8/13/2019 hased Array Ultrasonic Testing, Sectorial scan
1/7
Insight Vol 48 No 4 April 2006 1
Introduction
Ultrasonic phased arrays are a new technology which offers new
scan patterns and inspection procedures as well as emulations of
current ultrasonic procedures. Phased arrays have major advantages
over conventional ultrasonics: beams can be swept, steered and
focused. These advantages have been documented in the past(1).
However, phased array probes also offer some technical advantages
for crack sizing that have not yet been well-documented: variable
apertures for specic focusing applications, highly damped piezo-
composite arrays, frequency improvements, as well as controlled
focusing.
This paper describes the resolving capabilities of phased
arrays compared to established techniques. Specifically, the
paper investigates some resolution aspects of defect sizing using
backscatter tip diffraction techniques. Phased array probes are
particularly useful for this sizing technique since(2):
q No probe movement occurs; any slight skewing during scanning
may affect the image and results obtained;
q Coupling tends to be more constant;
q The variable angles of refraction from S-scans aid detection,
since optimum tip signals do not always correlate with xed
angle inspections; and
q S-scans allow the operator to image the corner reector and
the back-scattered tip diffracted signal at the same time. Thetwo biggest problems with back-scattered tip diffraction are
correctly identifying the tip signal from noise, and low signal-
to-noise ratio.
S-scans are the inspection procedure used in this study. An
S-scan is also termed a sectorial-scan, sector scan, swept angle
scan, or azimuthal scan; this may refer to either the beam movement
or the data display. As a data display, it is a 2D view of all A-scans
from a specific set of elements corrected for delay and refracted
angle. When used to refer to the beam movement, it refers to the
set of focal laws that sweep through a defined range of angles
using the same set of elements. Besides S-scans, phased arrays can
perform ASME raster-type electronic scans (for example E-scans at
45, 60 and/or 70), TOFD and tandem inspections; though thesescans are not the subject of this study, the same resolving capability
issues apply to them.
After an initial description of phased arrays, the paper describes
the key parameters which dictate resolution, followed by some
S-scan screen displays of typical defect sizing applications using
back-scattered tip diffracted signals.
Industrial phased arrays
Phased array probes use an array of elements, all individually
wired, pulsed and time-shifted on both pulsing and receiving.
These elements are typically pulsed in groups of ~16 and up to 32
elements at a time for weld inspections. With user-friendly systems,
a typical set-up calculates the time-delays from operator-input, oruses a pre-dened le calculated for the inspection angle, focal
distance, scan pattern etc, as illustrated in Figure 1. The time delay
values are back-calculated using time-of-ight from the focal spot,
and the scan assembled from individual Focal Laws. Time delay
circuits must be accurate to around two nanoseconds to provide the
accuracy required.
Due to the limited market, complexity, software requirements
and manufacturing problems, industrial uses of phased arrays have
been limited until the last few years, but are now becoming more
prevalent with phased array developments and training. Codes are
beginning to recognise the use and applications of phased arrays.
From a practical viewpoint, ultrasonic phased arrays are
primarily a method of generating and receiving ultrasound (though
with major imaging advantages). Consequently, many of the detailsof ultrasonic inspection remain unchanged; for example, if 5.0 MHz
is the optimum inspection frequency with conventional ultrasonics,
PHASED ARRAYS
Resolving capabilities of phased array sectorial scans
(S-scans) on diffracted tip signals
J M Davis and M Moles
This paper demonstrates the signicant improvementavailable with phased array Sectorial (S-) scans forresolving crack tip signals using back-scattered diffractionon thinner plates. Initially, phased arrays and the factorsaffecting resolution are described, such as aperture and
focusing. Another key feature is the use of piezo-compositearray probes, which have shorter pulses. A third factor isS-scan imaging, which signicantly improves crack tipsignal identication. Some experiments show the resolvingcapabilities of phased arrays.
Mark Davis is a graduate of the US Navy NDT Programme and is an ASNT
UT Level III with over 30 years of experience in Welding, Quality Assurance
and NDT. He has held Level IIIs in MT, PT, RT, ET, LT and VT. Davis has
taught ultrasonics for over 20 years, specialising in IGSCC detection, crack
sizing and weld overlay examinations. Formerly, Davis was the Inservice
Inspection Training Manager at the EPRI NDE Center in Charlotte, NC.
Currently, Davis is the Programme Administrator for the American
Petroleum Institute QUTE Shear Wave Examination, as well as the Crack
Sizing and Tank Bottom Thickness Programme. He has authored three
mathematic handbooks on UT, RT and ET, with one handbook on Advanced
Ultrasonic Crack Sizing Applications.
Davis NDE is an approved Training Organisation for Olympus NDT onadvanced phased array training for detection and sizing applications.
J Mark Davis is at Davis NDE, 4060 Bent River Lane, Birmingham, AL,
35216, USA. Tel: +1 (205) 733-0404; E-mail: [email protected]
Michael Moles is in Market Development with Olympus NDT. He has a
PhD and an MBA, and has worked in the eld of automated ultrasonics
for twenty-ve years. He was employed in the Canadian nuclear industry
for sixteen years, and has worked in petrochemical, aerospace and
manufacturing for a decade.
Michael has presented and published over one hundred papers on
ultrasonics, especially phased arrays, and edited the R/D Tech Introduction
to Phased Array Ultrasonic Technology Applications.
Michael is a member of ASME, ASNT, AWS, CINDE, and is a registered
engineer in Ontario.
Contact address: Michael Moles, Olympus NDT, 73 Superior Avenue,
Toronto, Ontario, M8V 2M7, Canada. Tel: +1 (416) 831 4428; E-mail:
8/13/2019 hased Array Ultrasonic Testing, Sectorial scan
2/7
2 Insight Vol 48 No 4 April 2006
then phased arrays would typically start with the same frequency,
focal length, and incident angle.
While it can be time-consuming to prepare the first set-up, theinformation is recorded in a file and only takes seconds to re-load.
Also, modifying a prepared set-up is quick in comparison with
physically adjusting conventional transducers. Using electronic
pulsing and receiving provides significant opportunities for a
variety of scan patterns.
Electronic scans (E-scans)
Multiplexing along an array generates electronic scans (see
Figure 2). Typical arrays have up to 128 elements, pulsed in groups
of 8 to 16. E-scanning permits rapid coverage with a tight focal
spot. If the array is at and linear, then the scan pattern is a simple
B-scan. If the array is curved, then the scan pattern will be curved.
E-scans are straightforward to program. For example, a phased
array can be readily programmed to inspect a weld using both
45, 60 and 70 shear waves, which mimic conventional manual
inspections or automated raster scans.
Sectorial (S-) scans
S-scans use a xed set of elements, but alter the time delays
to sweep the beam through a series of angles. Generally, this
is a straightforward scan to program. One application for
S-scans involves a stationary array, sweeping across a relatively
inaccessible component like a turbine blade root(3), to map out the
features and defects. Other applications (as in this paper) involve
scanning defects to image the corner reector, crack tips, facets,
crack branches etc. Depending primarily on the array frequency,
element spacing and wedge design, the sweep angles can vary from
+ 20 up to + 80.
S-scans are unique to phased arrays, and offer excellent imaging
and data interpretation with improved resolving power. For example,
Figure 3 shows a 10% ID-connected defect with tip signal clearlydetectable on a 9.5 mm plate, and Figure 4 a 10% OD-connected
defect in a 9.5 mm sizing bar. These images are true depth; all
refracted angles, for example 45 to 60, are focused and displayed
at the same depth on the screen. It is straightforward and quick
to analyse these particular S-scan images, based on knowledge
of the component thickness, because the signals are clean and
clear. Consequently, there is a big demand to use S-scans for weld
inspections.
Combined scans
Phased arrays permit combining electronic scanning, sectorial
scanning and precision focusing to give a practical combination
of displays. Optimum angles can be selected for welds and other
Figure 1. Schematic showing generation of electronic, sectorial and dynamic depth focusing scans using phased arrays
Figure 2. Schematic illustration of electronic scanning
Figure 3. A-scan and S-scan showing 10% ID-connected crackwith diffraction tip signal (arrowed)
Figure 4. A-scan and S-scan showing 10% OD-connected crackwith diffraction tip signal (arrowed)
8/13/2019 hased Array Ultrasonic Testing, Sectorial scan
3/7
Insight Vol 48 No 4 April 2006 3
components, while electronic scanning permits fast and functional
inspections. This introduces the concept of tailored inspections to
optimise detection, sizing and inspection time. The best example
of this approach is ASTM E-1961 for AUT of girth welds in
pipelines(4).
Key issues in phased array defect resolution
Array design features
Array design is a key issue, which dictates beam steering, focusingand sweeping. The key parameters are shown in Figure 5. These
factors are also described in the literature(5).
The key probe parameters are:
q Total number of elements in array (n).
qFrequency (f).
q Total aperture in steering or active direction (A).
q Height or elevation, aperture in mechanical or passive direction
(H).
q Width of an individual element (e). Typical element widths
range from 0.3 to 1 mm.
q Pitch, centre-to-centre distance between two successive
elements (p).
q Saw cut gap, or kerf, between two elements (g). Gaps are
typically ~0.05 mm.
Note:
1. The total number of elements in the array primarily dictates
coverage and is governed by the application.
2. The frequency is important for resolution, and is discussed
later.
3. The total (active) aperture dictates focusing ability, and again, is
discussed below.
4. H, the aperture in the passive direction, is ignored in this study
as we are using commercial arrays with xed H. The passive
aperture will affect focusing in the passive direction, as dened
by the equation for total (active) aperture.
5. The width of individual elements, e, is a critical parameter for
beam steering, and is discussed below.
6. p, the pitch of the array, is usually similar to e, the element
width as saw cuts are miniscule; however, sparse arrays are
possible, which introduce potential problems like grating lobes
(see sparse arrays below).
Normal phased array probe design is defined by frequency,
element width (e), number of elements (n) and pitch (p). First, this
paper will describe array manufacturing techniques, particularly
with respect to resolution.
Piezo-composite arrays
The piezo-composite manufacturing techniques for conventional
ultrasonic transducers have been transferred to arrays. Figure 6
shows a schematic of the lay-out of a matrix array. For mostapplications, a linear array is adequate, so only linear arrays were
considered in this study.
Piezo-composite arrays are made by using thin rods of ceramic
material embedded into a polymer. The signal-to-noise ratio
obtained from composite transducers is typically 10-30 dB greater
than obtained from piezo-ceramic probes. The array construction is
similar to conventional transducers, as shown in Figure 7.
Another major advantage of piezo-composites is their high
damping capability. Unlike conventional piezo-electric materials,
which ring for 2-4 cycles, piezo-composites can ring for as littleas half a cycle. In practice, 1-2 cycles are more normal, as shown
in Figure 8.
The shorter pulse from piezo-composite arrays permits better
crack tip resolution.
Aperture and focusing
The adjustable aperture with phased arrays allows focusing to be
tailored to the application. With a given array, aperture size can be
adjusted by selecting the number of elements activated. Table 1
shows the theoretical effects on focusing for a 5 MHz array with
1 mm pitch.
The focal beam width, d, shows that a larger aperture gives
a smaller spot size (but shorter depth of field), so the immediate
reaction is to use the largest aperture possible. However, fieldexperience indicates that this is not the optimum solution in most
applications. First, very large apertures give larger probe sizes,
which can present practical problems. Second, the highly-focused
beams tend to have less working range, which limits their use.
In practice, conventional transducers are sold in fixed diameters
for these practical reasons, and these sizes are emulated by standard
Figure 5. Diagram showing phased array probe manufacturingparameters
Figure 6. Schematic of piezo-composite matrix array lay-out
Figure 7. Schematic of array construction
Number of elements 10 16 32
Aperture (mm) 10 16 32
N Fresnel distance (mm) 84 216 865
Focusing depth (mm) 84 84 84
K 0.99 0.39 0.10
d (at focusing depth mm) 2.49 1.55 0.78
Table 1. Focal spot size for a selected aperture size
8/13/2019 hased Array Ultrasonic Testing, Sectorial scan
4/7
4 Insight Vol 48 No 4 April 2006
arrays and apertures. Consequently, for most practical applications,
the number of pulsers is not important, provided the active aperture
is appropriate. The exceptions occur when significant beam
steering is required (see beam steering later), but such applications
are limited in the manufacturing and petrochemical industries.
With phased array focusing capabilities, ultrasonic testing can
now be performed in the near field. Historically, the near field has
been an operators problem to overcome. Table 1 shows the effect
of aperture on the near field lengths.
Examples of three different apertures are shown in Figure 9, for
10, 16 and 32 elements of the same width e.
FrequencySince phased arrays are primarily a method of generating and
receiving ultrasound, and the physics remains unchanged, the simple
approach to selecting frequency is to use the same frequency as the
equivalent conventional transducer. So, if conventional ultrasonics
uses, say, 5 MHz, plan on using the same frequency for arrays.
The same applies to the aperture: if conventional ultrasonics uses a
10 mm diameter, plan on using similar aperture with phased arrays
(for example, 10 elements of 1 mm width).
In practice, phased arrays can often use higher frequencies (and
occasionally larger apertures) to provide better signal/noise and
give a tighter, optimised focal spot. Instead of a 5 MHz conventional
transducer, a 7.5 MHz (focused) array may be optimum. Both
higher frequency and tighter focus permit better defect resolution,
sensitivity, detectability and signal-to-noise ratios. Manufacturing
problems may occur with some arrays, but these are mostly at high
frequencies (>15 MHz), and not important for ferritic steels.
Beam steering
Beam steering ability dictates the angular range of the S-scan, and
is a key issue for selecting array element width (e), and array
frequency. The relationship between the steering angle at -6 dB, st,
frequency and element size is given in equation (1).
...................................(1)
As the element width decreases, or the wavelength increases
(frequency decreases), the steering ability increases. Again, practical
considerations rule here; as the frequency decreases, resolution
decreases accordingly. As element size decreases, steering will
increase but the number of elements (and instrumentation
requirements) increases rapidly; budget considerations may rule in
this instance.
Steering range can be improved using an angled wedge; for
example, shear wave inspections typically use a wedge with a
natural refracted angle of 45, which can sweep from, say, 30
to 70(6). For different angular ranges, different wedges may be
appropriate; for example with natural refracted angles of 55 or
60. In general, for most manufacturing applications, beam steering
is designed into the array-wedge combination.
Grating lobes
All ultrasonic transducers produce side lobes, which is part of thephysics of constructive-destructive interference that generates
the beam(5). Phased arrays also generate grating lobes, which
are similar in concept to side lobes, and occur from the regularity
of the elements and spacing. In effect, these grating lobes are
standing waves. With a regular arrangement of elements (as in
most arrays), the far-eld pattern of an array probe shows a main
beam and grating lobes at regular angular spacing. These grating
lobes are predictable, both in angular direction and in amplitude.
The position is determined by the array frequency and pitch, p, as
shown in equation 2.
.....................................(2)
Figure 10 shows a simulation of grating lobes for different
numbers of elements and pitches with the same aperture size.The general rule for grating lobes is to design them out.
q For element sizes (e) , side lobes will occur; this combination
is not recommended.
q If e < /2, no side lobes will occur.
q For element sizes /2 < e < , the lobe will depend on the
specic parameters.
Sparse arrays
One solution to the requirement of high beam steering with a
limited number of elements is to make sparse arrays. These
arrays are similar in principle to standard arrays, but have many
elements effectively taken out. The main problem is sparse arrays
will produce strong grating lobes, as simulated in Figure 11. While
the main beam is little affected in the sparse array, the grating lobesare now signicant.
One solution to making sparse arrays is to use quasi-random
element locations to effectively smear the grating lobes(7). However,
this approach is not recommended for standard defect sizing since
signal-to-noise ratio suffers, as illustrated in Figure 11.
Voltage
Though different instruments have different pulser voltages, in
practice pulser voltage is not important for resolution provided
the instrument has enough power to penetrate. Under normal
conditions with a suitable array and frequency, penetration is not
an issue. Figure 12 shows a series of scans on the same crack using
different pulser voltages and gains.
The low voltage has a double advantage: it increases probe lifeand produces less heat, so no extra cooling devices are needed. The
low voltage value may be compensated by increasing gain without
Figure 8. Typical waveforms (left, rf; right, rectied), showingshort pulse effect
Figure 9. Beam simulations showing 1 mm array with threedifferent apertures (left: 10 x 1 mm, middle: 16 x 1 mm, right32 x 1 mm)
8/13/2019 hased Array Ultrasonic Testing, Sectorial scan
5/7
Insight Vol 48 No 4 April 2006 5
a significant increase of the noise level. If the voltage is reduced by
50%, and the gain increased by 50% to compensate, the images are
effectively identical. The key issue with defect detection and sizing
is signal-to-noise ratio.
Experimental procedures
A standard OmniScan 16:128 phased array unit with a 5L16-A1
array (5 MHz, linear array, 16 elements) on a standard 45 wedge
was used for inspecting the defects.
A series of stainless and carbon steel plates were inspected
using S-scans. Stress corrosion cracks in the stainless steel were
grown by Davis NDE, and cracks in the carbon steel were multi-
branched fatigue cracks. Both types of defects are traditionally
hard to size and characterise. Both half-skip and skip analyses
were used, depending on the crack configuration. The inspections
were performed manually, and screen shots saved for comparisons.
Figure 13 shows a photo of typical samples.
Typical results
A 9.5 mm-thick stainless steel plate with a crack approximately
70-80% through-wall is shown in Figure 14. The crack is imaged in
direct reection (less than half skip), corner trap reector, and some
skip reection by bouncing off the plate bottom. The facets can
be individually resolved with this set-up and phased array probe.
In comparison, the A-scan at left (which correlates with the blue
line, or data cursor, at 54) is far harder to analyse. In contrast,
individual facets can be clearly seen in the S-scan image.
Figure 15 shows a 12.7 mm stainless steel plate, with an SCC
approximately 60% through wall. Imaging is direct, ie less than
half skip. Again, the phased array imaging is vastly superior to
Figure 10. Grating lobes as a function of element size and pitch
Figure 11. Simulation of two arrays: 5 MHz, focal point at50 mm, array width = 20 mm, element width (e) = 0.6 mm,wavelength = 0.3 mm in water. Left: 33 elements, no gaps.Right: sparse array with only nine elements. Note the yellowgrating lobes in the right image
Figure 12. Detection and sizing of a fatigue crack by a 3.5 MHz 1.0 mm pitch linear array probe using different voltage combinations(Courtesy of OPG)
8/13/2019 hased Array Ultrasonic Testing, Sectorial scan
6/7
6 Insight Vol 48 No 4 April 2006
standard A-scan imaging (at left in Figure 15), and the shortened
pulse length, appropriate focus and overall sectorial echodynamic
display permits much better analysis. Sizing is easily performed
using the cursors and scale at right.The 9.7 mm carbon steel plate in Figure 16 contains a multi-
branched fatigue crack approximately 6 mm deep. Here, the
signal-to-noise ratio is excellent due to the lower noise level in
ferritic steels, and various facets and the main crack tip are easily
distinguished. With short pulses and good focusing, sizing is
straightforward using cursors. With the S-scan imaging, the tip
signal is easily differentiated.
For comparison, Figure 17 shows a complex crack in a
carbon steel 9.9 mm bar, with multiple facets visible. The S-scan
echodynamic display shows considerable improvement over the
A-scan image at left.
Comparing resolution with different apertures
As mentioned above, resolution is dependant on focusing
capability, short pulses, and frequency. The simplest demonstration
of resolving capability is by altering the aperture. Figures 18, 19
and 20 show a series of three scans of the same crack at the samelocation using three different apertures of 16, 8 and 4 elements.
(Note that the start points of the scans are adjusted so that the array
is always in the same effective position). The effect of altering
focusing was less notable than altering aperture.
Comparing piezo-composites and piezo-ceramics
Figure 21 shows a typical RF signal from a standard single-channel
aw detector on a 10%-deep crack in a 10 mm carbon steel bar
using a standard piezo-ceramic transducer. The tip diffracted signal
is essentially invisible, ieresolution is much diminished.
Comparing Figure 21 with Figure 18 of the same sizing bar
shows that the S-scan approach using a piezo-composite phased
array probe has much better crack tip resolution than the piezo-
ceramic transducer. The situation is complicated as the fixed 45angle is not necessarily the optimum angle for back-scattered tip
diffraction, as in this case. The combination of highly damped
array and multiple angle imaging (S-scans) allows phased arrays to
provide more accurate crack sizing results.
Discussion
This paper has illustrated the excellent defect resolution obtainable
with phased arrays using piezo-composite probes on relatively thin
plates. Coupled with the S-scan imaging, the combination permits
vastly better defect characterisation and sizing than conventional
A-scan approaches. In addition, these scans are easy to set up,
and straightforward to perform. Another major advantage is the
speed of inspection, as detection and analysis are much faster.Similar results have been obtained on complex cracks on thicker
components by other workers(8).
Figure 13. Photo of typical SCC plate and sizing bars withcracks
Figure 14. Typical S-scan of stainless steel SCC showingmultiple reections
Figure 15. S-scan display of SCC in stainless steel plate
Figure 16. C-steel plate with multiple facets and tip imaged
Figure 17. Complex crack imaged in skip
8/13/2019 hased Array Ultrasonic Testing, Sectorial scan
7/7
Insight Vol 48 No 4 April 2006 7
In terms of improved resolution, the effects of aperture size
(and focusing capability) are clearly shown in Figures 18-20, and
predictable(9). These S-scans clearly demonstrate the importance of
appropriate resolution for back-scattered tip diffraction for defect
sizing. The spatial resolution of the S-scans permits the operator to
visually see all refracted angles displayed at the same time.The development of highly damped piezo-composite arrays has
certainly improved resolution. This is more difficult to demonstrate,
since equivalent piezo-ceramic arrays are not available. However,
comparing A-scans in Figures 18 and 21 shows the effect quite
clearly in general terms.
Conclusions
n Improved defect tip resolution and signal-to-noise ratios can
be obtained by using piezo-composite phased arrays, with
appropriate apertures and focusing.
n S-scan imaging and S-scan echodynamic displays make defect
characterising much simpler and more reliable, even for multi-
branched cracks.n Back-scattered tip diffraction is a viable technique for defect
sizing since S-scans permit correct tip identication, as well as
good resolution.
n With the individual facets and crack tips clearly imaged, sizing
should be much more reliable and accurate, even for small
(~1 mm defects).
References
1. G Lafontaine and F Cancre, Potential of ultrasonic phasedarrays for faster, better and cheaper inspections, NDT.net,
Vol 5, No 10, October 2000. http://www.ndt.net/article/v05n10/
lafont2/lafont2.htm
2. F Jacques, F Moreau and E Ginzel, Ultrasonic backscatter
sizing using phased array developments in tip diffraction awsizing, Insight, November 2003.
3. P Ciorau, D MacGillivray, T Hazelton, L Gilham, D Craig
and J Poguet, In-situ examination of ABB l-0 blade roots and
rotor steeple of low-pressure steam turbine, using phased array
technology, 15th World Conference on NDT, Rome, Italy,
11-15 October 2000.
4. M Moles, N Dub and E Ginzel, Pipeline girth weld inspections
using ultrasonic phased arrays, International Pipeline
Conference, IPC02-27393, Calgary, Alberta, 29 September
3 October 2002.
5. Introduction to phased array ultrasonic technology applications
R/D Tech guideline, published by R/D Tech Inc, 2004.
6. Olympus NDT, Phased-array probe ultrasonic catalogue, 2005-
2006.
7. J J Selman, J T Miller, O Dupuis, M D C Moles and P Herzog,
FASTFOCUS a novel ultrasonic phased array system for
fastener hole inspection, Aging Aircraft, San Francisco,
California, USA, 16-19 September 2002.
8. P Ciorau, Contribution to detection and sizing linear defects by
phased array ultrasonic techniques, 4th International Conference
on NDE in Relation to Structural Integrity for Nuclear and
Pressurized Components, London, UK, 6-8 December 2004.
9. J Mark Davis, Advanced ultrasonic aw sizing handbook,
Third Edition, Published by The Art Room Corporation, 2006.
Figure 18. S-scan of carbon steel plate with crack using sixteenelements. Crack tip is arrowed
Figure 19. Same S-scan as Figure 18, but with eight elements(half aperture). Note reduced resolution of crack tip
Figure 20. Same S-scan as Figures 18 and 19, but with fourelements (quarter aperture). Crack tip is barely resolved
Figure 21. A-scan of 10%-deep crack, with backscatter tipdiffracted signal undetectable