth welds was done using company specifications specially worded to allow mechanised UT. This was required because existing codes and standards were worded as though ultrasonic inspections would be by manual techniques only. In 1998 ASTM E-1961 was published describing the various aspects involved in pipeline inspection using mechanised UT and zonal discrimination. In the 19th edition of API 1104 they revised its description of UT requirements to include mechanised systems. It does not specify the zonal techniques but if one is to use the Alternative Acceptance Criteria in the appendix of API 1104 there is no practical option other than the zonal technique. Much of the acceptance of systems by the pipeline industry has been dictated by presentation. QA 9000 Ltd. introduced their Acuscan in the mid 1990s. Mechanically this looked very similar to the WeldSonix scanner but the report presentation had reverted to the top, side, and end views. This made it difficult to use with the ECA analysis by then common with the other systems.

16

Chapter 1

References to Chapter 1 1.

Goldman, Richard G. Corporation, 1962.

Ultrasonic

Technology.

Reinhold

Publishing

2.

McMaster, Robert C. Nondestructive Testing Handbook. Vol. II (first edition). ASNT, 1959.

3.

Nakayama, M., Y. Kato, and E. Isono. “Investigation of the improvement of speed and reliability in the inspection of field welded pipelines, Quality Control and Non-Destructive Testing in Welding.” International Conference, London, England, 1974.

4.

Glover, A. G., M. P. Fingerhut, and D. V. Dorling. Mechanised Ultrasonic Inspection of Pipeline Girth Welds. Part 4. Dept. of Supply & Services DSS File # 23SQ-23340-2-9027-4, Nov. 1988.

5.

Moles, M. D. C., and A. L. Allen. “Tandem Probe Ultrasonic Measurement of Cracks in Economizer Inlet Header Sections.” Materials Evaluation (May 1984).

6.

De Sterke, A. “Some aspects of radiography and ultrasonic testing of welds in steel with thicknesses from 100–300 mm.” Br. Journal of NDT, no. 9 (1967).

Acknowledgement and Thanks for Historic Notes To put together a collection of information and images spanning 40 years has required the help and cooperation of several people that have had ties to this subject stretching back even more years than I have been in the industry. I would like to extend thanks to the following: Anthony Richardson, of Inspectech Analgas Group, Scarborough, Ontario, Canada. Tony provided his original photos and the US patent that he obtained for the system he put together using the immersion probes. Robert van Agthoven, RTD Rotterdam, The Netherlands, provided the image of the earliest Rotoscan and commented extensively on its early development years B.C. (before computers). Rolf Diederichs has allowed use of his online collection of technical papers relating to the girth weld inspections (www.NDT.net). Olympus NDT Canada (www.rd-tech.com).

provided

copies

of

their early

brochures

Acknowledgement and Thanks for Historic Notes

17

Blaine Mitchell at CRC-Evans for providing information on the early developments of the GMAW process. Jan A. de Raad, RTD Rotterdam, The Netherlands, provided extensive proofreading, corrections, and opinions on several aspects of the details of RTD’s role in this history.

18

Chapter 1

2. Zones—How to Keep Them Apart

Nearly every great advance in science arises from a crisis in the old theory, through an endeavour to find a way out of the difficulties created. We must examine old ideas, old theories, although they belong to the past, for this is the only way to understand the importance of the new ones and the extent of their validity. —Albert Einstein (1879–1955) Although now usually known as AUT, the inspection method used on pipeline girth welds is probably best described by its earliest name, zone discrimination. This describes the method whereby the weld is divided into zones that can be defined by vertical intervals of the through-wall dimension. Using a multiple beam configuration from both sides of the weld allows for the complete volume plus the heat-affected zone (HAZ) to be inspected in a single pass. This technique can actually meet the requirements of ASME pressure vessel and piping weld inspection. When the GMAW process became the preferred option for pipeline girth welding, the techniques traditionally used for manual UT, similar to those used in ASME applications, were found to be less effective than they were for the shielded metal arc welding (SMAW). The main problem was that the primary flaw in GMAW was nonfusion and the fusion faces were not as well addressed using just a 45° and 60° refracted shear wave. But these fusion faces were also poorly addressed by radiography making the need to optimise the beam angle all the more important since the tried-and-true option of radiography now offered an even lower probability of detecting potentially serious flaws. The GMAW weld profile first encountered contained several angles, some of which were nearly vertical (5° off vertical in fact). This meant that a tandem probe configuration was required to optimise detection of nonfusion on that face. But tandem configurations are not useful for a raster scan because the bounce path for the beam is only effective for the receivertransmitter spacing at one point on the vertical face. Only since the late 1990s, with the advent of phased array probes, could such a dynamic scan be considered using a tandem technique. Until phased arrays made this Zones—How to Keep Them Apart

19

possible, to overcome this problem and to ensure that the entire fusion face (and weld volume) was interrogated, several probe pairs were used with the tandem technique when thickness required. Together with the pulse-echo probes (single element) for the other weld regions, the weld could be scanned in a single linear pass motion. The original concept can be summarised as explained in section 2.1.

2.1

Principles of the Zonal Technique The weld is divided into zones typical1y 1–3 mm high, and beam angles are selected to optimise response off the weld bevel fusion face as shown in Figure 2-1.

Figure 2-1 Schematic representation of weld zones and optimised beam paths for three weld bevel shapes commonly found in pipeline girth welds The probes are moved around the girth weld by a motorised carrier that moves along a track typically used by the welding apparatus. Ultrasonic signals received by the instruments are monitored using electronic gates. Both signal amplitude and time of flight in the gated region are monitored. The region gated is from just before the theoretical weld bevel preparation to just after the weld centreline. Gated output, both time and amplitude as well as selected waveforms, are digitised and displayed in a chart format. An operator evaluates the chart results and makes a decision as to weld acceptability based on the length of signals exceeding a threshold as set out in specifications and regulating codes. ID surface notches and flat-bottom holes 2–3 mm in diameter typically provide targets on which signal amplitude and travel time are set in a gate. 20

Chapter 2

These targets are machined in a specific section of pipe and arranged along the theoretical weld bevel profile. So-called volumetric channels are added to improve detection and identification of small and off-axis flaws such as porosity. As well, most systems now incorporate some form of TOFD as part of the configuration. Figure 2-2 illustrates many of the flaws and geometric conditions that an AUT system will be capable of detecting and identifying. These are shown on the CRC bevel and typical of the flaws that can form with the exception of the misfire. Misfire is a term coined by radiographers and AUT operators used here to denote a condition unique to the CRC GMAW processes with an internal welding head. It occurs when the arc is unsuccessfully initiated and no metal is deposited as a result. It is considered different from incomplete penetration (IP), which might be the term more appropriately used when the vertical land is not fused (but this is coined lack of cross penetration or LCP in the CRC jargon). Zone identification

Discontinuity

6

2nd fill and cap

Solidification crack 1st and 2nd fill

5

1st fill

Nonfusion 1st fill

4

Hot pass (upper)

3

Hot pass (lower)

2

Land for cross penetration

1

Root

Lack of cross penetration Root bead porosity

Fill porosity

Hi-low

Misfire

Burn through

Figure 2-2 Common flaws and geometric conditions in a weld made with the GMAW

Zones—How to Keep Them Apart

21

Examples of the zones targeted by the ultrasonic beam are identified by numbers on the left in Figure 2-2. It should be noted that these zones are not always identical with the welding passes that they may sometimes be named after. For example, the LCP zone straddles the root and hot passes, and the hot-pass region of the weld is covered using two zones. Also, the upper fill (2nd fill) and the cap are grouped together in a single zone. A true separation of weld passes identical to the fixed zones is not possible. In a welding process, especially where the weld is made for a vertically oriented joint, the weld puddle does not deposit an equal thickness of metal in all positions around the girth. Gravity pulls the weld puddle down. On the top the deposits tend to be thinner and on the bottom deposits thicker. Variations in weld pass thickness will be less pronounced when the pipe can be rotated under the arc or when the weld is oriented in the horizontal orientation as would be the case in an offshore J-lay. Although some association may be made of flaw location identified to a zone it is often an approximation. Because of its ability to follow the welding soon after the weld has been completed, AUT has also been able to provide a process control function. Process control is only possible where welding is mechanised because the parameters are much better controlled than in manual SMAW. AUT has been able to identify trending problems with root passes (e.g. one or more internal heads moving out of centre position or running out of cover gas) or it may be able to narrow the problem to a specific side and one or another weld operator (since the problem may straddle two fill passes). Zones targeted by the ultrasonic beams are fixed in their vertical position for the nominal wall thickness. Methods to ensure that a zone is both where it should be and covering only a limited surface area on the bevel has been the goal that has driven development of the zonal technique. Difficulties in maintaining beam position and size at the area of interest require an understanding of the parameters. Only then will the steps required to control or monitor the parameters make sense.

2.2

Beam Control To illustrate the issues that affect beam control, we will use several presentations of weld profiles and illustrate the effects of parameter changes. As described in Figure 2-1, the ultrasonic beam is intended to look at a specific area along the fusion line. Figure 2-1 uses a single line to represent the centre of the beam path that would be optimised to go from the transmitter to the centre of the zone and return along either a relatively direct return path (pulse-echo) or follow a skip path to a receiver set forward or behind the

22

Chapter 2

transmitter (tandem). We must ensure control on two main aspects to provide a separate signal from one zone to the next: 1.

Beam spot size

2.

Angle control

2.2.1

Beam Spot Size

Spot size is a function of the ultrasonic element dimensions. We will consider circular elements for the explanations but similar considerations apply to square or rectangular elements. With the advent of phased array probes it is also important to note that they too can be treated using the same mathematic principles since the pulse generated by a group of phased array elements is the same as a single element of the same dimensions. To assess the size of a beam, we normally define the size based on the dimensions outward from a centre (on-axis) maximum. Traditionally this has been assumed taken to the isobar (line demarking equal pressure) that is half the pressure of the maximum. This would relate to a 6-dB drop in a pulseecho signal. Two equations help us assess the spot size and position. The diameter of a beam can be approximated by equation (2.1). D B–6 dB = 0.2568DS F

(2.1)

Where DB –6 dB = beam diameter (at the 6-dB drop boundary) D = element diameter SF = normalized focal length This indicates that for a flat element (i.e. unfocused) the smallest beam diameter will occur at the near-field distance and it will be 25.7% of the element diameter. It is not possible to extend the near zone to a greater distance in a material without changing some aspect of the probe (the diameter or frequency being increased could push the near field to a greater distance). Only when we use a phased array probe could we adjust the near field and that only by using more elements since the frequency is fixed. However, we can reduce the distance to where the maximum pressure occurs.

Zones—How to Keep Them Apart

23

This is called focusing. By directing the side lobes of a beam toward the centre axis, the point at which the maximum pressure occurs is made to occur at some point less than the near-field distance. The new and reduced focal length not only increases the pressure available locally at the focal spot, it also reduces the spot dimension. When the new location of the focused pressure is given as a ratio of the natural focal distance (i.e. the near-field distance) it is sometimes called the normalised focal length. It is defined by the equations: F S F = ---N

(2.2)

Where SF = normalized focal length F = focal length N = near field From equation (2.1) we use the term SF = normalized focal length. But in equation (2.2) we can see that this is unity for the near field (i.e. for an unfocused beam). But if we were to focus a beam at half the near field the normalised focal length becomes 0.5. Putting this value in equation (2.1) we see that it reduces the spot diameter to half its diameter at the unfocused near field. It will be helpful to see how this looks graphically. Using a modelling programme, we have constructed a condition whereby a probe is put in contact with a steel surface. Only the compression mode is shown but the same principles apply to the shear mode with the differences in near field for the appropriate wavelength differences. A probe of 5 MHz, 12 mm in diameter is placed on a block of steel 50 mm thick, and the pressure distribution is plotted in Figure 2-3. On the left is the unfocused (flat) element, and on the right is a probe of the same dimensions but arranged to focus at 15 mm (i.e. half the near-field length of the flat probe). Below the crosssections are plots of the pressure taken at the sound paths of 30 mm and 15 mm respectively. The radial plots indicate the pressure over the probe 12 mm in diameter with the pressure normalised to a maximum at the focal distance so the amplitude indication of 1.0 is not to be compared from the focused to the unfocused condition.

24

Chapter 2

0 -10 -20 -30 -40 -50

12 mm 5 MHz flat probe

12 mm 5 MHz focus at 15 mm

Radial Beam Cross Section

Radial Beam Cross Section

1.0

1.0

0.8

0.8

0.6

0.6

0.4

0.4

0.2

0.2

0.0

0.0 -6.000

6.000

Flat beam at 30 mm is 3.2 mm in diameter

-6.000

6.000

Focused beam at 15 mm is 1.6 mm in diameter

Figure 2-3 Comparing flat and focused beams

This is now applied to our zone concept. Optimisation of a beam for a zone has been considered to occur if the spot size (i.e. the –6-dB beam pressure boundary) closely matches that of the zone. This thinking is primarily applicable to the fill zones where the zones are along the same plane and angle. To illustrate the considerations we will use a simple J-bevel on a 16-mm thick wall. Such a configuration usually uses a root land of approximately 1.4 mm and a hot-pass preparation of a radius of about 2.4 mm to the horizontal. This leaves 12.2 mm for the fill zones. If we use four zones of ultrasound in the fill region the zone height of each fill zone is 3.05 mm or near enough to 3 mm. A bevel angle of 3° will be used. These conditions will require a tandem configuration due to the steep fill angle. Four cylindrical targets (representing flat-bottom holes 3 mm in diameter) are arranged in Figure 2-4. The probes are arranged to indicate a transmitter (at 47° refracted angle) as the front probe in the tandem pair and the receiver would be set to 53°. The angle difference between a transmitter and receiver is double the angle of the bevel. For example, in this case the

Zones—How to Keep Them Apart

25

bevel is 3° so we add 6° to the transmitter angle to ensure optimum reception of the returned beam, that would be 53°.

Figure 2-4 Path for a tandem pair directed at fill-3 zone

The transmitter must be focused to limit the size of the beam to a reasonably small area when it reaches the bevel. In this case we have used a radius of curvature of 95 mm on the 7-MHz element, 12 mm in diameter. The receiver may be unfocused. When plotted as a group of rays we see a reasonable clustering at the target of interest (Figure 2-5).

Figure 2-5 Tandem configuration with the transmitter (front) beam focused on the fill-3 zone

But ray tracing does not provide an accurate representation of the beam that has a “fuzzy” zone around the focal spot. This results in overtrace (signals arriving from portions of the beam interacting with zones above or below the intended zone) due to the actual pressure field extending beyond the mere –6-dB drop region. This is illustrated in Figure 2-6 where the beam construction software programme is used to indicate the pressure distribution similar to that done in Figure 2-3.

26

Chapter 2

Figure 2-6 The beam near the target region

The beam calculated to provide the optimum focus for our tandem technique for the fill-3 zone used a radius of curvature (ROC) of 95 mm. This worked because of the calculations including the time delay in the wedge. But most companies cannot rationalise the costs of purchasing custom made focused probes that have increments of change in ROC every 5 mm or 10 mm, nor do they bother to make the necessary corrections by altering the travel time in the wedge by machining thicker and thinner wedges. The result is a slightly less than optimum condition. If, for example, we had only a 12.5-mm, 7.5-MHz probe with a 150-mm ROC (this curvature does not actually result in a focusing in the steel but is considered a beam correction), the path distance to the fusion-line target would occur well before the focal region. This means our beam is interacting at the intended area in the near zone and has a broad region of pressure variation (see Figure 2-7).

Figure 2-7 Beam focused after fusion line

Strong side lobes can be seen in Figure 2-7 as the beam meets the fusion line. These can produce annoying overtrace signals as they are seen to interact with strong (darker) components above and below the intended zone. In a tandem configuration the separation between the transmitter and receiver are optimised for a specific point of interest along the vertical plane. If the same spacing was maintained and the transmitter probe moved forward to have its centre ray locate the midpoint of the zone immediately below (fill-2 zone), the receiver would be well out of position to receive the returning pulse. This is shown using the beam centre in Figure 2-8. Zones—How to Keep Them Apart

27

Figure 2-8 Tandem configuration with the receiver (back) at the same position as fill 3 and transmitter now positioned to detect the fill-2 zone

For a given wall thickness the zones of the ultrasonic beams trace out a series of bands and at each point along the girth each band is sampled to see if any indications are present. The –6-dB beam boundary is the convenient beam assessment measurement but, as indicated above in the beam construction analysis, the beam does not suddenly end at the –6-dB boundary. Some overtrace will still occur due to the remaining sound pressure that exists beyond the –6-dB portions. Figure 2-9 illustrates the parallel fill zones along a girth weld. Also positioned on the illustration are the theoretical beam spots that would indicate the incident point of the beam on the fusion line and a typical flaw. The flaw indicated could be considered as nonfusion that occurred in the second fill pass. The illustration shows how the flaw crosses several ultrasonic beam zones. As explained earlier, this is typical of welding conditions where the joint is vertical and the weld pass thickness varies around the girth. In the illustration the flaw starts off in the upper portion of the fill-1 zone, and gradually moves across the fill-2 zone, and ends just inside the fill-3 zone.

Ultrasonic “spot” positions

F4

F3 Ultrasonic zones F2 F1

Second fill pass flaw Ultrasonic spot represented with two rings inner 0 dB to –6 dB outer –6 dB to –20 dB Figure 2-9 Spot size on ultrasonic beam zones

28

Chapter 2

Spot size is a result of beam focusing. Three practical options are available for focusing using contact probes: 1.

Lens focusing (flat element with curved plastic plates mounted in front— Figure 2-10).

2.

Curved element focusing (the element is curved and an insert in the gap—Figure 2-11).

3.

Phased array focusing (Figure 2-12).

Lens material 1

Wedge material 2

Figure 2-10 Lens focusing

Curved element Gap material (same as wedge) Wedge material

Figure 2-11 Curved element focusing

Zones—How to Keep Them Apart

29

Figure 2-12 Phased array focusing

2.2.2

Beam Angle

The other item requiring control in zone discrimination is the beam angle. By this we mean the nominal refracted angle. The refracted angle in any ultrasonic application follows Snell’s Law, which is given in the form of the equation: V1 sin θ ------ = ------------1V2 sin θ 2

(2.3)

Since the incident angle is fixed by the machining of the wedge angle then anything that changes the velocity in either the incident or refracting medium will result in altering the refracted angle. Problems encountered in the early stages of development led to the discovery that velocity variations were causing significant zone changes of the beams. These variations in velocity were from two sources: 1.

Steel type

2.

Temperature

2.2.2.1

Steel Type

A study by Silk1 points out the significance of crystallographic texture on acoustic properties in mild steels. He found crystallographic texture 30

Chapter 2

differences account for velocity differences on the order of 10–2 whereas differences due to stresses were only on the order of 10–4. Acoustic velocity is a birefringent quantity, i.e. different with respect to direction and polarization. If the stresses on the principal axes are different, then the velocities of the two polarisations are different. Acoustic velocity can be used to determine mechanical properties and dimensional attributes. Since it is a bulk measurement in metals, grain structure has significant influence on both the nominal value and the degree of birefringence. As a result of differences in both mill compositions and rolling processes, acoustic variations exist in line pipe steel. It is an easy matter to establish Poisson’s ratio by the ratio of the transverse to longitudinal velocity, e.g. Ct 2 1 ⎛ ------ – ⎝ -⎞⎠ 2 Cl µ = ---------------------Ct 2 ⎛ ⎞ 1 – ----⎝ C l⎠

(2.4)

where Ct is the transverse wave velocity and Cl the longitudinal wave velocity. This allows a means of determining material strengths. Schneider et al.2 have used the time difference between the two polarisations of SH wave signals to determine residual stress in steel. This held the advantage that no path length determination was needed. In 1994 Ginzel and Ginzel3 carried out the first study related to the velocity concerns for line pipe steels as it relates to AUT on girth welds. The cost of changing probes and manufacturing calibration pieces for each pipe variety is a significant factor for inspection companies providing girth weld inspection on construction projects. In preparation for an inspection a prior knowledge of acoustic velocity assists in determining if there is a need for a new calibration piece and/or probes (or wedges). This study was carried out on several different samples of line pipe and included materials from several manufacturers, different wall thicknesses, and different pipe diameters. See Table 2-1.

Zones—How to Keep Them Apart

31

Table 2-1 Part parameters used for the velocity study Sample no.

Manufacturer

1 2 3 4 5 6 7 8 9 10 11

A A B B C C C D D D D

Diameter (in.) 48 48 48 48 48 48 48 42 42 42 42

Wall (mm) 17.8 17.8 11.7 11.7 17.8 12.1 17.8 15.2 15.2 9.5 9.5

Long./Spiral seam Long. Long. Spiral Spiral Spiral Spiral Spiral Long. Long. Long. Long.

Acoustic velocities were determined for the following polarisation modes and planes (Table 2-2): Table 2-2 Modes and planes of polarisation Item 1 2 3 4 5 6 7

Mode Polarisation Longitudinal Radial Longitudinal Perpendicular to end, inclined 45° to surface and parallel to end SH Parallel to end SH Perpendicular to end SH Perpendicular to end (inclined 45° to surface) SV Radial SV Nominal 60° refracted

Items 2 and 5 required machining the sample at 45°. Item 7 is an important measurement as it most closely duplicates that mode used in girth weld inspections. To achieve the required accuracy for item 7, i.e. within 20 m/s, a tool equipped with a digital micrometer was built. For the most part velocity determination is fairly straight forward. You need only a known thickness (easily determined to ±10 micron accuracy with a micrometer or a vernier calliper) and a scope to monitor an ultrasonic pulse with a timing accuracy to about 10 ns. The details of the study developed techniques are now standardised in ASTM E-1961. To facilitate parallel surfaces for thickness and time measurements in the appropriate directions, slots were milled in each sample. Slot positions are illustrated in Figure 2-13. For measurements in the radial direction it was ensured that all paint was removed from the pipe outside surface.

32

Chapter 2

Circumferential axis

Radial axis

Longitudinal axis

Longitudinal axis of pipe 45° to longitudinal axis

Figure 2-13 Slot orientations: configuration of slots machined for velocity measurements

As noted in Table 2-2, several modes and directions were measured. The longitudinal mode (or compression mode) uses a method similar to the standard technique for thickness testing. Placing a normal, longitudinal beam probe on the surface of the test piece (using the necessary couplant) several multiples of the thickness are displayed on the scope and the time interval accurately measured between any two multiples gives the piece thickness. This can also be done using an immersion setup where the coupling distance is increased to ensure water path multiples are avoided. Since a pulse-echo technique is used, the travel time between multiples is composed of the time for the beam to traverse the thickness twice. Measuring the thickness accurately by means of a micrometer, velocity is found by dividing the thickness by half the time measured on the scope. The SH wave mode listed in Table 2-2 is not as familiar to most UT technicians as the shear mode used in weld testing. SH waves are also called horizontally polarised shear waves. The more familiar shear waves used in weld testing are referred to as SV waves (or vertically polarised shear waves). SH wave probes are polarised to provide particle displacements parallel to the test surface. Solids support two shear waves polarised at right angles to each other. By orienting the probe on the test piece in such a way as to excite either (or both) preferred vibrations in the lattice, the two transit times can be measured in essentially the same way as the longitudinal mode was. Special consideration is given to the couplant used. A non-Newtonian viscous fluid is required. Newtonian fluids (such as water) cannot support a shear-wave displacement. The most common non-Newtonian viscous fluid is honey and it is applied just as would be a jell couplant for compression wave assessment of thickness (see Figure 2-14).

Zones—How to Keep Them Apart

33

Probe

Honey used as couplant

Figure 2-14 Setup for horizontal polarization velocity determination

Both SH modes can be viewed simultaneously by orienting the probe to 45° from the maximum response. This is illustrated in Figure 2-15. The two modes of SH wave exist with different velocities when the stresses on the principal axes are not equal. This is seen as fast and slow modes on the scope.

Figure 2-15 Fast and slow wave through a 10.516-mm thick wall

Figure 2-15 shows the 2nd and 3rd multiples. Note the increasing separation in the two modes at the 3rd multiple compared to the 2nd one, resulting from increased travel time. To duplicate the conditions for generating SV waves used in automated inspections a transmitter-receiver arrangement was fabricated using standard probes and angulated wedges. Assembly and calculations are described

34

Chapter 2

below. To facilitate the precision motion required, the probes were assembled in a fixture whereby one probe was fixed while the other was moved using a threaded actuator with a micrometer to allow precise measurement of separation. A sketch in Figure 2-16 illustrates the prototype fixture used to make SV wave velocity measurements at a nominal refracted angle of 60°. Magnets

Digital micrometer

Transducer (fixed)

Transducer (moved by micrometer head)

Figure 2-16 Micrometer actuated probe fixture

A schematic of the transmitter-receiver probe pair and the principles used for the calculations is shown in Figure 2-17.

S Transmitter

θ1

S

l1

l2

a1

θ2 θ2

a2

θ1 Receiver h

Velocity in Lucite is V1 Velocity in steel is V2

Figure 2-17 Setup for angulated SV wave velocity determination

Zones—How to Keep Them Apart

35

Premises: •

Snell’s law: V 1 sin θ 2 = V 2 sin θ 1

•

d = vt (distance = velocity × time)

•

Symmetry of metal paths is assumed (i.e. a 1 = a 2 = a ).

•

Lucite® distances l1 and l2 may not be equal but both are known.

•

S is the distance between exit points.

•

θ1 is the incident angle and is the same and known for both probes.

Using standard unfocused probes for both transmitter and receiver, a signal is maximized at time T. Time T is the time travelled in l 1 + l 2 + 2a . If l1 and l2 are known distances and the velocity of the longitudinal mode in the Lucite is also known, then the portion of time spent in the Lucite is: l1 + l2 -------------- = t 1 v1

(time in Lucite)

A micrometer can easily and accurately determine the wall thickness h. The separation of the probe exit points is 2S and is determined using the micrometer.

Calculations V2 is required to be found in Snell’s law. V1, θ1, and l1 and l2 are characteristics of the Lucite wedges and are known prior to the test. Parameter h is the thickness determined at the time of the test, and S is also determined (by a micrometer) at the time of the test. The ratio of S to h allows determination of θ2, the refracted angle, which can be inserted into Snell’s law equation to determine V2. S--- = tan θ → θ = tan-1 ⎛ S ---⎞ 2 2 ⎝ h⎠ h ( V 1 sin θ 2 ) V 2 = -----------------------sin θ 1 Alternatively times can be used.

36

Chapter 2

T – t1 = ts Where ts is the time in steel. 2a t s = -----V2 Where a is found from Pythagoras. Therefore: S2 + h2 V 2 = 2-----------------------ts When tabulated, the results show that significant variations can occur from one steel to another and variation also exists within the same sample but varies as a function of direction. Tabulated results for some the eleven samples assessed are seen in Table 2-3. These may be considered typical of the range of velocities encountered in line pipe. Table 2-3 Average acoustic velocities for various modes and directions of propagation (m/s) Mode and Direction of Propagation

Sample Number 1

2

3

4

5

6

7

8

9

10

11

Longitudinal Radial Long. axis Circ. axis 45° and long.

5802 5945 5981 6026

5870 5930 5845 6073

5931 5913 5862 5932

5878 5910 5938 5914

5929 5874 5850 5903

5951 5914 5908 5876

5910 5885 5887 5915

5899 5936 6019 6074

5910 5918 5991 6046

5939 5901 5983 6025

5938 5931 5981 6060

SH slow Radial Long. axis Circ. axis 45° and long.

3191 3169 3099 3171

3183 3129 2977 3156

3182 3197 3232 3240

3176 3164 3270 3225

3148 3230 3205 3184

3189 3176 3251 3235

3170 3170 3278 3180

3184 3151 3145 3187

3195 3149 3186 3189

3210 3221 3184 3210

3206 3208 3176 3232

SH fast Radial Long. axis Circ. axis 45° and long.

3410 3362 3143 3178

3437 3359 3157 3165

3291 3258 3254 3312

3287 3236 3284 3307

3334 3387 3374 3315

3305 3236 3269 3322

3349 3402 3406 3312

3359 3356 3185 3187

3366 3353 3212 3209

3335 3307 3225 3225

3351 3343 3203 3232

SV radial SV at 60°

3367 3252

3385 3248

3227 3177

3215 3280

3289 3339

3220 3303

3303 3333

3325 3245

3330 3245

3265 3298

3293 3287

Zones—How to Keep Them Apart

37

Findings indicated that the SH wave fast velocity provided the closest approximation to the SV wave velocity. Since the accuracy required to obtain several velocity values from three or four angles in the inspection plane would be problematic for an SV wave setup, it was decided that the small samples prepared for the acoustic studies would be the standard required for field preparation and SH wave assessment would be the easiest option for reliable results. When the velocities are plotted in the plane of the longitudinal axis of the pipe (i.e. the axis along which the AUT would be carried out) the velocity variation, with respect to angle, is clearly evident. The plot clearly shows that trends are seen and therefore interpolations are feasible. This means that by plotting the velocities in a plane at three or four angles, the velocities at all the other angles can be reasonably approximated. Figure 2-18 is a polar-type plot of velocities from two of the samples already mentioned. Both axes are in meters per second (m/s), but the distance out from the origin to the arc is indicated along the angle at which the velocity was assessed. The presentation uses the horizontal as the direction along the pipe axis and moves upward to the vertical, which represents the velocities through the wall thickness of the pipe.

38

Chapter 2

Sample 8

Radial axis of pipe 6000

Longitudinal mode

5000

SH shear fast 4000 SH shear slow 3000 SV nominal 60º 2000 1000 0 0

Longitudinal axis of pipe 1000 2000 3000 4000 5000 6000

Sample 10

Radial axis of pipe 6000

Longitudinal mode

5000

SH shear fast 4000 SH shear slow 3000 SV nominal 60º 2000 1000 0 0

Longitudinal axis of pipe 1000 2000 3000 4000 5000 6000

Units along axes are in m/s

Figure 2-18 Velocity representation in the plane of pipe axis

Figure 2-18 compares samples 8 and 10. This comparison shows how differences still exist when the material is from the same manufacturer with the same diameter and seam type but thicknesses are different. This description of velocity variations was used to illustrate the fact that simply setting up on any piece of steel pipe having a nominal wall thickness the same as the project is not adequate to ensure that the zones achieved in the calibration steel would provide the same beam positioning when placed on a the project welded pipe. Zones—How to Keep Them Apart

39

Based on these studies recommendations were made for a tolerance of velocities between calibration and project pipe. This was based on the requirement that beam angle measured on the baseline inspection must not deviate from the nominal by more than ±1.5° for probes with a nominal refracted angle of less than or equal to 45°, and must not deviate from the nominal by more than ±2° for probes with a nominal refracted angle of greater than 45°. For the purpose of establishing tolerances for calibration block velocities, it would be appropriate to ensure the same limits are placed on refracted angles that would result due to velocity differences between the metal and wedge. This will require the operator to know the acoustic velocity of the metal and the acoustic velocity of the plastic used for the wedge along with the incident angle of the beam. Service companies provide a detailed procedure at the outset of each project. In these procedures, refracted angles to be used are illustrated for each calibration target. By selecting the greatest angle to be used a decision can be made as to the acceptability of a given probe on a calibration block. E.g.: The greatest nominal refracted angle used will be 74°, and the longitudinal velocity in the wedge material is 2350 m/s (polystyrene). The polystyrene is machined to provide an incident angle of 44.4° and the wedge angle is machined assuming metal velocity is 3230 m/s. From the information available and a tolerance of ±2° we can calculate the velocity maximum and minimum that the calibration block may have. In Snell’s law, we assign the parameters that are required. ( V 1 sin θ 2 ) V 2 = -----------------------sin θ 1 For the maximum velocity we put θ2 to 76°, θ1 is 44.4°, and V1 is the wedge material velocity of 2350 m/s. It remains to find V2 to see the maximum velocity for the new calibration block. Solution: Max. V2 = 3260 m/s By similar calculations we use 72° as the minimum refracted angle to obtain a minimum steel velocity for this probe. It is found to be 3195 m/s. We also suggested earlier that there are limits to the accuracy we can expect using the techniques described and that was ±20 m/s. For the worst case condition the assessment requirement is +30 m/s and −35 m/s. This is only slightly more 40

Chapter 2

than the accuracy we might expect from our measurements. A simple plus or minus velocity tolerance is not adequate due to the trigonometric relationship that exists. In the example just given, the velocity tolerance window is plus 30 m/s and minus 35 m/s. However, if the maximum angle used was 65° for the same conditions, the variation from the nominal 3230 m/s could be about plus or minus 55 m/s.

2.2.2.2

Temperature Effects on Velocity

The other factor affecting velocity we identified as temperature. It is a well-known fact that the temperature of the propagating material has an effect on the velocity of propagation. The relationship between temperature and velocity is usually quantified as a dV/dT value. By this we mean the rate of velocity change per temperature unit. It is generally considered that the dV/dT is a characteristic property of a material. There are not a lot of tables available on the dV/dT values for materials. Water is probably the best documented material. The CRC Handbook of Chemistry and Physics4 has long included equations for calculating the velocity at a specified temperature by relating it to a factor of the square root of the change in temperature. In the 1969 edition they indicate that the approximation is t V = V 0 1 + --------273 where V0 is the starting velocity and t the temperature in degrees Celsius. But this is an approximation and more suitable to compression waves in air. In the Nondestructive Testing Handbook,5 a table of velocities is provided illustrating changes with temperature and salinity. But references in the literature on dV/dT for solids are few. Alan Selfridge6 published an extensive acoustic properties table in a paper published in the IEEE Journal on Sonics and Ultrasonics, where a few dV/dT values can be found. But metals are not documented there. There is a paper by Chandrasekaran and Salama7 where they found for that for ASTM 533B steel typical compression mode dV/dT was –0.64 m/s/°C. For transverse wave velocities the assessments are much more difficult due to the problems of mounting the SH wave probes that provide the superior assessment compared to mode-converted SV wave. Work by R. Ginzel8 has found that the order of magnitude is similar to that found for the compression mode and reports dV/dT on the order of –0.3 m/s/°C for several steels. Published values of dV/dT for plastics indicate that quite a range can exist.

Zones—How to Keep Them Apart

41

Table 2-4 indicates the sort of range that exists for some of the common plastics. Table 2-4 dV/dT for some common plastics Material

Trade Name

Manufacturer

–dV/dT (m/s/°C)

Polycarbonate

Lexan®

GE

3.58

Polysulfone (PSO)

Polysulfone™ resin

Union Carbide

1.38

Crystal Methyl-acrylate

Lucite®

DuPont

5.07

Polypropylene

Profax™

Hercules, Inc.

9.55

Crystal polystyrene

Styrene 50

Foster Grant

5.44

Silicon rubber

RTV-60

GE

2.7

Acrylonitrilebutadine-styrene (ABS)

Cycolac™

Marbon Chemical Corp.

2.7

Cross-linked polystyrene

Rexolite®

C-Lec Plastics

1.58

Note that the dV/dT value is quoted as a negative item (i.e. –dV/dT). This means that the velocity decreases as the temperature increases. PMMA (polymethylemethacrylate, also called Lucite®) is one of the most commonly used wedge materials. When we compare the dV/dT value for it versus the approximate value for shear waves in steel, we see a factor of difference more than 15. Since the refracted angle at the wedge-steel interface is a function of the velocity ratio, the effect of the temperature ratio changing the velocities in the two materials is not uniform, even if the dV/dT was the same for the materials on both sides of the interface. E.g. if both the steel and the wedge had a dV/dT of –3 m/s/°C and the test was conducted with both at 40°C instead of the 20°C at which standard velocities are determined, the effect would be to reduce the velocities of both by 60 m/s. If we start with the velocity of the shear wave in steel at 3250 m/s and the velocity in the wedge plastic at 2650 m/s, the ratio of the steel-to-wedge velocity is 1.226. Making the temperature 20° warmer, the ratio changes to 42

Chapter 2

3190/2690 or 1.232. This is not much of a change (less than 0.5%); so if the wedge is machined to produce a refracted angle of 60° in steel at 20°C, the ratio difference due to the new temperature would mean it is increased to 60.4°. But the dV/dT difference is not identical in piece and wedge. In fact it is much greater in the wedge. Using typical values of –0.3 m/s/°C for shear wave velocity in steel and –3 m/s/°C for a typical plastic, the nominal refracted angle of 60° at 20°C changes to 62.1° when the temperature is increased to 40°C for the wedge and the piece (steel). If the probe was being used in a Canadian winter, by sitting outside on the back of the truck, the wedge could conceivably get to –40°C as it adjusted to ambient air temperature. The same probe with an angle of 60° would now have a refracted angle of 54.6°, with the wedge being at this low temperature. The effect of temperature on the refracted angle can therefore be significant even over a fairly small change away from the normal room temperature. The chart in Figure 2-19 illustrates the effect of temperature on the refracted angle over a range from –40°C to +50°C for three typical angles. The effects on velocity changes are again seen to be more pronounced for higher angles, e.g. the 90-degree range has only about a difference of 5 degrees for the nominal 45° beam but a 14° difference for the nominal 70° beam. Temperature (°C) –60

–40

–20

0

20

40

60

90

Refracted angle (°)

80 70 60

Nominal 60° Nominal 45° Nominal 70°

50 40 30 20 10 0

Figure 2-19 Refracted angle change with temperature

Zones—How to Keep Them Apart

43

The chart in Figure 2-19 uses the –dV/dT of 0.3 m/s/°C and 3 m/s/°C for steel and wedge plastic respectively. These concerns for velocity effects on refracted angle have led to procedures that require a monitoring of the wedge temperature between the time of calibration to the end of a weld scan. If the wedge temperature changes to a point greater than about 10°C from the calibration, then there is a risk that the beam positions will be outside of their intended zone positions. The same sort of rationalisation was used for temperature control as was used for velocities of calibration blocks, i.e. any change that results in an angle shift of more than 2° from the intended refracted angle is not acceptable. As an example, on a cool autumn day in a northern climate a land-lay project could start off the morning with a calibration run on a calibration block at 5°C. The probes are left on the calibration block between scans so the heat transfer balances the ambient temperature to all components and the block and wedge will be at 5°C for the first calibration of the morning. As the day progresses the probe tends to pick up heat from the warm pipe surface as the scanning keeps up to the welding process. Also, as the morning progresses the sun adds to the heat input of both the calibration block and the wedges. Monitors in the wedge material have indicated that the probe temperature can rise up to 12 to 15 degrees within a one-hour period. This could alter the angle of some beams outside the allowable tolerances as indicated above. A similar problem occurs in an offshore project. There, the pipe is constantly being cooled down by a quench but there is a significant heat transfer from the still very warm pipe that may be at approximately 70°C to 80°C, compared to the ambient temperature usually around 30°C. In both cases the solution has been to incorporate a warming pad under the calibration block. The function of this heating pad is not to match the metal temperature of the pipe weld being inspected, but instead to supply a heat source to transfer some heat to the probes while waiting on the calibration block between weld scans and thereby maintain the same velocity in the wedge during calibrations as during inspections. Some users have attempted to facilitate a wedge temperature control using a warm water couplant flow. If the temperature is closely regulated it may present an effective option to or addition to the heating pad. But if the heated water has thermal excursions the benefits are quickly lost. As well, the effects of heat transfer from pipe to wedge can be relatively subtle if no other heat is introduced from another direction. Simply altering the velocity of sound in the layers below the probe element will result in slight adjustment of only the exit point. However, if heat is added from the sides of the wedge, beam skew may be unpredictable. Figures 2-20a through 2-20c show the effects of changing polymer material 44

Chapter 2

temperatures. In Figure 2-20a a simple refracting wedge is used over a block of steel. The wedge is modelled with a velocity of 2350 m/s. Figure 2-20b illustrates the effect of a graduated change in velocity by inserting two layers. As in Figure 2-20a the pulse enters the wedge with a velocity of 2350 m/s but as it approaches the warmer surface of the pipe the temperature has begun to slow the sound and a layer at 2300 m/s is shown. A change of 50 m/s might be typical of a 10-degree change in temperature in a material with a dV/dT of 5 m/s/C°. A second layer at 2250 m/s is inserted below the first indicating the effect of a 20° change compared to the temperature at the probe. The net effect of the thermal gradient is seen to only alter the exit point where the beam enters the steel. However, the effect of adding heat to the sides of the probe wedge, as would be the case with a heated coupling flow, can be more deleterious to the intended beam shape if the effect is not uniform across the wedge. Figure 2-20c shows that effect by inserting a layer of plastic at 2250 m/s on either side of the wedge so that a portion of the beam, located directly under the probe, sees the change in velocity immediately upon entering the wedge. As the beam formation is retarded on the sides, the intended focusing effects are lost and the distribution of the beam is seen to deviate from its intended spot. a

Zones—How to Keep Them Apart

45

b

c

Figure 2-20 Effects of changing polymer material temperatures

An important factor in all this is the “intent” of the temperature control. If the regulating standard indicates that the temperature is to be monitored, it should be understood why the monitoring is required. We have explained here that the reason is to provide the operator with a reasonable working window in which the wedge velocity will not too significantly affect the refracted angle. Assessing the appropriate angles (wedge incident angle for a desired refracted angle) requires knowledge of the velocities. Since one must determine the acoustic velocity of the plastic used for the wedge, the dV/dT of the wedge materials should also be determined. Then the requirement for temperature measurement is obviously an indirect determination of velocity. Therefore, if a solution to this concern was to be able to measure the velocity directly, then the intent of the standard would also be met. At the time of writing this handbook none of the main users is making a direct velocity

46

Chapter 2

measurement of the plastic but it should not be ruled out as a viable solution to the parameter control concern. Attenuation is also affected by temperature. This has of yet not been quantitatively assessed for our applications in AUT. However, until attenuation is quantified with respect to temperature, the requirements to work within a small range close to the calibration conditions should provide adequate assurance of amplitude stability. The easiest way to assess the effects of temperature on the amplitude stability is to monitor the amplitude of calibration zone targets over the tolerance range in temperature. It would be difficult to separate the effects of temperature and angle change on the amplitude response under such conditions but results9 indicate that even the combined effects of attenuation and angle change due to temperature change allow calibration target signals to be maintained within 1 dB or 2 dB if the temperature is held to within ±10 °C.

Zones—How to Keep Them Apart

47

References to Chapter 2

48

1.

Silk, M. “Relationships Between Metallurgical Texture and Ultrasonic Propagation,” Metal Science, 15 (1981).

2.

Schneider, E., K. Goebbles, G. Hübschen, and H. J. Salzburger. “Determination of Residual Stress by Time-of-Flight Measurements with Linear Polarized Shear Waves.” 1981 IEEE Ultrasonics Symposium Proceedings, 1981.

3.

Ginzel, E. and R. Ginzel. “Study of Acoustic Velocity Variations in Line Pipe Steel.” ASNT. Materials Evaluation, vol. 53, no. 5 (1995).

4.

Lide, David R., ed. CRC Handbook of Chemistry and Physics 1999-2000: A Ready-Reference Book of Chemical and Physical Data (CRC Handbook of Chemistry and Physics, 80th ed.).

5.

American Society for Nondestructive Testing. Nondestructive Testing Handbook. 2nd edition. Vol. 7, Ultrasonic Testing. Columbus, OH: American Society for Nondestructive Testing, 1991.

6.

Selfridge, A. R. Approximate Materials Properties in Isotropic Materials, IEEE Transactions on Sonics and Ultrasonics, vol. SU-32, no. 3 (May 1985).

7.

Chandrasekaran and Salama (University of Texas). In Nondestructive Methods for Material Property Determination. Edited by Ruud and Green. New York: Plenum Press, 1983.

8.

Ginzel, R. Private communications.

9.

Stewart, D. Private communications.

Chapter 2