A STUDY ON CHARACTERISTICS OF ECHOES FROM PLANAR DEFECTS BY THE

ULTRASONIC TESTING OF ANGLE BEAM TECHNIQUE

Hiromi Shirahata
Department of Civil Engineering
Musashi Institute of Technology
1-28-1 Tamazutsumi Setagayaku
Tokyo 158 Japan

Chitoshi Miki
Department of Civil Engineering
Tokyo Institute of Technology
2-12-1 Ookayama Meguroku
Tokyo 152 Japan

INTRODUCTION

The application of field -welded butt joints of flanges and web plates for steel
girder bridges has been on the increase in Japan. Figure 1 is an illustration of such a
field-welded joint. The thickness ofthe flange plate may at times be in excess of 75 mm.
Radiographic testing has been used to inspect the quality of welded joints in the shop;
however, taking into account material having a thickness that is beyond the capacity of
the radiographic testing and the risk of using X-rays in the field, ultrasonic testing is
applied to field-welded joints[1,2].

Though ultrasonic testing by the angle beam method is utilized in field-welded

joints, there is a problem in the application of ultrasonic testing. That problem is the
difficulty of detecting hot cracks, lack of fusion and other planar defects by the single
probe method. The reason for the difficulty is the directivity of the echoes from planar
defects. As shown in figure 2, when ultrasonic wave is transmitted toward a planar
defect, the incident wave is reflected specularly and the reflected wave will not
necessarily be received by the probe. Only tip diffracted waves can be received by the
probe because oftheir lack of directivity[3].

Rl!View ofProgress in Quantitative Nondestructive Evaluation, Vol. 17

Edited by D.O. Thompson and D.E. Chimenti" Plenum Press, New York, 1998 1869
 If a planar defect is located near the back surface, there will be a path of corner
reflection. Incident waves are reflected specularly at the defect and reflected specularly
again at the bottom. These specularly-reflected waves may at times be received by the
probe. Specularly-reflected echoes such as these may hide tip echoes and, although it is
also possible to detect defects by receiving specularly-reflected echoes, this complicates
the reconstruction of an image of the defects. It is important to determine whether a
received echo is a tip echo or a specularly-reflected echo.

weld
Figure 1. Field welded joint of a plate girder bridge.

probe

Figure 2. Vector diagram of numerical simulation by finite element analysis.

The objective of this study was to investigate characteristics of ultrasonic

waves by the angle beam method. We focused in particular in this study on planar
defects located near the back surface. The approach that we adopted in this study is
numerical simulation of finite element analysis and experimentation. We first
investigated surface-breaking notches, simplifying the model of planar defects near the
back surface. Next, we investigated internal slits.

1870
 D!
flaw de:cctor

i
specimen
\~
el probe

1\ 1\

Figure 3. Experimental etup .

(a) (b)

!I"
~ 11-------'Ii""-------'~
!i
if
!!

22mn! ~
, - - - - - I
A
0 § B
D
8 .5mm
weld

~
A B 22mm Smm

icnglh
n
I.Omm 3.8cD
B.Smm

w,dlh 0.25... O.25om (reinforcement was ground)

Figure 4. Specimens (a) surface breaking notch model and (b) internal slit model.

EXPERIMENTAL SETUP AND SPECIMENS

All of the experiments in this study were conducted using the immersion
method. Figure 3 shows the experimental setup. The probe was moved by means of a
computer-controlled manipulator, having an accuracy of 0.5 mm in the x and y
directions and 0.01 around the axis ofrotation, as indicated in figure 3.
0

'l\vo types of specimens were used in this study, as shown in figure 4.0ne
contains a number of surface breaking notches, which were made by wire cut. The
notches have a length of 1.0 mm and 3.8 mm. The other specimen contains an internal
slit and was prepared by butt welding with a spacer inserted to make the slit. The width
and height ofthe slit is 0.25 mm and 5 mm, respectively. After welding, the weld
reinforcement was ground.

1871
 ECHO HEIGHT AND LENGTH OF SURFACE BREAKING NOTCHES

As the first step, we examined the relationship between the length of the
surface-breaking notches and echo height. The numerical simulation model is shown in
figure 5-(a). The center of incident wave was aimed to the corner of the notch and the
angle of incidence was 45° and 70° . Echo height is influenced by both the width of the
beam width and the length of the notch. The length of the notch is normalized by the
beam width obtained by the 6 dB dropping method.

(a)
center of
incident wave

(b) (c)
angle 01 irciderce 45 • angle 01 incIdence 70 ·

• numerical simuiahon
o elCperiment
·20 ~L..L.L.Wu.u.,..-'---'-.L.L.I...I..L.u...",..--J
10' 10
length of nOlchibeam widlh
length 01 nolcMleam WIdth

Figure 5. Relationship between echo height and length of surface notch (a) numerical
simulation model (b) in the case that angle of incidence is 45 degrees (c) 70 degrees

Figure 5-(b) indicates a plot of corner echo height and notch length when the
angle of incident is 45° .The corner echo height increases as the length of the notch
increases and echo height is proportional to the square root of the notch length.
Experimental data when the length of the notch is 1.0 mm and 3.8 mm conforms well
with numerical simulation as indicated in the diagram. The Ll2 level is -18dB. If -18dB
is selected as the detection level, it is possible to detect a notch length of 3 % of the
incident beam width. Figure 5-(c) shows a plot of 70° incident wave. It is possible to
detect a notch length of 5 % of incident beam width in this case. In order to detect small
defects, incident beam width should be narrower.

1872
AMPLITUDE PROFILE SCAN OF SURFACE BREAKING NOTCHES

We conducted amplitude-profile scans experimentally as indicated in figure 6.

Probes used in the experiment were 2 MHz, 5 MHz and 5 MHz focussed type. We first
determined the location of the maximum echo height. We scanned the notch by moving
the probe around that position in increments of 1 mm. Figure 6-(a) shows a plot when
the angle of incidence is 45' and the height ofthe notch is 3.8 mm. Probe position 0
represents the position of the maximum echo height and the center of the incident wave
is aimed at the corner. The highest echo was obtained when 5 MHz focussed probe was
used, which produced a sharp profile. On the other hand, a sharp profile was not
produced when the 2 MHz probe was used due to weak directivity.

(a) (b)

y QD ~t-F-~-
'-- ~~_,~__~_n ~v~_____
•.• __

b"IhlLImn,
" .. ~ l ;;'>. ___ ._.wtdtbO..26mru
.'" I.tdlhO.2f '--.. 1

angle 04 Incidence 45-

leng1h 01 notch
o 3.8mm jI4.
•
"06 -10 o

.
0" 0
60 o
o 0
. -., 0
•" 0 6 6
o

. -5. -20 angle 01 incidence

o o.
o
• 00 70- 00
06
1! lenglh 01 notch 38mm
.2
&l -30
° 2M Hz
-30 I ol5MHZ
• SMHz tocus!
-20 -10 0 10 20 -40 -30 -20 -10 0 10 20
posItion 01 probe (mm) posilion 01 probe (mm)

Figure 6. Amplitude scan plot (a) angle of incidence 45 degrees and (b) 70 degrees

Figure 6-(b) is also an amplitude scan plot. The angle of incidence is 70 ' and
the height of notch is 3.8 mm. There is not much change in echo height compared to an
incident wave of 45° .Since a 70' incident wave is closer to normal incidence than 45' ,
directivity is not strong. The appication of a 70' incident angle is preferable for the
sake of detectability, if there is a problem of distance between probe index and front end
of the probe housing. On the other hand, it is preferable to apply a 45' incident angle in
order to detect the corner of notches.

1873
 (a) (b)

IlIiiihl
+deplh 0(0,111

(c) (d)

angle of mc.dence .\S

lip ccho Ic\el

................. .. .............................

depth of htlb.:am ,,,dth deplh of .1It/beam \\ .dlh

Figure 7. Echo height and depth of an internal slit: (a) model (b) specularly reflected echo
in case of angle of incidence is (c) 45 degrees and (d) 70 degrees

SPECULARLY REFLECTED ECHO HEIGHT AND INTERNAL SLIT DEPTH

It is easier to detect flaw echoes in case of surface-breaking notches because

the maximum echo height is obtained when the center ofthe incident wave is aimed at
the corner. Take, for example, a case in which a planar defect exists near the back
surface. There will be a path of corner reflection echo in this case, as indicated in figure
7-(b). The height ofthe corner echo will become lower in proportion to the distance ofthe
location of the defect from the back surface. That makes it more difficult to evaluate
whether or not flaw echoes are specularly reflected echoes.

It is important to understand relationship between specularly reflected echo

height and tip echo height and we conducted numerical simulations in order to examine
this relationship using the model in figure 7-(a). The depth of the slit was defined as
shown in figure 7-(a). The center ofthe incident wave was aimed at the position on the
bottom surface beneath the slit. Echo height is influenced by both the depth of the slit
and the beam width. The depth of the slit is normalized by the beam width obtained by
using the 6 dB dropping method.

Figure 7-(c) shows the relationship in the case of 45° incident wave. The
specularly-reflected echo is higher than the tip echo when the depth ofthe slit is less
than 50 % ofthe beam width. Figure 7-(d) shows the relationship in the case of 70°
incident wave . Owing to large incident angle, specularly-reflected echo height does not
become as low as it does the case of 45° . Corner effect is much stronger.

1874
 S~ ~~ ____ ~¥~______

V·
' 70· -·····0 lil

(a) (b) . IC\-~-r---.---~--r---,

.-..... .•
- IQ---'---r-~-"-~---'
frequency 5MHz
,;..
<i5 -2e focu<Sed probe
of IncIdence 4S
•
.......... -.
~ nn~le

• •
frequ.:nq 5M III •
focussed pr be
. ngle of Incidence 70
·5d-_ _.l.-~_'---~--J'-.J

1'O'"lon of probe po ilIon of probe

(c) (d)

frequency 2 1Hz
,ngle of IncIdence 4~
••
.-~ ..
t }(.
....
.2 frequency 2 lH l
~ -40 angle of IncIdence 70

· SCl---'--~--'-----'--~-'
-5r1-~--'-~--'-~--'---'

po","on of probe poSlllon of probe

Figure 8. Amplitude scan plot of an internal slit

POSITION AMPLITUDE PLOT OF AN INTERNAL SLIT

Figures 8·(a)·(d) are amplitude scan plots of an internal slit obtained throught
experimentation. We first used the 5MHz focussed probe with an angle of incidence of
45° and 70° . Considering the focal length, beam width is 2 mm around the slit. The
beam width was sufficiently small compared to the depth of the slit and the ratio ofthe
depth of the slit to beam width was 2.5. Judging from figure 8·(b) and (c), we would
expect specularly reflected echo in this case to be lower than tip echo. As indicated in
figure 8·(a), echo height was extremely low in the case of 45° and a number of peaks in
echo height is evident, the distance between two of which is 6mm. Considering the
geometrical configuration ofthe specimen, these echoes may be tip echoes. As shown in
figure 8·(b), the echoes are higher when the angle of incidence is 70° than when it is
45° .Two echo peaks can be observed in the diagram. Taking into account the
geometrical configuration, these may also be tip echoes. It is easier to identify tip echoes
when the angle of incidence is 70° .

1875
 The 2 MHz probe was used in figure 8-(c) and (d). The width of the beam is
about 14 mm and the ratio of the depth ofthe slit to the beam width is 0.6. The beam
width is not small compared to the depth of slit. Specularly reflected echo height is the
same as that of tip echoes. As shown in figure 8-(c), many echo peaks can be observed;
however, it is very difficult to identify each of those echoes. Figure 8-(d) shows a case in
which the angle of incidence is 70° . Specularly reflected echoes are much higher than
tip echoes. Judging from figure 7-(b), it is expected that the maximum echo height will
be about -15 dB. At -15dB, a relatively higher echo level is obtained in the wide
scanning range. The numerical simulation conforms with experimental data.

CONCLUSIONS

We conducted a fundamental examination by the angle beam method to detect

planar defects, especially those located near the back surface. Corner echo height is
proportional to the square root of the height of surface breaking notches. Narrowing the
beam width of incident waves makes it possible to detect small surface breaking
notches.

It is better to apply 45° incident waves to identify the location of notches. On

the other hand, flaw echo can be obtained in a broad range by 70° incident waves,
although it is difficult to determine the location of notches by 70° incident waves.

Corner effect becomes weaker as the depth of the slit increases. Tip echo
becomes higher when the angle of incidence is 45° and the depth of the slit is more than
50% of the beam width, while corner effect is much stronger when the angle of incidence
is 70° .

We were able to obtain tip echoes of internal slits by using a 5MHz focused
probe. Larger flaw echoes can be obtained by means of a 70° incident wave of 2MHz
and that was the most beneficial in the detection of flaws.

REFERENCES

l.Chitoshi Miki: Applications ofNDI in Fracture Control of Honshu-Shikoku Bridge

Project, Review of Progress in QNDE, Vo1.l3, 1994 pp.2071-2082
2.Chitoshi Miki, Takao Narumiya, Touichi Morishita and Masahiko Kato: Application of
Automatic Ultrasonic Testing Systems to Welding Bridge Members, Proc. of JSSC
VoU No.4 December 1994
3.K. Harumi, Y. Ogura and M. Uchida: Japanese Tip Echo Handbook, Japanese Society
for Non-Destructive Inspection

1876