ECNDT 2006 - Th.2.6.

Ultrasonic Phased Array Inspection of

Turbine Components
Waheed A. ABBASI, Michael F. FAIR, SIEMENS Power Generation, Pittsburgh, USA
More info about this article: https://www.ndt.net/?id=4126

Abstract. The advent and proliferation of Ultrasonic Phased Arrays as a method of

choice for the advanced inspection of industrial components has brought many of
the benefits of this technology into the mainstream. Some of the benefits have been
the reduction in inspection time resulting in cost savings as well as reduction in the
logistics that the conventional method required. While using Phased Arrays offers a
variety of benefits to suit the needs of different applications, in certain types of
complex exams it is imperative that proper methods of examination and validation
be developed. This paper addresses the development of Ultrasonic Phased Array
methods for inspection of turbine rotor discs and turbine blade attachments and the
validation of such methods. It also discusses issues as they pertain to probe
positioning, ultrasound behavior, and component geometry.

Introduction

Ultrasonic Phased Array inspection of industrial components is quickly becoming the

method of choice for inspections that were formerly done by conventional ultrasonics. This
is especially true for applications where complicated geometry is involved resulting in
multiple probe positions and numerous variations in incident angles. The benefits that the
Phased Arrays offer to address these issues have resulted in methods and applications being
developed for the inspection of turbine and generator components [1].
While phased arrays offer numerous advantages, care must be taken when
developing methods and techniques so that a valid and accurate examination is performed.
The methods developed should be tested and validated before an examination is considered
reliable. In this paper, development and validation of inspection methods for turbine disc
bores and turbine blade attachments is discussed.
Ultrasonic inspection of Westinghouse and other OEM nuclear Low Pressure (LP)
disc bores has been an established (and required) technique to ensure safe and reliable
operation of nuclear power generation units. Typically, conventional ultrasound has been
used to detect cracks on the disc bore and determine if any radial axial depth of the crack is
present. Based on the information gained from the ultrasonic inspection and other methods,
re-inspection interval and remaining life can be determined for the nuclear LP rotor discs.
Siemens PG has been developing ultrasonic phased array techniques to reduce inspection
times and improve data evaluation for the Westinghouse design LP disc bores as well as
other design disc bores [2, 3].
Corrosion attack on blade attachments can adversely affect the life expectancy of
steam turbines. Pitting is the first stage of corrosion attack and it occurs in the steeple
attachment serrations. If allowed to progress pitting can develop into more serious stress
corrosion cracking. Depending on the turbine design, one or more rows of blade
attachments may operate in an environment with wet steam that is subject to high stress
from blade loading. These are the ingredients necessary for stress corrosion cracking.

1
Siemens PG has developed a phased array technique to detect stress corrosion damage in
blade attachments while the blades are installed in the turbine disc (in-situ examination).

Modeling of Phased Arrays

As mentioned earlier, it is of extreme importance that the inspection methods that

are developed accurately addresse the requirements of the ultrasonic examination. To
fulfill this need the inspection process is modeled using a variety of methods and is then
tested and validated.
These methods must include the modeling of the inspection component geometry
and Phased Arrays to develop accurate control (focal) laws. In addition, the methods must
include means of determining the behavior of the incident beam as well as the behavior of
the reflected beam.
The modeling of the beam is even more critical when dealing with a compound
curvature considering the effect the surface has on the sound beam. In the case of a
continuously changing surface with
complex geometry, there are instances
when the curvature is positive in one
direction and negative in another.
This affects the sound beam
differently and has to be modeled to
get a sensible and accurate set of focal
laws for an examination.
Developing accurate focal
laws is the first step in formulating a
technically sound inspection
methodology. The next step is
determining the sound field that is
transmitted and verifying that it is
available at the area of interest. This Figure 1 Focal law simulation on a blade attachment.
is done by simulating the field of the
beam at the focal depth and
determining the quality of the sound
that is available for reflection. Based
on this information important
decisions can be made regarding the
scanning strategy of the area of
interest.
The preceding step is only one
side of the coin, though. What is also
of importance is the determination of
the quality of the signal received back
at the transducer from the area of
interest. There are various methods of
accomplishing this. A traditional
method is testing this experimentally Figure 2 Focal law simulation on a disc bore.
and evaluating the complete exam on test specimens. While this method is ideal, it may not
be a feasible or realistic method for reasons ranging from availability of sample specimens
to the time and cost involved. Another approach is the accurate simulation of the
examination and then analysis of the results. This provides enough information to devise a

2
scan strategy with the confidence that the results obtained would be meaningful and
reliable.
Figures 1 and 2 illustrate the simulation of focal laws on blade attachments and disc
bores respectively. The laws are developed using a CAD model so it is representative of a
true 3D space. No matter what the application, when dealing with complex geometry,
simulation of laws provides a valuable tool in determining the accuracy of the laws and
provides feedback to optimize the scans. Further, once the focal laws are developed for a
desired scan, analysis is required on the beam properties to ensure the validity of the focal
laws. Figures 3 and 4 illustrate the field of a beam at a focal point. This field diagram
provides information about the sensitivity of the sound beam at the focal point and is a
feedback tool for the user to vary the scan strategy. It should also be noted from looking at
these figures that the field shape is a function of the complex geometrical interface that the
sound goes through.
As is obvious from the above
discussion, these modeling and
simulation tools are of prime
importance establishing sound and
thorough inspection methods.

Inspection of Westinghouse Design

Blade Attachments

Until recently, a reliable

inspection of the blade attachment to
detect corrosion damage could only
be achieved by removing the blades
and performing a magnetic particle Figure 3 Beam Field at a Focal point (loss of dB’s).
(MP) inspection. This process is
highly labor intensive, time
consuming, and in terms of the
inspection results, difficult to
interpret due to limited access to the
blade attachment serrations.
Westinghouse blade
attachments are unlike other OEM
designs. Depending on the turbine
design the blade attachment
serrations can be straight, curved or
skewed side entry configuration,
which makes them extremely
difficult to examine unless the correct
geometry and dimensions are known.
Siemens PG has developed a Figure 4 Beam Field at a Focal Point (loss of dB’s).
phased array technique to detect
stress corrosion damage in blade attachments while the blades are installed in the turbine
disc. The obvious advantage of using phased arrays is not having to remove the turbine
blades. This saves an enormous amount of time in terms of manually removing the turbine
blades, not having to abrasively clean the blade attachment serrations (grit blasting) and
eliminating the chance of damaging the turbine blades while removing them from the disc
blade attachment.

3
 To completely inspect the blade attachment requires phased array sound to be
directed from under the disc rim at changing inspection angles; this is a difficult hurdle to
overcome due to the compound curvature geometry found underneath the disc rim that
complicates the phased array sound steering in some Westinghouse turbine designs.
The first step in developing a phased array inspection on the blade attachments is to
model the geometry of the attachment. The drawing of the blade attachment is made into a
3D CAD model. The model is inserted into phased array software being developed by
Siemens PG that generates the phased array focal laws by knowing the type of the phased
array probe, the wedge (incident angle), wedge position and CAD model characteristics, as
illustrated in Figure 5. This can determine if an inspection is realistic by showing if the
phased array sound can indeed be directed to the inspection area of the blade attachment.

Figure 5 Modeling of a Blade Attachment and development of Focal Laws.

To further verify that the focal laws that have been developed are accurate and will
yield the desired feedback, test blocks of the blade attachment were manufactured. EDM
notches were placed at the critical inspection areas and verification of the phased array
focal laws for the particular style blade attachment were documented, as illustrated in
Figure 6. The phased array focal laws, wedge design, and wedge positioning were
confirmed. An inspection plan was developed and a mock-up of a rotating disc was made to
test the manipulator scanning arm.

Figure 6 Modeling of a Blade Attachment and testing using a test block.

4
 Over the past year, Siemens PG has performed a number of phased array
inspections on Westinghouse design blade attachments. In particular, an LP rotor was
inspected that was over 40 years old and had a mixture of T-root and curved side entry
design blade attachments. The phased array inspection was part of a series of NDE exams
to extend the running life of this LP rotor. All type of blade attachments were inspected
according to the methods developed.
Nearly all of the blade attachments were inspected in a single revolution during the
actual field inspection of the LP rotor. The geometry of each blade attachment occurred as
predicted from the phased array models and from the verifications on the blade attachment
test blocks. This produced highly confident and reproducible test results. Figure 7 shows a
display from a curved side entry blade attachment. The illustration on the left shows a
section of the scan with the geometry from the blade attachment and an indication in the
bottom serration, while the illustration on the right shows an overlay of the geometry and
indication on the blade attachment.

Figure 7 Inspection display from a curved side entry blade attachment and overlaid data on geometry.

Phased array inspection on Westinghouse style blade attachment can be very

challenging. With the right approach using phased array modeling software, verification of
the phased array focal laws and a developed inspection plan the blade attachments can be
examined with high reproducibility and sensitivity. By knowing where geometry signals
will occur in complex blade attachments interrupting the results are made easier for the
inspector thus increasing the confidence of the inspection from false calls. And lastly the
phased array application can help place older equipment back into service and have
significant time savings where in the past this couldn’t be achieved by conventional NDE
methods.

Inspection of Westinghouse Design LP Disc Bores

A turbine disc bore inspection is a complex matter. The disc is machined with complex
contoured surfaces where the ultrasonic probe must be placed to direct the ultrasonic sound
to the disc bore surface. Westinghouse nuclear LP disc have special challenges compared to
other manufacturers’ disc styles. Typically, Westinghouse discs feature asymmetrical
geometry from the steam inlet and outlet sides of the disc (Figure 8), which in turn requires

5
more ultrasonic wedges (positions and incident angles) to inspect both sides. Many are also
heavy hub key-plated designs that have very long metal paths and “blind spots” caused by
the key-plate drive keys. Additional geometry from the key-plates introduces additional
signals that the inspector must interpret. Depending on the LP design there can be an axial
opening as small as 38mm (1.5 inches) between discs to insert the ultrasonic probe through
and perform the inspection.

Figure 8 A typical configuration of an LP Turbine.

Perhaps the best tool Siemens PG has is the experience and knowledge developed
over the years in nuclear LP disc bore inspection; this knowledge has been used in unison
with developing the phased array technique. The approach to deal with the Westinghouse
disc geometric complexities was to determine where the phased array sound could be
directed by using modeling methods developed in house for Westinghouse LP disc (Figure
9). Phased Array sound can be
directed over a large area of the
disc bore from a single position,
however, much of the coverage is
not useful detecting a crack. The
model can determine the most
effective “positions” according to
guidelines established from past
disc bore development programs
and knowledge based on the
present conventional ultrasonic
disc bore method. The model
also generates the phased array
focal laws to steer the phased
array sound to the area of interest
using previously determined
required coverage and wedge Figure 9 Focal law development for a turbine disc bore.
design.

6
 Figure 10 Test Disc with artificial flaws to test and verify phased array inspection techniques.

Two nuclear LP discs with artificial flaws were manufactured to confirm the phased
array transducer design, wedge design and coverage; one of the discs is illustrated in Figure
10. The two discs selected represent the most complex conditions for ultrasonic
examination found in the Westinghouse disc design due to thickness (long metal paths) and
geometry.
The test discs were inspected with the present conventional ultrasonic disc method
using the wedges, transducers and time of flight tables that would be used during an actual

Figure 11 Phased Array scan (left) and the examination model (right).

7
inspection. The same exams were repeated with the phased array technique that has been
devised. The results using the phased array method confirmed that the results obtained by
the phased array technique were at least equal to and in most cases better than the present
conventional inspection method.
Figure 11 shows a phased array scan (on left) and the phased array model (on right)
of a Westinghouse LP disc. The phased array scan is inspecting a region on the disc bore
that would require 6 conventional wedges as compared to one wedge for the phased array
scan. Obviously a great reduction in time for the inspection can be achieved with phased
array by reducing the scanning time due to a reduction in wedges required by conventional
ultrasonic method. Figure 12 shows the wedges required to inspect an outlet side of one of
the Westinghouse test disc while Figure 13 shows the wedges required to perform the same
inspection using a phased array transducer.

Figure 12 Wedges required to inspect an area of the Figure 13 Wedges required to inspect the
disc using conventional ultrasound. same area of the disc using Phased Arrays.

The phased array method is a powerful tool to reduce the number of scans and
address the difficulties due to geometry and disc design on Westinghouse nuclear LP bore
exams. Obstructions caused by disc hardware can be handled by wedge design, phased
array transducer design, and scan locations. Spurious indications caused by disc hardware
and geometry are less complicated to interpret because of screen presentations of the signal
evolution, relationships between different phased array angles, and phased array models of
the disc. Detection and sizing is equal to or better than the conventional ultrasonic method.
The phased array technique developed for Westinghouse type disc can be applied to other
OEM style disc bore exams.

Conclusions

Phased array ultrasonic methods have been developed for the inspection of various
turbine components, specifically the blade attachments of turbine discs and the bores of
turbine discs. These methods offer significant improvements over the conventional
ultrasonic inspections and provide significant savings in time and cost. These methods are
based on a sound development cycle where the inspection technique is developed, verified,
and tested to prove its validity and worthiness for field use. It should be noted that for a
reliable inspection it is imperative that the thorough development cycle of part geometry
modeling (especially for complex geometries), optimum transducer design, accurate focal
law calculation, and beam characterization be followed. Failing to follow this process can
result in an inadequate inspection.

8
References

[1] Poguet, Jerome, et. al, “Phased Array Technology: concepts, probes, and applications.” NDT.net,
Vol. 7, No. 05, May 2002.
[2] Fair, M. F. et. al, “Advanced In-Service Inspections of Turbine Generator Components using Phased
Array Technology.” Seventh EPRI Steam Turbine/Generator Workshop & Expo, Baltimore, MD,
August 20-24, 2001.
[3] Clossen, Michael, et. al, “Advanced NDE Inspection Methods for Detection of SCC in Blade
Attachments and Blade Roots.” Ninth EPRI Steam Turbine/Generator Workshop & Expo, Denver,
CO, August 22-24, 2005.