Nondestructive Evaluation: CIVA Software

Comparisons for Ultrasonic Techniques

1022949

10290640
10290640
Nondestructive Evaluation: CIVA Software Comparisons
for Ultrasonic Techniques

1022949

Technical Update, October 2011

EPRI Project Managers

M. Dennis
G. Connolly

This document does NOT meet the requirements of

10CFR50 Appendix B, 10CFR Part 21,
ANSI N45.2-1977 and/or the intent of ISO-9001 (1994)

ELECTRIC POWER RESEARCH INSTITUTE

3420 Hillview Avenue, Palo Alto, California 94304-1338 ▪ PO Box 10412, Palo Alto, California 94303-0813 ▪ USA
800.313.3774 ▪ 650.855.2121 ▪ askepri@epri.com ▪ www.epri.com

10290640
DISCLAIMER OF WARRANTIES AND LIMITATION OF LIABILITIES
THIS DOCUMENT WAS PREPARED BY THE ORGANIZATION(S) NAMED BELOW AS AN ACCOUNT OF
WORK SPONSORED OR COSPONSORED BY THE ELECTRIC POWER RESEARCH INSTITUTE, INC. (EPRI).
NEITHER EPRI, ANY MEMBER OF EPRI, ANY COSPONSOR, THE ORGANIZATION(S) BELOW, NOR ANY
PERSON ACTING ON BEHALF OF ANY OF THEM:

(A) MAKES ANY WARRANTY OR REPRESENTATION WHATSOEVER, EXPRESS OR IMPLIED, (I) WITH
RESPECT TO THE USE OF ANY INFORMATION, APPARATUS, METHOD, PROCESS, OR SIMILAR ITEM
DISCLOSED IN THIS DOCUMENT, INCLUDING MERCHANTABILITY AND FITNESS FOR A PARTICULAR
PURPOSE, OR (II) THAT SUCH USE DOES NOT INFRINGE ON OR INTERFERE WITH PRIVATELY OWNED
RIGHTS, INCLUDING ANY PARTY'S INTELLECTUAL PROPERTY, OR (III) THAT THIS DOCUMENT IS
SUITABLE TO ANY PARTICULAR USER'S CIRCUMSTANCE; OR

(B) ASSUMES RESPONSIBILITY FOR ANY DAMAGES OR OTHER LIABILITY WHATSOEVER (INCLUDING
ANY CONSEQUENTIAL DAMAGES, EVEN IF EPRI OR ANY EPRI REPRESENTATIVE HAS BEEN ADVISED
OF THE POSSIBILITY OF SUCH DAMAGES) RESULTING FROM YOUR SELECTION OR USE OF THIS
DOCUMENT OR ANY INFORMATION, APPARATUS, METHOD, PROCESS, OR SIMILAR ITEM DISCLOSED IN
THIS DOCUMENT.

REFERENCE HEREIN TO ANY SPECIFIC COMMERCIAL PRODUCT, PROCESS, OR SERVICE BY ITS

TRADE NAME, TRADEMARK, MANUFACTURER, OR OTHERWISE, DOES NOT NECESSARILY
CONSTITUTE OR IMPLY ITS ENDORSEMENT, RECOMMENDATION, OR FAVORING BY EPRI.

THE FOLLOWING ORGANIZATION PREPARED THIS REPORT:

Electric Power Research Institute (EPRI)

THE TECHNICAL CONTENTS OF THIS DOCUMENT WERE NOT PREPARED IN ACCORDANCE

WITH THE EPRI NUCLEAR QUALITY ASSURANCE PROGRAM MANUAL THAT FULFILLS THE
REQUIREMENTS OF 10 CFR 50, APPENDIX B AND 10 CFR PART 21, ANSI N45.2-1977 AND/OR
THE INTENT OF ISO-9001 (1994). USE OF THE CONTENTS OF THIS DOCUMENT IN NUCLEAR
SAFETY OR NUCLEAR QUALITY APPLICATIONS REQUIRES ADDITIONAL ACTIONS BY USER
PURSUANT TO THEIR INTERNAL PROCEDURES.

This is an EPRI Technical Update report. A Technical Update report is intended as an informal report of
continuing research, a meeting, or a topical study. It is not a final EPRI technical report.

NOTE
For further information about EPRI, call the EPRI Customer Assistance Center at 800.313.3774 or
e-mail askepri@epri.com.

Electric Power Research Institute, EPRI, and TOGETHER…SHAPING THE FUTURE OF

ELECTRICITY are registered service marks of the Electric Power Research Institute, Inc.

Copyright © 2011 Electric Power Research Institute, Inc. All rights reserved.

10290640
 ACKNOWLEDGMENTS
The following organization prepared this report:
Electric Power Research Institute (EPRI)
Nondestructive Evaluation (NDE) Program
1300 West W.T. Harris Blvd.
Charlotte, NC 28262
Principal Investigators
M. Dennis
T. Seuaciuc-Osorio
G. Connolly
W. Ratliff
F. Yu
This report describes research sponsored by EPRI.
EPRI would like to acknowledge Roy Salisbury and Thomas Stafford of System One, whose
efforts led to the successful completion of this report.

This publication is a corporate document that should be cited in the literature in the following
manner:
Nondestructive Evaluation: CIVA Software Comparisons for Ultrasonic Techniques. EPRI, Palo
Alto, CA: 2011. 1022949.

10290640 iii
10290640
 ABSTRACT
This report documents the status of an independent validation of CEA’s CIVA simulation
software for ultrasonics performed by EPRI. A general overview of the CIVA program inputs is
presented, along with some comparisons between experimental and CIVA simulated results for
both conventional ultrasonics and phased array methods. The examples given range from a very
simple yet often numerically challenging scenario analyzing the multiple responses from a steel
block to a dissimilar metal pipe weld containing several interfaces. Overall, the CIVA ultrasonic
module output compared well with the experimental results.

Keywords
CIVA
CEA
Abaqus
NDE simulation
NDE modeling

10290640 v
10290640
 CONTENTS
1 INTRODUCTION ....................................................................................................................1-1
Motivation.............................................................................................................................1-1
Background ..........................................................................................................................1-1
2 CIVA SOFTWARE..................................................................................................................2-1
Generic Simulation Procedure .............................................................................................2-1
Specimen .......................................................................................................................2-1
Probe..............................................................................................................................2-2
Inspection.......................................................................................................................2-2
Array Settings.................................................................................................................2-2
Flaws..............................................................................................................................2-2
Computation Parameters ...............................................................................................2-2
3 COMPARISON OF RESULTS ...............................................................................................3-1
Step Block ............................................................................................................................3-1
Equipment ......................................................................................................................3-1
Experimental Procedure.................................................................................................3-2
Simulation Procedure.....................................................................................................3-2
Comparison of Results...................................................................................................3-5
Abaqus Simulation Procedure........................................................................................3-8
Notch Flat Block .................................................................................................................3-10
Equipment ....................................................................................................................3-10
Experimental Procedure...............................................................................................3-11
Simulation Procedure...................................................................................................3-13
Comparison of Results.................................................................................................3-16
706 Series Dissimilar Metal Pipe Weld ..............................................................................3-16
Equipment ....................................................................................................................3-18
Experimental Procedure...............................................................................................3-19
Simulation Procedure...................................................................................................3-21
Comparison of Results.................................................................................................3-23
Flexible Array Simulations on Tapered Aluminum Reference Block..................................3-23
Equipment ....................................................................................................................3-25
Experimental Procedure...............................................................................................3-25
Simulation Procedure...................................................................................................3-31
Comparison of Results.................................................................................................3-31
4 SUMMARY .............................................................................................................................4-1

10290640 vii
10290640
 LIST OF FIGURES
Figure 3-1 Stainless steel step block ........................................................................................3-1
Figure 3-2 1.5-MHz probe atop the fourth step of the stainless steel step block ......................3-2
Figure 3-3 Experimentally measured input pulse for CIVA simulations at 1.5-MHz frequency.3-3
Figure 3-4 CIVA screen shot of output at 1.5 MHz ...................................................................3-4
Figure 3-5 CIVA screen shot of output at 3.5 MHz ...................................................................3-4
Figure 3-6 CIVA screen shot of output at 5.0 MHz ...................................................................3-5
Figure 3-7 CIVA and UV data comparison with attenuation value of 0.142 dB mm-1, plotted
against time (µs) with normalized vertical axis (at 5.0 MHz)......................................................3-6
Figure 3-8 CIVA and UV data comparison with attenuation value of 0.082 dB mm-1, plotted
against time (µs) with normalized vertical axis (at 5.0 MHz)......................................................3-6
Figure 3-9 CIVA and UV data comparison with attenuation value of 0.044 dB mm-1, plotted
against time (µs) with normalized vertical axis (at 5.0 MHz)......................................................3-7
Figure 3-10 FFT of first echo from backwall of step 4 from 5.0 MHz data ................................3-8
Figure 3-11 CIVA screen shot of sound field visualization........................................................3-8
Figure 3-12 Comparison of Abaqus 2D (top) and 3D (bottom) simulations (blue) with
experimental data (Ex) in red.....................................................................................................3-9
Figure 3-13 Photograph of 5.0-MHz, 0.5-inch diameter transducer on notch block................3-10
Figure 3-14 End view of notch block with dimensions ............................................................3-11
Figure 3-15 Model of CIVA simulation; end of block and notches shown in red and probe
path shown in purple................................................................................................................3-12
Figure 3-16 Unprocessed UltraVision data flaw images at 1.5 MHz (transducer orientation:
looking toward the deeper notch).............................................................................................3-12
Figure 3-17 Unprocessed UltraVision data flaw images at 5.0 MHz (transducer orientation:
looking toward the deeper notch).............................................................................................3-13
Figure 3-18 CIVA simulation screen shot of the 1.5-MHz results (transducer orientation:
looking toward the shallower notch).........................................................................................3-14
Figure 3-19 CIVA simulation screen shot of the 1.5-MHz results (transducer orientation:
looking toward the deeper notch).............................................................................................3-14
Figure 3-20 CIVA simulation screen shot of the 5.0-MHz results (transducer orientation:
looking toward the shallower notch).........................................................................................3-15
Figure 3-21 CIVA simulation screen shot of the 5.0-MHz results (transducer orientation:
looking toward the deeper notch).............................................................................................3-15
Figure 3-22 Illustration of the 706 stainless steel weld ............................................................3-16
Figure 3-23 Cross-sectional dimensions of the EPRI 706 configuration stainless steel weld..3-17
Figure 3-24 Flaw information for the EPRI 706 configuration stainless steel weld ..................3-17
Figure 3-25 Zetec Dynaray phased array UT system ..............................................................3-18
Figure 3-26 MultiX LF phased array UT system ......................................................................3-18
Figure 3-27 R/D Tech MCDU-02 motion controller ..................................................................3-19
Figure 3-28 ATCO LPS-1000 2-axis automated scanner ........................................................3-19
Figure 3-29 Scanning procedure around the pipe section, showing array probe and wedge ..3-20
Figure 3-30 Unprocessed experimental data output, showing reflections from the two flaws .3-21
Figure 3-31 Location of two flaws perpendicular to the profile plane.......................................3-22
Figure 3-32 Unprocessed CIVA output for a circumferential scan of the first flaw...................3-22
Figure 3-33 Unprocessed CIVA output for a circumferential scan of the second flaw .............3-23
Figure 3-34 Photograph of aluminum tapered reference block................................................3-24
Figure 3-35 CIVA model of aluminum tapered reference block ...............................................3-24
Figure 3-36 Schematic of aluminum tapered reference block .................................................3-25
Figure 3-37 Imasonic 1.5-MHz 24-element 1D linear flexible phased array probe ..................3-26

10290640 ix
Figure 3-38 Zetec phased array calculator for the front 12 elements ......................................3-26
Figure 3-39 Zetec phased array calculator for the middle 12 elements ...................................3-27
Figure 3-40 Zetec phased array calculator for the back 12 elements ......................................3-28
Figure 3-41 Tapered aluminum reference block automated scan setup with a probe skew of
270°..........................................................................................................................................3-29
Figure 3-42 Ultrasonic response from flaw A using the back 12 elements and a 45°
longitudinal wave .....................................................................................................................3-29
Figure 3-43 Ultrasonic response from flaw D using the middle 12 elements and a 41°
longitudinal wave .....................................................................................................................3-30
Figure 3-44 Tapered aluminum reference block automated scan setup with a probe skew of 90°
.................................................................................................................................................3-30
Figure 3-45 Ultrasonic response from flaw A using the back 12 elements with a 46° longitudinal
wave.........................................................................................................................................3-31
Figure 3-46 Unprocessed CIVA output for a scan of flaws A and D with a probe skew of 270° ..3-
32
Figure 3-47 Unprocessed CIVA output for a scan from the taper with a probe skew of 90°....3-32

10290640 x
1
INTRODUCTION
Motivation
In response to frequent requests, this work is being performed to initiate an independent
assessment of the accuracy of CEA’s CIVA simulation software for nondestructive testing
(NDT) inspections. The results of this evaluation will be used by EPRI members to make
informed decisions regarding the relevance of the CIVA software output to their particular
applications.
Although EPRI has the capabilities to perform such a validation exercise, this work has never
been completed by EPRI. EPRI members would prefer that this software be tested prior to their
acceptance of the CIVA results. If successful, the results could be used as technical justification
to reduce or possibly eliminate the need for site-specific procedure demonstrations.

Background
The importance of mathematical models to assist in predicting inspection results has increased
with the need to properly demonstrate nondestructive evaluation (NDE) procedures prior to field
use. Due to advances in computing technologies, a handful of companies are now including NDT
inspection software tools as part of their product offerings. One of the most commonly used
programs is the CIVA software developed by CEA. CIVA allows users to perform simulations
for ultrasonic, eddy current, and radiographic inspection methods. It also provides a three-
dimensional graphical user interface for improved visualization and familiar data displays
consistent with each NDE technique.
Although the CIVA software development specialists have performed internal validations, an
independent assessment regarding the accuracy of this software is needed. EPRI owns a license
to use the CIVA software and also has access to component mockups, instruments, probes, and
existing data required to conduct an independent verification of the CIVA software output. This
report documents ultrasonic simulation results of a CIVA software validation for review and use
by EPRI members.

10290640 1-1
10290640
2
CIVA SOFTWARE
CIVA is a purpose-built computer software program for nondestructive testing and evaluation
(NDT&E) purposes. It is developed and maintained by the Commissariat à l’Energie Atomique
(CEA), a technological research organization funded by the French government, headquartered
in the southwestern suburbs of Paris. CEA has over 15,000 employees and 10 major research
centers. CIVA is distributed by Extende to customers in Europe and by Bercli to customers in
North America. EPRI has used CIVA for several years and continues to build working
knowledge and expertise in its features and applications.
CIVA contains modules for ultrasonic (UT) NDE, eddy current (ET) NDE, and radiographic
(RT) NDE, although this report is only concerned with UT. Being a complete NDE package, it
allows the user to follow a working procedure from defining an inspection geometry and
inserting flaws to producing results and extracting various images and scans.
CIVA continues to be extended and upgraded over time. In particular, recent upgrades have
tackled:
• Semi-analytical expressions to more accurately model transmission fields and reception
fields after interaction with a defect
• Adding compatibility with CAD/CAE such that parts and defects can now be imported
• Allowing the modeling and simulation of flexible phased array probes
• Improved visualization for anisotropic materials, visualizing both their slowness surfaces and
their elastic constant orientations within the specimen
• Incorporation of ray-tracing tools

Generic Simulation Procedure

CIVA has a standardized workflow under the following series of tabs within the graphical user
interface (GUI):
• Specimen
• Probe
• Inspection
• Array Settings
• Flaws
• Computation Parameters
Specimen
First, a specimen geometry is defined. The required geometry can either be built from standard
geometries provided by the CIVA software (cylinder, cone, sphere, nozzle, elbow, or bore), or it
can be imported as a 2D or 3D CAD model. (If imported as a 2D CAD model, it can then be

10290640 2-1
revolved or extruded.) The material properties are also defined, to include properties such as
anisotropy, attenuation, and noise, and they are assigned to regions of the model.

Probe
The probe parameters are entered under this tab. The crystal shape is chosen, the probe type is
selected, and the pattern and dimensions and pitch of the elements (if applicable) are defined. If
necessary, the focusing of the crystal(s) is defined and the wedge properties and dimensions are
also entered.

Inspection
The movement of the probe is defined in this section. It can move in three dimensions with a
given number of steps in each dimension. If applicable, the material properties of the coupling
medium and the bottom medium are supplied. A scan preview allows the user to check the range
and nature of the defined inspection pattern.

Array Settings
Once the wedge and the probe or array properties have been defined, in this section how they are
used to introduce the sound into the specimen is detailed. The initialization and the sequencing
information are given, in addition to the delay laws for both transmission and reception. Delay
laws can be calculated within CIVA via a ray-tracing routine, or they can be imported by the
user. The GUI allows the user to quickly visualize the delay laws as a function of element index,
and the rays are drawn to the main window, allowing a visualization of the ray path.

Flaws
Flaws are defined in this section. Like the specimen geometry, they can be built from CIVA
radicals (flat-bottom hole, side-drilled hole, hemispherical-bottom hole, spherical, plane
cylindrical, or ellipsoidal), or imported separately as a CAD-contoured plane. Flaws can be
highlighted separately in the GUI.

Computation Parameters
The types of rays that are used for computation are chosen: backwall, mode conversion, creeping
waves, and echoes can be activated or deactivated. It is possible to specify which modes (or
combination of modes) are allowed to propagate along the ray’s journey.

10290640 2-2
3
COMPARISON OF RESULTS
In this section, the comparison between the CIVA simulated ultrasonic results and the
experimental results are presented for four different specimens:
• A steel step block of five steps
• A flat block with two cut notches of varying depths
• A 706 Series dissimilar metal pipe weld (generic BWR N2 replacement nozzle design)
• An aluminum reference block

Step Block
The step block, pictured in Figure 3-1, is a structure whose base measures 0.75 in. × 3.75 in. The
height of each step varies from 0.1 in. to 0.5 in. in 0.1-in. increments. The longitudinal wave
velocity was measured to be 5772.72 m s-1. For the CIVA simulations, it was required that the
transverse wave velocity and the material density be provided, and these values were 3060 m s-1
and 7900 kg m-3, respectively.

Figure 3-1
Stainless steel step block

Equipment
• R/D Tech Tomoscan III/PA
• Imasonic 1.5-MHz, 0.5-in. conventional ultrasonic probe
• Imasonic 3.5-MHz, 0.375-in. conventional ultrasonic probe
• Imasonic 5.0-MHz, 0.375-in. conventional ultrasonic probe
• Stainless steel step block

10290640 3-1
Experimental Procedure
Experimental data were collected using the R/D Tech Tomoscan III/PA system via several types
of conventional ultrasonic probes placed atop each step in turn, as illustrated in Figure 3-2. Data
were visualized and extracted using the Zetec UltraVision (UV) 3.2R9.

Figure 3-2
1.5-MHz probe atop the fourth step of the stainless steel step block

Simulation Procedure

The specimen was imported as a 2D CAD geometry and then extruded to form the final 3D
structure. The probe is a circular single-element crystal with flat focusing, simulated to emit the
experimentally measured pulse shown in Figure 3-3 for the 1.5-MHz case. The coupling medium
was defined as water and the bottom medium as air. No defect was inserted into the specimen,
where up to 20 round trips (all longitudinal waves) were simulated. In total, five simulated
a-scans were collected corresponding to each of the five steps. A typical CIVA output is shown
in Figure 3-4 for 1.5 MHz, Figure 3-5 for 3.5 MHz, and Figure 3-6 for 5.0 MHz.

10290640 3-2
Figure 3-3
Experimentally measured input pulse for CIVA simulations at 1.5-MHz frequency

10290640 3-3
Figure 3-4
CIVA screen shot of output at 1.5 MHz

Figure 3-5
CIVA screen shot of output at 3.5 MHz

10290640 3-4
Figure 3-6
CIVA screen shot of output at 5.0 MHz

Material attenuation was applied to the CIVA model and the 5.0-MHz transducer. Three methods
were used to calculate an appropriate level of attenuation:
1. An exponential curve fitted to measured decay from consecutive peaks. The yielded value
was 0.142 dB mm-1.
2. A two-point measurement using an alternative method. The yielded value was
0.082 dB mm-1.
3. The first method was slightly adjusted to subtract the decay measured in a non-attenuated
CIVA simulation—that is, only the geometric beam scattering was taken into account. This
method yielded a much lower value of 0.044 dB mm-1.

Comparison of Results
Simulated and experimental results are compared against one another in this section. For brevity,
only the results for the fourth step are illustrated. Figures 3-7, 3-8, and 3-9 show the comparison
between the CIVA simulated results and the experimental results extracted from UV using the
attenuation values yielded by the three methods above, respectively.

10290640 3-5
Figure 3-7
CIVA and UV data comparison with attenuation value of 0.142 dB mm-1, plotted against time (µs)
with normalized vertical axis (at 5.0 MHz)

Figure 3-8
CIVA and UV data comparison with attenuation value of 0.082 dB mm-1, plotted against time (µs)
with normalized vertical axis (at 5.0 MHz)

10290640 3-6
Figure 3-9
CIVA and UV data comparison with attenuation value of 0.044 dB mm-1, plotted against time (µs)
with normalized vertical axis (at 5.0 MHz)

In the case of the highest level of attenuation, the peak values associated with each simulated
CIVA echo are more heavily attenuated than their UV counterparts, although the phase is
respected, with no more than half a cycle difference over the recorded length of the signals. The
most favorable comparison is for the thickest step, with the discrepancies being due mainly to
electronic noise and a difference in the shape of the envelope. The comparisons are the poorest at
the thinnest step, where significant wave paths other than direct echoes clutter the experimental
data but are not handled by the CIVA simulation.
Where an attenuation value of 0.082 dB mm-1 is used, the peak values of CIVA and UV at each
reflection compare more favorably, although the CIVA values still do attenuate more strongly
than the corresponding experimental values. This is especially noticeable at the steps of higher
thickness. In the case of the lowest level of attenuation, it was generally observed that the UV
signals attenuated less strongly than the simulated CIVA signals.
The frequency spectra (via Fast Fourier Transform [FFT]) of the first echo from the backwall
were also calculated for the UV and the unattenuated CIVA data, as shown in Figure 3-10 for the
fourth step. They are compared against the spectrum of the input signal. It is seen that in spite of
the differences in the shape of the echo in the time domain, the frequency spectrum of the input
is retained well.
CIVA is capable of modeling and displaying the resulting sound field, illustrated in Figure 3-11.
A 2D slice of the image field is illustrated in the top left window and is also drawn in the correct
location within the specimen in the bottom right window. This allows the user to understand to
which part of the specimen most energy arrives and therefore would give greater interpretation as
to the complex returning signal when confronted with experimental data.

10290640 3-7
Figure 3-10
FFT of first echo from backwall of step 4 from 5.0 MHz data

Figure 3-11
CIVA screen shot of sound field visualization

Abaqus Simulation Procedure

For the purposes of comparison against CIVA and against experimental results, simulations
using the finite element (FE) software package Abaqus were also executed in 2D and 3D at 1.5
MHz, and in 2D at 5.0 MHz. In Abaqus, a regular mesh of quadrilateral elements (2D) or
hexahedral elements (3D) in a rectilinear mesh were assembled. To eliminate unwanted extra
waves at the simulated transducer, the input was apodized according to the envelope of a sine
function.
Using the appropriate criterion of elements per wavelength, an elemental dimension of 115 µm
was chosen for 1.5 MHz, where the longitudinal wavelength in steel is approximately 4 mm. If
5.0-MHz simulations were modeled, then the elemental dimension was modified to 35 µm.

10290640 3-8
Comparison between the 2D simulated Abaqus results and the experimental data at 1.5 MHz, and
between 3D simulated Abaqus results and the experimental data, are shown in Figure 3-12. In
both cases, the phasing compares well except for the input cycle (at 0–4 µs), where the phase is
inverted due to the different nature of data collection as the input signal is being formulated.
Abaqus, as part of its simulation, naturally models geometric dispersion in both 2D and 3D,
although CIVA does not; in CIVA geometric dispersion must be artificially introduced.
Abaqus simulations do produce a considerable amount of noise due to wave scattering, which in
some cases does seem to agree well with simulation (see 2D at 11µs) but not at other times (see
3D at 12µs). A fuller study is also available from EPRI, incorporating some results from other
FE software packages such as COMSOL, LS-DYNA, PZFlex, and Wave3000.
CIVA has the advantage of being able to run simulations more quickly and with greater memory
efficiency than Abaqus. EPRI’s current hardware was not able to run Abaqus simulations in 5.0
MHz in 3D.

Figure 3-12
Comparison of Abaqus 2D (top) and 3D (bottom) simulations (blue) with experimental data (Ex) in
red

10290640 3-9
Notch Flat Block
Comparisons are also made between the simulated CIVA data and experimental UV data
collected from a stainless steel flat block with two notches. The block, which has planar
geometry, measures 255.6 mm × 152.4 mm and stands 25.298 mm tall. The material properties
used in this CIVA simulation were 5772.72 m s-1, 3086 m s-1 and 7900 kg m-3 for the wave
velocities and the density.
There are two notches, each 40 mm away from its respective end of the block and parallel to the
end face. Each notch was cut such that its depth was stepped in five steps, thus ten steps in total,
of slowly increasing thickness from 2.53 mm to full thickness, in 2.53-mm increments.

Equipment
• R/D Tech Tomoscan III/PA
• Imasonic 1.5-MHz, 0.5-in. conventional ultrasonic probe
• Imasonic 5.0-MHz, 0.5-in. conventional ultrasonic probe (see Figure 3-13)
• Stainless steel notch flat block (see Figure 3-14)

Figure 3-13
Photograph of 5.0-MHz, 0.5-inch diameter transducer on notch block

10290640 3-10
Figure 3-14
End view of notch block with dimensions

Experimental Procedure
A circular probe with a diameter of 0.5 in. was used to scan the block via a plexiglass wedge so
as to introduce a transverse wave into the specimen at an angle of 45°. The probe assembly was
moved across the top surface of the block in nine parallel rows of 456 linear steps at 0.5-mm
increments (i.e., 457 data points were collected per row). There was 15 mm between each pair of
rows, so chosen since the notch steps were 30 mm wide. Thus a row was positioned such that it
would lie directly atop the transition in notch depth and the two either side of it would lie atop
the middle of the step. The path of the probe upon the top surface of the specimen is illustrated in
Figure 3-15.
Examples of unprocessed UltraVision data are shown in Figure 3-16 for 1.5 MHz and in Figure
3-17 for 5.0 MHz.

10290640 3-11
Figure 3-15
Model of CIVA simulation; end of block and notches shown in red and probe path shown in purple

Figure 3-16
Unprocessed UltraVision data flaw images at 1.5 MHz (transducer orientation: looking toward the
deeper notch)

10290640 3-12
Figure 3-17
Unprocessed UltraVision data flaw images at 5.0 MHz (transducer orientation: looking toward the
deeper notch)

Simulation Procedure
The geometry was constructed and defined within CIVA. As before, the coupling medium was
modeled as water and the bottom medium as air. An exponential delay law was used for both
wave modes in the stainless steel: a wave attenuation of 0.004 db mm-1 at the fourth power of
attenuation.
The computation was modeled in 3D with backwall echoes active for accurate computation. Up
to three reflections were simulated, all in transverse mode, when composing the resulting images,
which are shown in Figure 3-18 (looking toward the shallower notch) and Figure 3-19 (looking
toward the deeper notch) for an inspection frequency of 1.5 MHz. In each figure, the top left
window shows the corrected c-scan, the top right window displays the corrected d-scan, and the
bottom left window shows the corrected b-scan.
In the corrected b-scans of Figures 3-18 and 3-19, it can be seen that the corners where the
notches meet the backwall and the top of the notches reflect strong amounts of energy due to
corner trap mechanisms and edge diffraction, respectively. The full-skip echo can be faintly seen
as a reflected image of the edge diffraction echo.
Figures 3-20 and 3-21 are the corresponding CIVA screen shots at 5.0 MHz.

10290640 3-13
Figure 3-18
CIVA simulation screen shot of the 1.5-MHz results (transducer orientation: looking toward the
shallower notch)

Figure 3-19
CIVA simulation screen shot of the 1.5-MHz results (transducer orientation: looking toward the
deeper notch)

10290640 3-14
Figure 3-20
CIVA simulation screen shot of the 5.0-MHz results (transducer orientation: looking toward the
shallower notch)

Figure 3-21
CIVA simulation screen shot of the 5.0-MHz results (transducer orientation: looking toward the
deeper notch)

10290640 3-15
Comparison of Results
The UltraVision and CIVA results compare well where a corner echo is predicted. The main
differences are:
• That CIVA produces stronger edge diffraction, or that the experimental edge diffraction is
too weak to be seen above the noise level in UltraVision.
• That CIVA produces a weaker echo from the upper corner. This is explained by the fact that
CIVA handles ray reflections from the top surface differently from defects and sidewalls.
These differences are due to the phenomenological nature of CIVA calculations.

706 Series Dissimilar Metal Pipe Weld

The simulated CIVA data and experimental results collected from the Zetec DYNARAY system
are compared in this section. Two short sections of circular pipe are joined in a dissimilar metal
weld (DMW), shown in Figure 3-22; the dimensions of the weld cross-section are shown in
Figure 3-23, and the flaw truth information is shown in Figure 3-24.

Figure 3-22
Illustration of the 706 stainless steel weld

10290640 3-16
Figure 3-23
Cross-sectional dimensions of the EPRI 706 configuration stainless steel weld

Figure 3-24
Flaw information for the EPRI 706 configuration stainless steel weld

10290640 3-17
Equipment
• Zetec DYNARAY phased array UT system (Figure 3-25)
• MultiX LF phased array UT system (Figure 3-26)
• R/D Tech MCDU-02 motion controller (Figure 3-27)
• ATCO LPS-1000 2-axis automated scanner (Figure 3-28)
• Wedge
• 706 series stainless steel DMW sample

Figure 3-25
Zetec Dynaray phased array UT system

Figure 3-26
MultiX LF phased array UT system

10290640 3-18
Figure 3-27
R/D Tech MCDU-02 motion controller

Figure 3-28
ATCO LPS-1000 2-axis automated scanner

Experimental Procedure

To acquire the experimental data, the probe, a dual-element in a matrix phased array pattern of
64 elements in total with flat focusing, was moved axially 115 steps in increments of 1 mm, and
circumferentially 360 steps in increments of 500 µm so as to cover a half revolution of the pipe
on one side of the weld (as illustrated in Figure 3-29). An example of the output of the M2M
system is presented in Figure 3-30, showing an uncorrected c-scan in the top left window, an
uncorrected b-scan in the bottom left window, and a corrected d-scan in the top right window.

10290640 3-19
Figure 3-29
Scanning procedure around the pipe section, showing array probe and wedge

10290640 3-20
 Flaw

Flaw

Figure 3-30
Unprocessed experimental data output, showing reflections from the two flaws

Simulation Procedure
The specimen geometry was imported as a 2D CAD and revolved 180° to make the 3D form.
The geometry was divided into eight isotropic sections, each composed of one of four different
types of steel with slightly varying ultrasonic wave velocities. The materials were:
• Carbon steel: cL=5522.278 m s-1; cT=3188.289 m s-1; ρ=7870 kg m-3
• Alloy 600: cL=6056.422 m s-1; cT=3088.616 m s-1; ρ=8470 kg m-3
• 316 stainless steel: cL=5400.439 m s-1; cT=3104.359 m s-1; ρ=7990 kg m-3
• 309 stainless steel: cL=5779.877 m s-1; cT=3096.618 m s-1; ρ=8030 kg m-3
The wedge is cylindrical concave to fit the outside diameter of the pipe section. The CIVA
results of the scanning of two flaws inserted into the simulated model are shown here. The flaws
are perpendicular to the profile plane and are shown in Figure 3-31 in red. The scanning results
are shown in Figure 3-32 for the first flaw and in Figure 3-33 for the second flaw.

10290640 3-21
Figure 3-31
Location of two flaws perpendicular to the profile plane

Flaw

Figure 3-32
Unprocessed CIVA output for a circumferential scan of the first flaw

10290640 3-22
 Flaw

Figure 3-33
Unprocessed CIVA output for a circumferential scan of the second flaw

Comparison of Results
When performing simulations with the software, great care must be taken to properly input
geometric and material property quantities that accurately represent those of the specimen.
Failure to do so would add more uncertainty as to the comparison between experimental and
simulated results. Exercising due care in this regard also makes quantitative comparison easier
and more useful.
From visual observation of the unprocessed outputs of CIVA, the major differences are:
• The nature of the backwall echo; the real specimen is much more complicated in nature than
the simulated weld, and the backwall echo is much more irregular.
• The flaw echoes, which are sharper in CIVA, possibly as a result of not modeling multiple
ray reflections from internal boundaries; the experimental results show much more noise
from lesser echoes surrounding the main echo.

Flexible Array Simulations on Tapered Aluminum Reference Block

The aluminum tapered reference block is illustrated in Figure 3-34, and shown in CIVA
simulation space in Figure 3-35. A schematic of the block is shown in Figure 3-36.

10290640 3-23
Figure 3-34
Photograph of aluminum tapered reference block

Figure 3-35
CIVA model of aluminum tapered reference block

10290640 3-24
Figure 3-36
Schematic of aluminum tapered reference block

Equipment
• Zetec DYNARAY phased array UT system
• MultiX LF phased array UT system
• R/D Tech MCDU-02 motion controller
• ATCO LPS-1000 2-axis automated scanner
• Imasonic 1.5-MHz 24-element 1D linear flexible phased array probe (Figure 3-37)
• Aluminum tapered reference block
Experimental Procedure
This tapered reference block has two circumferential flaws and three axial flaws. Inspections
were made of the two circumferential flaws (flaws A and D) using the 24-element flexible
phased array but only using 12 active elements in any one scan. They could be the front 12,
middle 12, or back 12 as illustrated in Figure 3-38, Figure 3-39, and Figure 3-40, respectively.

10290640 3-25
Figure 3-37
Imasonic 1.5-MHz 24-element 1D linear flexible phased array probe

Figure 3-38
Zetec phased array calculator for the front 12 elements

10290640 3-26
Figure 3-39
Zetec phased array calculator for the middle 12 elements

10290640 3-27
Figure 3-40
Zetec phased array calculator for the back 12 elements

In Figure 3-41, the experimental positioning of the scanner and probe upon the wider section of
the pipe (at 6 in. diameter) is shown. The unprocessed results of the scan of flaw A are shown in
Figure 3-42, and the results of flaw D are shown in Figure 3-43. In addition, one scan was made
from the tapered part of the specimen, as illustrated in Figure 3-44. The experimental results are
shown in Figure 3-45.

10290640 3-28
Figure 3-41
Tapered aluminum reference block automated scan setup with a probe skew of 270°

Figure 3-42
Ultrasonic response from flaw A using the back 12 elements and a 45° longitudinal wave

10290640 3-29
Figure 3-43
Ultrasonic response from flaw D using the middle 12 elements and a 41° longitudinal wave

Figure 3-44
Tapered aluminum reference block automated scan setup with a probe skew of 90°

10290640 3-30
Figure 3-45
Ultrasonic response from flaw A using the back 12 elements with a 46° longitudinal wave

Simulation Procedure
The simulated specimen was generated from a fully rotated 2D CAD geometry. A flexible probe
with a linear phased array is simulated to sit at the outside diameter of the pipe. The probe
assembly is simulated to perform a full revolution of the pipe in 400 steps. Only two flaws were
present in the CIVA model, having been imported individually as 2D CAD drawings. When
generating the results, only one reflection was taken into account, with no mode conversion.
The unprocessed output from CIVA is shown in Figure 3-46, with Figure 3-47 representing
output for the case of the scan from the taper.

Comparison of Results
These results compare in a similar manner to those from the previous section; the unprocessed
CIVA output reveals indications for both flaws whether the probe is upon the taper or upon the
wider section of the pipe. The UV results show a much higher signal to noise ratio to the extent
that flaw D is difficult to observe when scanned from the taper with a 46° longitudinal wave.
Flaw A, however, is easily observed in all cases at the predicted location in UV results.

10290640 3-31
 Flaw D
Flaw A

Figure 3-46
Unprocessed CIVA output for a scan of flaws A and D with a probe skew of 270°

Flaw D
Flaw A

Figure 3-47
Unprocessed CIVA output for a scan from the taper with a probe skew of 90°

10290640 3-32
4
SUMMARY
The CIVA software performs well from a phenomenological standpoint. Significant defect
responses are predicted and displayed; their accuracies are dependent upon the input material
properties and geometrical parameters provided by the user. If used wisely, CIVA is capable of
bringing understanding to the ultrasonic inspection of even a complex component.
The qualitative comparisons shown in this report show that every flaw has been accounted for in
the simulated environment. The least favorable comparison came from the dissimilar metal weld,
where multiple materials and different material properties were employed, compounding the
possible error if even one of the properties of the materials were misrepresented. CIVA, like all
modeling and simulation tools, can produce inaccurate results if:
• Geometries are misrepresented (internal geometries and boundaries from one material to
another).
• Material properties are erroneous.
• Scanning parameters do not match those of the experiment.
• Flaws are not correctly set up in the simulated model.
It is of course not to be expected that exactly the same result would be derived from experiment
and CIVA simulation, for the following reasons:
• CIVA does not model every possible wave path; desired mode conversions and reflections
are selected by the user.
• Probe and transducer intricacies are ignored by CIVA.
• Systematic and random noise is ignored by CIVA.
• CAD geometry imported to CIVA is approximated—that is, unnoticed imperfections or
surface roughness is ignored.
In spite of the differences, a modeling tool can be powerful when used correctly. For instance,
the simulation can present the modeler with results from hypothetical scenarios, not limited to:
• Different array types for known geometry
• Removal of certain reflected wave types or wave modes (but leaving others present)
• Removal of certain defects or geometrical features (but leaving others present)
• Small change in geometry from the reference or the norm
These can help facilitate understanding of ultrasonic phenomena within the specimen. Another
advantage provided by simulation software in general, and by CIVA, is that every simulation of
a set of given parameters will give the same result, although there are many factors, some
controllable and others uncontrollable, that will affect the outcome of an experiment. If a small

10290640 4-1
change is made to the input parameters of a set of parameters for a simulation, the change in
outcome, no matter how small, must have been as a result of that change in input. This is another
learning aid and may help the user to understand a very complex system of interaction between
ultrasonic rays and the component.

10290640 4-2
10290640
Export Control Restrictions The Electric Power Research Institute Inc.,
Access to and use of EPRI Intellectual Property is granted (EPRI, www.epri.com) conducts research and
with the specific understanding and requirement that development relating to the generation, delivery
responsibility for ensuring full compliance with all applicable and use of electricity for the benefit of the public.
U.S. and foreign export laws and regulations is being An independent, nonprofit organization, EPRI
undertaken by you and your company. This includes an
brings together its scientists and engineers as well
obligation to ensure that any individual receiving access
hereunder who is not a U.S. citizen or permanent U.S. as experts from academia and industry to help
resident is permitted access under applicable U.S. and address challenges in electricity, including
foreign export laws and regulations. In the event you are reliability, efficiency, health, safety and the
uncertain whether you or your company may lawfully obtain environment. EPRI also provides technology, policy
access to this EPRI Intellectual Property, you acknowledge and economic analyses to drive long-range
that it is your obligation to consult with your company’s legal
research and development planning, and supports
counsel to determine whether this access is lawful.
Although EPRI may make available on a case-by-case research in emerging technologies. EPRI’s
basis an informal assessment of the applicable U.S. export members represent more than 90 percent of the
classification for specific EPRI Intellectual Property, you and electricity generated and delivered in the United
your company acknowledge that this assessment is solely States, and international participation extends to 40
for informational purposes and not for reliance purposes. countries. EPRI’s principal offices and laboratories
You and your company acknowledge that it is still the
are located in Palo Alto, Calif.; Charlotte, N.C.;
obligation of you and your company to make your own
assessment of the applicable U.S. export classification and Knoxville, Tenn.; and Lenox, Mass.
ensure compliance accordingly. You and your company
understand and acknowledge your obligations to make a Together…Shaping the Future of Electricity
prompt report to EPRI and the appropriate authorities
regarding any access to or use of EPRI Intellectual Property
hereunder that may be in violation of applicable U.S. or
foreign export laws or regulations.

© 2011 Electric Power Research Institute (EPRI), Inc. All rights reserved.
Electric Power Research Institute, EPRI, and TOGETHER…SHAPING THE
FUTURE OF ELECTRICITY are registered service marks of the Electric
Power Research Institute, Inc.
1022949

Electric Power Research Institute

3420 Hillview Avenue, Palo Alto, California 94304-1338 • PO Box 10412, Palo Alto, California 94303-0813 • USA
10290640800.313.3774 • 650.855.2121 • askepri@epri.com • www.epri.com