sensors Article

Monitoring Concrete Deterioration Due to Reinforcement Corrosion by Integrating Acoustic Emission and FBG Strain Measurements Weijie Li 1 , Changhang Xu 2 , Siu Chun Michael Ho 1 , Bo Wang 3, * and Gangbing Song 1 1 2 3

*

Department of Mechanical Engineering, University of Houston, Houston, TX 77204-4006, USA; [email protected] (W.L.); [email protected] (S.C.M.H.); [email protected] (G.S.) College of Mechanical and Electronic Engineering, China University of Petroleum, Qingdao 266580, China; [email protected] Key Laboratory of Transportation Tunnel Engineering, Ministry of Education, Southwest Jiaotong University, Chengdu 610031, China Correspondence: [email protected]; Tel.: +86-28-8760-1585

Academic Editor: Vittorio M. N. Passaro Received: 29 January 2017; Accepted: 17 March 2017; Published: 22 March 2017

Abstract: Corrosion of concrete reinforcement members has been recognized as a predominant structural deterioration mechanism for steel reinforced concrete structures. Many corrosion detection techniques have been developed for reinforced concrete structures, but a dependable one is more than desired. Acoustic emission technique and fiber optic sensing have emerged as new tools in the field of structural health monitoring. In this paper, we present the results of an experimental investigation on corrosion monitoring of a steel reinforced mortar block through combined acoustic emission and fiber Bragg grating strain measurement. Constant current was applied to the mortar block in order to induce accelerated corrosion. The monitoring process has two aspects: corrosion initiation and crack propagation. Propagation of cracks can be captured through corresponding acoustic emission whereas the mortar expansion due to the generation of corrosion products will be monitored by fiber Bragg grating strain sensors. The results demonstrate that the acoustic emission sources comes from three different types, namely, evolution of hydrogen bubbles, generation of corrosion products and crack propagation. Their corresponding properties are also discussed. The results also show a good correlation between acoustic emission activity and expansive strain measured on the specimen surface. Keywords: reinforcement corrosion; reinforced concrete; acoustic emission; fiber Bragg grating

1. Introduction Reinforced concrete (RC) structures constitute the majority of civil infrastructures of the world. Corrosion of concrete reinforcement members has been recognized as one of the predominant deterioration mechanisms for steel reinforced concrete structures. The two major factors responsible for corrosion of steel reinforcement are carbonation of concrete and chloride penetration. Of the two, chloride induced corrosion is a far more serious problem. Under chloride attack, the passivation of steel within the concrete is broken down and indicates the initiation of reinforcement corrosion. The corrosion products are usually two to six times more voluminous than the original steel consumed [1], resulting in the generation of expansive pressure exerted by corrosion products to the surrounding concrete. Damage due to reinforcement corrosion occurs in the form of cover cracking, reduction of the steel cross-sectional area, and deterioration of a concrete–steel interface bond [2]. Failure of the structures leads to heavy economic loss, environmental pollution and increased risks

Sensors 2017, 17, 657; doi:10.3390/s17030657

www.mdpi.com/journal/sensors

Sensors 2017, 17, 657

2 of 12

of injury. It is therefore of paramount importance to monitor and characterize steel corrosion in RC structures and assess their performance and safety. Structural health monitoring for concrete structures has attracted a lot of attention [3–5]. A wide variety of techniques have been reported in the literature that may be suitably employed for monitoring and locating corrosion nondestructively. Some commonly used reinforcement corrosion monitoring techniques are electrochemical measurements [6–8], ultrasonic testing methods [9,10], radiography [11] and infrared thermography [12,13]. Methods based on electrochemical measurements are the most popular ones, which directly take advantage of the electrochemical nature of steel corrosion reaction. They can only provide information about the possibility of a corrosion event, and fail to describe the rate of corrosion and the accumulated amount of corrosion. Ultrasonic testing relies on the variation of wave propagation properties due to structural damage caused by steel corrosion, and is limited by its short propagation distance. Radiography relies on the passage of radiation (i.e., X-rays or gamma rays) through the structure and the consequent degrees of attenuation depending on the state of the structure (e.g., defects). The accuracy of radiography techniques is high; however, to accomplish radiographic measurements, both sides of the structure in question must be accessible. Depending on the type of radiation used, additional safety precautions are required for the operator. Infrared thermography may be a good choice for corrosion monitoring due to the strong correlation between the apparent diffusion coefficient of chloride ion and the heat dissipation characteristics of concrete. However, the interpretation of infrared thermography is only qualitative, incapable of quantitatively indicating the chloride content inside the concrete. Concrete expansion and cracking are among the most significant and obvious physical features related to concrete deterioration due to reinforcement corrosion, and thus monitoring of these two features is critical for evaluating the safe performance of reinforced concrete structures. These two features can be readily detected using acoustic emission (AE) techniques and fiber optic sensing, both of which have recently emerged as new methods to measure and characterize corrosion in concrete structures. Acoustic emission is comprised of high frequency elastic waves or stress waves generated when there is a rapid release of energy within a material that is undergoing deformation [14–16]. In concrete structures, as corrosion products are formed around steel reinforcement, pressure is built up in the surrounding concrete. When the pressure is high enough to break the interface bonding, micro-cracking and crack growth will induce stress waves that are immediately captured by AE sensors mounted on the concrete surface. Corrosion characteristics can be interpreted from AE parameter analysis. Li et al. showed that AE technique can be used in rebar corrosion detection [17]. Idrissi and Limam demonstrated a high correlation between the progressive characteristics of acoustic emission and the current density of corrosion [18]. Kawasaki et al. identified different corrosion stages such as onset of corrosion and nucleation of corrosion-induced cracking, all through AE parameter analysis [19]. The results were further verified by comparing with results from scanning electron micrograph and electron probe micro analyzer. Fiber optic sensors carry numerous advantages over traditional sensors, such as small size, immunity to electromagnetic interference, corrosion resistance, and real-time monitoring [20,21]. Therefore, they are very suitable for corrosion monitoring in reinforced concrete structures. Extensive investigation has been conducted concerning the suitability of fiber optic sensors to monitor strain in reinforced concrete structures under corrosive environment. Grattan et al. used a fiber Bragg grating (FBG) strain sensor to monitor the production of rusts during corrosion process [22]. Zheng et al. reported wrapping an FBG around a reinforcement member to capture surface strains of the member due to corrosion-induced expansion [23]. Zhao et al. examined the feasibility of a Brillouin optical time domain analysis (BOTDA) distributed sensing technique for measuring steel corrosion expansion [24]. Mao et al. investigated the combination of BOTDA and FBG sensors for monitoring corrosion expansion cracking [25]. Li et al. examined the rebar corrosion using the combined carbon fiber and FBG active thermal probe [15]. McCague et al. designed a novel corrosion sensor based on

Sensors 2017, 17, 657 Sensors 2017, 17, 657

3 of 12 3 of 12

Leung et al. proposed a steel corrosion detection method using a simple light reflection technique, in which one end of the fiber was coated with an iron thin film to interrogate steel corrosion [27]. birefringent photonic crystal fibers for pressure/force measurement [26]. Leung et al. proposed a steel Li et al. devised a metal loss detector for corrosion measurement based on fiber optic macro-bend corrosion detection method using a simple light reflection technique, in which one end of the fiber was light loss [28]. coated with an iron thin film to interrogate steel corrosion [27]. Li et al. devised a metal loss detector In the present study, an approach based on the combination of the AE technique and FBG strain for corrosion measurement based on fiber optic macro-bend light loss [28]. measurements is presented with the aim to monitor the progression of reinforcement expansion and In the present study, an approach based on the combination of the AE technique and FBG strain concrete cracking due to reinforcement corrosion in concrete structures. Acoustic emission measurements is presented with the aim to monitor the progression of reinforcement expansion measurement helps to detect corrosion initiation and crack growth while FBG strain measures the and concrete cracking due to reinforcement corrosion in concrete structures. Acoustic emission extent of corrosion and identify time of cracking. Such a combined approach has not yet been measurement helps to detect corrosion initiation and crack growth while FBG strain measures the reported in the literature, and the results of such a study can provide a correlation between acoustic extent of corrosion and identify time of cracking. Such a combined approach has not yet been reported emission events and concrete surface strains. A reinforced mortar specimen was prepared and an in the literature, and the results of such a study can provide a correlation between acoustic emission accelerated corrosion test was performed in laboratory to verify the proposed method. The rest of events and concrete surface strains. A reinforced mortar specimen was prepared and an accelerated the paper is organized as follows: Section 2 briefly describes the corrosion-induced cracking corrosion test was performed in laboratory to verify the proposed method. The rest of the paper mechanism and Section 3 illustrates the experimental setup of the corrosion tests. Experiment results is organized as follows: Section 2 briefly describes the corrosion-induced cracking mechanism and are shown and discussed in Section 4 and conclusions are drawn in Section 5. Section 3 illustrates the experimental setup of the corrosion tests. Experiment results are shown and discussed in Section 4 and conclusions are drawn in Section 5. 2. Corrosion-Induced Cracking Process 2. Corrosion-Induced Cracking Process members leads to the desolvation of metallic iron into Corrosion of steel reinforcement corrosion products as a reinforcement result of the oxidation gives of risemetallic to an increase in Corrosion of steel membersprocess. leads toThis theprocess desolvation iron into volume, which, depending on the level of oxidation, may be up to about six times the original iron corrosion products as a result of the oxidation process. This process gives rise to an increase in volume. which, The composition of the expansive corrosion products exist in the form of volume, depending on the level of oxidation, may be up tousually about six times the original iron∙ Fe OH + ∙ Fe OH + ∙ H O , where a, b and c are the variables that depend on the alkalinity volume. The composition of the expansive corrosion products usually exist in the form of pore the oxygen and the moisture content. Fe(OH + b·Fesolution H2 O}, where a, b andsupply c are the variables that depend on the alkalinity of )2water (OH)3 +ofc·concrete, {ofa·the According to the corrosion cracking model for reinforced concrete described by Zhao and the pore water solution of concrete, the oxygen supply and the moisture content. Jin [29], the steel corrosion process can be divided into three stages, namely, the free expansion the According to the corrosion cracking model for reinforced concrete described by Zhao and Jinof[29], corrosion products, stress build-up in the into concrete and the cracking concrete cover. Figure 1a the steel corrosion process can be divided threecover stages, namely, the freeofexpansion of the corrosion illustrates the initial unrestrained condition for a reinforced concrete cylinder with a steel bar of products, stress build-up in the concrete cover and the cracking of concrete cover. Figure 1a illustrates initial diameter d and a clear cover of thickness . There exists a porous zone around the the initial unrestrained condition for a reinforced concrete cylinder with a steel bar of initial diameter interface that has a mean thickness δ .around Once the steel starts corroding, the dsteel–concrete and a clear cover of thickness c. There exists a porousofzone the steel–concrete interface that corrosion products will expand freely until the porous zone is filled and then the uniform expansive has a mean thickness of δ0 . Once the steel starts corroding, the corrosion products will expand freely pressure begins thethe surrounding concretepressure wall. AsP the corrosion progresses until the porous zonetois develop filled andon then uniform expansive r begins to develop on the further, the expansive pressure willcorrosion induce tensile stresses and strains on the internal concrete wall, surrounding concrete wall. As the progresses further, the expansive pressure will induce as shown in Figure 1b. In the meantime, the corrosion products will also be restrained tensile stresses and strains on the internal concrete wall, as shown in Figure 1b. In the meantime, and the condensed by the surrounding concrete. When the hoop tensile stress at any part of an inner corrosion products will also be restrained and condensed by the surrounding concrete. When the hoop circumference strength of the concrete, will occur tensile stress atexceeds any partthe of tensile an inner circumference exceeds the the corrosion-induced tensile strength of cracks the concrete, the firstly at the steel–concrete interface and then gradually expand towards the outer cover surface, as corrosion-induced cracks will occur firstly at the steel–concrete interface and then gradually expand depicted the by outer Figure 1c. surface, The concrete cracking dueThe to concrete reinforcement corrosion towards cover as depicted by process Figure 1c. cracking process will due be to monitored by an acoustic emission technique, whereas the concrete expansion process will be reinforcement corrosion will be monitored by an acoustic emission technique, whereas the concrete measured by FBG strain sensor. expansion process will be measured by FBG strain sensor.

Figure Figure 1. 1. Schematic Schematic diagram diagram of of corrosion-induced corrosion-induced concrete concrete cracking cracking process: process: (a) (a) initial initial unrestrained unrestrained reinforced concrete; (b) internal expansive pressure for concrete in restrained condition; (c) reinforced concrete; (b) internal expansive pressure for concrete in restrained condition; (c) corrosion corrosion products products deposited depositedwithin withinopen opencracks. cracks.

Sensors 2017, 17, 657

4 of 12

Sensors 2017, 17, 657 3. Experimental Setup

4 of 12

3. Experimental Setup approach for reinforcement corrosion monitoring through a combined AE To verify the proposed and FBG detection system, a cylindrical mortarcorrosion specimenmonitoring was prepared anda an accelerated To verify the proposed approachreinforced for reinforcement through combined corrosion test FBG was detection conducted underalaboratory Both AE transducers and the FBG sensor AE and system, cylindrical conditions. reinforced mortar specimen was prepared and an were used in the corrosion same corrosion sounder as to investigate the correlation between acoustic accelerated test wasspecimen conducted laboratory conditions. Both AE transducers andemission the weresurface used instrain. the same corrosion specimen so as to investigate the correlation between eventsFBG andsensor concrete acoustic emission events and concrete surface strain.

3.1. Preparation of Specimen 3.1. Preparation of Specimen

The cylindrical reinforced mortar specimen was 80 mm in diameter and 200 mm in height, cylindrical reinforced mortar specimen was 80 mm in diameter and 200 mm in height, as as shown The in Figure 2. The cement used was ordinary Portland cement (OPC) and the cement, water, shown in Figure 2. The cement used was ordinary Portland cement (OPC) and the cement, water, sand mix ratio was 1:0.5:2 by weight. One deformed rebar 16 mm in nominal diameter was embedded sand mix ratio was 1:0.5:2 by weight. One deformed rebar 16 mm in nominal diameter was longitudinally into the center of the mortar cylinder. The rebar was 300 mm long with one end embedded longitudinally into the center of the mortar cylinder. The rebar was 300 mm long with embedded approximately 20 mm away the bottom the mortar A smallA steel one end embedded approximately 20 to mm away to theofbottom of thecylinder. mortar cylinder. smallplate steel was welded to was the welded exposed of rebarend forofthe convenience of AE transducer installation. TheThe mortar plate to end the exposed rebar for the convenience of AE transducer installation. was removed from the PVC mold day afterone being It wascasted. curedItinwas moist environment mortar was removed from theone PVC mold daycasted. after being cured in moist for for seven and dried20 indays air forunder another 20 days under room temperature. sevenenvironment days and dried in airdays for another room temperature.

Figure 2. Reinforcedmortar mortarspecimen specimen and Figure 2. Reinforced andsensor sensorinstallation. installation.

3.2. Setup of Monitoring Systems 3.2. Setup of Monitoring Systems Two AE transducers were mounted on the specimen. One transducer was on the top of the steel

Two AE transducers were mounted on the specimen. One transducer was on the top of the plate, and the other was on the top surface of the mortar, as shown in Figure 2. The purpose of such steel plate, and the other on the surface of the mortar, shown in Figure 2. The purpose an arrangement was was to study thetop characteristic of AE signalaspropagating via different media. of such an arrangement wasapplied to study themounting characteristic ofbefore AE signal propagating via different Lubricant grease was on the surface attaching AE transducers in order media. to Lubricant was applied on detection the mounting surface before attaching AE transducers in order ensuregrease good contact for signal between transducer surface and specimen surface. The AE to transducers werefor then mounted on thebetween surface and tightenedsurface using adhesive tape. An FBG strain ensure good contact signal detection transducer and specimen surface. The AE sensor (central wavelength: nm, strain pm/με) was tape. circumferentially transducers were then mounted1544.878 on the surface andcoefficient: tightened1.15 using adhesive An FBG strain around the mortar surfacenm, in the middle cross section, as depicted Figure 2. The middle sensorbonded (central wavelength: 1544.878 strain coefficient: 1.15 pm/µε) wasincircumferentially bonded section was chosen in order to record a more uniform expansion. The FBG strain sensor was around the mortar surface in the middle cross section, as depicted in Figure 2. The middle section was pre-stressed about 300 pm. A thin layer of epoxy (Epoxy GEL, Devcon, Danvers, MA, USA) of chosen in order to record a more uniform expansion. The FBG strain sensor was pre-stressed about 17 MPa strength was then applied for sensor protection. To eliminate the influence of temperature 300 pm. A thin layer of epoxy (Epoxy GEL, Devcon, Danvers, MA, USA) of 17 MPa strength was then variation, an FBG temperature sensor (central wavelength: 1549.335 nm, temperature coefficient: applied sensorwas protection. To eliminate influence ofsensor. temperature variation, an FBG temperature 10.3forpm/℃) also connected in line the with the strain The FBG temperature sensor was ◦ C) was also connected in sensorprepared (centralby wavelength: 1549.335 nm, temperature coefficient: 10.3 pm/ putting a bare FBG into a small stainless steel tube, followed by sealing both ends of the line with strain sensor. Thethat FBG sensor was prepared variation. by putting a bare FBG into a tube the using epoxy, to ensure thetemperature FBG is only sensitive to temperature The corrosion monitoring systems consists of two systems, namely, thetoAE corrosion small stainless steel tube, followed by sealing both endsseparate of the tube using epoxy, ensure that the monitoring system the FBG strain monitoring system. The layout of these two systems are FBG is only sensitive to and temperature variation. depicted in Figure 3. The AEsystems instrumentation up of AE transducers, AE The corrosion monitoring consistsmade of two separate systems,preamplifiers namely, theand AEan corrosion system (Micro-ΙΙ Digital AE System, Physical Acoustic Corp. (PAC), Princeton Junction, NJ, USA). monitoring system and the FBG strain monitoring system. The layout of these two systems are depicted in Figure 3. The AE instrumentation made up of AE transducers, preamplifiers and an AE system (Micro-II Digital AE System, Physical Acoustic Corp. (PAC), Princeton Junction, NJ, USA).

Sensors 2017, 17, 657

5 of 12

Sensors 2017, 17, 657

5 of 12

The transducers used were R15 type, which resonate at 150 kHz. The sampling frequency for The transducers used were type, which resonate at 150 kHz. Theset sampling frequency the recording waveforms was R15 5 MHz. The threshold level was at 40 dB, whichforisthe sensitive recording waveforms was 5 MHz. The threshold level was set at 40 dB, which is sensitive enough to enough to capture corrosion induced signal while allowing the acquisition of all the significant data. capture corrosion induced signal while allowing the acquisition of all the significant data. The The waveforms was amplified with 40 dB gain by a preamplifier. The AE waveforms and parameters, waveforms was amplified with 40 dB gain by a preamplifier. The AE waveforms and parameters, such as such events, hits, counts, energyenergy and duration, werewere recorded on the computer, and cancan bebe exported as events, hits, counts, and duration, recorded on the computer, and as ASCIIexported files forasfurther processing. ASCII files for further processing.

Figure 3. Reinforced mortar specimen specimen and sensor installation. Figure 3. Reinforced mortar and sensor installation.

For the FBG strain monitoring system, the wavelength variation due to corrosion expansion

Forwas the interrogated FBG strain by monitoring system, the(sm130, wavelength due to corrosion expansion was an FBG interrogator Micronvariation Optics Inc., Atlanta, GA, USA). The interrogator to a (sm130, laptop viaMicron Ethernet, and the wavelength data USA). was stored the interrogated by an was FBGconnected interrogator Optics Inc., Atlanta, GA, Theininterrogator laptop with saving frequency of one sample every 100 s. data was stored in the laptop with a saving was connected to aalaptop via Ethernet, and the wavelength frequency of one sample every 100 s. 3.3. Pencil Lead Break Test The Break attenuation 3.3. Pencil Lead Test rate of acoustic emission signal transmitted in concrete and steel materials are different. To check the sensitivity of each sensor, a pencil lead break test was conducted on the

Themortar attenuation rate acoustic transmitted in concrete and steel materials are specimen. Theof test consists emission of breakingsignal a 0.5 mm diameter pencil lead by pressing it against theTo specimen surface. The testing points werea along longitudinal and the spacing different. check the sensitivity of each sensor, pencilthe lead break testdirection was conducted on the mortar between wasof 40 breaking mm. Thus, a there six testing pencil points inlead total.by pressing it against the specimen. Thethese test points consists 0.5 will mmbediameter specimen surface. The testing points were along the longitudinal direction and the spacing between 3.4. Accelerated Corrosion Test these points was 40 mm. Thus, there will be six testing points in total. To corrode the reinforced mortar in a reasonable amount of time, an accelerated corrosion test

was performed. TheTest specimen was placed in a bucket filled with 3.5% NaCl solution, which was 3.4. Accelerated Corrosion

used to simulate seawater. The specimen was immersed in the NaCl solution for three days before a

To constant corrode current the reinforced mortar a reasonable amount accelerated was applied. The in constant current was set at of 80 time, mA byan a power supply.corrosion The anodetest was was connected to the rebar and the in cathode was filled connected the sacrificial stainless steel plate, asused to performed. The specimen was placed a bucket withto3.5% NaCl solution, which was shown in Figure 3. The corrosion test was conducted in a quiet room. In addition, the bucket was simulate seawater. The specimen was immersed in the NaCl solution for three days before a constant placed on a sponge cushion to reduce the noise coming from the environment. current was applied. The constant current was set at 80 mA by a power supply. The anode was connected to the and rebar and the cathode was connected to the sacrificial stainless steel plate, as shown 4. Results Discussion in Figure 3. The corrosion test was conducted in a quiet room. In addition, the bucket was placed on a Resultstofrom Pencilthe Leadnoise Break coming Test sponge 4.1. cushion reduce from the environment. Attenuation of the generated acoustic emissions was considered as having a major influence on

4. Results Discussion the and sensitivity of the transducer and accuracy of the collected data. A pencil lead break test was used to produce similar AE events and check the sensitivity of each transducer. The signal amplitude

4.1. Results fromfrom Pencil Lead Break Test received both transducers during the pencil lead break test is shown in Figure 4. Each column Attenuation of the generated acoustic emissions was considered as having a major influence on the sensitivity of the transducer and accuracy of the collected data. A pencil lead break test was used to produce similar AE events and check the sensitivity of each transducer. The signal amplitude received from both transducers during the pencil lead break test is shown in Figure 4. Each column in the

Sensors 2017, 17, 657

6 of 12

Sensors 2017, 17, 657

6 of 12

graph represents a different pencil lead breaking location starting from the top to the bottom of the in with the2017, graph representsofa 40 different pencil location starting from topsensitive to the bottom specimen increments mm. As canlead be breaking seen, both transducers are the quite Sensors 17, 657 6 ofto 12 the AE of the specimen with increments of 40 mm. As can be seen, both transducers are quite sensitive to the source located at any place of the specimen, capable of reaching 60 dB in amplitude. Transducer A, AE source at any place of the specimen, capablelocation of reaching 60 dB in amplitude. Transducer in the graphlocated represents a different pencil lead breaking starting from the top to the bottom which was mounted on the reinforcement, is more sensitive than Transducer B, which was attached to A, which was mounted on the reinforcement, is more sensitive than Transducer B, which of the specimen with increments of 40 mm. As can be seen, both transducers are quite sensitive to was the mortar. AE The difference of highest amplitude between the transducers is approximately 10 dB. attached mortar.atThe of highest amplitude between the transducers is approximately 10 dB. sourcetolocated anydifference place of the specimen, capable of reaching 60 dB in amplitude. Transducer A, which was mounted on the reinforcement, is more sensitive than Transducer B, which was 100 attached to mortar. The difference of highest amplitude between the transducers is approximately 10 dB.

Amplitude (dB) Amplitude (dB)

80 100 60 80 40 60 20 40 200

Transducer A Transducer B 0

4 Transducer A Transducer B

8 12 Distance (cm)

16

20

0

Figure 4. Amplitude from both at different locations. 0 4 8 transducers 12 20 Figure 4. Amplitude from both transducers at16different locations. Distance (cm)

Figure 5 showsFigure the recorded waveforms to events of highest amplitude at the 4. Amplitude from bothcorresponding transducers at different locations.

Amplitude (mV) (mV) Amplitude (mV)Amplitude Amplitude (mV)

same5testing location. The amplitude from Transducer A is muchto greater thanofthat from Transducer B. at the Figure shows the recorded waveforms corresponding events highest amplitude The waveform that propagated through the mortar was more attenuated as compared to the one that 5 shows recorded waveforms corresponding to events of highest amplitude at the same testingFigure location. Thethe amplitude from Transducer A is much greater than that from Transducer B. propagated through The the amplitude steel reinforcement, that isAtoissay, the mortar provides greater damping same testing much greater than thatcompared from Transducer The waveform thatlocation. propagated throughfrom theTransducer mortar was more attenuated as to theB.one that effect on the acoustic signal than the steel reinforcement. pencil lead break testtoconfirms that The waveform that propagated through the mortar was moreThe attenuated as compared the one that propagated through the steel reinforcement,inthat is to say, the mortar provides greater damping effect attenuation through of the AEthe signal mortar is greater that in provides reinforcement. propagated steeltransmitted reinforcement, that is to say, than the mortar greater damping on the acoustic than the steel reinforcement. The pencil break confirms that attenuation effect on signal the acoustic signal than the steel reinforcement. Thelead pencil lead test break test confirms that 500 of the AE signal transmitted in mortar is greater than that inthan reinforcement. attenuation of the AE signal transmitted in mortar is greater that in reinforcement. Transducer A

0 500 Transducer A -500 0-100 -500 500 -100

-80

-60

-80 -60B Transducer

-40

-40

-20

-20

0 20 Time (s) 0 20 Time (s)

40

40

60

60

80

80

100

100

0 500 Transducer B -500 0-100

-80

-60

-40

-20

0 20 Time (s)

40

60

80

100

-500

-100 5. Waveforms -80 -60 from -40 pencil -20 lead 0 break 20 tests40 60 transducers. 80 100 Figure for both Time (s)

4.2. Acoustic Emission Study Figure 5. Waveforms from pencil lead break tests for both transducers.

Figure 5. Waveforms from pencil lead break tests for both transducers.

AE hits from both transducers during accelerated corrosion of the reinforced mortar specimen 4.2. Study areAcoustic shown Emission in Figure 6. The hits number measured by Transducer B is far less compared to 4.2. Acoustic Emission Study that the mortar attenuated the AE signal at a much greater rate than did the Transducer A, indicating AE hits from both transducers during accelerated corrosion of the reinforced mortar specimen steel, much in in the same6.way in the earlier pencil-break Since only an amount of Figure Theashits number measured bytests. Transducer Breinforced is insignificant far lessmortar compared to AEare hitsshown from both transducers during accelerated corrosion of the specimen are AE events was captured that by Transducer B (only 30 events), the interpretation from Transducer B the will Transducer A, indicating the mortar attenuated the AE signal at a much greater rate than did shown in Figure 6. The hits number measured by Transducer B is far less compared to Transducer A, be very inaccurate. Thus, inas theinfollowing we will focus our discussion on results obtained steel, much in the same way the earlieranalysis, pencil-break tests. Since only an insignificant amount of indicating thatTransducer the mortarA.attenuated the AE signal at a much greater rate similarity than did the steel, much from TheTransducer curves from A interpretation show a highfrom with the AE events was captured by B (onlyTransducer 30 events), the Transducer B will in the same way as in the earlier pencil-break tests. Since only an insignificant amount of AE events phenomenological modelinof corrosion in marine environments [30]. The obtained corrosion be very inaccurate. Thus, thereinforcement following analysis, we will focus our discussion on results was captured by Transducer B (only 30from events), theaninterpretation Transducer B will activity is strong in early stage, followed by evident decrease in the corrosionwith rate after from Transducer A.theThe curves Transducer A show a from high similarity thebe very around seven days. In this period, the flow of oxygen and water, which are essential to steel inaccurate. Thus, in themodel following analysis, we will focus our environments discussion on results obtained from phenomenological of reinforcement corrosion in marine [30]. The corrosion corrosion, iscurves inhibited thestage, buildup corrosion products, thuswith slowing down the rate corrosion activity is strong in from the by early followed byaan evident decrease in the the phenomenological corrosion after model Transducer A. The Transducer Aofshow high similarity around seven days. Ininthis period, the flow of oxygen and corrosion water, which are essential to steel of reinforcement corrosion marine environments [30]. The activity is strong in the early corrosion, is inhibited by the buildup of corrosion products, thus slowing down the corrosion stage, followed by an evident decrease in the corrosion rate after around seven days. In this period, the flow of oxygen and water, which are essential to steel corrosion, is inhibited by the buildup of corrosion products, thus slowing down the corrosion process. The corrosion proceeded further after

Sensors 2017, 17, 657

7 of 12

Sensors 2017, 17, 657 7 of 12 process. The corrosion proceeded further after 22 days with the assistance of anaerobic corrosion 17, 657 7 of 12 until Sensors cracks2017, were observed on the concrete surface on around the 36th day.

22 days with The the assistance of anaerobic corrosion were observed on the concrete surface process. corrosion3500 proceeded further after 22until dayscracks with the assistance of anaerobic corrosion 35 on around the 36th until cracks wereday. observed on the concrete surface on around the 36th day. 35

3000

30

20

2500

25

1500

2000

20

1000

1500

15

5001000 0 500 0 0

5 0

10 5

15

10

20 Time (day)

15

AE Hits - Transducer B

2000

25

15 10

AE Hits - Transducer B

30

3500

2500 AE Hits - Transducer A

AE Hits - Transducer A

3000

AE Hits - Transducer A10 5 AE Hits - Transducer B AE Hits - Transducer A 5 0 25 30 35 40 AE Hits - Transducer B

20 25 Time (day)

30

0 40

35

Figure 6. AE hits during the accelerated corrosion test. Figure 6. AE hits during the accelerated corrosion test. Figure 6. AE hits during the accelerated corrosion test.

Peak frequency is the frequency with the maximum magnitude as determined by a fast Fourier frequency is the domain frequencysignal. with the maximum magnitudecan as determined a fastindicator Fourier for transformPeak (FFT) of the time The peak frequency serve as aby good Peak frequency is the the time frequency with theThe maximum magnitude as determined by a fast for Fourier transform (FFT) of domain signal. peak frequency can serve as a good indicator different AE sources arising from reinforcement corrosion. Other indicators of frequency, like transform (FFT) the time domain The peak frequency serve asofa frequency, good indicator different AE of sources arising from signal. reinforcement corrosion. Othercan indicators like for average or central frequency, share the same properties. Figure 7 shows the peak frequency different AE sources arising from reinforcement corrosion. Other indicators frequency, like average average or central frequency, share the same properties. Figure 7 showsofthe peak frequency corresponding to each hits from Transducer A. It can be seen that the population can be divided into corresponding to each hits from Transducer A. It Figure can be seen that the can be divided into or central frequency, share the same properties. 7 shows thepopulation peak frequency corresponding three three majormajor parts: oneone below 5050 kHz (denoted as Type I), one onearound around 110 kHz (denoted as Type II) parts: below kHz (denoted as Type I), 110 kHz (denoted as Type to each hits from Transducer A. It can be seen that the population can be divided into threeII)major and another part with peak frequency higher than 240 kHz (denoted as Type III). Specifically, the with(denoted peak frequency higher than 240 kHz (denoted as Type as III).Type Specifically, the parts:and oneanother belowpart 50 kHz as Type I), one around 110 kHz (denoted II) and another majority (75%) of the hitshits exhibit peak than50 50kHz, kHz,while while 11% around 100 kHz majority (75%) of the exhibit peakfrequency frequency lower lower than 11% areare around 100 kHz part with peak frequency higher than 240 kHz (denoted as Type III). Specifically, the majority (75%) and 14% in the highest kHz.These Thesethree three types of frequency bands and are 14%scattered are scattered in the highestband bandabove above 200 200 kHz. types of frequency bands are are of the hits exhibit peak frequency lower than 50 kHz, while 11% are around 100 kHz and 14% are in close relation to different AE sources due to reinforcement corrosion. Mazille and Rothea [31] in close relation to different AE sources due to reinforcement corrosion. Mazille and Rothea [31] scattered in that the highest band above 200 kHz. threegave types of frequency frequency AE bands are in close relation bubble evolution theliquid liquidThese medium signals (~50 kHz); statedstated that bubble evolution in inthe medium gave low low frequency AE signals (~50 kHz); to different AE sources due to reinforcement corrosion.ofMazille and Rotheafrom [31]electrochemical stated that bubble therefore, I can attributed the evolution evolution therefore, TypeType I can be be attributed totothe of hydrogen hydrogenbubbles bubbles from electrochemical evolution in the liquidanode medium low frequency AE by signals (~50 kHz); therefore, Typefrom I can be reaction between andgave cathode. As suggested Yoon’s work [32], AE signals reaction between anode and cathode. As suggested by Yoon’s work [32], AE signals from attributed to the evolution of hydrogen from electrochemical between anode micro-cracking and macro-cracking are bubbles of short duration and of relativelyreaction high frequency (180 kHz and micro-cracking and macro-cracking are of short duration and of relatively high frequency (180 kHz to 350 and by the Yoon’s AE signals with debonding are of longer and theirare of cathode. As kHz), suggested workassociated [32], AE signals from micro-cracking andduration macro-cracking to 350 kHz), and the AE kHz. signals associated with that debonding are of longer duration and their frequencies peak at 120 Li et al. [17] concluded the peak frequencies for a typical concrete short duration and of relatively high frequency (180 kHz to 350 kHz), and the AE signals associated frequencies peakwere at 120 kHz. Li et al. [17] concluded peak frequencies for a typical concrete crack signal frequencies, from 275that kHzthe topeak 350 kHz. It is reasonable state that with debonding are ofhigh longer durationranging and their frequencies at 120 kHz. Li et al.to[17] concluded crackType signalII were high frequencies, ranging from 275 kHz to 350 kHz. It is reasonable to state appears to correspond to the generation and movement of corrosion products and that that the peak frequencies for a typical concrete crack signal were high frequencies, ranging from Type reinforcement/mortar II appears to correspond to the generation movement ofcomes corrosion products and interface debonding while Type and III peak frequencies from micro-crack 275 kHz to 350 kHz. It is reasonable to state that Type II appears to correspond to the generation and development and crack propagation. All while of theseType threeIII types peak frequencies occurred for the reinforcement/mortar interface debonding peakoffrequencies comes from micro-crack movement corrosion productstest. and reinforcement/mortar debonding Type III peak whole of span of crack the corrosion in our case,types weinterface can that the while process of final development and propagation.Therefore, All of these three of surmise peak frequencies occurred for the frequencies comes from micro-crack development and crack propagation. All of these three types of mortar cover cracking is an accumulative process, which is comprised of a large amount of final whole span of the corrosion test. Therefore, in our case, we can surmise that the process of peak micro-cracking frequencies occurred for the whole span of the corrosion test. Therefore, in our case, we can events before final cover cracking occurs. Figure 8 presents typical waveforms of mortar cover cracking is an accumulative process, which is comprised of a large amount of these three andof their corresponding fast Fourier is transform (FFT). surmise that thetypes process final mortar cover cracking an accumulative process, which is comprised micro-cracking events before final cover cracking occurs. Figure 8 presents typical waveforms of of a large amount of micro-cracking events before final cover cracking occurs. Figure 8 presents typical these three types and their corresponding fast Fourier transform (FFT). waveforms of these three types 400 and their corresponding fast Fourier transform (FFT). 350 Peak Frequency (kHz)

400 300

Peak Frequency (kHz)

350 250 300 200 250 150 200 150

100 50 0

100

0

5

10

15

20 25 Time (day)

30

35

40

50

Figure 7. Peak frequency content from Transducer A. 0

0

5

10

15

20 25 Time (day)

30

35

40

Figure A. Figure 7. 7. Peak Peak frequency frequency content content from from Transducer Transducer A.

Sensors 2017, 17, 657 Sensors 2017, 17, 657

8 of 12 8 of 12

0

-20 -1000

Type I

-50

-20

Amplitude (mV)

20 -100

0 Time (s)

-50

Type II

50

0 Time (s)

100

50

Amplitude (mV)

Type II

-20 0 -100 -20 -100

20

Amplitude (mV)

-50

0 Time (s)

-50

50

0 Time (s)

100

50

Type III Amplitude (mV)

Type III 0

-20 -100

-50

-20 -100

0 Time (s)

-50

50

0 Time (s)

0.0010 0.0000

100

50

0

200

400 600 Frequency (kHz)

0.0005 0.0000

0.00100

200

400 600 Frequency (kHz)

800

800

1000

1000

0.0005 0.0010 0.0000 0.0005

0.0000

100

20

0

0.0015 0.0005

100

0 20

8 of 12

0.0010

MagnitudeMagnitude

Amplitude (mV)

20

Magnitude Magnitude

0.0015 Type I

0

200

0

200

0.0015 Magnitude Magnitude

Amplitude (mV)

20

Sensors 2017, 17, 657

400 600 Frequency (kHz) 400 600 Frequency (kHz)

800 800

1000 1000

0.0010

0.0015

0.0005

0.0010

0.0000 0.0005 0.0000

100

0

200

0

200

400 600 Frequency (kHz) 400 600 Frequency (kHz)

800 800

1000 1000

Figure Figure 8. 8. Typical Typical waveforms of different types and their fast Fourier transform (FFT). Figure 8. Typical waveforms of different types and their fast Fourier transform (FFT).

The effective technique AE signal signal for for the the The cross-plot cross-plot is is an an effective technique to to analyze analyze the the features features of of an an entire entire AE The cross-plot is an effective technique to analyze the featuresthe of an entireofAE signal for the of purpose of distinguishing signal characteristics. Figure 9 shows results a 3D cross-plot purpose of distinguishing signal characteristics. Figure 9 shows the results of a 3D cross-plot of purposeenergy of distinguishing signal characteristics. Figure obtained 9 shows the results of a 3D cross-plot of duration, andsignal signal strength of signals AE signals from Transducer A. The circles duration, energy and strength of AE obtained from Transducer A. The circles represent duration, energy and signal strength of AE signals obtained from Transducer A. The circles represent events I andindicate squaresAE indicate AE events from Type II,diamonds whereas diamonds AE eventsAE from Typefrom Ifrom andType squares eventsAE from Type II, Type whereas are those represent AEType events Type is I and squares indicate events from II, whereas diamonds are those from III. Duration defined as the time from the first threshold crossing to the endlast of from Duration defined is asdefined the time thefrom firstthe threshold crossing to thetoend the areType thoseIII. from Type III.isDuration as from the time first threshold crossing the of end of the last threshold crossing of signal. the AE signal. Energy is defined as the integral of the rectified voltage threshold the AE is defined as the as integral of theofrectified voltage signal the last crossing thresholdof crossing of the AEEnergy signal. Energy is defined the integral the rectified voltage signal over the duration of the AE hit. Signal strength is defined as the measured area under the oversignal the duration of the AE hit. Signal strength is defined as the measured area under the amplitude over the duration of the AE hit. Signal strength is defined as the measured area under the amplitude timeIt envelope. It that can be noted that Type I AE over events are spread over the time envelope. be noted I AE events are spread the characterized by plot, long amplitude timecan envelope. It canType be noted that Type I AE events areplot, spread over the plot, characterized by long duration, high energy and strong signal strength and Type II AE events characterized by long high strength energy and strength andshorter Type IIduration, AE events duration, high energy andduration, strong signal and strong Type IIsignal AE events feature lower feature shorter duration, and signal strength, whereas Type III events are feature shorter duration, lowerenergy energy and weaker weaker strength, Type III AEAE events energy and weaker signallower strength, whereas Type IIIsignal AE events arewhereas more tightly clustered andare have more tightly clustered and have the value of the thethree three parameters above. Overall, these tightly clustered and have the smallest smallest valueOverall, of above. Overall, these the more smallest value of the three parameters above. theseparameters three types of AE events, due to three types of AE events, due to reinforcement corrosion, can be easily recognized. three types of AE events, due to reinforcement corrosion, can be easily recognized. reinforcement corrosion, can be easily recognized. 5

Type I Type Type II Type Type III III Type

5

4 4 2

2

0

Energy (10 Vs/count)

2000

0

0

0

00 0

Duration (s)

Duration (s)

5

6

2

2

0

0

1000

1000

2000 3000 4000 Duration (s)

2000

3000

4000

5000

5000

6000

6000

Signal Strength (pVs) Signal Strength (pVs)

4

0

20

20

x 10

40 60 40 60 Energy

80

80

100

100

Energy

5

x 10

Signal Strength (pVs)

Signal Strength (pVs)

x510

0

4

0

2000

5000

50 50 Energy (10 Vs/count)

100

6

4000 4000

5000

0 0 100

6

6000 6000

Duration(s) (s) Duration

6 6

Signal Strength (pVs)

Signal Strength (pVs)

x 10x 10

5

x 10

6 4

4 2

2 0

0

0

0

20

20

40 60 Energy

40

80

60

100

80

100

Duration of (s) Energy Figure cross-plot duration, energy energy and its perspective plots. Figure 9. 9. 3D3D cross-plot of duration, and signal signalstrength strengthand and its perspective plots.

Figure 9. 3D cross-plot of duration, energy and signal strength and its perspective plots.

Sensors 2017, 17, 657

9 of 12

Sensors 2017, 17, 657

9 of 12

4.3. FBG Strain Measurement

The 2017, circumferential strain variation of the reinforced mortar was measured using an9FBG Sensors 657 of 12 strain 4.3. FBG Strain17,Measurement sensor. Measurement commenced the moment that the current was applied all the way to when the 4.3. Strain Measurement specimen cracked. The strain variation data overof the span ofmortar about was 36 days is shown in an Figure The TheFBG circumferential thetime reinforced measured using FBG10. strain originalMeasurement wavelength variation was converted to strain variation by adopting a strain sensitivity sensor. commenced the moment that the current was applied all the way to when The circumferential strain variation of the reinforced mortar was measured using an FBG strain coefficient ofcracked. 1.15 pm/με, which data was calibrated beforehand. Asapplied can thetostrain increased the specimen The strain over the time span ofwas about 36 be days shown in Figure 10. sensor. Measurement commenced the moment that the current allseen, theisway when the specimen cracked. The variation strain datawas over the time span of exerted about 36internal daysby is adopting shown in Figure The The with time aswavelength the corrosion products accumulated pressure the10.mortar. The original converted toand strain variation aon strain sensitivity original variation strainand variation by down adopting a strain sensitivity strain grew approximately 1000was με converted every fivetobeforehand. days, slowed tothe about 500 με everywith five coefficient ofwavelength 1.15 pm/µε, which was calibrated As can be seen, strain increased coefficient of 1.15 pm/με, which was calibrated beforehand. As can be seen, the strain increased daysasinthe thecorrosion later period. The change in the and strain rate can be attributed of cracks, time products accumulated exerted internal pressureto onthe theformation mortar. The strain withprovides time as the corrosion products and exerted internal pressure the mortar. The before which new space for corrosion The corrosion products fills the grew approximately 1000 µε every fiveaccumulated days,products. and slowed down to about 500 µεonevery fivecrack days in the strain grew approximately 1000 με every five days, and slowed down to about 500 με every five building upThe stress on the when generated. Notice that, 24which days, provides the strain later period. change in mortar the strain rate cracks can be were attributed to the formation of after cracks, days in the later period. The change in the strain rate can be attributed to the formation of cracks, curve exhibited a jagged variationThe thatcorrosion results from the redistribution the corrosion new space for corrosion products. products fills the crackofbefore building products up stress into on which provides new space for corrosion products. The corrosion products fills the crack before themortar widening cracks as the crack front propagates further into the cover. This period can be the when cracks were generated. Notice that, after 24 days, the strain curve exhibited a jagged building up stress on the mortar when cracks were generated. Notice that, after 24 days, the strain identified as the crack propagation stage. The embedded plot partially shows an enlarged view variation results from the redistribution of from the corrosion products the widening cracks curvethat exhibited a jagged variation that results the redistribution of into the corrosion products intoas theof the the jagged curve.cracks The further crack reached the outer surface on can around the cover. 36th day and waspropagation marked crack front propagates into cover. This period be identified as the crack widening as the crackthe front propagates further into the This period can be by a sudden inthe strain reading. After many days of internal buildup the accumulation identified as crack propagation stage. The embedded plot partially shows an from enlarged view of the stage. Thedrop embedded plot partially shows an enlarged view ofpressure the jagged curve. The crack reached the jagged curve. The crack reached the outer surface on around the 36th day and was marked by of corrosion products, the36th whole cylinder Whenreading. the pressure in the outer surface on around the daymortar and was markedhas by abecome suddenstressed. drop in strain Aftera many sudden drop in strain reading. After many days of internal pressure buildup from the accumulation mortar surface exceeds tensile from strength, surface cracking occurs and the pressure on themortar mortar days of internal pressureits buildup the accumulation of corrosion products, the whole of corrosion products, theresulting whole mortar cylinder has become stressed. the pressure in the that surface suddenly releases, drop in stress strain on theWhen surface. Ititsistensile noteworthy cylinder has become stressed. When in thea pressure in theormortar surface exceeds strength, mortar surface exceeds its tensile strength, surface cracking occurs and the pressure on the mortar surfacecracking microcracking occurred throughout allmortar of the stages long as the stressresulting exceeds the surface occurs and the pressure on the surfaceassuddenly releases, in atensile drop surface suddenly releases, resulting in a drop in stress or strain on the surface. It is noteworthy that ofstrain the on concrete. As reinforcement corrosion continued, microcracks developed into instrength stress or the surface. It is noteworthy that surface microcracking occurred throughout surface microcracking occurred throughout all of the stages as long as the stress exceeds the tensile macrocracks. The macrocrack, which had a width of 1.391 mm, is shown in Figure 11. The crack all ofstrength the stages as long as theAsstress exceeds the tensile strength the concrete. As reinforcement of the concrete. reinforcement corrosion continued,ofmicrocracks developed into width estimated from strain measurement was 1.141 mm (4540 με ), which shows good corrosion continued, microcracks developed into macrocracks. The macrocrack, which had a width of macrocracks. The macrocrack, which had a width of 1.391 mm, is shown in Figure 11. The crack agreement with results from direct measurement. 1.391 mm, is shown in Figure 11. The crack width estimated from strain measurement was 1.141 mm width estimated from strain measurement was 1.141 mm (4540 με ), which shows good with results from good direct agreement measurement. (4540agreement µε × πD), which shows with results from direct measurement. 5000 5000

4500

4500

4000

4000

3500

Strain ()

Strain ()

3500

3000

3000

2500

4800

2500

4800

2000 2000

46004600

1500 1500

44004400

1000 1000

42004200 40004000

500 500 0

00 0

28

5

5

10

10

15

20 15 Time (day) 20

28

30

25

32

30

25

30

34

36

32

30

34

35

36

35

40

Time (day)

Figure 10. Strain history from FBG strain measurement. Figure Figure10. 10.Strain Strainhistory historyfrom fromFBG FBGstrain strainmeasurement. measurement.

Figure 11. Illustration of cracking of the specimen.

Figure11. 11.Illustration Illustrationofofcracking crackingofofthe thespecimen. specimen. Figure

40

Sensors Sensors 2017, 2017, 17, 17, 657 657

10 10 of of 12 12

The above results showed that FBG strain measurement is effective in monitoring reinforcement The above results showed that FBG strain measurement is effective in monitoring reinforcement corrosion induced mortar expansion and cracking. For the case investigated in this study, a single corrosion induced mortar expansion and cracking. For the case investigated in this study, a single strain curve provides corrosion information from these three aspects. First, the tensile strain around strain curve provides corrosion information from these three aspects. First, the tensile strain around the mortar specimen due to expansive pressure exerted by the formation of corrosion products was the mortar specimen due to expansive pressure exerted by the formation of corrosion products was continuously monitored. Second, the crack initiation and propagation period, featured by notable continuously monitored. Second, the crack initiation and propagation period, featured by notable jagged variation of strain curve, can be easily recognized. Third, the moment of the first appearance jagged variation of strain curve, can be easily recognized. Third, the moment of the first appearance of macrocracks on the mortar surface was detected and the crack width can be estimated from strain of macrocracks on the mortar surface was detected and the crack width can be estimated from strain value. In future work, more samples will be tested and additional experiments will be designed to value. In future work, more samples will be tested and additional experiments will be designed to validate the efficacy of FBG strain measurement. validate the efficacy of FBG strain measurement. 4.4. Discussion 4.4. Discussion Figure 12 is a plot that shows both acoustic emission hits and FBG strain measurement. These Figure 12 is a plot that shows both acoustic emission hits and FBG strain measurement. These two different corrosion monitoring methods interpret the phenomenon of reinforcement corrosion two different corrosion monitoring methods interpret the phenomenon of reinforcement corrosion from two different aspects; one is from acoustic emission aspects and the other is from expansive from two different aspects; one is from acoustic emission aspects and the other is from expansive strain strain development aspects. As can be seen, the AE hits and FBG strain show a similar trend and development aspects. As can be seen, the AE hits and FBG strain show a similar trend and confirm that confirm that the level of acoustic emission activity and the development of surface strain generally the level of acoustic emission activity and the development of surface strain generally had a positive had a positive correlation as the corrosion of the reinforced concrete progressed. This agreement is correlation as the corrosion of the reinforced concrete progressed. This agreement is qualitative, and, qualitative, and, for a more detailed understanding and analysis of the AE signals, the adoption of for a more detailed understanding and analysis of the AE signals, the adoption of advanced signal advanced signal processing techniques and pattern recognition algorithms would be helpful. processing techniques and pattern recognition algorithms would be helpful. 5000

3000

4000

2000 3000 1500 2000

Strain ()

AE Hits - Transducer A

2500

1000 500 0

AE Hits - Transducer A FBG Strain 0

5

10

15

20 25 Time (day)

30

35

1000 0 40

strain measurement. measurement. Figure 12. Simultaneous evaluation of acoustic emission activity and strain

The presented presented integrated integrated corrosion corrosion monitoring monitoring technique technique was was performed performed under under controlled controlled The laboratory conditions. conditions. As integrated corrosion corrosion monitoring monitoring technique, technique, laboratory As for for field field application application of of the the integrated there will be many practical challenges. Since the reinforced concrete structure is generally in there will be many practical challenges. Since the reinforced concrete structure is generally largelarge in size, size, surface mounting AE sensors may reduce the sensitivity and efficacy of the AE sensors. surface mounting AE sensors may reduce the sensitivity and efficacy of the AE sensors. Alternatively, Alternatively, we recently proposedembedded the low-cost, embedded aggregate as a typeAE of sensor embedded we recently proposed the low-cost, smart aggregatesmart as a type of embedded [16]. AE sensor [16]. The embedded AE sensors can be strategically installed at the locations that are The embedded AE sensors can be strategically installed at the locations that are prone to damage in prone to damage in order to increase the efficacy of the AE method. As for the FBG strain order to increase the efficacy of the AE method. As for the FBG strain measurement, the authors will measurement, the authors willthe instrument FBG sensors the reinforced samples in instrument the FBG sensors on reinforcedthe concrete samples on in such a way that concrete a ring shaped cavity such a way that a ring shaped cavity is created around the rebar and FBG sensors are installed is created around the rebar and FBG sensors are installed within the cavity. In this way, the stress within the cavity. In this way, the is stress transfer direction is ensured andfrom the transfer in circumferential direction ensured and in thecircumferential fragile FBG sensors are well-protected fragile FBG sensors are well-protected from external interruption during construction. external interruption during construction. 5. Conclusions Conclusions 5. In this this paper, paper, we we present present an an integrated integrated monitoring monitoring technique technique for for reinforced reinforced concrete concrete corrosion corrosion In based on on the the novel novel coupling coupling of of acoustic acoustic emission emission and and FBG FBG strain strain sensing. sensing. A A reinforced reinforced mortar mortar was was based prepared and the corrosion process was accelerated through the injection of constant current. The prepared and the corrosion process was accelerated through the injection of constant current. The setup setup allowed the study of corrosion characteristics at different stages. Corrosion monitoring can be

Sensors 2017, 17, 657

11 of 12

allowed the study of corrosion characteristics at different stages. Corrosion monitoring can be divided into two aspects: the corrosion initiation and cracking propagation can be captured by an acoustic emission system, whereas the mortar expansion due to corrosion product generation will be monitored by a fiber Bragg grating strain sensor. It was found from the pencil lead break test that the attenuation of AE signal transmitted in mortar material is greater than that in steel material. Analysis from AE hits show that the hit history resembles the phenomenological model of reinforcement corrosion loss in a marine environment. AE signals are generated from three types of sources, including hydrogen bubble evolution, generation and movement of corrosion products, and micro-crack development and crack propagation. These three types of AE sources distinguish themselves in parameters like duration, energy and signal strength. Through continuous monitoring of tensile strain using FBG strain sensors, crack initiation and the propagation period were identified, and the time of concrete cover cracking was detected. Good agreement was obtained between acoustic emission activity and expansive strain measured on the specimen surface. Specifically, a similar trend in AE activity and strain measurement was observed as corrosion progressed. The AE events characterize the corrosion development based on corrosion induced AE events, and FBG strain measurement characterized the corrosion in terms of concrete expansion induced by rebar corrosion. Results have shown that the integration of these two nondestructive techniques is promising for reinforced concrete corrosion monitoring and characterization. Acknowledgments: This research reported in this paper was partially supported by the Major State Basic Research Development Program of China (973 Program, No. 2015CB057704), National Natural Science Foundation of China (No. 51378434, 51578456) and the China Scholarship Council (No. 201306060086). Author Contributions: W.L. developed the idea, designed the experiments, and wrote the paper; C.X. helped perform the experiments; S.C.M.H. helped write the paper; B.W. helped analyze the data; and G.S. made critical comments to the paper. Conflicts of Interest: The authors declare no conflict of interest.

References 1. 2. 3. 4. 5. 6.

7. 8. 9. 10. 11.

Liu, Y.; Weyers, R.E. Modeling the time-to-corrosion cracking in chloride contaminated reinforced concrete structures. ACI Mater. J. 1998, 95, 675–681. Mehta, P.K. Concrete. Structure, Properties and Materials; Prentice-Hall: Englewood Cliffs, NJ, USA, 1986. Ji, Q.; Ho, M.; Zheng, R.; Ding, Z.; Song, G. An exploratory study of stress wave communication in concrete structures. Smart Struct. Syst. 2015, 15, 135–150. [CrossRef] Feng, Q.; Kong, Q.; Huo, L.; Song, G. Crack detection and leakage monitoring on reinforced concrete pipe. Smart Mater. Struct. 2015, 24, 115020. [CrossRef] Liang, Y.; Li, D.; Parvasi, S.M.; Kong, Q.; Song, G. Bond-slip detection of concrete-encased composite structure using electro-mechanical impedance technique. Smart Mater. Struct. 2016, 25, 095003. [CrossRef] Lemoine, L.; Wenger, F.; Galland, J. Study of the Corrosion of Concrete Reinforcement by Electrochemical Impedance Measurement. Corrosion Rates of Steel in Concrete; ASTM STP: West Conshohocken, PA, USA, 1990; Volume 1065, pp. 118–133. Lambert, P.; Page, C.; Vassie, P. Investigations of reinforcement corrosion. 2. Electrochemical monitoring of steel in chloride-contaminated concrete. Mater. Struct. 1991, 24, 351–358. [CrossRef] Andrade, C.; Alonso, C. Test methods for on-site corrosion rate measurement of steel reinforcement in concrete by means of the polarization resistance method. Mater. Struct. 2004, 37, 623–643. [CrossRef] Yeih, W.; Huang, R. Detection of the corrosion damage in reinforced concrete members by ultrasonic testing. Cem. Concr. Res. 1998, 28, 1071–1083. [CrossRef] Song, H.-W.; Saraswathy, V. Corrosion monitoring of reinforced concrete structures-A. Int. J. Electrochem. Sci. 2007, 2, 1–28. Duffó, G.; Gaillard, N.; Mariscotti, M.; Ruffolo, M. Application of gamma-ray radiography and gravimetric measurements after accelerated corrosion tests of steel embedded in mortar. Cem. Concr. Res. 2015, 74, 1–9. [CrossRef]

Sensors 2017, 17, 657

12. 13. 14. 15. 16.

17. 18. 19. 20. 21.

22.

23.

24.

25. 26.

27. 28. 29. 30. 31.

32.

12 of 12

Aggelis, D.; Kordatos, E.; Soulioti, D.; Matikas, T. Combined use of thermography and ultrasound for the characterization of subsurface cracks in concrete. Constr. Build. Mater. 2010, 24, 1888–1897. [CrossRef] Kobayashi, K.; Banthia, N. Corrosion detection in reinforced concrete using induction heating and infrared thermography. J. Civ. Struct. Health Monit. 2011, 1, 25–35. [CrossRef] Li, W.; Ho, S.C.M.; Patil, D.; Song, G. Acoustic emission monitoring and finite element analysis of debonding in fiber-reinforced polymer rebar reinforced concrete. Struct. Health Monit. 2016. [CrossRef] Li, W.; Ho, S.C.M.; Song, G. Corrosion detection of steel reinforced concrete using combined carbon fiber and fiber bragg grating active thermal probe. Smart Mater. Struct. 2016, 25, 045017. [CrossRef] Li, W.; Kong, Q.; Ho, S.C.M.; Mo, Y.; Song, G. Feasibility study of using smart aggregates as embedded acoustic emission sensors for health monitoring of concrete structures. Smart Mater. Struct. 2016, 25, 115031. [CrossRef] Li, Z.; Li, F.; Zdunek, A.; Landis, E.; Shah, S.P. Application of acoustic emission technique to detection of rebar corrosion in concrete. ACI Mater. J. 1998, 95, 68–81. Idrissi, H.; Limam, A. Study and characterization by acoustic emission and electrochemical measurements of concrete deterioration caused by reinforcement steel corrosion. NDT E Int. 2003, 36, 563–569. [CrossRef] Kawasaki, Y.; Wakuda, T.; Kobarai, T.; Ohtsu, M. Corrosion mechanisms in reinforced concrete by acoustic emission. Constr. Build. Mater. 2013, 48, 1240–1247. [CrossRef] Zhao, X.; Li, W.; Song, G.; Zhu, Z.; Du, J. Scour monitoring system for subsea pipeline based on active thermometry: Numerical and experimental studies. Sensors 2013, 13, 1490–1509. [CrossRef] [PubMed] Zhao, X.; Li, W.; Zhou, L.; Song, G.-B.; Ba, Q.; Ou, J. Active thermometry based DS18B20 temperature sensor network for offshore pipeline scour monitoring using K-means clustering algorithm. Int. J. Distrib. Sens. Netw. 2013, 2013. [CrossRef] Grattan, S.; Basheer, P.; Taylor, S.; Zhao, W.; Sun, T.; Grattan, K. Fibre bragg grating sensors for reinforcement corrosion monitoring in civil engineering structures. In Proceedings of the Journal of Physics: Conference Series, Liverpool, UK, 11–13 September 2007. Zheng, Z.; Lei, Y.; Sun, X. Measuring corrosion of steels in concrete via fiber bragg grating sensors—Lab experimental test and in-field application. In Proceedings of the Earth and Space Conference, Honolulu, HI, USA, 14–17 March 2010; pp. 2422–2430. Zhao, X.; Gong, P.; Qiao, G.; Lu, J.; Lv, X.; Ou, J. Brillouin corrosion expansion sensors for steel reinforced concrete structures using a fiber optic coil winding method. Sensors 2011, 11, 10798–10819. [CrossRef] [PubMed] Mao, J.; Chen, J.; Cui, L.; Jin, W.; Xu, C.; He, Y. Monitoring the corrosion process of reinforced concrete using botda and fbg sensors. Sensors 2015, 15, 8866–8883. [CrossRef] [PubMed] McCague, C.; Fabian, M.; Karimi, M.; Bravo, M.; Jaroszewicz, L.R.; Mergo, P.; Sun, T.; Grattan, K.T. Novel sensor design using photonic crystal fibres for monitoring the onset of corrosion in reinforced concrete structures. J. Lightwave Technol. 2014, 32, 891–896. [CrossRef] Leung, C.K.; Wan, K.T.; Chen, L. A novel optical fiber sensor for steel corrosion in concrete structures. Sensors 2008, 8, 1960–1976. [CrossRef] [PubMed] Li, W.; Ho, S.C.M.; Luo, M.; Huynh, Q.; Song, G. Fiber optic macro-bend based sensor for detection of metal loss. Smart Mater. Struct. 2017, 26, 045002. [CrossRef] Zhao, Y.X.; Jin, W.L. Modeling the amount of steel corrosion at the cracking of concrete cover. Adv. Struct. Eng. 2006, 9, 687–696. [CrossRef] Melchers, R.E.; Li, C.Q. Phenomenological modeling of reinforcement corrosion in marine environments. ACI Mater. J. 2006, 103, 25–32. Mazille, H.; Rothea, R. The use of acoustic emission for the study and monitoring of localized corrosion phenomena. In Modelling Aqueous Corrosion: From Individual Pits to System Management; Springer: Dordrecht, The Netherland, 1994; Volume 266, pp. 103–127. Yoon, D.-J.; Weiss, W.J.; Shah, S.P. Assessing damage in corroded reinforced concrete using acoustic emission. J. Eng. Mech. 2000, 126, 273–283. [CrossRef] © 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).