sensors Article

A Monitoring Method Based on FBG for Concrete Corrosion Cracking Jianghong Mao 1 , Fangyuan Xu 1, *, Qian Gao 1,2 , Shenglin Liu 1,2 , Weiliang Jin 1 and Yidong Xu 1 1 2

*

Ningbo Institute of Technology, Zhejiang University, Ningbo 315100, China; [email protected] (J.M.); [email protected] (Q.G.); [email protected] (S.L.); [email protected] (W.J.); [email protected] (Y.X.) College of Civil Engineering, Chongqing Jiaotong University, Chongqing 400074, China Correspondence: [email protected]; Tel.: +86-574-8822-9512

Academic Editor: Gangbing Song Received: 11 February 2016; Accepted: 1 July 2016; Published: 14 July 2016

Abstract: Corrosion cracking of reinforced concrete caused by chloride salt is one of the main determinants of structure durability. Monitoring the entire process of concrete corrosion cracking is critical for assessing the remaining life of the structure and determining if maintenance is needed. Fiber Bragg Grating (FBG) sensing technology is extensively developed in photoelectric monitoring technology and has been used on many projects. FBG can detect the quasi-distribution of strain and temperature under corrosive environments, and thus it is suitable for monitoring reinforced concrete cracking. According to the mechanical principle that corrosion expansion is responsible for the reinforced concrete cracking, a package design of reinforced concrete cracking sensors based on FBG was proposed and investigated in this study. The corresponding relationship between the grating wavelength and strain was calibrated by an equal strength beam test. The effectiveness of the proposed method was verified by an electrically accelerated corrosion experiment. The fiber grating sensing technology was able to track the corrosion expansion and corrosion cracking in real time and provided data to inform decision-making for the maintenance and management of the engineering structure. Keywords: Fiber Bragg Grating; reinforced concrete; structural health monitoring; reinforcement corrosion; corrosion cracking

1. Introduction One of the main causes of structure durability failure is reinforcement corrosion. The maintenance of corrosive concrete structures costs an estimated $100 billion annually around the world [1]. The occurrence and propagation of reinforcement corrosion deteriorate the usability and safety of reinforced concrete structures. For example, if a bridge was used in a chloride environment for a long period of time, the volume expansion caused by the corrosion of reinforcement could lead to concrete cover cracks that expose the reinforcement to the corrosive environment. Therefore, monitoring reinforcement corrosion is important for guaranteeing the safety of reinforced concrete structures. At present, commercial sensors based on electrochemical principles, such as the anode ladder sensor [2] and electrode array sensor [3], can also be used to detect reinforcement rusting. In practice, resistance strain gauge is vulnerable to high humidity while the vibrating wires strain gauge is not small enough to embed it into concrete cover. For these reasons, the above sensors cannot discern the extent of the reinforcement corrosion and its influence on the concrete. Optical fiber sensing technology has many advantages, including strong implantation, anti-electromagnetic interference, and good durability. Optical fiber sensing is widely used for monitoring reinforcement corrosion by detecting chloride concentrations [4], pH values [5], and light intensity attenuation [6]. Abu [7] developed an optical fiber corrosion sensor that detects corrosion based on the Bragg reflected wavelength Sensors 2016, 16, 1093; doi:10.3390/s16071093

www.mdpi.com/journal/sensors

Sensors 2016, 16, 1093

2 of 11

spectra; via an etched-cladding fiber Bragg grating, it can detect the production of corrosion waste. Zheng [8], using FBG sensors packaged with fiber-reinforced plastics (FRP) to wrap around the steel Sensors 2016, 16,accelerating 1093 2 of 11 was bar, performed corrosion tests in lab experiments to show that the corrosion sensor feasible for monitoring the early corrosion of rebar in concrete. developed an optical fiber corrosion sensor that detects corrosion based on the Bragg reflected Although these sensors can detect information about grating, the corrosive the detection or wavelength spectra; via an etched-cladding fiber Bragg it can environment, detect the production of monitoring ofwaste. concrete strain caused bysensors reinforcement has not been achieved. Brillouin corrosion Zheng [8], using FBG packagedcorrosion with fiber-reinforced plastics (FRP) The to wrap Optical Time Domain Analysis (BOTDA) [9,10] is a new structural health monitoring technology around the steel bar, performed accelerating corrosion tests in lab experiments to show that the and can perform spatial synchronous a complex environment. In recent corrosionmulti-dimensional sensor was feasible for monitoring the earlymonitoring corrosion of in rebar in concrete. Although sensors can detect the corrosive environment, theThe detection years, BOTDA has these also been applied to theinformation monitoringabout of reinforced concrete durability. researchers or monitoring of concrete strain caused by monitoring reinforcementsystem corrosion been achieved. The designed the distributed optical fiber sensor for has the not entire process of reinforced Brillouin Optical Time Domain Analysis (BOTDA) [9,10] is a new structural health monitoring concrete corrosion [11] and established the corresponding relationship between the optical fiber strain technology and can perform multi-dimensional spatial synchronous monitoring in a complex and the concrete expansion and presence of cracks. However, BOTDA demodulation is costly, limiting environment. In recent years, BOTDA has also been applied to the monitoring of reinforced concrete its wide use as a durability monitoring sensor. durability. The researchers designed the distributed optical fiber sensor monitoring system for the Compared Fiber Braggcorrosion Grating [11] (FBG) technology has the advantages of high entire processtoofBOTDA, reinforced concrete andsensing established the corresponding relationship precision and low cost. Therefore, it can be used as an effective method for the monitoring of reinforced between the optical fiber strain and the concrete expansion and presence of cracks. However, BOTDA concrete corrosioniscracking. Lee [12] and Lo a detectionsensor. method to judge reinforcement demodulation costly, limiting its wide use[13] as a proposed durability monitoring Compared to was BOTDA, Fiber Bragg (FBG) sensing technology the advantages of high rusting; the criterion collected from Grating a bare FBG sensor pasted on thehas surface of the reinforcement. and low bare cost. FBG Therefore, it can betoused as an effective method for the moment monitoring of crack Theyprecision also embedded into concrete determine the corrosion cracking and reinforced concrete corrosion cracking. Lee [12] and Lo [13] proposed a detection method to judge location [11]; however, the results were only qualitative. In order to obtain quantitative data, the reinforcement rusting; the criterionand was the collected a bare sensor pasted on the surface of the coefficient between the wavelength strainfrom must firstFBG be established. reinforcement. They also embedded bare FBG into concrete to determine the corrosion cracking In this paper, packaged concrete corrosion monitoring sensors based on FBG (CCM-FBG) were moment and crack location [11]; however, the results were only qualitative. In order to obtain embedded into concrete. The coefficient between the grating wavelength and strain was calibrated quantitative data, the coefficient between the wavelength and the strain must first be established. by an equal beam test. Thecorrosion effectiveness of the proposed method was verified In thisstrength paper, packaged concrete monitoring sensors based on FBG (CCM-FBG) wereby an electrically accelerated corrosion experiment. was found that CCM-FBG was able to track the embedded into concrete. The coefficient betweenItthe grating wavelength and strain was calibrated corrosion and beam corrosion in real-time and also provided decision-making by an expansion equal strength test. cracking The effectiveness of the proposed methoddata wasfor verified by an electrically accelerated corrosion experiment.ofItthe was found that structure. CCM-FBG was able to track the regarding the maintenance and management engineering corrosion expansion and corrosion cracking in real-time and also provided data for decision-making

2. Monitoring of Reinforced Concrete regarding thePrinciple maintenance and management of theCorrosion engineeringCracking structure.Based on FBG 2.1. Mechanical Mechanism Expansion Cracking 2. Monitoring Principle of ofCorrosion Reinforced Concrete and Corrosion Cracking Based on FBG Concrete elastically deforms beforeExpansion cracking;and therefore, 2.1. Mechanical Mechanism of Corrosion Cracking elastic mechanical principles were used to explain the mechanism of corrosion-induced concrete expansion. In this paper, three main assumptions Concrete elastically deforms before cracking; therefore, elastic mechanical principles were used were made:

to explain the mechanism of corrosion-induced concrete expansion. In this paper, three main assumptions were made:Concrete behaved as an elastic material, such that non-linear deformation (1) Elastic deformation:

wasElastic not considered; (1) deformation: Concrete behaved as an elastic material, such that non-linear deformation was notcorrosion: considered;Corrosion products uniformly arranged around the steel; therefore, the (2) Uniform (2) Uniformstress corrosion: productscover uniformly arranged around the steel; therefore, the expansion actingCorrosion on the concrete was uniform; expansion stress acting on the concrete cover was uniform; (3) No leakage of corrosion products: Corrosion products did not leak out of the concrete cover; (3) No leakage of corrosion products: Corrosion products did not leak out of the concrete cover; therefore, expansion stress increased with increasing amounts of corrosion products. therefore, expansion stress increased with increasing amounts of corrosion products. TheThe mechanical model ofofreinforced corrosionisisillustrated illustrated in Figure 1 [11]: mechanical model reinforced concrete concrete corrosion in Figure 1 [11]: d' d p

d : Initial diameters ' d : New diameters after expansion p : Internal pressure induced by corrosion Figure 1. Mechanical model of reinforced concrete corrosion.

Figure 1. Mechanical model of reinforced concrete corrosion.

Sensors 2016, 16, 1093

3 of 11

Sensors 2016, 16, 1093

3 of 11

According to the theory of elasticity, the equation for the uniform internal stress in a circular ring is described in Equations (1)–(3): According to the theory of elasticity, the equation for the uniform internal stress in a circular ring a2 p d 2 (1)   2 2 ( 2  1) is described in Equations (1)–(3): d 2 a r2 a p d p ` 1q (1) σ “ '2 (2)  dd ´ad2 (1r2  ) πd1 “ πdp1 ` εq

a22 pd d2 2 d  d  d σ  2a pd 2 d( 2  1) ∆d “ dε “ d E “ (d2  a2 ) Ep 2r` 1q E pd ´ a qE r



(2)

(3) (3)

where σ is the hoop stress of concrete, p is the internal pressure, d is the external diameter of the circular ring, a is the internal diameter of the the circular circular ring, r is the distance from the analysis point to the centre, εε is is the thehoop hoopstrain strainatatthe theanalysis analysispoint; point; and E the is the elastic modulus of concrete. From and E is elastic modulus of concrete. From the the equations above, is clear that there linearrelationship relationshipbetween betweencorrosion-induced corrosion-induced expansion equations above, it isit clear that there isisa alinear (Δd) and concrete concretestrain strain(ε). (ε).Therefore, Therefore, estimate concrete expansion by monitoring hoop (∆d) and oneone cancan estimate concrete expansion by monitoring hoop strain, strain, which be recorded by CCM-FBG. which can be can recorded by CCM-FBG. 2.2. Principle Principle of of FBG FBG Sensing Sensing Technology Technology 2.2. The structure of FBG fiber grating is shown in Figure 2.

Fiber cladding Fiber core Incident light

Transmitted light

Reflected light Bragg grating Figure Figure 2. 2. Schematic Schematic of of FBG FBG sensor sensor technology. technology.

FBG is a periodic and permanent modification of the core refractive index value along the optical FBG is a periodic and permanent modification of the core refractive index value along the optical fiber axis [14]. The FBG sensor operates by measuring the changes in the reflective signal from the fiber axis [14]. The FBG sensor operates by measuring the changes in the reflective signal from the grating, and the reflective signal is influenced by the external parameters of the surrounding material. grating, and the reflective signal is influenced by the external parameters of the surrounding material. The Bragg wavelength (λB) reflected at the sensor is determined by The Bragg wavelength (λB ) reflected at the sensor is determined by B  2nB (4) λ B “ 2n B Λ (4) where nB is the reflective index and Λ is the spatial pitch. When projecting a wide spectrum light where nonto index and Λ is the spatial pitch. satisfying When projecting a wide light source optical fibers, the narrow band spectrum, Equation (4), is spectrum reflected back B is the reflective source ontoeffect the optical fibers, the narrow band spectrum, satisfying Equation (4), is reflected back due due to the of the gratings. to theThe effect of the gratings. relationship between the reflective Bragg wavelength shift (ΔλB) and changes in strain and The relationship between the(ε reflective Bragg as wavelength shift (∆λB ) and changes in strain and temperature in the grating region g) is expressed temperature in the grating region (εg ) is expressed as

B

B ∆λ B

 (1   ) g      T “ p1 ´ ρε q∆ε g ` pα ` βq ∆T

(5) (5)

λB where ρε is the effective light pressure coefficient, and α and β represent the thermal expansion and where ρε is thecoefficients, effective light pressure coefficient, and αin and β represent the thermal and thermo-optic respectively. Any changes external parameters (e.g.,expansion temperature, thermo-optic coefficients, respectively. Any changes in external parameters (e.g., temperature, pressure, pressure, etc.) alter the grating characteristics and result in a shift of the reflected Bragg wavelength. etc.) alter the grating characteristics and result in a shift of the reflected Bragg wavelength.

2.3. Design and Embedding of CCM-FBG According to the concrete corrosion cracking mechanism and the FBG sensing principle, a concrete corrosion monitoring sensor, as shown in Figure 3, was designed based on the FBG (CCM-

Sensors 2016, 16, 1093

4 of 11

2.3. Design and Embedding of CCM-FBG Sensors 2016, 16, 1093

4 of 11 According to the concrete corrosion cracking mechanism and the FBG sensing principle, a concrete corrosion monitoring sensor, as shown in Figure 3, was designed based on the FBG (CCM-FBG). FBG). The FBG sensing units were arranged along the circumference. They were directly connected The FBG sensing units were arranged along the circumference. They were directly connected with the with the transmission fiber. The FBG sensing units and transmission fiber were fixed on a platform transmission fiber. The FBG sensing units and transmission fiber were fixed on a platform composed composed of ropes. The ropes were passed through holes in the concrete formwork and then pulled of ropes. The ropes were passed through holes in the concrete formwork and then pulled taut to make taut to make the platform. the platform. z

z Steel bar Epoxy resin

y Concrete

x

sensor installation

Transmission fiber

CCM-FBG

Steel bar

CCM-FBG

CCM-FBG

y

y

x

Transmission fiber

x

Axonometric view

CCM-FBG

Veritcal view

Figure 3. Design and embedding of CCM-FBG.

Figure 3. Design and embedding of CCM-FBG.

3. Packaging and Calibration of CCM-FBG 3. Packaging and Calibration of CCM-FBG 3.1. Packaging Principle of CCM-FBG 3.1. Packaging Principle of CCM-FBG It is difficult to guarantee that the bare FBG sensor remains intact and functional after embedding into the concrete. coefficient be established unlessand the packaging It is difficultAlternatively, to guaranteethe that the barecould FBGnot sensor remains intact functional was after finished. Thus, a specific packaging process andthe calibrating testcould were performed. When embedding embedding into the concrete. Alternatively, coefficient not be established unless the CCM-FBG thefinished. concrete,Thus, the size of the sensor was a serious Thetest packaged CCM-FBG packaginginto was a specific packaging processconsideration. and calibrating were performed. needs meet the following requirements: Whentoembedding CCM-FBG into the concrete, the size of the sensor was a serious consideration. The packaged CCM-FBG needs to meet the following requirements: (1) Suitable size: In order to avoid stress concentration around the CCM-FBG, the size should (1) be Suitable size: In order to avoid stress concentration around the CCM-FBG, the size should be minimized. minimized. (2) Sufficient stiffness: In order to ensure high rate of successful embedding, the CCM-FBG should (2) be Sufficient stiffness: orderthe to impact ensure force high rate of successful embedding, the CCM-FBG should sufficiently stiff toInresist during concrete casting. be sufficiently stiff to resist the impact force during concrete casting. (3) Temperature compensation: Because the wavelength variation of FBG is also dependent on (3) Temperature compensation: Because the wavelength variation of FBG is also dependent on temperature, the coefficient between the wavelength and the temperature should be determined. temperature, the coefficient between the wavelength and the temperature should be determined. 3.2. Packaging Process of CCM-FBG 3.2. Packaging Process of CCM-FBG Based on the above principles, the size of the CCM-FBG designed as 10 mm ˆ 5 mm ˆ 3 mm Based on the above principles, the sizeand of the CCM-FBG designed 10 mm × 5are mm × 3 mm (length ˆ width ˆ thickness). The structure packaging process of theasCCM-FBG shown in (length × width × thickness). The structure and packaging process of the CCM-FBG are shown Figure 4. The Bragg grating of the fiber was set at a certain length of the casing and encapsulatedin The Bragg grating of thebreaks fiber was set bare at a certain lengthatofthe thefree casing encapsulated inFigure epoxy4.resin in order to prevent in the fiber grating end.and Fiber gratings forin epoxy resin in order to prevent breaks in the bare fiber grating at the free end. Fiber gratings forstrain strain strain and temperature embedded in the sensor were directly cast into the epoxy resin. The and temperature embedded in the sensor were directly cast into the epoxy resin. The strain fiber grating, marked CCM-FBG (S), was in contact with only a tiny amount of epoxy resin so thatfiber the grating, marked CCM-FBG (S), was in contact with only a tiny amount of epoxy resin so that the sensor accurately reflected the structural strain. The temperature fiber grating, marked CCM-FBG (T), sensor accurately reflected the structural strain. The temperature fiber grating, marked CCM-FBG was kept in a relaxed state in a thin tube so that the sensor was free from structural strain. (T), was kept in a relaxed state in a thin tube so that the sensor was free from structural strain.

Sensors 2016, 16, 1093

5 of 11

Sensors 2016, 1093 Sensors 2016, 16,16, 1093 Sensors 2016, 16, 1093

5 of511 of 11

3.03.0 3.0

1. 1. 1.25 25 25 1 1. 1. . 25 25 25

5. 5. 0 50 .0

2 2. 2. .5 5 5

5 of 11

10.0 10.0 10.0

Epoxy resin Epoxy resin Protective Epoxy resintube Protective tube Protective4. tube Figure Schematic

CCM-FBG (S)

CCM-FBG (S)

CCM-FBG (T)

CCM-FBG (T)

Embedded of Protective tube CCM-FBG (T) CCM-FBGsegment (S)

Embedded segment of Protective tube

Embedded segment of Protective tube of the structure of CCM-FBG. Figure 4. Schematic of the structure of of CCM-FBG. CCM-FBG. Figure 4. Schematic of the structure

Figure 4. Schematic of the structure of CCM-FBG.

3.3. Calibration of CCM-FBG 3.3. Calibration Calibration of of CCM-FBG CCM-FBG 3.3. 3.3. Calibration of strength CCM-FBG An equal beam test was performed in order to determine the coefficient between the An equal equaland strength beam test was in order to the coefficient between the An strength test was performed performed in order resistance to determine determine the coefficient betweentothe wavelength strain.beam The CCM-FBG and the traditional strain gauges were bonded An equal strength beam test was performed in order to determine the coefficient between the wavelength and strain. The CCM-FBG and the traditional resistance strainat gauges were bonded to wavelength strain. CCM-FBG the traditional resistance strain gauges were bonded the surfaceand of the equalThe strength beam and at the same time. Loads were added the end of the equal to wavelength and strain. The CCM-FBG and the traditional resistance strain gauges wereofbonded to thestrength surfacebeam, of the equal strength beam at the same time. Loads were added at the end the equal the surface of the5.0 equal strength beam at the same time. Loads were added at the end of the equal N each time until 50.0 N was added. CMM-FBG was connected to the SM 130 MOI the surface of the equal strength beam same time. Loads were added at thetoend of the strength beam, 5.0 each time until 50.0atof Nthe was added. CMM-FBG was connected theAsSM SM 130equal MOI strength beam, and 5.0 N N each time until 50.0 N was added. CMM-FBG to the 130 MOI interrogator, then, the wavelengths CCM-FBG were recordedwas for connected each load level. well, the strength beam, 5.0 N each time until 50.0 N was added. CMM-FBG was connected to the SM 130 MOI and then, the were recorded for load level. strain of the equal strength beam was of recorded by the resistance strain gauges. The layout of thethe interrogator, and then, the wavelengths wavelengths of CCM-FBG CCM-FBG were recorded for each each load level. As well, interrogator, then, the wavelengths of CCM-FBG were recordedstrain for each load level. As well, strain of the and equal strength recorded by the resistance gauges. The layout ofisthe the calibration test is shown in beam Figurewas 5. The connection between the sensor and equal strength beam by the resistance strain gauges. strain of the equal strength beam was recorded by the resistance strain gauges. The layout of the achieved by thin glue adhesive can resist highbetween tension force but lowand shear force. After testing, calibration testa is shown in Figurewhich 5. The connection the sensor equal strength beam is calibration test is shown in Figure 5. The connection between the sensor and equal strength beam is the sensor beglue removed by knocking its resist lateral side tension lightly. by acan resist high tension force force but low low shear shear force. force. After After testing, testing, achieved thin adhesive which can achieved a thin glue adhesive which can resist high force but low shear force. After testing, the sensorbycan be removed by knocking its lateral sidetension lightly. lightly.

the sensor can be removed by knocking its lateral side lightly.

Figure 5. Layout of the calibration test.

Figure 3.3.1. Wavelength/Strain Coefficient Figure 5. 5. Layout Layout of of the the calibration calibration test. test. Figure 5. Layout of the calibration test.

The calibration test results are shown in Figure 6. 3.3.1. Wavelength/Strain Coefficient 3.3.1. Wavelength/Strain Wavelength/StrainCoefficient Coefficient 3.3.1. 1.6are shown The calibration test results in Figure 6. CCM-FBG(S)-1 The calibration test results are shown in Figure 6. The CCM-FBG(S)-2

Wavelength variation /nm Wavelength variation /nm /nm Wavelength variation

1.4 CCM-FBG(S)-3 1.6 CCM-FBG(S)-1 Fitting for CCM-FBG(S)-1 1.6 1.2 CCM-FBG(S)-1 CCM-FBG(S)-2 Fitting for CCM-FBG(S)-2 1.4 CCM-FBG(S)-2 CCM-FBG(S)-3 Fitting for CCM-FBG(S)-3 1.0 1.4 CCM-FBG(S)-3 Fitting for CCM-FBG(S)-1 1.2 Fitting 0.8 Fitting for for CCM-FBG(S)-1 CCM-FBG(S)-2 1.2 Fitting Fitting for for CCM-FBG(S)-2 CCM-FBG(S)-3 1.0 0.6 Fitting for CCM-FBG(S)-3 1.0 0.8 0.4 0.8 0.6 0.2 0.6 0.4 0.0 0.4 0 100 200 300 400 500 600 700 800 900 0.2 Strain variation / 0.2 0.0 Figure Calibration results strain600 for CCM-FBG. 200 300 400of 500 700 800 900 0.0 0 6.100 0 100 200 300 400 500 600 700 800 900

Strain variation / variation / the wavelength variation of CCMAs shown in Figure 6, there was a linearStrain relationship between 6. Calibration results of straininfor CCM-FBG. FBG and strain variation. Figure The fittings of the data are shown Table 1. Figure Figure 6. 6. Calibration Calibration results results of of strain strain for for CCM-FBG. CCM-FBG.

As shown in Figure 6, there was a linear relationship between the wavelength variation of CCMAs shown in Figure 6, there was a linear relationship between the wavelength variation of CCMFBG and strain variation. The fittings of the data are shown in Table 1. FBG and strain variation. The fittings of the data are shown in Table 1.

Sensors 2016, 16, 1093

6 of 11

As shown in Figure 6, there was a linear relationship between the wavelength variation of CCM-FBG and strain variation. The fittings of the data are shown in Table 1. Sensors 2016, 16, 1093

6 of 11

Table 1. Simple linear regression model of the wavelength/strain coefficient.

Table 1. Simple linear regression model of the wavelength/strain coefficient. Senor Wavelength/Strain Coefficient Correlation Coefficient

Senor(S)-1 CCM-FBG CCM-FBG(S)-2 (S)-1 CCM-FBG CCM-FBG CCM-FBG(S)-3 (S)-2

Wavelength/Strain 0.0016Coefficient 0.0016 0.0007 0.0007 0.0007 0.0007

CCM-FBG (S)-3

Correlation Coefficient 0.998 0.998 0.996 0.999 0.996 0.999

The correlation coefficients between the wavelength and strain were close to 1, indicating The correlation coefficients between the wavelength and strain were close to 1, indicating that that the strain of structure can be calculated with high precision from the data of CCM-FBG. the strain of structure can be calculated with high precision from the data of CCM-FBG. The The wavelength/strain coefficient of CCM-FBG (S)-1 was twice that of CCM-FBG (S)-2 and CCM-FBG wavelength/strain coefficient of CCM-FBG (S)-1 was twice that of CCM-FBG (S)-2 and CCM-FBG (S)(S)-3. The main reason was that the mixture proportion of resin and curing agent influenced 3. The main reason was that the mixture proportion of resin and curing agent influenced the elastic the elastic modulus of the hardened epoxy resin. However, it is difficult to precisely control the modulus of the hardened epoxy resin. However, it is difficult to precisely control the mixture mixture proportion. proportion. 3.3.2. Wavelength/Temperature Coefficient 3.3.2. Wavelength/Temperature Coefficient The wavelength of FBG is also affected by temperature, and thus temperature compensation The wavelength of FBG is also affected by temperature, and thus temperature compensation should thispaper, paper,thethecalibration calibration shouldbebeconsidered consideredfor forlong-term long-term corrosion corrosion monitoring. monitoring. InInthis testtest of of wavelength/temperature coefficientof ofCCM-FBG CCM-FBGwas was conducted. CCM-FBG immersed wavelength/temperature coefficient conducted. TheThe CCM-FBG was was immersed in ˝ C water, and then the temperature was gradually decreased to 10 ˝ C. During this time, the in 80 80 °C water, and then the temperature was gradually decreased to 10 °C. During this time, the wavelength ofofCCM-FBG was recorded recordedfor forevery every1010°C˝ C reduction. These results wavelength CCM-FBG(S) (S)and andCCM-FBG CCM-FBG (T) (T) was reduction. These results areare shown ininFigure shown Figure7.7. 1.0

CCM-FBG(S)-1 CCM-FBG(S)-2 CCM-FBG(S)-3 Fitting for CCM-FBG(S)-1 Fitting for CCM-FBG(S)-2 Fitting for CCM-FBG(S)-3

2.0 1.8 1.6 1.4

Wavelength variation /nm

Wavelength variation /nm

2.2

1.2 1.0 0.8 0.6 0.4 0.2 0.0

0

10

20

30

40

50

60

70

80

Temperature variation /℃ (a)

CCM-FBG(T)-1 CCM-FBG(T)-2 CCM-FBG(T)-3 Fitting for CCM-FBG(T)-1 Fitting for CCM-FBG(T)-2 Fitting for CCM-FBG(T)-3

0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0

0

10

20

30

40

50

60

70

80

Temperature variation /℃ (b)

Figure 7. Calibration results of temperature for CCM-FBG: (a) CCM-FBG (S); (b) CCM-FBG (T). Figure 7. Calibration results of temperature for CCM-FBG: (a) CCM-FBG (S); (b) CCM-FBG (T).

There was a linear relationship between the wavelength variation of CCM-FBG and temperature There was linear relationship the wavelength variation. The adata were fitted withbetween a linear regression model,variation as shownofinCCM-FBG Table 2. and temperature variation. The data were fitted with a linear regression model, as shown in Table 2. Table 2. Simple linear regression model of the wavelength/temperature coefficient. Table 2. Simple linear regression model of the wavelength/temperature coefficient.

Sensor CCM-FBG Sensor(S)-1 CCM-FBG (S)-2 CCM-FBG (S)-1 CCM-FBG CCM-FBG (S)-3 (S)-2 CCM-FBG CCM-FBG (T)-1 (S)-3 CCM-FBG (T)-2 (T)-1 CCM-FBG CCM-FBG (T)-3 (T)-2 CCM-FBG CCM-FBG (T)-3

Wavelength/Temperature Coefficient 0.01944 Wavelength/Temperature Coefficient 0.02316 0.01944 0.02593 0.02316 0.01107 0.02593 0.01107 0.01247 0.01247 0.00972 0.00972

Correlation Coefficient 0.998 Coefficient Correlation 0.995 0.998 0.997 0.995 0.999 0.997 0.999 0.996 0.996 0.999 0.999

The correlation coefficients between the wavelength and temperature were close to 1, indicating that temperature compensation could be easily applied. According to the calibration test, the strain caused by the concrete corrosion cracking can be calculated by the following formula:

Sensors 2016, 16, 1093

7 of 11

The correlation coefficients between the wavelength and temperature were close to 1, indicating Sensors 2016, 16, 1093 compensation could be easily applied. According to the calibration test, the strain 7 of 11 that temperature caused by the concrete corrosion cracking can be calculated by the following formula:

ε “

  T   T ∆λε ´ ∆λ T αε {α T ε β

(6) (6)

thestrain strainofofCCM-FBG; CCM-FBG;∆λ Δλ thewavelength wavelengthvariation variation of of CCM-FBG CCM-FBG (S); (S); ∆λ ΔλTT is the where εisisthe ε εisisthe wavelength variation of CCM-FBG (T); α ε is the wavelength/temperature coefficient of CCM-FBG wavelength variation of CCM-FBG (T); αε wavelength/temperature coefficient of CCM-FBG (S); (S); α the wavelength/temperature wavelength/temperature coefficient and ββεε isis the the wavelength/strain wavelength/strain αTT is the coefficient of CCM-FBG (T); and coefficient of CCM-FBG (S). 4. Application Applicationof ofCCM-FBG CCM-FBGto toConcrete ConcreteCorrosion CorrosionMonitoring Monitoring 4.1. Experiment Design Design 4.1. Experiment Because in ainnatural environment occurs occurs over a long of time, accelerated Because concrete concretecorrosion corrosion a natural environment overperiod a long period of time, corrosion achieved with the use of electro chemistry is generally induced for concrete accelerated corrosion achieved with the use of electro chemistry is generally induced fordurability concrete research. evolution of reinforcement concrete corrosion cancorrosion still be observed accelerated durabilityThe research. Theprocess evolution process of reinforcement concrete can stillinbe observed corrosion and thus the method was adopted in this study. A steel bar was connected to the positive in accelerated corrosion and thus the method was adopted in this study. A steel bar was connected electrode, and the stainless steel net was connected to the negative electrode of a DC power supply. to the positive electrode, and the stainless steel net was connected to the negative electrode of a DC 2 The specimen a 5% NaCl solution. direct currentThe density set todensity 100 A/cm power supply.was Theimmersed specimenin was immersed in a 5%The NaCl solution. directwas current was to soslow that corrosion abundantso data was generated forwas analysis. 2 to ensure setensure to 100 slow A/cmcorrosion that abundant data generated for analysis. The 150 mm. mm. The Theconcrete concrete strength strength grade grade The size size of of the the concrete concrete specimen specimenwas was150 150mm mmˆ× 150 150 mm mm ˆ × 150 was C30, and the ratio of water: cement: sand: stone was 1:0.52:1.94:3.19. The concrete cover was was C30, and the ratio of water: cement: sand: stone was 1:0.52:1.94:3.19. The concrete cover was 30 30 mm [15] for the steel bar with a diameter of 18 mm. The top and bottom surfaces of the concrete mm [15] for the steel bar with a diameter of 18 mm. The top and bottom surfaces of the concrete specimen were coated The purpose purpose of of the the epoxy epoxy resin resin coating coating was was to specimen were coated with with epoxy epoxy resin. resin. The to prevent prevent permeation of chloride ions into the interface between the steel bars and concrete; otherwise, the steel permeation of chloride ions into the interface between the steel bars and concrete; otherwise, the steel bar around the interface would corrode firstly. bar around the interface would corrode firstly. The The concrete concrete corrosion corrosion cracking cracking due due to to reinforcement reinforcement rusting rusting normally normally occurs occurs at at the the bottom bottom surface or the lateral side of the RC beam, but the specific location of the cracking is random. surface or the lateral side of the RC beam, but the specific location of the cracking is random. Therefore, three FBG FBGsensors sensorswere wereapplied applied consider randomness. Three CCM-FBG Therefore, three to to consider thethe randomness. Three CCM-FBG units units were were distributed along the reinforcement, and the diameter of the ring was 40 mm. All sensors were distributed along the reinforcement, and the diameter of the ring was 40 mm. All sensors were connected to the the SM connected to SM 130 130 MOI MOI interrogator interrogator and and the the data data was was collected collected in in real-time. real-time. The The layout layout of of the the experiment inin this paper, connected to experiment is is shown shown in inFigure Figure8.8.The Thethree threeFBG FBGsensors sensorswere wereindependent independent this paper, connected different channels of the interrogator. to different channels of the interrogator.

Figure experiment. Figure 8. 8. Layout Layout of of the the accelerated accelerated corrosion corrosion experiment.

The experiment lasted for two months, and the specimen had not been moved during this whole experimental period. In order to observe the surface of the concrete specimen, the NaCl solution was extracted with a small pump regularly every day. With accelerating corrosion, the rust overflowed into the solution along the crack and made the solution turbid and difficult to pump. With the purpose of reducing any disturbance to the corrosive specimen, corrosion products were not cleaned until the end of the experiment.

the concrete corrosion cracking occurs, the corrosion product outflows from the crack and the rust will turn up at concrete surface. By observing the surface of the beam, it was found that the crack appeared around CCM-FBG (S)-3. The cracks inside concrete were detected by splitting the specimen. After splitting, it was found that the rust was located between CCM-FBG (S)-1 and CCM-FBG (S)-3 (but close (S)-3). Sensors 2016,to 16,CCM-FBG 1093 8 of 11 800

Concrete strain/

CCM-FBG(S)-1 The experiment lasted for two months, and the specimen had not been moved during this whole 700 CCM-FBG(S)-2 experimental period. In order to observe the surface of the concrete specimen, the NaCl solution was 600regularlyCCM-FBG(S)-3 extracted with a small pump every day. With accelerating corrosion, the rust overflowed into 500and made the solution turbid and difficult to pump. With the purpose of the solution along the crack reducing any disturbance 400 to the corrosive specimen, corrosion products were not cleaned until the end of the experiment.

300

200Accelerated Corrosion Experiment 4.2. Monitoring Results of the

CCM-FBG(S)-3

Concrete strain was 100 calculated using Equation (6). The concrete strains recorded by CCM-FBG Crack (S)-1, CCM-FBG (S)-2, and 0CCM-FBG (S)-3 during the entire experiment are shown in Figure 10. CCM-FBG(S)-2 Once the concrete corrosion cracking occurs, the corrosion product outflows from the crack and the -100 0 5 10 15 20 25 30 35 40 45 50 55 60 rust will turn up at concrete surface. By observing the surface of the beam, it was found that the crack Experiment time /day appeared around CCM-FBG (S)-3. The cracks inside concrete were detected by splitting the specimen. After splitting, it was found that the rust was located between CCM-FBG and CCM-FBG (S)-3 Figure 9. Concrete strain over the course of the accelerated corrosion(S)-1 experiment. (but close to CCM-FBG (S)-3). Asobserved observed in in Figure Figure 9, 10,allallthree threeCCM-FBG CCM-FBGunits unitswere wereable abletototrack trackthe theentire entirecorrosion corrosionprocess. process. As Eachdisplayed displayedthe thesame sametrend trendofofthree three stages: expansion stage, initial cracking stage, Each stages: thethe expansion stage, thethe initial cracking stage, and and the the crack propagation stage. three CCM-FBG sensors obtained thethree threeparticular particularstages stages with with crack propagation stage. All All three CCM-FBG sensors obtained the somewhat different different strain strain values. values. somewhat CCM-FBG(S)-1

(1) Expansion Expansionstage. stage. (1) Thisstage stagelasted lastedabout about nine ninedays dayswithout without cracking. cracking. The Therust rustkept keptfilling fillingthe thepore poreand andcaused causedthe the This concrete corrosion corrosionexpansion expansionforce forceto to increase increaserapidly. rapidly. The The details details of of strain strain variations variations are are shown shown in in concrete Figure 10. 9. Figure 400

355.28

Concrete strain/

350 300

239.68

247.58

250 200 150 CCM-FBG(S)-1 CCM-FBG(S)-2 CCM-FBG(S)-3

100 50 0

0

2

4

6

8

10

12

14

16

18

Experiment time /day Figure 10. Stage of electrically accelerated corrosion cracking. Figure 9. Stage of electrically accelerated corrosion cracking.

During the first three days, the concrete strains obtained by all CCM-FBG (S) were mostly During the the firstincreasing three days, the concrete strains obtained by all(S)-3 CCM-FBG (S) were higher mostly consistent. Then, concrete strain obtained by CCM-FBG was significantly consistent. Then, the increasing concrete strain obtained by CCM-FBG (S)-3 was significantly higher than those recorded by CCM-FBG (S)-1 and CCM-FBG (S)-2. Finally, the maximum strain captured by CCM-FBG (S) occurred at different time. The concrete corrosion crack also appeared near the side of CCM-FBG (S)-3. Alternatively, the corrosion expansion force inside of the concrete was not uniform. (2)

Initial cracking stage.

Once the expansion force of concrete exceeded the tensile strength of concrete, the crack of the concrete appeared at the surface. At the same time, the stress of concrete redistributed and tensile stresses were partially released. As shown in Figure 9, there was a region of decreasing concrete strain.

Sensors 2016, 16, 1093

9 of 11

Among them, the concrete strains at CCM-FBG (S)-1 and CCM-FBG (S)-2 significantly decreased, and there was no significant decrease at CCM-FBG (S)-3. The main reason was that there was a crack in the vicinity of CCM-FBG (S)-3. The crack provided a passageway for harmful ions to gather around 8steel Sensors 2016, 16, 1093 of 11 bar, and thus, the relevant high corrosion rate of the reinforcement close to CCM-FBG (S)-3 provided a lasting expansion force.of the Accelerated Corrosion Experiment 4.2. Monitoring Results

Concrete strain wasstage. calculated using Equation (6). The concrete strains recorded by CCM-FBG (3) Crack propagation (S)-1, CCM-FBG (S)-2, and CCM-FBG (S)-3 during the entire experiment are shown in Figure 9. Once Although some corrosion flowed out during stage, the by the corrosion the concrete corrosion crackingproducts occurs, the corrosion productthis outflows fromfilling the crack and the rust product thesurface. concreteBy corrosion force; increasing obviously will turnstill upincreased at concrete observing thehowever, surface ofthe the beam, it rate was was found that thelower crack than beforearound cracking, as shown in Figure 10. Therefore, the growth rate of the (S)-1, which appeared CCM-FBG (S)-3. The cracks inside concrete were detected byCCM-FBG splitting the specimen. was farsplitting, from theitcrack, was slower, and thewas growth ratesbetween of CCM-FBG (S)-2 and (S)-3 were After was found that the rust located CCM-FBG (S)-1CCM-FBG and CCM-FBG (S)-3 much higher. (but close to CCM-FBG (S)-3). 800

CCM-FBG(S)-1 CCM-FBG(S)-2 CCM-FBG(S)-3

Concrete strain/

700 600 500 400 300 200

CCM-FBG(S)-3

100

CCM-FBG(S)-1

0 -100

Crack

CCM-FBG(S)-2

0

5

10 15 20 25 30 35 40 45 50 55 60

Experiment time /day Figure10. 9. Concrete experiment. Figure Concrete strain strain over over the the course course of of the the accelerated accelerated corrosion corrosion experiment.

As observed in Figure 9, all three CCM-FBG units were able to track the entire corrosion process. 5. Discussion Each displayed the same trend of three stages: the expansion stage, the initial cracking stage, and the (1) cost of monitoring was low, CCM-FBG and the packaging wasthree simple. crackThe propagation stage. All three sensors technology obtained the particular stages with somewhat different strain values. A kind of concrete corrosion monitoring sensor had been previously developed based on (1) Expansion stage. BOTDA [11]. The BOTDA technology was able to obtain the corrosion force and the crack width of concrete. BOTDA could also identify the position of cracking when combined with non-packaged This stage lasted about nine days without cracking. The rust kept filling the pore and caused the FBG. However, it is difficult for researchers or infrastructure management to use the technology due concrete corrosion expansion force to increase rapidly. The details of strain variations are shown in to its high costs. The sensor developed in this paper is based on FBG. Therefore, this sensor had the Figure 10. advantages of lower cost and easy packaging, both of which can increase application in scientific research and real-projects. 400

Concrete strain/

(2)

355.28 350 High precision could be beneficial to study and solve scientific problems. 300

239.68

247.58

Traditional resistance strain gauges cannot be buried within internal concrete. The BOTDA 250 technology has a relevant lower precision (˘20 µε); moreover, it is short of high spatial resolution 200 (the minimum spatial resolution is 0.5 m). In this paper, the size of the CCM-FBG was 10 mmˆ 5 mm ˆ 3 mm, and 150 it could obtain the concrete corrosion force with high precision. The test CCM-FBG(S)-1 results show that the concrete corrosion strain behaved heterogeneously in all three stages, and these 100 CCM-FBG(S)-2 results asserted the mechanism of the corrosion. CCM-FBG(S)-3

50 0

0

2

4

6

8

10

12

14

16

18

Experiment time /day Figure 10. Stage of electrically accelerated corrosion cracking.

During the first three days, the concrete strains obtained by all CCM-FBG (S) were mostly

Sensors 2016, 16, 1093

(3)

10 of 11

The packaging technology needs to be improved for application in practical engineering.

The simple packaging method was an acceptable approach for short-term research. Once applied to practical engineering, some problems such as service life and material creeping would need to be addressed. Some packaging methods for other sensors based on FBG have already been developed and therefore could be applied to CCM-FBG. (4)

The application of the proposed sensor needs to overcome some challenges.

Although the packaging of the sensor can improve the strength of FBG, the transmission optical fiber cannot be armored and is still brittle; therefore we need to solve the problem of embedding the sensors with low survival. Also, the procedure of embedding the sensor is so complicated that it requires professional staff and a specialized interrogator to obtain data, which is difficult for routing inspection. 6. Conclusions and Forecast Chloride is one of the main causes of failure of reinforcement concrete structures. It is critical to evaluate the remaining service life of structures and make decisions on potential maintenance. In this paper, a CCM-FBG sensor with a size of 10 mm ˆ 5 mm ˆ 3 mm was investigated. The coefficients of wavelength/strain and wavelength/temperature were calibrated. The validity of the CCM-FBG was proved by conducting a two-month accelerated corrosion test. The CCM-FBG accurately reflected the expansion strain, and the real-time monitoring data can be beneficial for determining the mechanism of the corrosion. The technology could be used in practical engineering after further design and improved encapsulation. The sensors provided an effective, long-term, and stable detection and monitoring method of a reinforced concrete structure in a chloride environment. The collected data can inform decision-making for the maintenance and management of engineering structures. Acknowledgments: This paper was supported by the National Natural Science Foundation of China (Grant No. 51408544), Natural Science and technology support program (2015BAL02B03), the Natural Science Foundation of Zhejiang Province (Grant No. LQ14E080007, LY15E080025) the Science and Technology Project of Ningbo (2016C51024) and the Ningbo Scientific and Technological Innovation Team (Grant No. 2011B81005). Author Contributions: Among the authors, Jianghong Mao initiated the research and designed the experiments, Fangyuan Xu., Qian Gao, and Shenglin Liu performed the experiments. Jianghong Mao and Fangyuan Xu wrote the paper. Weiliang Jin and Yidong Xu derived several important mathematical equations. Conflicts of Interest: The authors declare no conflict of interest.

References 1. 2. 3.

4. 5. 6. 7.

Li, C.Q.; Melchers, R.E. Time-dependent Risk Assessment of Structural Deterioration Caused by Reinforcement Corrosion. ACI Struct. J. 2006, 102, 754–762. Gan, W.Z.; Jin, W.L.; Lu, Z.D. In-situ Monitoring and Early Warning System for Durability of Concrete Structure of Hangzhou Bay Sea-crossing Bridge. China J. Highw. Transp. 2010, 23, 31–35. Sudarshan, S.; Basheer, P.A.M.; Mao, J.H.; Jin, W.L.; McCarter, W.J. An Integrated Sensing System for Monitoring the Corrosion Activity in Concrete Blocks Exposed at Hangzhou Bay Bridge. In Proceedings of the Second Conference on Smart Monitoring, Assessment and Rehabilitation of Civil Structures, Istanbul, Turkey, 9–11 September 2013. Ghandehari, M. Ingress Monitoring in Concrete Structures. In Proceedings of the 15th ASCE Conference on Engineering Mechanics, New York, NY, USA, 2–5 June 2002. Melhorn, K.; Flachsbarth, J.; Kowalsky, W. Novel Sensors for Long-term Monitoring of pH and Humidity in Concrete. In Proceedings of the 6th IWSHM Conference, Stanford, CA, USA, 11–13 September 2007. Leung, C.K.Y.; Wan, K.T.; Chen, L.A. Novel Optical Fiber Sensor for Steel Corrosion in Concrete Structures. Sensors 2008, 8, 1960–1976. [CrossRef] Abu Hassan, M.R.; Abu Bakar, M.H.; Dambul, K.; Adikan, F.R. Optical-based sensors for monitoring corrosion of reinforcement rebar via an etched cladding Bragg grating. Sensors 2012, 12, 15820–15826. [CrossRef] [PubMed]

Sensors 2016, 16, 1093

8. 9. 10. 11. 12. 13. 14. 15.

11 of 11

Zhang, Z.P.; Sun, X.N.; Lei, Y. Monitoring corrosion of reinforcement in concrete structures via fiber Bragg grating sensors. Front. Mech. Eng. China 2009, 4, 316–319. [CrossRef] Zou, L.; Bao, X.; Ravet, F.; Chen, L. Distributed Brillouin Fiber Sensor for Detecting Pipeline Buckling in an Energy Pipe Under Internal Pressure. Appl. Opt. 2006, 45, 3372–3377. [CrossRef] [PubMed] Zhang, W.; Shi, B.; Zhang, Y.F.; Liu, J.; Zhu, Y.Q. The Strain Field Method for Structural Damage Identification Using Brillouin Optical Fiber Sensing. Smart Mater. Struct. 2007, 16, 843–850. [CrossRef] Mao, J.H.; Chen, J.Y.; Cui, L.; Jin, W.L.; Xu, C.; He, Y. Monitoring the Corrosion Process of Reinforced Concrete Using BOTDA and FBG Sensors. Sensors 2015, 15, 8866–8883. [CrossRef] [PubMed] Lee, J.; Yun, C.; Yoon, D. A Structural Corrosion–monitoring Sensor Based on a Pair of Prestrained Fiber Bragg Gratings. Meas. Sci. Technol. 2010, 21, 1–7. [CrossRef] Lo, Y.L.; Shaw, F.Y. Development of Corrosion Sensors Using a Single-pitch Bragg Grating Fiber with Temperature Compensations. SPIE 1998, 3325, 64–72. Othonos, A.; Kalli, K. Fiber Bragg Gratings: Fundamentals and Applications in Telecommunications and Sensing; Artech House: New York, NY, USA, 1999. Ministry of Housing and Urban-Rural of the People’s Republic of China. Code for Design of Concrete Structure (GB 50010-2010); China Architecture & Building Press: Beijing, China, 2010. © 2016 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/).