Proceedings

ProceedingsofofPVP2005:
PVP2005
2005ASME
ASMEPressure
Pressure Vessels
Vessels &
andPiping
PipingDivision
DivisionConference
Conference
July
July17-21,
17-21,2005,
2005,Denver,
Denver,Colorado
Colorado USA
USA

PVP2005-71408
PVP2005-71408

NOVEL ULTRASONIC TESTING OF COMPLEX WELDS

Barry Messer, Jose R. Fuentes, Bart Tarleton, Fluor Canada, Ltd., Calgary, Alberta, Canada

Peter den Boer, RTD Corp., Edmonton, Alberta, Canada

ABSTRACT precautions must be taken. Furthermore, in cases where RT is

Fluor and RTD have recently applied an ultrasonic testing used to inspect welds, a delay of 24 to 48 hours occurs before
(UT) technique that incorporates a phased array to field and interpretation results are obtained. However, in these cases
shop operations. The new technique allows verification of weld improper welding practices may not be effectively remedied as
integrity for difficult to access welds such as branch connection production welding is carried forward during the film
fittings and full penetration groove welds with fillet developing time. Although ultrasonic testing (UT) with contact
reinforcements. Verification of these types of welds is a or immersion probes has been used for weld inspection, as
necessity for power, oil and gas facilities, in particular, those referred to in this work as conventional UT, several limitations
operating under high pressure, high temperature, and corrosive occur with this method. Conventional UT, for example, is
environments. Historically, visual inspections of welds and limited to use on homogeneous parts with simple geometries
radiograph testing (RT) have been used, but these methods are and smooth surfaces. The accuracy of this method is also
costly, time-consuming, and cannot match the benefits of the limited by the experience and knowledge of the operator.
new UT phased array (UT-PA) method. The UT-PA technique Weld integrity for difficult to access areas such a branch
has an arrangement of multiple piezoelectric elements that are connection fittings and groove welds can now be NDE tested
independently controlled for developing synchronized and with a new application of ultrasonic testing using a phased
manageable sonic waves. The technique requires less time than array (UT-PA) technique. When inspecting carbon, alloy, and
conventional UT, is not hazardous as compared to RT, and stainless steels, the UT-PA method provides an efficient means
allows for 100% volumetric inspection. Other advantages of of focusing on potential problem areas and scanning 100% of
UT-PA include its ease of use, increased accuracy, and the weld volume. This technique offers a weld-and-check
development of instantaneous digital inspection records for methodology with instantaneous digital results and thus,
tracking defect propagations in the future. The present work reduces the need for costly repairs. The present work describes
describes the application of this nondestructive examination the basic operation of the UT-PA method and its field
(NDE) technique to a branched connection of an ASTM B564 application to the inspection of cast stainless steels. A
outlet fitting to both an ASTM A608 modified 20Cr-32Ni-Nb comparison of the UT-PA technique with other examination
statically cast header and an HP45 modified tee. An outline of methods is also discussed.
the advantages for the UT-PA method is also included which
explains the rationale that, in the future, will cause the welding BACKGROUND
industry to rely more on modified UT advanced imaging.
Conventional UT
KEYWORDS Austenitic stainless steel weld areas are normally very
NDE, Phased Array, Ultrasonics, UT-PA, Weld Inspection difficult to inspect with conventional UT due to the beam
distortion and scattering effects of their typically large
INTRODUCTION anisotropic grain structure. These grains are generally
In the past, nondestructive examination (NDE) testing of characterized by course columnar shapes and provide numerous
very difficult to access areas required visual checks or costly interfaces with a range of orientations. A variation in the sound
examination methods that subjected the users to various velocities amongst the crystals results in several adverse effects
hazards (e.g., radiation, chemicals, etc.). Commonly, the commonly associated with conventional UT. These effects are
manufacturing industry has used radiography testing (RT) to described as follows:
inspect welds. This examination method requires the use of a 1. Signal scattering is the deflection of small amounts of
dangerous radioactive source such that expensive safety acoustic energy out of the main ultrasonic beam. The

1 Copyright © 2005 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/conferences/pvp2005/71770/ on 05/11/2017 Terms of Use: http://www.asme.org/ab

 deflection is the result of interactions between the sound an infinite number of individual point sources, shifted in time
wave and discontinuities in the material such as grain or space. Phased array probes use an array of piezoelectric
boundaries, inclusions, and defects. Scattering is highly elements rather than just one. The elements are contained in a
dependent on the relation between grain size and ultrasonic signal probe; however, with computer controls the individual
wavelength [1]. Both attenuation and beam scattering are elements can be manipulated and synchronized to produce a
well established problems when using UT to inspect stainless focused ultrasonic beam with steering capabilities as shown in
steel welds. Figure 2. These multi-element probes are able to adjust their
2. Mode conversion is an effect that is common in ultrasonics focal point and steer the ultrasonic beam to detect and size
and typically occurs when the ultrasonic beam strikes an cracks of most orientations and depths [3]. With electronically
interface at an oblique angle. The interface typically controlled ultrasonic sound waves a wider inspection range can
separates two materials with differing acoustic velocities. be developed, thereby allowing 100% volumetric inspection.
When the beam impinges on the interface the beam is split Focusing of the beam also enhances space resolution with
into reflected and refracted beams consisting of different better sizing and mapping characteristics. Furthermore, these
modes and wave classifications (i.e. longitudinal, transverse, expanded capabilities of UT-PA can be performed in a fraction
and surface waves). Splitting the incident beam reduces its of the time required when using conventional UT methods. A
strength and produces undesired reflections that can create thorough comparison of UT-PA with other testing methods is
erroneous UT indications. shown in Table 1.
3. Attenuation is the absorption of the sound energy as it
passes through the material, converting the sound energy into
heat. When the sound is absorbed, the signal-to-noise ratio is
reduced making it difficult to distinguish the signal from the
background noise. The anisotropic characteristics of stainless
steels that produce mode conversion can also contribute to
beam distortion, which causes attenuation and scattering of
the ultrasonic beam as it moves through the material [2].
These effects can be addressed and minimized by using the
proper probes and techniques, but generally are done at the
expense of accuracy and cost. For example, the effects of mode Figure 1: Differing detection capabilities of single-element
conversion and beam distortion are reduced by avoiding the use (left) and multi-element (right) probes.
of shear wave probes and using longitudinal probes instead.
However, when using longitudinal probes it is important to note
that longitudinal signals lose considerable energy when
reflected from an internal surface of a component and are
normally limited to a half skip. The consequences of
attenuation can be countered by using lower frequency probes,
but this is done at the expense of reduced sensitivity and
resolution. Low signal-to-noise ratios created by scattering can
be alleviated by using focusing probes. Although, it should be
noted that if focused beams are to be used with standard twin
crystal probes, the inspection can be a time consuming process
requiring many probes of differing angles and focal points.
Should conventional ultrasonic inspection be used for weld
areas, using historical techniques and equipment, it requires
highly skilled and experienced personnel and has several
limiting disadvantages. As indicated by Figure 1, a loss of
detection sensitivity can occur if the contact transducer is not
perfectly aligned to the specimen defect. This is because a
single-element probe has a fixed focal point and is limited to
detecting flaws with orientations that provide a sufficient back
reflection. Conventional UT inspection times are also
considered proportional to weld section thicknesses, requiring
more time for thicker specimens. Phased array ultrasonics
offers an alternative to conventional UT that provides most of
the benefits of ultrasonics without many of its limiting
drawbacks.

UT-Phased Array NDE Technology

Phased array ultrasonic inspection is an extension of
conventional ultrasonic inspection. The design of phased array
technology is modeled on the Huygens principle. According to
this principle, each wave front can be formed by the addition of Figure 2: Beam steering and focusing capabilities of phased
array probes.

2 Copyright © 2005 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/conferences/pvp2005/71770/ on 05/11/2017 Terms of Use: http://www.asme.org/ab

 Table 1: Comparison summary of various nondestructive examination methods for subsurface flaw detection.

Characteristic RT Conventional UT UT-TOFD(1) UT-PA

Very good detection Very good detection

Detection Capabilities No vertical sizing Limited vertical sizing
and sizing and sizing
Some limitations at
Coverage Full Full Full
ID/OD surfaces

Reproducibility Good May vary with operator Good Good

Planar Limited Good Good Good

Defect Type
Volumetric Good Limited Good Average

Limited to t ≤ 50.8 mm Limited on Less suitable for coarse

Material limitations Generally no limitations
(Ir. 192) austenitic materials grained materials and clad

Data Presentation Top view None Side view Top, side, end view

Generally limited Generally limited Can be adopted to fit

Geometric Considerations Generally, no limitations
to butt welds to butt welds geometric conditions

Probability of Detection(2) 66% 52% 82% 89%(3)

Well trained operators Well trained operators

Operator Experience Qualified technicians Qualified technicians
are essential are essential
Yes, to some degree No, due to No, due to
Operator Dependence Yes, based on experience
during interpretation digital archiving digital archiving
(1)
UT-TOFD refers to ultrasonic testing with time-of-flight-diffraction.
(2)
Probability of Detection (POD) results are from a Dutch Institute of Welding survey [4].
(3)
Estimated from related work.

FIELD APPLICATION

Site Situation
The new UT-PA technique was used on a large scale
refinery project in Fort McMurray, Alberta, Canada. For this
case, accurate and fast verification results for the reinforced
fittings were critical to the success of the project. A single
welded connection is shown in Figure 3. The fittings of the
equipment and materials under inspection consisted of:
• ASTM B564 MSS (UNS N08811) MSS SP-97 branch
outlet fittings with bevel design and dimensions, according to
ASME B16.25: ID = 28.5 mm, ODMAX = 89.5 mm, and
average wall thickness t = 30.5 mm.
• ASTM A608 modified 20Cr-32Ni-Nb pipe headers with an
ID = 279.4 mm and wall thickness t = 38.1 mm.
• HP45 modified tees.
At site, there were considerable challenges in examining
weld integrity for the branch connections. These welds were in Figure 3: One of the 624 branch connections to be inspected
locations where pipe configurations made examination by using UT-PA.
traditional RT methods very difficult and time consuming. In
addition to these challenges, poor results from sample connected to a branch pipe by butt welding, socket welding,
radiograph tests motivated a search for an alternative threaded, or flanged joint, including a branch outlet fitting
examination method that could be applied to the welded conforming to MSS SP-97 [5]. For RT examination, this meant
connections. that acceptance criteria for the welds were not allowed to have
ASME B31.3, Paragraph 300.2 defines branch connections any cracks, lack of fusion, incomplete penetration, internal
as an integrally reinforced fitting welded to a run pipe and

3 Copyright © 2005 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/conferences/pvp2005/71770/ on 05/11/2017 Terms of Use: http://www.asme.org/ab

 porosity, exposed slag inclusions, undercutting, or concave root suitable for most stainless steel applications. The UT-PA
surfaces. requires only a few days of training before on-screen
During these RT sample trials, a total of 25 branch outlet interpretations of results can be made.
fittings were tested with the double wall exposure technique. With conventional UT, the selected probe angle should be
Six or 24% were rejected per ASME B31.3, Table 341.3.2 suitable for the expected defect orientation. This means that the
Category M fluid under the column Girth, Miter Groove & incident angle of the ultrasonic beam should strike a defect
Branch Connection. Challenges associated with doing this type perpendicular for maximum echo amplitude. However,
of double wall exposure was related to the extensive time and determining the proper angle with conventional UT is difficult
effort involved; approximately 6 hours to examine a single when flaw orientation is not known before hand and, as a result,
connection. With 624 branch connections to inspect, RT the probe must be maneuvered about the part surface. With
methods would have required more than five months to phased array technology, the angle of the beam can be varied
complete with crews working non-stop. The phased array
technique proved to be a highly desirable alternative by
accomplishing the inspection in ten working days.
Another RT examination employed the single wall
exposure technique. This technique was permissible at certain
locations due to the connection configuration and piping wall
size. Results of these examinations showed a total of 29
Movement of Probe
rejected connections out of 45 inspected for a 64.4% rejection
rate. The geometric unsharpness factor for these welds,
however, was not compliant to ASME Section V Article 2,
Paragraph T-274.2 for material thickness less than 50 mm [6].

Suitability of UT-PA
There were two main issues to consider for use of UT-PA.
One was the compliance of the UT-PA technique to Code and
the other was its applicability for use with tools currently
available for the welding configurations.
The design conditions involved pressures at 485 psig and
temperatures at 1625°F (885°C). Design was based on API 560 Two Side Drilled Holes
Fired Heaters for General Refinery Service and ASME B31.3
Process Piping. Acceptance criteria for these welds using RT
fell under ASME B31.3, Table 341.3.2. The ASME Section
VIII Code Case 2235-6 allowed use of UT in lieu of RT for
welds in material 12.7 mm or greater in thickness [7]. This case
restricts UT use to the conformance of ASME Section V, which
references UT-PA in Paragraphs T-452 and E-474.
Furthermore, in cases where weldment geometries or materials
prohibit conventional UT, the statements of API-560 and Figure 4: Sample connection used as a “calibration block” for
ASME B31.3 support UT-PA. API-560 Code, Paragraph the UT-PA equipment.
14.2.2.7 states: In cases where weld or material configuration
makes radiographic examination difficult to interpret or
impossible to perform, such as nozzles welds, ultrasonic
examination may be substituted [8]. Additionally ASME B31.3,
Paragraph 341.5.3, Examinations to Resolve Uncertainty, states
any method may be used to resolve doubtful indications [5].
With these supporting provisions, the issue UT-PA compliance
to Code is satisfied. Complying with the second issue of
application of UT-PA to instruments currently available in the
welding industry was done through the calibration procedure.

Calibration
To calibrate the UT-PA device, actual reinforced outlet
fitting samples were taken from the site facilities and used as
calibration blocks. Figure 4 illustrates the use of these samples
for calibration. The calibration piece with side-drilled holes was
prepared from a welded branch to tee connection sample. This
sample was representative of all the connections examined, in
terms of weld preparation, process, and heat treatments. As
shown in Figure 5, an R/D Tech Omniscan instrument was used
Figure 5: Sample welded connection used to calibrate signal
with a 16-element probe at 5.0 MHz and an Ultra Gel couplant,
with a 3.2 mm diameter drilled hole.

4 Copyright © 2005 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/conferences/pvp2005/71770/ on 05/11/2017 Terms of Use: http://www.asme.org/ab

 such that the part can be scanned with the probe held at a fixed 10.2mm1. This equals to a 10 mm length for field applications
position on the part surface. For this situation, the weld was where there is a 0.2 mm measurement uncertainty.
scanned from the flat surface of the welded outlet fitting as A trial correlation of RT versus UT-PA was established in
shown in Figures 4 and 5. a sample of 25 branched connections. This resulted in a 100%
Commonly, only transverse (shear) wavelengths are used correlation of signal or, in some cases, an indication providing
for weld inspection. However due the uniqueness of this evidence that UT-PA located more flaws than RT. When a
inspection and the issue of coverage, the client and engineering strict code criterion was applied to these 25 connections, UT-
personnel in the project agreed to the use of longitudinal PA rejected 8% more than RT (28% for UT-PA in comparison
(straight) wavelengths such that the required sectional scan to 20% for RT). It should be noted that UT-PA rejection was
width could be achieved. Transverse wavelengths have about based on flaw length, as specified by ASME B31.3, Paragraph
half the wavelength of longitudinal waves and can not be 344.6.2; whereas RT rejection was based on defect length and
spread open to such a wide coverage; 30º to 35º at the most. width, as indicated by ASME B31.3, Table 344.3.2. Therefore,
Using longitudinal wavelengths allows for a twice as wide a decision was made to continue testing all the other welds with
sectorial coverage of 20º to 75º as illustrated in Figure 4. The the UT-PA method. Data samples of the correlation tests for the
scan was programmed in 1º increments of the ultrasonic beam. UT-PA and RT methods are shown in Figures 7 and 8.
Figure 6 is a cross-section from a welded connection Analysis of the correlation between UT-PA and RT
sample with defects demonstrating that the method effectively provided evidence of the efficiency and accuracy of phased
can detect flaws for these challenging examination array based on locating and sizing of flaws. Table 2
environments. The weld material was determined to have an summarizes the characteristics of both the UT-PA and RT
average grain size No. 8 based on ASTM E-112 [9]. The pieces nondestructive examination methods used in this case study.
were cut perpendicular to the weld axis to show the structure
and the weld fusion faces. The holes were respectively drilled UT-PA Results
at the fusion zones to compensate for the effects of the weld Of the 624 connections tested, the UT-PA method
structure. indicated a 20.5% rejection rate with a total of 128 connections
The sensitivity is mainly determined by the size of the failing to meet acceptance criteria. There were 97 outlet fittings
side-drilled holes, which were 3.2 mm in diameter per ASME on branches and 31 on tee locations. As a result of these
B31.3 requirements for a weld thickness of 30.5 mm. findings, 94 (15.1%) connections were replaced and 46 (7.4%)
Sensitivity calibration was carried out on the 3.2 mm diameter were repaired.
holes, scanning with an extra 6 dB of sensitivity in order to size In addition to the tested and identified connections with
the indications. For evaluation of indications and data defects warranting rejection, a complete scan was requested and
collection, the extra 6 dB was removed. performed on all remaining connections. Only rejected
connections were repaired or replaced. There were a total of
442 outlet fittings with lengths under 10 mm that were not
rejected. These identified indications will be monitored in the
future using the same UT-PA technique. At that time, the future
ultrasonic results will be cross checked with the records
obtained during this UT-PA examination. Therefore, a behavior
or propagation of defects can be identified and monitored, with
actions taken accordingly for repairs or replacements.

ADVANTAGES & BENEFITS

The advantages and benefits of the UT-PA method as
discussed in this work can be summarized as follows:
• The UT-PA technology allows for ultrasonic beam
steerage and focus to cover the weld area from a limited
scanning surface. In this study, this was done from 45°
degree tapered area of a reinforced branched connection.
• A 100% volumetric assessment can be obtained.
• The digital outputs provide three-dimensional cross
sectional images of defect indications and permanent records
Figure 6: Cross-section of branched connection showing actual
of collected data for future reference.
welding defects.
• Faster results relative to RT which requires considerable
developing and exposure times, especially for thick walls.
Correlation of UT-PA versus RT
According to the ultrasonic acceptance criteria of the 1
The 10.2 mm length comes from the letter “F” = Tw/3 where Tw is the
ASME B31.3, Paragraph 344.6.2, for this situation a linear-type nominal wall thickness of the thinnest component joined by the weld. For
discontinuity was unacceptable if the amplitude of the linear-type discontinuities, if the amplitude of the indication exceeds the
indication exceeded a reference level and its length was above reference level and its length exceeds Tw/3 for thickness between 19 to 57 mm
(actual thickness considered was 30.5 mm), the sample will be considered a
reject. Should ASME Section VIII Division 1 have been taken as the main code
for the acceptance criteria, then the criteria would have been the same as ASME
B31.3 [10].

5 Copyright © 2005 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/conferences/pvp2005/71770/ on 05/11/2017 Terms of Use: http://www.asme.org/ab

 Table 2: Comparison of the UT-PA and RT methods used for field site inspections.

RT
Characteristic UT-PA
Single-wall Double-wall

Inspection Time for

1 Hour 6 Hour 10 min
Single Connection
N/A, Since it did not
Total Inspection Time 5 Months (estimated) 10 Days (actual)
comply with code
Day after Day after
Delay Time for Results Immediate
(or minimum developing time) (or minimum developing time)
Impact to Surrounding Work Total isolation due to Total isolation due to
None
Area radiation hazard radiation hazard

Code Compliant No Yes Yes

Greater than 4× Greater than 4× Less than 0.25×

Cost
the cost of UT the cost of UT the cost of RT

Coverage & Resolution Poor resolution of weld root Difficult to interpret 100% Clear volumetric coverage

RT film & NDE RT film & NDE CD with Color Images &
Proof of Test
Level III Report Level III Report NDE Level III Report

Figure 7: Flaw indication on UT-PA output screen (left) and the same defect as indicated by a single wall RT shot (right).

Figure 8: Left: Flaw indication on UT-PA output screen (left) and the same defect as indicated by a double wall RT shot (right).

6 Copyright © 2005 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/conferences/pvp2005/71770/ on 05/11/2017 Terms of Use: http://www.asme.org/ab

 • Technique is not hazardous to persons conducting the REFERENCES
examination and thus, requires less downtime with less 1. Bar-Cohen, Y. & Mal, A. K., 1989, “Ultrasonic
impact on the working area. Inspection,” Nondestructive Evaluation and Quality
• UT-PA achieves excellent testing results in carbon, alloy, Control, ASM Handbook, Formerly Ninth Edition, 17, pp.
and stainless steels. 231-277.
• General user friendliness and ease of interpreting results. 2. AWS, 1986, “Handbook on the Ultrasonic Examination of
Austenitic Welds,” The International Institute of Welding,
CONCLUSION pp. 5-44, American Welding Society Publication, USA.
Ultrasonic inspection of complex geometry components 3. R/D Tech inc., 2004, Introduction to Phased Array
and austenitic stainless steels with conventional UT methods Ultrasonic Technology Applications, Canada.
can lead to limited performances. Similarly, other NDE 4. Stelwagen, U., 1996, “NIL Project: Non Destructive
methods have several drawbacks when applied to irregular Testing of Thin Plate, Part End Report,” Netherlands
parts with anisotropic grain structure. Phased-array technology Institute of Welding, Netherlands.
allows for the UT inspection of these materials without 5. ASME, 2002, “Process Piping ASME B31.3-2002”,
sacrificing the benefits of conventional UT. American Society of Mechanical Engineers, NY, USA.
On a field project case study with 624 complex connection 6. ASME, 2004, “ASME 2004 Section V Nondestructive
examinations, UT-PA revealed that 128 (20.5%) failed to meet Examination”, American Society of Mechanical Engineers,
ASME B31.3 acceptance criteria. The materials inspected NY, USA.
consisted of a cast modified chromium-iron-nickel alloy, 7. ASME, 2003, “Case 2235-6 Supplement #10: Cases of
traditionally not suitable for ultrasonic examination. UT-PA ASME Boiler and Pressure Vessel Code”, American
also provided the client with a method of tracking over 400 Society of Mechanical Engineers, NY, USA.
non-rejected welds for future preventative inspections. 8. MSS, 1987, “MSS SP-97-2001 Integrally Reinforced
In summary, UT-PA has proven to be a reliable and safe Forged Branch Outlet Fittings – Socket Welding,
substitute for RT. The UT-PA technique provides detailed Threaded, and Butt welding Ends”, Manufacturer
results with an excellent presentation of permanent records that Standardization Society of the Valve and Fitting Industry,
are easy to read and interpret. There are minimal costs Inc.
associated with training and code compliance can be achieved 9. ASTM, 2004, “ASTM E-112 Standard Test Methods for
with greater certainty than with other inspection methods. Determining Average Grain Size”, American Society for
Further, its ease of use allows for a weld-and-check Testing and Materials.
methodology where crews can verify joints and repairs shortly 10. ASME, 2004, “ASME 2004 Section VIII Rules for
after welding. Construction of Pressure Vessels Division 1”, American
Society of Mechanical Engineers, NY, USA.

7 Copyright © 2005 by ASME

Downloaded From: http://proceedings.asmedigitalcollection.asme.org/pdfaccess.ashx?url=/data/conferences/pvp2005/71770/ on 05/11/2017 Terms of Use: http://www.asme.org/ab