Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
SYSTEM AND METHOD FOR EFFICIENTLY REVIEWING WELD
SCAN DATA BY A WELD INSPECTOR
Field of the Invention
The present invention relates generally to software tools for accelerating and
enhancing the data analysis of non-destructive testing using ultrasonic sound
waves in base
metal and welded materials. In particular, the present invention relates to
improving
efficient analysis of ultrasonic weld data pertaining to pipe, plate, and
structural weld
seams. In even greater particularity, the present invention relates to
assisted data analysis
of ultrasonic phased array testing data to increase the speed, reliability,
evaluation, and
lo reporting efficiency for code compliance of weld inspectors.
Back2round of the Invention
Bridges, buildings, railroads, pipes, vessels, tanks, and other metal or steel
welded
structures are a vital part of modern infrastructure. These structures are
typically
fabricated from sets of discrete metal sub-components that are welded together
to form a
is critical component and/or a total system. Welded seams and other weld
joining points
must be fused into a welded unit having satisfactory strength to ensure
building code
compliance and to achieve proper structural integrity of the entire system,
and/or to ensure
meeting the purpose of the design for which it was intended. For example, a
pressure
vessel or fluid tank must have water tight exterior as well as provide
structural support for
20 the entire vessel or tank system. Pipes and vessels similarly must have
water or gas sealed,
welded seams to ensure the integrity of the pipe/vessel and to properly
isolate the fluid or
gas held by the pipe from the environment. In addition, these systems
deteriorate over
time due to operational and environmental factors such as, residual and
applied stresses,
CA 3077555 2020-03-31
vibration, rain, snow, strong winds, temperature variance, earthquakes,
oxidation, material
fatigue, and other changes that occur over the passage of time. Hence,
nondestructive
testing of welds and metal structures, and their components, are utilized
after initial
fabrication, installation, and periodically thereafter, to ensure a
structure's integrity.
Further, both for new construction and routine periodic maintenance, careful
analysis of
the weld joints in each structure is necessary to ensure satisfaction of
various weld
specifications, industry codes, and construction regulations. For example, the
American
Society Mechanical Engineers ("ASME"), the American Welding Society ("AWS"),
and
the American Petroleum Institute ("API"), among others, each have their own
welding
codes, procedures, and specifications.
Modern inspectors use non-destructive test ("NDT") equipment to inspect
constructed metal (e.g. steel) structures and their weld joints. These
inspection devices use
ultrasonic wave generators to take digital "snap-shots" of welds from which an
inspector
may verify weld integrity and to ensure compliance with welding codes and
specifications.
Ultrasonic technology is used to detect internal and surface breaking flaws in
the weld and
the base metal, which are not visible externally, and is based on the
principle that a gap or
defect in the weld changes the propagation of ultrasonic sound through the
metal. One
common method of NDT testing uses conventional, single-probe ultrasonic
testing
requiring an operator's interpretation of a screen similar to an oscilloscope
screen that
presents time and amplitude information. Another method uses an array of
ultrasonic
phased array sensors to test a structure. Such methods can be combined into a
single
digital piece of inspection equipment that uses phased array ("PA") and time-
of-flight
("TOFD") diffraction methodologies to provide a three dimensional image of a
weld
2
CA 3077555 2020-03-31
displayed on a color screen. An inspector then evaluates the potential for a
flaw or defect
in the weld by reviewing the screen. Such ultrasonic testing ("UT") equipment
is typically
highly mobile, and allows for the recordation of ultrasonic data for the
analysis of welded
areas in joined metal pieces. For example, Olympus NDT, Inc. markets and sells
ultrasonic units through its OmniScanTM and EpochTM lines of weld flaw
detectors. The
Olympus ultrasonic inspection systems include conventional ultrasonic flaw
detectors,
which use ultrasonic waveforms to detect flaws, and advanced ultrasonic phased
array flaw
detectors, which create internal cross-section images of the areas being
inspected. These
ultrasonic flaw detectors may also be configured to do a phased array
ultrasonic testing
io ("PAUT") inspection that produces encoded digital data points on welds
which may be
further processed at a later time. The data may also be reviewed later at a
time convenient
for an inspector and in an environment away from the component or structure's
location,
which is typically more suitable for detailed analysis work. Such ultrasonic
data is
recorded and saved in large data files which may be retrieved for evaluation
by an
inspector using specialized software applications. While these data files are
readable by
the ultrasonic testing device, software applications also exist that assist in
the evaluation
and visual display of such inspection data on common computing devices, such
as a PC.
An example of such applications is the OmniPCTM analysis software also
available from
Olympus NDT, Inc., along with an additional analysis tool set called
TomoViewTm. Both
of these tools allow for a more precise and reliable review of the three
dimensional PA data
by an inspector.
A phased array data file consists of captured data representative of
continuous
A-scans along a weld which may be processed to create a three dimensional data
set
3
CA 3077555 2020-03-31
representative of the top, side, and end views of a weld joint. Within each A-
scan, a series
of recorded data points record an intensity or amplitude value from 0% to 100%
of
ultrasonic signal reflections. These data points are then saved as OPD, RTD or
TV file
formats for subsequent analysis. Since the recordation of ultrasonic data is
correlated to
the exact PA probe position on the structure being tested, the inspector may
use the
analysis software including various sets of sophisticated analysis tools to
review the
recorded data in a number of geometric views and orientations (e.g. top view,
side view,
end view) to improve inspection review accuracy.
Further information regarding the use of phased array UT equipment,
configuring
0 such UT equipment, establishing a test scan plan applicable to a
particular inspection or
weld joint design situation, the recording of that data in various file
formats, the physics
and geometries of the ultrasonic sound beams and resulting scanning views in
UT, the use
and applicability of "data" libraries, the storing of testing data files, the
usage of different
types of scan views, the visual analysis of weld flaw indications, and the
generation of
is inspection reports based upon UT shall not be discussed herein as such
information is well
known known in the NDT industry and not necessary for a complete understanding
of the
disclosed invention. However, Applicant references.
the treatise UT Classroom Training Book, Paul T. Marks, ISDN No. 978-1-57117-
345-4
(e-book), published by The American Society for Nondestructive Testing, and
two treatises
published by Olympus NTD, Inc: (1) Introduction to Phased Array Ultrasonic
Technology
Applications, third printing 2007, ISBN No. 0-9735933-4-2; and (2) Advances in
Phased
Array Ultrasonic Technology Applications, 2007, ISBN No. 0-9735933-4-2. The
books
may be obtained at the ASNT website www.asnt.org or the Olympus NDT resources
4
Date Recue/Date Received 2021-10-18
website www.olympus-ims.com. These treatises explain the above subjects in
detail and
the general theory of UT using modern equipment. Further, Applicant references
U.S. Pat. Nos. US8156813B2, US8577629B2,
US9032802B2, US9081490B2. These patents discuss and disclose background
information regarding the electronics and theory behind PA ultrasonic testing.
Nevertheless, even with modern PA and time-of-flight UT devices, and even when
inspection analysis is conducted in an environment conducive for careful
study, the data
analysis and reporting process can be a tedious and fatiguing task for
inspectors. For
example, metal pipe and plate structures typically have girth welds and long
seams that
must be inspected. PA ultrasonic and time-of-flight, diffraction inspection
for those welds
seams can produce extremely large data files requiring many hours of data
review and
analysis of all data points along the weld seams by an inspector. Usually most
of the data
points are nominal, satisfactory welds, creating a monotonous review period
and
potentially reducing weld flaw recognition by an inspector due to fatigue.
Hence, what
would enhance the inspection process would be a system for focusing an
inspector's
attention on actual weld defect indications, by excluding data points that
present
satisfactory and acceptable weld characteristics. Such a system would improve
an
inspector's efficiency and accuracy in conducting new or reviewing prior weld
inspections,
thereby saving time and money.
Summary of the Invention
In summary, the invention reads data from PAUT or TOFD data files that capture
ultrasonic testing results from a series of metal welds, or a set of
continuous welded metal
seams, and creates a table of target indications requiring an inspector's
review and
5
Date Recue/Date Received 2020-10-16
evaluation. The invention reads one or more testing data files and collects
weld indications
from the data files, applies a series of filters to each recorded indication,
and ranks the
indications in a meritorious order, including the application of color coding
to these
indications to assist the inspector in analysis of the testing data file. The
invention acts as
s a weld analysis concentrator by focusing the attention of an inspector on
weld targets that
merit their attention, thereby increasing the efficiency of the inspector. The
process
typically excludes 95% to 98% of the weld data, while retaining significant
flaw
indications necessary for an inspector's review and in conformity with
inspection
regulations from various organizations establishing weld codes and
specifications.
Other features and objects and advantages of the present invention will become
apparent from a reading of the following description as well as a study of the
appended
drawings.
Brief Description of the DrawinEs
The patent or application file contains at least one drawing executed in
color.
is Copies of this patent or patent application publication with color
drawing(s) will be
provided by the Office upon request and payment of the necessary fee.
A system incorporating the features of the invention is depicted in the
attached
drawings which form a portion of the disclosure and wherein:
Figure 1 a side view of a typical ultrasonic testing system for a pipe weld
seam
using a phased array or time-of-flight inspection system;
Figure 2 is a perspective view of a typical ultrasonic testing scenario for a
pipe
weld seam using a phased array system or time-of-flight inspection system;
6
CA 3077555 2020-03-31
Fig. 3 is an S-scan view showing a two dimensional scan matrix overlay of
refracted sound wave angles in a typical ultrasonic testing scenario for a
pipe weld seam
showing a potential weld flaw;
Figure 4 depicts a plan view with a matrix overlay showing the locations and
recordations of ultrasonic data on the information displayed in the graph of
Fig. 3;
Figure 4A is a diagram of a data cuboid;
Figure 4B depicts an end view of a weld with a matrix overlay showing the
locations and recordations of ultrasonic data from the information displayed
in the graph of
Fig. 3;
Figure 5 is a process flow diagram of the overall invention process;
io
Figure 6 is another process flow diagram of the invention process;
Figure 7 is a process flow diagram showing the data filtration process
utilized in
the invention; and,
Figure 8 is a representative depiction of a spreadsheet chart produced by the
invention after processing of weld indications showing color coding.
Description of the Preferred Embodiments
Referring to the drawings for a better understanding of the function and
structure of
the invention, Fig. 1 shows a side view 10 of a typical ultrasonic testing
scenario for a steel
pipe weld seam using a phased array ultrasonic tester. The pipe 11 includes
two sections
12a and 12b joined together via weld seam 23. Each piece of the pipe separates
the
environment in which it is positioned into an interior 13 and an exterior 16,
with each pipe
section having an interior surface 17 and exterior surface 18 contacting the
interior and
exterior spaces. As may be noted, this arrangement would be similar for other
weld joint
7
CA 3077555 2020-03-31
designs such as vessels, tanks, and structural members. Each section 12a and
12b extends
away from weld seam 23 curving downwards on each side 21, 22 to enclose
interior 13.
The weld seam 23 is typically "V" shaped having an upper width 24 larger than
bottom
width 27, but may have other shapes as is known. The top of weld 23 typically
has a
portion of excess weld material 26 that extends slightly above the exterior of
the pipe 18
for reinforcement.
Positioned adjacent to the pipe 11 is a phased array UT probe 31 including a
transducer probe 32 having an angled probe wedge 34, an ultrasonic array
transmitter and
receiver element 36, and a connection cord 33 connected to a testing device 35
(not
io shown). In addition to an electrical signal wire, such as a coaxial
cable, the cord 33 may
also include a tube for the transmission of water or other fluid (known as
"couplant") in
which to couple the phased array probe and wedge to the base material and in
order to
efficiently transmit the ultrasonic sound waves into the base metal and weld
metal.
The probe 36 emits ultrasonic waves from elements held in probe 31 that are
15 controlled by testing device 35 at timed intervals to create various
types of wave fronts or
sound beams that are propagated through the metal of the base metal 11 and the
weld
material 23. This results in a series of longitudinal wave beams or shear wave
beams 39.
The sonic beams 39 vary in angle generally from about 45 degrees to 70
degrees, refracting
through the weld 23 and surrounding base metal 12a,12b, and produce ultrasonic
20 reflections received by probe 31 that are sent back to testing device
35. The combination
of the wave segments 39, software that controls the timing, frequency,
emission position,
and other qualities of those sound wave beams, and the position of the probe
31 adjacent to
the weld seam 23, is designed to reveal weld flaws 41 that may be present in
or around the
8
CA 3077555 2020-03-31
weld seam. The wave beams 39 will typically generate reflections off the
interior surface
17 and exterior surface 18 (not shown).
Referring to Fig. 2 probe head 31 is positioned adjacent to weld seam 23 and
scanned along the length of the weld seam direction 29 joining two sections
12a, 12b of
pipe 11. Since weld seam 23 has approximately a V-shape as shown in Fig. 1,
the
extension of seam 23 along 29 creates a wedge shaped weld seam. Testing probe
31
positioned adjacent to weld seam 23 is moved in direction 37 along the length
of seam 29
while transmitting ultrasonic sound waves to the interior of the seam weld and
surrounding
weld metal. At all times, probe 31 tracks its position along seam 23 and its
index position
io from the centerline of the weld seam as it scans along the weld
seam. The weld seam may
be scanned manually, semi-automatically, or automatically through encoded or
motorized
means, as is understood in the art. While the probe 31 is shown to be
positioned on one
side of weld seam 23, the probe may be positioned on the other side with
equally effective
results. In addition, multiple probes may be utilized at the same time.
5
Referring now to Fig. 3 an ultrasonic data sampling system 45 of wave segments
1
and samples is shown in a sectorial diagram similar to Fig. 1. The system 45
shows a
typical sampling arrangement used in recording data from phased array or time-
of-flight
diffraction UT scanning of a weld seam. The diagram 45 approximates an example
of a
sectorial or "S-scan" view of a weld, non-destructively tested with ultrasonic
waves as may
20 be understood. Weld 23 is positioned in the middle of the system 45
having sound beam
segments 39 with probe head 31 positioned at the origin of the sampling system
45. Due
to the proximity to the probe, area 44 just under the probe 31 is either not
included during
scanning or excluded later during processing. The arrangement of the sound
beam
9
CA 3077555 2020-03-31
segments and sampling data points produces a two dimensional scan matrix 50 of
each
scan slice of data taken along weld seam 23. Each matrix 50 is not
rectangular, but has a
radial shape as shown, with each data point having a coordinate location of
(a) scan offset,
(b) beam, and (c) sample index, as will be further discussed. Hence, each data
point
sample 53 is composed of its three dimensional coordinate and an amplitude
value
representing the strength of the reflection from beams 39 at that point in
space. That
information can also be organized in a three dimension rectangular array in
the form of a
"cuboid" or a "voxel" as will be discussed in Fig. 4A.
As ultrasonic waves 39 are transmitted into the weld area 23 from probe 31 any
potential weld defects 41 will produce ultrasonic reflections back to probe 31
which are
received and recorded as a group of data points having elevated amplitude
values. Sound
beam segment 47 is approximately 45 degrees from normal, and beam segment 48
is
approximately 70 degrees from normal incidence. Other angles of sound waves 39
span
the angular distance between segments 47 and 48 such that the entire weld area
for any
current segment is covered by these scanning segments, including wave segments
49 that
impinge upon potential flaw 41. Since the entire area has angles of sound
waves
transmitted through the base material and weld seam, any potential flaws 41 in
the weld
seam 23 will be contacted by one or more of the sound waves. Within the system
45,
potential flaw 41 is within a bounded area 46 of data points that surrounds
the potential
flaw. As may be understood, the cells shown in Fig. 3 are not to scale and
provided for
illustrative purposes only to show the workings of the herein disclosed
invention. The
resolution of a typically overlay would normally be many times finer than that
shown by
the matrix 50.
CA 3077555 2020-03-31
System 45 also includes three zones of interest in any scan. Zone 1 56 is
positioned
along the interior surface 17 of the weld 23 and includes all data cells along
that surface.
Zone 3 58 is positioned along the exterior surface 18 of weld 11 and includes
all data
sample cells along that surface. Reflections from zones 1 and 3 may be
filtered for more
s in-depth data analysis, depending upon the content of that data and the
type of filter being
applied. Zone 2 57 is positioned between zones 1 and 3 and includes all data
cells not in
zones 1 and 3 and are part of an analyzed set of data points in later
processing.
Fig. 4 shows a rectangular coordinate system overlaid on a top-down or "C-
scan"
view of the seam shown in Fig. 2. As indicated above, the coordinate system 50
is not
lo rectangular, however because each data point has a point in three
dimensional space in the
scanned workpiece a corresponding point in space may be recorded in a
coordinate matrix
overlay 60 of data cells. In particular, the combination of rows 61 and
columns 62
produces a rectangular matrix of cells that will contain each data point
sample 53 in the
coordinate system 60 as the probe 31 is moved along weld seam 23 (rows R3 and
R4) in
15 direction 29. Potential weld flaw 41 would be recorded in at least 6
cells, such as 63 in this
depiction (i.e. R4, C4-R4,C6; R5,C4-R5,C6). A similar end-view or "B-scan"
view
looking into the weld having from one end with a coordinate overlay might
similarly be
produced and is shown in Fig. 413. In particular, the combination of rows 161
and columns
162 produces a coordinate matrix overlay 160 of data cells viewed from the end
of the
20 weld seam 23 in either direction along 29 at a particular scan offset.
Potential weld flaw
163 would be recorded in at least 6 cells 166 shown surrounding potential flaw
163. Each
cell has a two dimensional measurement as depicted in Fig. 4A.
11
CA 3077555 2020-03-31
As shown in Fig. 4A, an example of each data cell referred to hereinafter as a
"cuboid" (or "voxel") or "data cuboid" 65 has length 66 of about .04 inches, a
width 67 of
about .02 inches, and height 68 of about .02 inches. As will be understood,
these values
may change based upon a variety of factors associated with each weld scan.
While the
s coordinate systems shown in Figs. 3 and 4 are two dimensional, in
reality each scan
snapshot taken along seam 23 has a sample width of .02 inches due to the way
in which
ultrasonic waves are transmitted into the material. Hence, each cuboid 65 is
in reality three
dimensional. UT probe 31 takes sectional scan snap-shots along weld seam 23
along travel
path 37 (see Fig. 2) with each section scan recorded as a matrix of cuboid
data values in a
file, adding a new matrix section of data with each new sectional snap-shot.
Each recorded data cuboid has a set of values associated with it, comprising
the
prior noted three dimensional coordinate set of system 45 (see Fig. 3) and an
amplitude
value. The coordinate portion includes the values of depth from the top
surface of the pipe
also referred to as (1) the "ultrasonic axis" distance; (2) an index axis (or
"index offset")
is which is the distance from the center line of the weld 23; and (3) a
"scan axis" value (or
"scan offset") which is the distance along the weld as recorded by the probe
as it travels
along path 37. The amplitude value is a reflection value (i.e. a sound
intensity value)
recorded by the probe 31 normalized to a relative value of between 0% and
100%.
In actuality, each scan section includes raw data position values of (1) scan
positon
of the probe head as it tracks along the weld seam, (2) an angle value
representing the
angle of a wave segment as it is emitted from the probe emitter, typically
between 45 and
70 degrees, and (3) a sample index value which is the distance from the probe
emitter.
Using a known set of geometric calculations for polar coordinates, any scan
data value
12
CA 3077555 2020-03-31
position in any data cuboid may be determined when data values are retrieved
from the
scan data file ("SDF") as will be discussed.
Referring now to Fig. 5 a system is presented 70 that reduces the number of
weld
scan indications that an inspector must review in order to produce a report
meeting
applicable welding examination codes and requirements, such as those published
by
ASME, AWS, or other organizations. The system 70 is a standard software
application
that may run on a standard WindowsTM operating system, such as for example
Windows 7
sold by Microsoft Corporation, running on a standard PC configuration. The
system may
also be incorporated as a module directly into existing testing and/or scan
analysis
software. Initially, an inspector assesses a weld situation and then
configures their UT
equipment for a scan, including the positioning of the UT probe 72 adjacent to
a target
weld. The weld is scanned by the inspector 73 and a data file recording the
weld scan data
saved 74. The data may be saved locally on the UT device, transferred to a
connected
drive storage 76, or uploaded to a network drive via Wi-Fi or other data
connection,
is depending upon the size of the data file. The scan file is then
processed 77 by extracting
all data cuboids that include potential weld flaw indications, essentially
extracting all
cuboids that have amplitudes greater than 0, and then creates a file recording
those
indications and saves it in a local, fast access storage location 78. Further
processing
occurs on the indications file by applying a series of filters 81 that ranks
and categorizes
the indications into a usable form. In particular, a ranked list of
indications is created in a
table based on a ranking value for each indication which consists of multiple
data cuboids.
That priority listing of indications is then produced 82 and displayed 83 for
the inspector's
analysis at a place and time of their choosing. The process shown in 70,
typically removes
13
CA 3077555 2020-03-31
over 95 percent of the non-relevant data stored in a scan data file, and
presents a focused
list of only a fraction of the overall indications held by a scan data file,
without degrading
an inspector's ability to properly review the scan data in accordance with
applicable code
requirements.
Fig. 6 shows the benefit of this system 70 in a typical operation 90 by an
inspector.
A phased array UT is arranged, configured, and then used to scan a weld under
review 92.
A scan data file is created and that file is saved locally, or at a remote
storage location such
as a cloud based drive 95 via the internet 94. The data file is then accessed
at a later time
and processed 97 to extract its weld indications and filtered to remove
indications that do
o not merit review. That processing 97 may occur at the initiation of
an inspector, or it may
be automated upon the arrival of the data file at storage location 95. The
inspector will
review the published report of target indications 98 and using a compatible
scan data
visualization application or device 100 review each indication listed in the
published
indications report that merits careful study. The inspector may then produce
their own
inspection report 99, saving it in a local storage location, and may take
remedial action 102
to correct any noted weld defects.
Turning now to Figs. 7 and 8, it may be seen the detailed operation of the
system
70 along with an example report table produced by the system in a typical pipe
weld seam
situation. A typical industry UT device is an Olympus OmniScanTM MX2 phased
array
flaw detector. That device produces a scan data file in the form of an
OmniScan Phased
Array Data File or "OPD" file. The OPD file may be saved in a storage location
for later
review and accessed by various visualization applications to graphically
review any
portion of the scan along a pipe weld seam. Each OPD file includes
configuration and
14
CA 3077555 2020-03-31
setup information such as weld type, weld bevel angle, thickness of the pipe
material, the
ultrasonic velocity utilized, probe scan and index offset, and probe skew, and
also includes
information on how many scan jobs are included in the file. Additionally the
overall scan
length is recorded and a sound path minimum value recorded. That information
is utilized
by the process 105 as will be discussed.
As shown in the process 105 of Fig. 7 an OPD file is read 107 from a storage
location 108. The accessing of data in the OPD file may be accomplished with
the help of
an intervening software module known as a data access library or "DAL." The
DAL is
typically written or authorized to be written by the author of the OPD file
specification,
which in this example is Olympus NDT. Whether read directly from the OPD file,
or
indirectly through the use of a DAL, the data is retrieved in the form of a
scan position, a
beam number (i.e. beam angle), a sample index (i.e. sample number along the
beam), and a
data point value (i.e. magnitude) at the corresponding sample. System 105
accesses the
OPD file by making a data retrieval request pursuant to an instruction format
specified by
DAL 111 which allows the extraction of all data cuboids in the scan data file.
Any data
scan cuboids accessed but having no amplitude are not saved, so only cuboids
that are
potentially part of a weld indication are retained. A local database file is
created to hold
those indications in a three dimensional matrix referenced by position and
amplitude
values, which is then stored 112.
The system further examines the amplitude of each data cuboid in the database
and
excludes any cuboids that have a value less than or equal to 16%. Sixteen
percent is a
value derived from weld code specifications below which scan reflection data
does not
typically need to be reviewed by an inspector. Each data cuboid remaining in
the database
CA 3077555 2020-03-31
is then analyzed to determine if any adjacently positioned cuboids to the one
being
analyzed also have been retained (i.e. their intensity values were also above
16%, see e.g.
Fig. 3, 46, Fig. 4, 63). A first cuboid is considered adjacent to a second
cuboid if they
differ by only one position in scan, beam, and/or sample coordinate(s). Such
remaining
contiguous cuboids are then grouped together and referred to as an individual
"indication"
to signify a potential flaw indication that requires an inspector's review.
Each cuboid is
similarly analyzed until each cuboid is assigned into a unique weld indication
109. Several
properties of each weld indication are also calculated and utilized in further
processing of
each indication. Namely, information such as depth, index position, and scan
position
range limits are recorded; the maximum intensity position within the
indication is
calculated to give a sense of a "center of gravity" for each indication; and
the maximum
amplitude of all cuboids in the indication is recorded 112.
Each indication is also given a unique ID based on a hashing of cuboid data
contained in the indication. As may be understood, any identical A-scan data
will always
produce the same unique hash ID for each indication. Further, the process 105
also assigns
a significance score to each indication to produce a priority ranking order of
indications
needing attention 124 as will be discussed. That significance score is a value
assigned to
each indication between 0 and 1, with 1 being the highest value, and
represents a best
estimation on the part of the process 105 to draw the attention of the
inspector to an
indication that is more likely to be a flaw.
Once the weld indications have been extracted 109 and saved 112, the saved
data
file may be pre-processed 110 to optimize the data file for filter
application. For example,
a data smoothing processing algorithm may be applied to the data 113 to
increase the
16
CA 3077555 2020-03-31
signal to noise ratio to facilitate the recognition of meaningful indications.
Other
preprocessing steps may be performed for different data enhancement results
114 prior to
filtering the data.
A series of filters 115 are applied to each indication that affects the
significance
s score and which is used to form a ranking 124 of each unique ID. Each
flaw indication is
ranked in a table based on this significance value as will be discussed in
Fig. 8. Each
indication starts with an initial value of 1, and resulting filter values are
multiplied together
and then multiplied by 1. Hence, and as will be seen, some filters are
essentially fully
exclusionary because they only return either a 1 or a zero. Other filters
return a value
io between 0 and 1 that may reduce the significance score of the indication
or leave it at 1.
Filter 116 is a minimum size filter 116 which is applied to each indication.
If an
indication is below a minimum threshold size, that indication is assigned a
significance
score of zero. Specifically, if an indication includes a number of contiguous
cuboids less
than a threshold value, a zero is assigned to that unique indication ID.
Alternatively, a 1 is
is returned if the filter 116 does not apply. Currently, the ASME code
specifies threshold
values of 0.08 inches length (scan dimension), 0.06 inches width (index
dimension), and
0.06 inches height which might change with each particular scan situation.
Also, any flaw indications that are too close to the probe emitter 117 (see
Fig. 3,
44) returns a value of zero which would result in a significance value of
zero. Indications
20 are discarded (i.e. assigned a zero score) due to the distortions that
are experienced near the
probe head in the pipe material. So, any indications that are within 25% of
the scanned
workpiece thickness from the probe head are assigned a zero significance score
117.
Alternatively, a 1 is returned if the filter 117 does not apply.
17
CA 3077555 2020-03-31
Reflection of sound waves is timed in each UT device and any indications that
are
more distant than a predetermined percentage of the pipe steel thickness
distance (e.g.
indications in Zones 1 and 3 of Fig. 3), and also as set by welding code
specifications, are
assigned a zero significance score 118. Alternatively, a 1 is returned if the
filter 118 does
not apply.
Also, any reflections that are more distant than a set lateral distance from
the weld
23 are assigned a value 119 between zero and 1 representing a reduction in
significance in
that indication the farther away an indication is it positioned away from the
weld
centerline. The significance factoring for filter 119 is shown in the
processing pseudo code
is of Table 1 below.
Let MAX = maximum distance from weld before an indication may be
disregarded by the code (currently this value is 2" if the part thickness is
8"
or more, otherwise the minimum of I" or the part thickness).
If an indication is farther away than MAX from the weld it is considered
not in the "heat affected zone" or HAZ.
Let DIST = the lateral distance (i.e., in the index offset dimension) of the
indication from the weld surface.
Let DI = DIST / MAX
If DI <= 0.25, return score = 1.0
If DI >= 1.0 return score = 0Ø
Otherwise return score = (D1 - 0.25) / 0.75 (i.e. linearly interpolate).
Table 1
As part of the UT scanning process, it is typical that some inherent noise
echoes or
"hot spots," also known as "root or weld cap geometry signals," will be
created for each
sectorial scan as the probe traverses along the weld seam. Essentially, these
false returns
18
CA 3077555 2020-03-31
are artifacts that are produced by the geometries of the top and bottom weld
caps. The
nature of these hot spots is that they are periodically repeated along the
seam at regular
intervals and typically have common characteristics between one another,
namely they
appear at substantially the same index offset and depth locations. Since these
hot spots
represent false flaw indications, filter 121 assigns a reduction in
significance score to all
such indications which are likely to be false returns due to these ultrasonic
geometries.
The significance factoring sub-process for filter 121 is shown in the
processing pseudo
code of Table 2 below.
19
CA 3077555 2020-03-31
Let S= {sl, bl, ml}, Is2, b2, m2},. , be the set of 3D
coordinates (in Scan, Beam, and saMple) of all voxels/data points
within the indication. This will represent a contiguous region/blob
of data points in 3 dimensions. Each data point will have an
associated amplitude measured from 0 to 100.
Let MAX = maximum amplitude of all of the amplitudes of all the
coordinates in S.
Let P = {bl, ml), {b2, m2}, . be the set of 2D coordinates in
beam and sample dimensions that results when you take all of the
3D coordinates in S and simply discard their scan coordinate; if
there are any duplicates, discard the duplicates. (i.e. P is the
projection of S along the scan dimension).
For each coordinate Ib', in P. let TYPICAL(b', m') be the
average amplitude over all data points {s', b', m'}, where s' ranges
over all possible values except for those values where {s', b', m'} is
a member of S.
For each {b', m'} in P, let RATIO(b', m') = MAX / TYPICAL(b',
m').
Let AVGRATIO = average of RATIO(W, m') over all points {b',
m'} in set P.
If AVGRATIO >= 2.0 then return score 1.0
If AVGRATIO <= 0.25 then return score 0.0
Otherwise, return score = (AVGRATIO - 0.25) / 1.75 (i.e. linearly
interpolate).
Table 2
As may be understood, the filters incorporated in 115 may be altered in their
variables and operation to lesson or increase their individual effects on an
indication's
significance score. In addition, more filters might be incorporated in the
future as weld
specifications and codes are updated, or resulting from better understanding
of flaw
indications present in an ultrasonic weld scan.
CA 3077555 2020-03-31
After filters 115 are applied to the indications held in database 112, the
indications
are ranked 124 based upon their resultant significance values. Color
designations are then
optionally applied 126 to each ranked indication to assist an inspector in
correlating the
same instance of an indication over multiple scanning data files. A
designation ranking
table is then published 127 for the inspector's use in evaluating an OPD data
file.
Fig. 8 shows an abridged ranking report of flaw indications that can be relied
upon
by an inspector to focus their attention on important indications. The chart
130 is used for
example purposes only and as will be understood any real weld scan data file
would likely
include hundreds of indications, but chart 130 is abridged to show only 45
entries for
illustration purposes but still show the usefulness of the report. The chart
130 is a spread
sheet and displays 45 rows each representing a flaw indication ranked in order
of
significance from 2 to 45 in rows, row 2 being the most important flaw
indication and row
46 the least significant indication. Each indication is assigned a unique ID
created via a
hashing function as shown in column A 132. Also, a series of columns with
headings 140
provides characteristic information on each indication. Individual entries in
column A 132
may be hyperlinked to views in a cooperative data scan visualization program
allowing for
direct selection of any indication in the chart for instant viewing in the
scan visualization
application. Column B 133 displays the scan file from which the indication was
produced,
and as can be seen at least 4 different scan files were processed by system
105 to produce
the table 130. The chart also indicates the probe group from which the samples
were taken
134. Rankings shown in Column D 135 list all indications retrieved from the
scan data file
in descending order from the highest value of 1.00 to the lowest value of
0.00.
Alternatively, the chart may be reordered to arrange indications in a
direction running
21
CA 3077555 2020-03-31
along the weld length. Column F 138 shows the maximum amplitude of any cuboid
present in any indication, typically close to the center of gravity of any
indication. The
refracted angle of the indication is shown in Column J 141. The scan position,
index
position, and surface distance are shown in columns K 142-M 144. The sound
path is
shown in column N 146, and the indication depth is shown in column 0 147.
Color coding is also optionally utilized to assist an inspector in recognizing
certain
attributes of indications. First, rankings in column D 135 are assigned a
graduated scale
from red to blue using the color spectrum to show a high ranking to a lower
ranking. For
example, the top 10 entries in rows 2-11 have a bright red color indicating
importance,
)0 j importance. Location match column E 137 displays a color based on
location for each
indication by assigning color values from a color palette to the center of
gravity value of
each indication and using depth of the center of gravity indication as a red
value, using the
scan position of the center of gravity indication as a green value, and using
the index
positon of the center of gravity indication as a blue value. Hence, even if
multiple scan
files are produced and analyzed in the process 105, a color location match
will assist the
inspector in recognizing repeated indications for the same position along the
weld seam.
So, for example, entries on rows 13 and 18, which come from different scan
data files, are
likely from the same three dimensional location along the seam because they
have similar
colors listed in column E 137. Finally, amplitudes are color coded to indicate
their relative
position between 0% and 100% similarly to the assignments of ranking. However,
because
high amplitudes do not necessarily lead to high ranking, colors do not show as
a graduated
scale in the present example, as in column D 135 for instance, because the
indications
ranking of 130 is sorted on ranking and not on amplitude value. Nevertheless,
the color
22
CA 3077555 2020-03-31
does provide a visual indication of amplitude that may be useful when
correlating a
particular indication in a scan visualization application to the chart 130.
As may be seen, entries 2-18 149 show rankings from 1.00 to 0.71. These
rankings
would likely be the most important indications for an inspector to review on
the scan
visualization application to determine whether remedial action is necessary.
Lower entries
152 and 151 may not need to be reviewed or reviewed with less scrutiny than
usual.
Entries 40-46 151 probably need no or only cursory attention from the
inspector. Also, as
might be understood, the listed indications might be re-sorted in additional
tabs under chart
130 to further assist in focusing the inspector on indications of interest.
For example, a
io subsequent tab might list only the top 30 indications.
Overall an inspector having to review the four listed scan files in scan
visualization
software would take a great deal of time to review each and every indication,
but using the
chart shown in Fig. 8 an inspector can quickly focus their attention on
indications that
merit detailed scrutiny.
While I have shown my invention in one form, it will be obvious to those
skilled in
the art that it is not so limited but is susceptible of various changes and
modifications
without departing from the spirit thereof.
23
CA 3077555 2020-03-31
