The reliability of critical components in industrial plants directly and
significantly affect
the economic viability of the plant. Therefore, it is vital that the
condition of critical
process equipment be carefully monitored and controlled during operation to
assure
that forced, unscheduled outages are minimized or eliminated and scheduled
maintenance outages are well planned and efficiently executed.
One way the process industry can achieve these goals is to provide
nondestructive
evaluation (NDE) for detecting corrosion/erosion, hydrogen damage, and
cracking on
critical process equipment while the equipment is on line during regular
plant operations. Ultrasonics has been used for detecting these types of
defects for many years
when the components are at or near ambient temperature. However, during regular
plant operations, these critical components may reach temperatures as high
as 1200°F.
Conventional ultrasonic transducers cannot operate at these high
temperatures for long
periods of time since the piezoelectric crystals, in effect, depolarize and
no longer
produce mechanical oscillations at temperatures above the Curie temperature
(for most
piezoelectric materials, the Curie temperature is less than 700°F).
Another approach for introducing ultrasonic energy into a part at high
temperature is
the use of electromagnetic acoustic transducers (EMATs). An EMAT generates
ultrasonic energy in a conductive material in the following way. A magnetic
field is set up in
the part, and a high-frequency RF pulse is coupled into the part with a
coil. The RF
pulse interacts with the magnetic field to produce a mechanical force. This
force produces mechanical waves at the RF frequency, which is usually in
the ultrasonic range.
This electromagnetic force can be applied to the part under inspection
without using
any liquid couplant; however, the coupling is dependent upon liftoff between the
EMAT and the part, so it is important to keep the EMAT and part as close as
possible
(usually within 1 millimeter).
Since EMATs can be used to introduce ultrasonic energy into the part
without liquid
coupling and direct contact with the part, they have great potential to
allow ultrasonic
inspection at temperatures above the Curie temperature. EMAT development
has been
under way for approximately 20 years, and their capabilities have been
improving consistently. However, the capability of the present state of the
art in EMAT technology for
performing ultrasonic inspection at ambient and operating plant
temperatures was not
clearly understood.
Materials Technology Institute (MTI) funded Southwest Research Institute
(SwRI) to
perform an unbiased evaluation of the present state of the art on EMAT
technology for
high-temperature NDE of remaining wall thickness. The project goal was to
provide a
standard set of test specimens, containing representative corrosion-like
defects and
remaining wall thicknesses, and have several EMAT vendors inspect them
under controlled conditions so that a benchmark could be made on the state
of the art. To
accomplish this goal, SwRI (1) developed a test protocol, (2) designed and
fabricated
test specimens for use during EMAT evaluations, (3) designed and fabricated
a high
temperature test fixture, (4) conducted laboratory evaluations, and (5)
assimilated the
results. The work associated with each of these areas is described in the
following
sections.
SwRI also communicated directly with each prospective EMAT contractor to (1) establish the goals of the evaluation, (2) develop contractual arrangements, (3) set up dates
for their participation in the evaluation, (4) supply laboratory assistance
as needed, (5)
monitor the inspections, and (6) review the final results. The EMAT
contractors who
were contacted were Babcock and Wilcox (B&W); Quality Network on behalf of the
Institute for Material Testing (IzfP); Welding Inspection Systems, Inc.
(WIS); Industrial
Sensors; Sonic Sensors; and Tecrad. Although all of these contractors
established contracts with SwRI to participate in the MTI evaluation, only
B&W, Quality Network, and
WIS had participated by the July 31, 1996, deadline. Sonic Sensors and
Tecrad were
developing EMAT systems for the evaluation, but stated that they were not
ready for
the high-temperature testing at the present time. Industrial Sensors
decided not to participate due to a heavy workload and no time to dedicate
to the evaluation.
To provide a totally unbiased review of EMAT capabilities, SwRI worked with MTI
from the viewpoints of industrial application and the need to develop a
test protocol
that would provide information about the inspection sensitivities and
effects of material
and temperatures on the inspection capabilities. This protocol was also
influenced by
SwRI's understanding of the present state of the art of EMAT technology.
The development of the test protocol, design of the test specimens and test
fixtures, and basic
ground rules for conducting the evaluations are discussed in Sections 2.1
through 2.4.
2.1Development of Test Protocol and Contractor Guidelines to the Test
Protocol
The Test Protocol and Contractor Guidelines to the Test Protocol were
developed
with the intention of providing an objective evaluation of the state of the
art of EMAT
technology. The Test Protocol was prepared for limited distribution to
representatives
of SwRI and members of MTI's EMAT Resource Group, while the Contractor
Guidelines to the Test Protocol was distributed to the EMAT contractors.
The Test Protocol
provided (1) the objective of the EMAT evaluation, (2) general testing
rules to the vendors, (3) a description of the test specimens and the
defects that each contained, (4) the
temperatures at which each test specimen would be tested, and (5) the
recording procedure. The Contractor Guidelines to the Test Protocol
provided (1) a general discussion of the objective of the evaluation, (2) a
general description of the test specimens, (3)
testing rules, and (4) a discussion of the data recording and analysis
procedures.
2.2 Design and Fabrication of Test Specimens for the EMAT
Evaluation
To meet the goal of this project, test specimens were designed and
fabricated. They
provided corrosion-like defects and remaining wall thicknesses in the
ranges expected
in the industry. These specimens consisted of a 0.75-inch-thick A 36 carbon
steel plate, a
0.75-inch-thick In600 plate, and a 0.5-inch-thick carbon steel plate. The
types of defects
contained in the test specimens consisted of round-bottom holes (RBHs) with
diameters
ranging from 0.125 to 1 inch and flat-bottom holes (FBHs) that were 1.5
inches in diameter. The RBHs had depths ranging from 0.25 to 0.65 inch and
provided remaining wall
thicknesses ranging from approximately 0.5 to 0.1 inch. The FBHs had depths
ranging
from 0.55 to 0.675 inch and provided remaining wall thicknesses ranging
from 0.2 to
0.075 inch.
Each specimen was inspected by SwRI using conventional 0-degree
longitudinal
wave ultrasonic inspection technology at ambient temperature as a control
for the
EMAT results. A 0.25-inch-diameter, 5-MHz, dual-delay line transducer was
used with
a Sonic Mark I as the ultrasonic instrument. The results obtained for each
specimen are
listed in Table 1.
Table 1
COMPARISON OF REMAINING WALL THICKNESS MEASURED MECHANICALLY (M), MEASURED
ULTRASONICALLY (UT), AND AS DESIGNED (D) FOR EACH BLOCK
(The design remaining wall thickness is based upon using the actual
material
thickness obtained for each block and subtracting the depth of the designed
hole.) |
A practice block was designed and six were fabricated. A practice
block was sent
to each potential EMAT contractor prior to their evaluation at SwRI so that
they could
prepare for the evaluation.
2.3 Design and Fabrication of the High-Temperature Test Fixture
To conduct the evaluation at high temperatures, a test fixture was
designed and
fabricated that would allow the test specimen to be heated to and
maintained at the
proper temperature during the high-temperature test. Some of the EMAT
contractors
requested approval to conduct the high-temperature test with the specimens
in the vertical orientation, while others preferred the horizontal
orientation.
The heating coils on the test fixture were controlled by a Reliant
system that consisted of a TR800 Power Supply, TR600 DC Rectifier, SCD8000
Program Temperature
Controller, and SCC75-3004 Contactor Box. The feedback loop was
accomplished using
a thermocouple placed in a hole in the side of the test specimen.
Temperature crayons
were used to touch the surface and verify that test temperatures (ranging
from 700 to
1200°F in 100-degree increments) had been achieved and maintained.
This fixture at
high temperatures is shown in the horizontal and vertical orientations in
Figures 1 and
2, respectively.
The top and left edges of the test fixture were marked with the
letters A through G
and the numbers 1 through 9, respectively. This was to assist the teams in
properly
marking their results on the data sheets.
2.4 Laboratory EMAT Evaluations
The laboratory evaluations were set up to accommodate the three teams. The
ambient tests were conducted in a laboratory setting, and the
high-temperature tests
were conducted in a welding shop. The teams were given data sheets for each
test specimen prior to starting the tests. Each team was asked to mark the
data sheets with the
locations of each defect detected as well as provide the remaining wall
thicknesses.
Also, if the teams could provide an estimate of the defects' diameters,
this information
was to be illustrated on the data sheet. An example of a data sheet is
shown for one
specimen in 
Figure 3.
The goal was to allow each team 2 days to set up their EMAT equipment
and conduct the evaluation on three ambient-temperature plates and two
high-temperature
plates. The first day was to be used to set up the equipment and inspect
the three test
specimens at ambient temperature. The completed data sheets were given to
the SwRI
monitor by the following morning before the team could conduct the
evaluation on the
high-temperature test specimens. After completing all the evaluations, the
data sheets
for the high-temperature specimens were given to the SwRI monitor. After
all the tests
were completed, each team was debriefed regarding the defects in each test
specimen.
3.1 Equipment and Setup Time
Of the EMAT contractors who did participate, the complexity and
potential fieldability varied widely. One had numerous boxes, one had all
the components in a single
large box, and one had two boxes. All of them digitized data for later
review. Two used
manual scanning and had to record the data manually. One used a semimechanical
scanning system and was able to construct C-scan images.
3.1.1 Contractor No. 1
Contractor No. 1 arrived on the morning of the first day
expecting their
equipment (shown in Figure 4) to be at SwRI. However, it did not arrive
until approximately 2:00 p.m. They began unpacking and setting it up. Their
equipment consisted of
the EMAT probe, pulser/receiver instrumentation, and a data processing
unit. They
began initial equipment testing but encountered many problems getting their
equipment to function properly. They stopped at approximately6:45 p.m. and
decided to
begin again the following morning. On the second day, they encountered a
number of
problems. However, by 3:00 p.m. they had completed their initial equipment
evaluation
and calibration. After approximately 3 days of difficulties, they requested
a second
chance. SwRI agreed, since no information about the actual test specimens
had been
given to them.
Approximately 3 weeks later, Contractor No. 1 returned to conduct
their evaluations. During this evaluation, they had their equipment running
within approximately
3 hours. They conducted tests on Specimens CS-1, CS-2, and CS-3 in 2 days.
3.1.2 Contractor No. 2
Contractor No. 2 set up their equipment (shown in Figure 5)
within approximately 2-1/2 hours. The initial equipment evaluation and
calibration were completed
within approximately 1 hour. They completed tests on Specimens CS-1 and
CS-3 within
3 days.
3.1.3 Contractor No. 3
Contractor No. 3 arrived around 4:00 p.m., unpacked their
equipment, and
completed their initial equipment setup by 5:30 p.m. Their equipment (shown
in Figure 6) consisted of a manually operated scanner (with potentiometer
position readout
along one axis), the pulser/receiver instrument, and the EMAT probe. They
conducted
their calibration the next morning and were ready to begin the EMAT tests
by 10:00 a.m.
They completed data collection on Specimens CS-1, CS-2, and CS-3 within 2 days.
3.2 Data Presentation by Test Specimen
This section provides a general discussion of the test results and
data obtained by
the teams for each test specimen.
3.2.1 Test Specimen CS-1
Test specimen CS-1 was a 0.75-inch-thick A36 carbon steel plate
with 18
machined defects and two natural flaws. It was tested at ambient
temperature. The
ground truth data for defect locations, approximate defect sizes, and
remaining wall
thicknesses are shown in Figure 7. (The term "ground truth data" is defined
as mechanically measured with calipers.)
Table 2
LIST OF DEFECTS CONTAINED IN TEST SPECIMEN CS-1 (INSPECTION OF 0.75
INCH CARBON STEEL AT AMBIENT TEMPERATURE) WITH ASSOCIATED
DEFECT DIAMETER, MECHANICALLY MEASURED REMAINING WALL, EMAT
TEAM RESULTS, AND CONVENTIONAL UT RESULTS |
The results obtained by each team are summarized in Table 2.
Contractor
No. 1 detected a total of 18 defects, including 17 machined defects and one
natural
defect. They missed one 0.125-inch-diameter RBH (which had approximately
0.2 inch
remaining wall thickness, located in E8 of the test specimen) and one
natural defect
(which was estimated to be about 0.125 inch in diameter and had a remaining
wall thickness of approximately 0.4 inch, located in D8). The results for
remaining wall thicknesses are also given. In cases where the remaining
wall thicknesses were not determined, the notation Loss of BW
(backwall) was used to notate defect detection and "no
data" for remaining wall. The data showed that most of the remaining wall
thickness
values were within 1 to 20 percent. However, there were a number of
remaining wall
thickness errors that were much larger, i.e., 65 (D5) and 165 percent (D3).
They also
made a false call, i.e., they found a defect, shown in F4, where no defect
was intentionally placed or found by conventional ultrasonics.
Contractor No. 2 detected all 18 machined flaws. They did not
detect the
two naturally occurring flaws, but did have a questionable indication where one natural
defect was located in D8. The error of the remaining wall thickness data
ranged from 6
to 150 percent. They used a different ultrasonic wave mode to determine
remaining
wall thicknesses than the other two contractors.
Contractor No. 3 detected all 18 machined flaws, but they did not
detect the
two naturally occurring flaws. They reported remaining wall thicknesses for
eight
defects. For these defects, the errors in the remaining wall thickness data
were between
0.3 and 9 percent.
3.2.2 Test Specimen CS-2
Test specimen CS-2 was the same as CS-1, except that it was
rotated 180
degrees, had the name tag located on the upper left side, and was tested at
800°F. Contractor No. 2 did not have EMAT equipment designed for the
high-temperature tests.
The actual inspection approaches used by the other two contractors at high
temperatures were different. Contractor No. 1 positioned the test specimen
horizontally and
used water cooling. Contractor No. 3 positioned it in a vertical
orientation and used air
cooling. Both teams found the heat difficult to work around. The ground
truth data for
defect locations, approximate defect sizes, and remaining wall thicknesses
for CS-2 are
shown in Figure 8. The results obtained for Contractors No. 1 and No. 3 are
summarized in Table 3.
Table 3
LIST OF DEFECTS CONTAINED IN TEST SPECIMEN CS-2 (INSPECTION OF 0.75-INCH CARBON
STEEL AT 800°F) WITH ASSOCIATED DEFECT DIAMETER, MECHANICALLY MEASURED
REMAINING WALL, AND EMAT TEAM RESULTS (CONVENTIONAL UT WAS NOT
CONDUCTED AT HIGH TEMPERATURE) |
Contractor No. 1 detected 7 defects. They missed one
0.25-inch-diameter
RBH (with remaining wall thickness of 0.119), shown as B3. They also did
not detect the
two naturally occurring flaws. It appears that the machined flaw was missed
because of
the difficulty associated with scanning the block at high
temperatures‹not because of
any EMAT signal information. They reported remaining wall thickness data
for 7 of the
16 defects, with an accuracy of 20 percent or better (except for one, which
was at 30
percent). (It should be noted that the defect at location C3 was originally
reported to
have a remaining wall thickness of 0.43 inch. However, they later presented
supporting
information showing they had provided the correct data but had misread the
data, and
that the remaining wall thickness should have been 0.196 inch. This value
was within 11
percent of the actual remaining wall thicknesses.)
Contractor No. 3 detected all 18 machined flaws, but did not
detect the two
naturally occurring flaws. They reported remaining wall thicknesses for
five defects,
with an accuracy between 1 and 9 percent.
3.2.3 Test Specimen CS-3
Test specimen CS-3 was a 0.5-inch-thick A36 carbon steel plate,
which contained 10 machined flaws. All three teams inspected this specimen
at ambient temperature. A conventional ultrasonic inspection was also
conducted. The results are summarized in Table 4.
Table 4
LIST OF DEFECTS CONTAINED IN TEST SPECIMEN CS-3 (INSPECTION OF 0.5-INCH CARBON STEEL AT AMBIENT TEMPERATURE) WITH
ASSOCIATED DEFECT DIAMETER, MECHANICALLY MEASURED REMAINING WALL, EMAT TEAM
RESULTS,
AND CONVENTIONAL UT RESULTS |
Contractor No. 1 detected all 10 flaws. This team reported
remaining wall
thickness data for 8 of the 10 flaws. Most of the errors in the remaining
wall thickness
data ranged from 2 to 18 percent; however, two defects had errors of
approximately 60
percent.
Contractor No. 2 detected all 10 flaws and reported remaining
wall thickness data for all flaws. The errors in the remaining wall
thickness data ranged from 0 to
300 percent. The explanation for the high degree of error was that they had
not had a
0.5-inch-thick test plate with representative defects to practice on prior
to testing.
Contractor No. 3 detected all 10 flaws and reported remaining
wall thickness data for only two of the flaws. The errors in the remaining
wall thickness data were
low, i.e., 0 and 5 percent.
3.2.4 Test Specimens IN-1 and IN-2
Test specimen IN-1, a 0.75-inch-thick In600 steel plate, contained 18
machined flaws. It was to be tested at ambient temperature. Test specimen
IN-2 was the
same as test specimen IN-1 except that it was rotated 180 degrees and was
to be tested
at 1200°F. However, none of the contrctors chose to inspect these
specimens.
Figure 3.
Figure 7
Figure 8