DETECTION AND IDENTIFICATION OF CONCRETE CRACKING IN
REINFORCED CONCRETE BY ACOUSTIC EMISSION
MasayasuOhtsu
Graduate School of Science and Technology, Kumamoto University, Kumamoto 860-8555, JAPAN
ABSTRACT. Cracking in concrete due to corrosion of rebars in reinforced concrete is one of critical problems
in concrete structures. To clarify cracking process, acoustic emission (AE) measurement is applied In an
accelerated corrosion test, AE events are detected and monitored continuously. Comparing with permeation
of chloride ions, it is found that onset of corrosion and nucleation of cracking can be qualified from AE activity.
Applying SiGMA procedure, nucleation mechanisms of cracks due to expansion of corrosive product are
identified During extension of the surface crack, tensile cracks are nucleated dominantly For the spalling
crack, both the tensile and the shear cracks are generated, as the former dominates the latter approaching to a
stress-fiee surface. In contrast, it is found that the internal crack is nucleated mainly due to shear-crack motioa
EVTRODUCHON
It is seriously recognized in concrete engineering that concrete structures are no longer
maintenance-free. Consequently, a variety of activities for maintenance and repair of the
concrete structures are reported. According to the Standard Specification for Concrete
Structures on Maintenance by the Japan Society for Civil Engineers (JSCE) [1], six mechanisms
of deteriorations are prescribed. These are the deteriorations due to salt attack, neutralization,
chemical attack, freezing and thawing, alkali-aggregate reaction, and fatigue. In reinforced
concrete structures, alkali surface layer on the reinforcement in concrete is broken due to ingress
of chloride ions. Thus, salt attack leads to corrosion of reinforcement. Because almost of all
concrete structures are reinforced by rebars, the corrosion is the most critical deterioration of the
structures.
In the present paper, monitoring the corrosion by acoustic emission (AE) is investigated.
This is because in situ monitoring techniques by AE are going to be standardized to estimate the
structural integrity of the existing concrete structures [2-3]. AE techniques are extensively
studied in reinforced concrete and are applied to estimation of corroded members [4]. Here, an
applicability of AE technique to reinforced concrete is investigated by two tests. In an
accelerated corrosion test, AE events are detected continuously. Comparing with permeation
of chloride ions, onset of corrosion and nucleation of cracking are estimated from AE
observation. In a crack-expansion test, cracks due to expansion of corrosive product are
simulated by using expansive agent. Then, cracking mechanisms of a surface crack, a spalling
crack, and an internal crack are identified by applying SiGMA (Simplified Green's functions for
Moment tensor Analysis) procedure.
CP657, Review of Quantitative Nondestructive Evaluation Vol. 22, ed. by D. O. Thompson and D. E. Chimenti
© 2003 American Institute of Physics 0-7354-0117-9/03/S20.00
1455
EXPERIMENTS
Mixture proportions of concrete are given in Table 1 . In the accelerated corrosion test, two
types of mixture were employed, as the water-to-cement ratios (W/C) were 45 % and 55%. In
the crack-expansion test, the ratio W/C of concrete was 50%. In all mixes, sand and gravel
were of the same kinds, and the maximum gravel size was 20 mm. Air-entraining admixture
was added to control slump values and air contents of concrete. Casting cylindrical samples of
10 cm radius and 20 cm height, compressive strengths were determined at 28 days after
moisture-cured in the standard room. The averaged values of three samples were 40.9 MPa for
= 45%,34.6MPaforW/C = 55%and39.1MPaforW/C =
TABLE 1. IVfixturepixpoitionsofconcnste.
W/C
(%)
Test
Accelerated
test
Crack-expansion
test
contents
(%)
Slump
values
(cm)
5.2
6.8
5.5
3.5
7.7
4.1
Air
45
55
50
Unit weight per volume (kg/m3)
Water
Cement
Sand
Gravel
389
321
323
175
176
162
AE sensors
/ I \ \
/
core-drilled
rebars
1
On
1
^
\
i .
.............. ........................j...............f
I
i
!
!
i
1
i. 1. _ J__
....
CO
^
-± ,
1
1
50
150
150
50
FIGURE 1. Sketch of reMorced concrete slab for an accelerated test
3<Wi NaCl solution
+
S7
~
1 100mA
ii
I
1
1iT"*
1 *
Tank
+
/
100 mJI
1
/
7 TP
-"
Tank
_
/
Copper
plate
/
*
FIGURE 2. Accelerated corrosion test
1456
686
741
853
1138
1179
1072
Cmm]
FIGURE 4. Set-up for the crack-expansion test
FIGURE. 3 Sample for a crai±-expansion test
In the accelerated corrosion test, a reinforced concrete slab of dimensions 10 cm x 25 cm x 40
cm was cast. A sketch of the slab is shown in Fig. 1. The test was conducted as shown in Fig.
2. In a tank, a copper plate was placed at the bottom, and 100 mA electric current was charged
between rebars and the copper plate. To maintain electrical conductivity, the tank was filled with
3% NaCl solution. AE sensors of 50 kHz resonance (RA5; Physical Acoustics Corp.) were
placed on the top surface as shown in Fig. 1. Amplification was 40 dB gain in total and the
frequency range was set fromlO kHz to 200 kHz. AE hits generated during accelerated corrosion
were detected continuously. After the tests, core samples located in Fig. 1 were taken and
crashed. Then, chloride contents were measured by the potentiometric titration.
In the crack-expansion test, a concrete plate of dimensions 10 cm x 25 cm x 25 cm in Fig. 3
was cast. In order to simulate radial pressure due to corrosive product, expansive agent of
dolomite paste was poured into a hole of 30 mm diameter in the figure. The specimen
corresponds to a cross-section of a reinforced concrete member. 6-channel system in Fig. 4 was
employed to record AE waveforms.
SiGMA PROCEDURE
In the theory of AE, the moment tensor consists of two vectors, which are crack motion
vector b and unit normal vector n to crack surface F, and is defined as,
MM- f F (Wib(y)lknidF = C^m [ J Fb(y)dF] =
(1)
Where Cpqki is the elastic tensor and 1 is the unit direction vector of crack motion. AV represents
the crack volume, integrating the crack-motion displacement b(y) over the crack surface F The
moment tensor is defined by the product of the elastic constants [N/m2] and the crack volume [m3],
leading to the moment of physical unit [Nm]. By the use of the moment tensor, AE wave u(x,t)
is theoretically formulated,
uk(x,t)-Gkp,q(x,y,t)Mpq*S(t).
(2)
Glp,q(x, y, t) means the spatial derivatives of Green's functions. Multi-channel observation of the
first motions at more than sk channels is required to determine the moment tensor components,
and SiGMA (simplified Green's functions for moment tensor analysis) is developed [5].
In SiGMA procedure, two parameters of the arrival time (PI) and the amplitude of the first
motion (P2) in Fig. 5 are read from the waveform on CRT screen. By applying the source
location procedure, crack location y is determined from the arrival time differences. Thus,
distance R and its direction vector r are determined. From the amplitudes of the first motions at
more than sk channels, a simplified version of eq. 2 is solved to determine the moment tensor.
1457
256
Time (usec)
384
512
FIGURES. DetectedAE\^vefoimandwavefoimparametersPlandP2.
Since the SiGMA code requires only relative values of the moment tensor components, the
relative calibration of the sensors is sufficient enough. Then, the classification of a crack is
performed by the eigenvalue analysis of the moment tensor. Setting the ratio of the maximum
shear contribution as X, three eigenvalues for the shear crack become X, 0, -X Likewise, the
ratio of the maximum deviatoric tensile component is set as Y and the isotropic tensile as Z.
Then, the eigenvalues of the moment tensor are normalized and decomposed as,
1.0
= X+ Y +Z,
the intermediate eigenvalue/the maximum eigenvalue = 0 -Y/2+Z,
the minimum eigenvalue/the maximum eigenvalue
= -X -Y/2 + Z,
(3)
where X, Y, and Z denote the shear ratio, the deviatoric tensile ratio, and the isotropic tensile ratio,
respectively. In the present SiGMA code, AE sources of which the shear ratios are less than 40%
are classified into tensile cracks.
The sources of X > 60% are classified as shear cracks. In
between 40% and 60%, cracks are referred to as mixed mode.
In the eigenvalue analysis,
three eigenvectors are also determined, and then the two vectors 1 and n, which are interchangeable,
are recovered.
RESULTS AND DISCUSSION
Accelerated Corrosion Test
Three specimens were prepared for each W/C ratio, and the potentiometric titration tests were
conducted after the accelerated corrosion tests. For W/C = 45%, core samples were taken after 4
days, 10 days, and 12 days elapsed, while those of 4 days, 8 days, and 10 days were prepared for
W/C = 55%. The, distribution of chloride contents in depth were determined. Results are
given in Fig. 6. At the location of concrete cover-thickness, chloride ions per m3 becomes higher
than 1.2 kg after 12 days in the case W/C = 45%. In contrast, it reaches over 1.2 kg after only 8
days for W/C = 55%. This is because the permeability becomes high with increase in W/C ratio.
According to the Standard Specification [1], lower and upper bounds for triggering the
corrosion are prescribed as 0. 3 kg/m3 and 1.2 kg/m3, respectively. Consequently, amount of
chloride contents, C(t), was computed at the cover-thickness based on Pick's law,
(4)
1458
10 days
WC = 55%
cover-thickess
2
3
4
5
2
3
4
5
Depth (cm)
Depth (cm)
FIGURE 6. Distribution of chloride ions in depth.
Ci
°
5
(kg) 4
1 UUUU
:
AE
y»
32
!
:
2
4
6
8
2
> 6000 £
' 6000
4000
^ chloride •
1
°iLi.iJL
10 1Z
.
activity
2000
^..^Urv4 6 8 1
I
Time (days)
Time (days)
b)W/C=55%
a)W/C=45%
FIGURE 7. Con^arison between AE activities and chloride contents.
Deterioration of reinforced concrete
Nucleation of cracking
of corrosion
Incubation
period
[Time]
Development,
period
Accelerated
period
,r
Dur ability of r eiiifor c ed c oner ete
FIGURES. Deterioration process due to salt attack.
Where Co is the surface intensity and erf moms Gauss's error function. Results are compared
with AE activities in Fig. 7. Right after the chloride contents become higher than the lower
1459
bound (0.3 kg), high AE activity is observed in the both cases of W/C = 45% and 55%. Then,
around the stage where the chloride contents surpass the upper bound (1.2 kg), another AE activity
is observed. These results imply that two stages of high AE activities are observed in the
accelerated corrosion test. One is the stage where the chloride content becomes over 0.3 kg/m3,
and the other is the stage where the chloride level surpasses 1.2 kg/m3 in concrete. These are in
remarkable agreement with the deterioration process due to salt attack, which is prescribed in the
Standard Specification as shown in Fig. 8. There exist the first stage for onset of corrosion from
the incubation period to the development, and the second stage for nucleation of cracking from the
development period to the accelerated. This suggests that these two stages are readily identified
by AE monitoring.
Corrosion Cracks by SiGMA
Corrosion cracking was simulated by using expansive agent. A hole of rebar location was made
at 5 cm cover-thickness and with 3 cm diameter. Casting expansive agent of dolomite paste into
the hole, cracks were observed after two days. During extension of these cracks, AE events were
detected by six AE sensors as shown in Fig. 4, and then SiGMA analysis was conducted. Three
crack-patterns observed are shown in Fig. 9. These are rebelled as crack traces (a), (b), and (c).
Crack trace (a) corresponds to a surface crack, which is normally observed as corrosion
cracking. Actually, crack trace (b) is more often observed due to corrosion as a spalling crack.
The internal crack (crack trace (c)) is generally not taken into account, because inspection is
mostly carried out up to the cover.
AE events analyzed by SiGMA procedure were located, and then were classified as three
clusters responsible for crack traces (a), (b) and (c), comparing with locations of crack traces in Fig.
9. All results of SiGMA analysis are plotted in Fig. 10. AE events analyzed are located at their
locations with symbols. The arrow symbol represents a tensile crack, of which opening direction
is identical to that of the arrow. Shear cracks are denoted by the cross symbol, of which two
orientations correspond to the crack motion vector and the crack normal vector, respectively.
Concerning crack trace (a), tensile cracks vertically opening to the surface crack are mostly
observed. In contrast, shear cracks seem not to be explicitly associated with extension of the
surface crack. For crack trace (b), tensile cracks are primarily observed at the locations far from
the reinforcement or close to the stress-free surface, while shear cracks are mostly observed near
the reinforcement. In crack trace (c), tensile and shear cracks are fully mixed up.
In order to elucidate cracking mechanisms on these three crack patterns, activities of AE
events are plotted in Figs. 11 and 12, after classifying into tensile cracks and shear cracks It is
obviously observed that in the beginning crack trace (a) is created as propagation of the tensile
cracks. For both crack traces (b) and (c), tensile cracks are nucleated as well as shear cracks. In
(b)
FIGURE 9. Crack patterns observed after the test
1460
(c
CbJ*
~»®?s&4mtkiJit$»!K&z,.*Hi
FIGURE 10. Results of SiGMA analysis classified in three clusters.
particular, activity of shear cracks is higher than that of tensile cracks in the case of crack trace (c).
Thus, it is found that cracking mechanisms are different, depending on crack patterns. The
surface crack (a) is first observed due to corrosion and is generated due to tensile cracks along the
final crack surface. The spalling crack (b) is nucleated as mixture of tensile and shear cracks.
Approaching to a stress-free surface, tensile cracks dominate shear cracks. The internal crack
(c) is generated diagonally to the surface and consists mostly of shear cracks.
20
10
0)
0
1000
800
1200 1400
Time (min.)
FIGURE 11. AE activities during crack nucleatioa
Tensile
..*"
15
20
15
Time(hours)
FIGURE 12. AEactmtiesclassifiedintoshearandtensilecracks.
1461
20
Time(hours)
25
In the actual case of visual inspection, estimation of the surface crack and the spalling crack is
the main target for inspection. Monitoring the nucleation of tensile cracks by applying SiGMA
procedure, it is possible to estimate and predict extension of these two types of cracks. It is noted
that the internal crack could be generated following the surface crack, and the main mechanism of
the internal crack is of shear motion. This must be taken into account in visual inspection.
Even though the cover concrete and the reinforcement are repaired, the internal crack may still
exist and could result in the loss of durability.
CONCLUSION
Cracking of concrete due to corrosion of rebars in reinforced concrete is one of critical
problems in the concrete structures. Consequently, an application of acoustic emission (AE) is
studied. Results obtained are summarized, as follows:
(1) In the accelerated corrosion tests, AE occurrence is monitored continuously. Comparing
with the permeation of chloride ions, a relationship with chloride concentration and AE
activity is clarified. Right after the chloride contents become higher than 0.3 kg, high AE
activity is observed. Then, around the stage where the chloride contents surpass 1.2 kg,
another AE activity is observed.
(2) These two stages of high AE activities correspond remarkably to the deterioration process due
to salt attack, where two stages for onset of corrosion from the incubation period to the
development, and nucleation of cracking from the development period to the accelerated are
prescribed. This suggests that these two stages can be identified by AE monitoring.
(3) Applying SiGMA procedure, nucleation mechanisms of a surface crack, a spalling crack, and
an internal crack due to expansion of corrosive product are identified. The surface crack is
nucleated dominantly by tensile cracks. For the spalling crack, both the tensile and the
shear cracks are generated, as the former is domination the latter approaching to a stress-free
surface. In contrast, the main mechanism of the internal crack is shear-crack motion.
(4) In the actual case of inspection, estimation of the surface crack and the spalling crack is the
main target. Monitoring the nucleation of tensile cracks by applying SiGMA procedure, it is
possible to estimate and to predict extension of these two crack patterns. It is noted that the
internal crack could be generated following the surface crack, and the main mechanism of the
internal crack is shear-crack motion. Even though the cover concrete and the reinforcement
are repaired, this implies that the internal crack may still exist and could result in the loss of
durability.
REFERENCES
1. Concrete Committee of JSCE, Standard Specification for Concrete Structures on
Maintenance (2001).
2. Yuyama, S, Okamoto, T, Shigeishi, T, Ohtsu, M. and Kishi, T," A Proposed Standard for
Evaluating Structural Integrity of Reinforced Concrete Beams by AE," Acoustic Emission :
Standards and Technology Update, ASTM STP 1353, (1998), p. 25.
3. Recommended Practice for In-Situ Monitoring of Concrete Structures by Acoustic Emission,
NDIS 2421, Japanese Society for Non-Destructive Inspection, (2000).
4. Yoon, D. J., Weiss, W. and Shah, S. P.,"Assessing Damage in Corroded Reinforced Concrete
using Acoustic Emission," J. Engineering Mechanics, ASCE, Vol. 26, No. 3, (2000), p. 189.
5. Ohtsu, M., Okamoto, T. and "Vuyama, S.,"Moment Tensor Analysis of Acoustic Emission
for Cracking Mechanisms in Concrete," ACI Structural Journal, Vol. 95, No. 2, 1998,
87-95.
1462