TELKOM
NIKA
, Vol.11, No
.3, March 2
0
1
3
, pp. 1251 ~ 1257
ISSN: 2302-4
046
1251
Re
cei
v
ed O
c
t
ober 1
2
, 201
2; Revi
se
d Ja
nuar
y 7, 201
3
;
Accepte
d
Ja
nuary 19, 20
1
3
Simulation Research o
n
Ultrasonic Guided Waves
Detection of Metal Rod Buffer System Bonding Quality
Yang Hu*
1
, Wang
Chen
g
2
1
Nation
al Ke
y
Lab
orator
y for Electron
ic Mea
s
ureme
n
t T
e
ch
nol
og
y,
North Univ
ersi
t
y
of Chi
na, T
a
iyu
an 0
3
0
051,
Chin
a;
2
State Ke
y
La
b
o
rator
y
for Expl
osio
n Scienc
e
and T
e
chno
log
y
,
Beiji
ng Institute
of
T
e
chnolo
g
y
, Beijin
g 10
008
1, Chin
a
*Corres
p
o
ndi
n
g
author, e-ma
i
l
:
y
a
ngh
u@n
u
c
.
edu.cn
A
b
st
r
a
ct
The judgm
e
nt
of m
e
t
a
l rod buffe
r
system
bonding qualit
y
depends on the bond
of the
m
e
tal rods
w
i
th the surrou
ndi
ng
me
diu
m
.
System bo
nd
i
ng are
a
w
ill le
ad to differe
nt reflectio
n
a
m
pl
itude. Co
mpar
e
d
the si
mu
latio
n
w
i
th the exp
e
ri
me
nt, it can
b
e
c
onc
lu
ded t
hat differ
ent b
ond
ing
are
a
c
an
me
asur
e b
o
n
d
qua
lity, i.e, the
greater
t
he
b
ond
ing
are
a
is
the
mor
e
exc
e
lle
nt b
ond
in
g
qua
lity w
ill
be
. This conc
lusi
o
n
provides the basis for the
system
ultras
onic testing.
Ke
y
w
ords
: ultr
ason
ic testing;
bon
din
g
qu
ality
;
simul
a
tio
n
Copy
right
©
2013 Un
ive
r
sita
s Ah
mad
Dah
l
an
. All rig
h
t
s r
ese
rved
.
1. Introduc
tion
The judg
men
t
of metal rod buffer sy
stem bondi
ng
quality depe
n
d
s on the b
o
nd of the
metal ro
ds
wi
th the su
rrou
nding m
ediu
m
. In this
pa
per, the
co
ncept of bon
din
g
area
whi
c
h
will
lead to
different refle
c
tion
amplitud
e is empl
oye
d
to
mea
s
ure th
e bon
ding
qu
ality. Resp
on
se
curve
can be
measu
r
ed o
n
site and it
can sho
w
the amplitude of each refle
c
t
ed wave, whi
c
h
indicates the
quantitative possibility to distin
g
u
ish
the metal rod bon
d qu
ality. From the
perspe
c
tive o
f
simulating fi
nite element
nume
r
ic
al, this pap
er p
e
rf
orm
s
thoroug
h analysi
s
an
d
elabo
ration o
n
the time domain re
spo
n
se of t
he metal rod buffer system, and then obtain
s
the
relation
shi
p
betwee
n
the bondin
g
area,
bonding q
u
a
lity of the metal rod and
its surrou
ndi
ng
medium [1-3].
2. Model Establishment
2.1. Model Descriptio
n
Figure 1 is the free meta
l pole model. Figure 2 is a metal rod buffer syste
m
model,
whi
c
h is divid
ed into a stee
l layer, epoxy resin layer a
nd the ceme
n
t
layer (simul
ated ro
ck lay
e
r).
The rod le
ng
th is 1.5m, t
he diam
eter i
s
2
c
m,
the b
ond le
ngth i
s
40cm. Ba
si
c para
m
eter
a
r
e
sho
w
n in Ta
b
l
e 1.
2.2. Model Fla
w
De
scripti
ons
It is uncertain
to identify fla
w
types of the
sce
ne meta
l rod cu
shio
n system, but as to its
basi
c
m
ode, i
t
is po
or
bon
ding of m
e
tal
pole
with th
e su
rroun
din
g
ro
ck o
r
the
falling off of
the
binde
r that ca
use
s
the sy
stem to malfun
ction or fail.
Figure 1. Fre
e
metal rod m
odel
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 2302-4
046
TELKOM
NIKA
Vol. 11, No
. 3, March 20
13 : 1251 – 1
257
1252
Table1. Ba
sic param
eters of metal rod
s
buffer sy
stem
’s nume
r
i
c
al
model
Elastic
modulusE
(
Gpa
)
Poisson
rati
o
μ
Densit
y
ρ
(k
g/m
3
)
Diameter
Ф
(c
m)
Length
L(cm)
Metal rod
206
0.25
7900
2
150
Resin level
14
0.30
2000
4(Out
e
r diamete
r
)
40
La
y
e
rs of
cement
40 0.29
2520
4(Out
e
r
diamete
r
)
40
In this paper,
the flaw degree ca
n be obtai
ned by calcul
ating the area of bonding of
metal rod wit
h
its surro
u
n
g
ing ro
ck, the metal
rod bu
ffer system of
different bon
ding quality can
be simulate
d
by establish
i
ng four kind
s of fi
nite element model
s with differe
nt bonding a
r
ea
buffer syste
m
[4- 7]. Bonding quality is categ
o
ri
zed i
n
to: excellent
, good, fair, poor, rep
r
e
s
e
n
te
d
respe
c
tively by BG, MG, CC
and EE. And metal
ro
d model
s a
r
e
also
created
, represented
by
FG. As sho
w
n in Figure 3
,
after removing the ceme
nt outer layer, four finite element molde
d
grap
hs of the
resin level a
nd the metal rod with different bondin
g
quality are left
behin
d
, name
l
y
BG model (to
simulate the
best bondi
n
g
quality,
100% of the area is co
mplet
e
ly bonded
), MG
model (to sim
u
late the better bondi
ng quality, 70%
of the area is completely bo
nded
), CC model
(to sim
u
late t
he go
od bo
n
d
ing q
uality, 40% of
the a
r
ea i
s
compl
e
tely bond
ed
), EE model
(to
simulate po
o
r
bondin
g
qu
ality, 20% of the area is completely bond
ed), an
d FG model
(to
simulate the f
r
ee metal rod
,
20% of the
area i
s
co
mpl
e
tely bonde
d).
3. Simulation Anal
y
s
is
Figure 4 i
s
the sy
stem a
c
celeration ti
me
-d
omain resp
on
se curve.
Excited
guide
d
wave go
es
along the m
e
tal rod an
d
spre
ad
s out
ward at the central h
e
m
i
sph
e
ri
cal ra
dial
dire
ction. It is
quite comple
x to
sprea
d
at the
top. When the wave goes alon
g the rod to a
certain
distan
ce
s
(a
ccordi
ng to th
e Saint-Ve
na
nt prin
cipl
e, S > 1
~
2D,
D i
s
the m
e
tal ro
d diam
eter), the
Figure 2. Metal rod buffe
r system model
1 metal rod 2
resi
n level 3 layers of
cem
ent
2
3
1
a
BG Model
MG Mode
l
EE Mode
l
FG Model
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Sim
u
lation Rese
arch on
Ultraso
n
ic G
u
id
ed Wa
ve
s De
tection of Met
a
l Rod … (Y
a
ng Hu
)
1253
wavefro
n
t then is approxim
ate to plane dissemin
at
ion
.
Therefore, whe
n
gathe
ri
ng sign
als fro
m
the rod’
s ce
ntral point, the front sig
nal sh
ould be
zero.
Figure 4. Acceleratio
n
time
-dom
ain re
sp
onse gra
p
h
Table 2. Data
point a in Figure 4
Data (point a
)
BG Model
MG Model
CC Model
EE Model
Ti
me(
s)
438 439
438
439
Amplitude (ms
-2
)
5.26941
5.24098
5.22515
4.94928
In Figure 4, point a is the reflection p
o
int
on the interface of the solid sid
e
.
We can
cal
c
ulate the
spe
ed of wav
e
transmissio
n in the rod a
s
follows:
1
32
.
5034
4
37
.
4
2
1
.
1
ms
e
c
a
Point b is the free metal rod at the botto
m
of the refle
c
ted wave, b
=
62
3s, the speed of
wave tran
smi
ssi
on in the rod is
1
43
.
5136
4
23
.
6
2
6
.
1
ms
e
c
b
And, accordi
ng to the
material p
r
o
perti
e
s
, the cal
c
ulate
d
theoreti
c
al ro
d
speed
is
1
0
46
.
5106
7900
11
06
.
2
ms
e
E
c
. Relative error is
%
46
.
1
46
.
5106
32
.
5034
46
.
5106
,
%
58
.
0
46
.
5106
43
.
5136
46
.
5106
, which a
r
e
very small. In order t
o
observe the wave
transmissio
n in different m
odel
s, we enl
arge the
refle
c
tion point a
r
ea of the third solid si
de o
n
the interfa
c
e,
as
sho
w
n in
Figure 5. It can see
that
with mod
e
l b
ondin
g
qualit
y decrea
s
ing,
the
waveform is in turn clo
s
e
to the free rod wavefo
rm
and its pha
se ba
ck offset. This obvious
cha
nge of varying with different bon
dng
quality re
veal
s the relia
bility of finite ele
m
ent model.
a
b
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 2302-4
046
TELKOM
NIKA
Vol. 11, No
. 3, March 20
13 : 1251 – 1
257
1254
Figure 5. Each model refle
c
tion re
gion e
n
larg
eme
n
t of the interface
on the third solid sid
e
3.1. Relation
ship Be
t
w
e
e
n Bonding
Area and
Amp
litude Atten
u
a
tion
Model bon
di
ng length is
set to 40 CM (FG Mode
l Glue size i
s
0). The different
bondi
ng area
leads to different am
plitude attenuat
io
n. There mu
st exist a relationship betwe
en
the bond
ed a
r
ea a
nd am
pl
itude attenua
tion. We coll
ect the curve
data that at equidi
stant th
e
cente
r
point o
f
the interface
on the solid
side
a
nd bott
o
m cente
r
poi
nts of each m
odel, put the
m
together, and
then we get
a resp
on
se chart of
corre
s
pondi
ng time domain, as
shown in Figure 6
and 7.
Figure 6. Time-do
m
ain ma
p of the equid
i
stant ce
nter
point on the
model fixed e
nd interfa
c
e
In Figure 6 waveform o
s
cil
l
ation amplitu
de is
relativel
y
large, while
in Figure 7,
the
waveform oscillation am
plitude is sig
n
ificantly sma
ll,
whi
c
h refle
c
ts the attenuation of the ene
rgy
to rea
c
h th
e
bottom of the p
r
o
c
ess.
a, b ar
e the
first wave spots of
each
model.
We
can
cal
c
ulate the
attenuation of
bondin
g
se
g
m
ent of
each
model, as
sh
own in Fig
u
re
8 and Table
3.
a
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Sim
u
lation Rese
arch on
Ultraso
n
ic G
u
id
ed Wa
ve
s De
tection of Met
a
l Rod … (Y
a
ng Hu
)
1255
Figure 7. Time-do
m
ain
cha
r
t of the model bottom cen
t
er point
Table 3. Each model bo
nd
ing se
gment
amplitude att
enuatio
n
From th
e fitting curve i
n
Figure 8, it c
an see that t
he bo
ndin
g
segment
ampl
itude
attenuation i
s
nonline
a
r, simila
r to th
e att
enuated
form of
the exponential
function. When
bondi
ng are
a
is between
0% -40%, amplitude atte
nuation chan
ges si
gnifica
ntly, but when th
e
bondi
ng are
a
betwe
en is 4
0
% -100%, its cha
nge is in
signifi
cant, which explai
ns
why the BG and
MG model cu
rve in Figure 4 are clo
s
e to good distin
guishing; wh
erea
s CC, RE, and FG model
of the three
curves
so
signi
ficantly differ. It c
an be con
c
lud
ed that th
e accele
ratio
n
domai
n cu
rve
in Figure 4 can ba
sically separate the
s
e two se
ts of
bondi
ng qu
ali
t
y of the bonding a
r
ea 1
0
0%
-
40% and 40
% -0%, therefore, verify
ing that it is feasible to tell the
pro
s
and
con
s
of the bondi
ng
quality by bo
nding
are
a
a
nd furthe
rmo
r
e ca
n refle
c
t
the quality st
atus of the m
e
tal rod
bon
d
i
ng
on site.
Figure 8. Amplitude attenu
ation
and the
bondi
ng area
relation g
r
a
p
h
Data
BG Model
MG Model
CC Model
EE Model
FG Mo
del
Amplitude
attenuation (
%
)
99.2%
98.97%
98.37%
93.28%
25.17%
Bonding area
(%
)
100%
70%
40%
20%
0
b
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 2302-4
046
TELKOM
NIKA
Vol. 11, No
. 3, March 20
13 : 1251 – 1
257
1256
4. Comparis
on of Simulation and Te
s
t
Da
ta
4.1. Test Des
c
ription
The fre
e
met
a
l rod
expe
ri
ment u
s
e
s
th
e field u
s
ed
buffer rod at
lengh
L=1.5
m
and
diamete
r
φ
=2
cm. Wh
en in
experim
ent, it is at the free
state. Metal rod b
u
ffer sy
stem test ta
kes
th
e
r
e
pr
es
enta
t
ive
me
d
i
um’s
s
t
ab
le
ro
ck
la
ye
r
as the simul
a
tion obj
ect. T
he ro
ck ma
ss is
desi
gne
d as
0.6×0.5
×
0.5
m
3
con
c
rete
cub
e
, with bo
nding le
ngh
4
0
cm, metal
rod leng
h L
=
1
.
5m,
and the intermediate de
ck use
s
epoxy resin.
4.2. Results Analy
s
is
As sho
w
n in Figure 9, the
sup
p
o
s
ed me
tal rod bottom end reflect
i
on time is T
0
; the
metal rod len
g
th is L0, the bondin
g
length is L;
the fixed end interface refle
c
tion is T. If
the
averag
e velo
city of guid
e
d
wave
in m
e
tal ro
d is v0, then
0
0
0
2
v
L
T
,
0
0
2
v
L
L
T
.
Thro
ugh the experim
ent
1
0
5161
ms
v
.
Simulation an
d experiment
al data obtained are sho
w
n
in Table
4. It can
se
e that
the nume
r
i
c
al
simulati
o
n
value
s
of the
critical
refle
c
tion poi
nt and
the
experim
ental
value are b
a
sically in agree
ment.
Figure 9. Metal rod sy
stem
fluctuation di
agra
m
Table 4. Time
value of the solid
side inte
rface
refle
c
tio
n
point and th
e bottom refle
c
tion
point
5. Conclusio
n
Comp
arative analysis of the simulatio
n
re
sult
s an
d experiment
al data com
e
s to th
e
f
o
llowin
g
con
c
lu
sion
s:
1)
The size of the bondin
g
area ca
n be a
true re
fle
c
tion
of
the size of the bond quality on site.
The differe
nce of the ene
rgy amplitud
e attenuat
ion
resulting fro
m
the differe
nce of th
e
bondi
ng quali
t
y provides a
sci
entific cali
bration m
e
tho
d
for simul
a
tion study.
2)
Finite elem
en
t simulation
a
nd theo
retical
re
sults
ca
n
clea
rly identif
y the interfa
c
e refle
c
tion
on the fixed end point but fail to identify the bottom reflecti
on point. Simulation and
experim
ental
result
s coi
n
cide
well. Th
e subtle
diff
eren
ce i
s
that the experi
m
ental cu
rve
Time value of nu
merical simulatio
n
T(
μ
s)
Time value of experimental test T
(
μ
s)
622
621
623
620
466
468
L
L
0
Botton end
Fixed end
T
t
T
0
Ce
me
nt
lay
e
r
Me
ta
l rod
Resin lay
e
r
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Sim
u
lation Rese
arch on
Ultraso
n
ic G
u
id
ed Wa
ve
s De
tection of Met
a
l Rod … (Y
a
ng Hu
)
1257
appe
ars to jitter, whi
c
h is caused
by the interferen
ce
of the fiel
d te
st instru
ment.
Hence, the
simulation results are reliable .
3)
Energy atten
uation is the reason why the bo
ttom reflection wave ca
n not be clearly identified,
but it can cl
early identify the reflected wave
on the interface of the solid
side. We ca
n
determi
ne the
bondin
g
qual
ity of
the system thro
u
gh the amplitud
e of the reflecte
d wave.
4)
It can be ideal to simulate by using the fini
te element software. This will pro
v
ide a stron
g
tec
hnic
a
l
s
u
pport for more
in-depth s
t
udy.
Referen
ces
[1]
MD Beard, MJS Lo
w
e
. No
n-destructiv
e
testi
ng of rock
bolts usin
g gui
ded u
l
traso
n
ic
w
a
ves
.
Internatio
na
l Journ
a
l of Rock
Mecha
n
ics, Mi
nin
g
Scienc
es
. 200
3; 40: 527-
536.
[2
]
Be
a
r
d
MD
, Low
e MJS, C
a
w
l
ey
P. U
l
t
ra
so
nic g
u
i
d
ed
w
a
ve
s for the inspe
c
tion of gro
u
te
d tend
ons a
n
d
bolts.
J Mater Civil En
g
. 200
2
,
in press.
[3]
BN Pavlak
ovic
, MJS Lo
w
e
,
P Ca
w
l
e
y
. H
i
g
h
-frequ
enc
y l
o
w
-
l
o
ss ultr
ason
ic modes i
n
im
bed
de
d bars.
Journ
a
l of App
l
ied Mec
han
ics
. 2001; 6
8
: P67
-
65.
[4]
Paul
us Insap Santosa.C
o
st and Ben
e
fit of In
formation Search usi
n
g
T
w
o Differen
t
Strategies.
T
E
LKOMNIKA Indon
esi
an Jou
r
nal of Electric
al Eng
i
ne
eri
n
g
.
2010; 8(3):
196
-1
99
.
[5]
Roy
BVB Sim
o
rangkir, Achmad Munir.
N
u
merica
l Desi
gn of Ultra-W
ideb
an
d Printe
d Antenn
a for
Surface Pen
e
tr
ating R
a
d
a
r Ap
plicati
on.
TEL
K
OMNIKA
. 2011
; 9(2): 341-35
0
.
[6]
W
u
Bin, Sun
Yaxin, He
Cu
nfu, W
ang
Xi
u
y
a
n
ect.
Appli
c
ation
of hig
h
freque
nc
y ultra
s
onic g
u
i
d
e
d
w
a
ves to
insp
e
c
tion of ful
l
-le
n
g
th-bo
ndi
ng
bo
lt.
Chin
ese Jo
u
r
nal
of Rock M
e
cha
n
ics a
nd
Engi
neer
in
g
.
200
6; 25(6): 1
240-
1 24
5.
[7]
F
Q
Guo and
C
S
Z
hang.
Revi
e
w
of res
earc
h
on n
on-
destru
c
tive testing
of
bon
di
ng q
u
a
lit
y
of grout
ed
rock bolt (in C
h
ines
e).
Journa
l
of T
a
iyuan u
n
i
v
ersity of T
e
chnol
ogy
. 20
05; 36: 11-1
4
(Sup
.).
[8]
Chu
ng-Hs
in L
i
u, Po-Chi
ng T
eng. T
he Stud
y for
Hom
e
Su
rveill
anc
e F
u
zzy C
ontrol S
y
st
em of Attacks
Anal
ys
is.
JCIT
: Journa
l of Con
v
erge
nce Infor
m
ati
on T
e
ch
no
logy
. 20
11; 6(1
1
): 30- 40.
[9]
J Alamelu Ma
nga
i and S Sa
meen F
a
thim
a
Appavu al
ias
Balam
u
rug
an. F
eature Sel
e
ction for Larg
e
Scale Data U
s
ing Cl
ass Associati
on Rul
e
Minin
g
.
JCIT
: Journal of Converg
enc
e
Informatio
n
T
e
chno
logy
. 2
011; 6(1
1
): 371
- 377.
[10]
W
r
itten b
y
JL
Rose. T
r
anslat
ed b
y
He
Cu
n-
fu
, W
u
Bin, W
ang
Xiu-Y
an,
Check
ed
b
y
Y
ang Gu
i-ton
g
.
Ultrasonic Wav
e
s in Solid Media
. Bei
jin
g: Scienc
e Press. 2004; 3: 14-
16.
[11]
JL Ross.
Soli
d
s
Ultrason
i
c
. Scienc
e Press, Beiji
ng C
h
in
a. 200
4: 14-1
7
.
Evaluation Warning : The document was created with Spire.PDF for Python.