sensors Article

Evaluation of Early-Age Concrete Compressive Strength with Ultrasonic Sensors Hyejin Yoon 1 , Young Jin Kim 1 , Hee Seok Kim 1 , Jun Won Kang 2, * and Hyun-Moo Koh 3 1

2 3

*

Structural Engineering Research Institute, Korea Institute of Civil Engineering and Building Technology, 283 Goyangdae-Ro, Ilsanseo-Gu, Goyang-Si, Gyeonggi-Do 10223, Korea; [email protected] (H.Y.); [email protected] (Y.J.K.); [email protected] (H.S.K.) Department of Civil Engineering, Hongik University, 94 Wausan-ro, Mapo-gu, Seoul 04066, Korea Department of Civil and Environmental Engineering, Seoul National University, 1 Gwanak-ro, Gwanak-gu, Seoul 08826, Korea; [email protected] Correspondence: [email protected]; Tel.: +82-2-320-1601

Received: 17 June 2017; Accepted: 2 August 2017; Published: 7 August 2017

Abstract: Surface wave velocity measurement of concrete using ultrasonic sensors requires testing on only one side of a member. Thus, it is applicable to concrete cast inside a form and is often used to detect flaws and evaluate the compressive strength of hardened concrete. Predicting the in situ concrete strength at a very early stage inside the form helps with determining the appropriate form removal time and reducing construction time and costs. In this paper, the feasibility of using surface wave velocities to predict the strength of in situ concrete inside the form at a very early stage was evaluated. Ultrasonic sensors were used to measure a series of surface waves for concrete inside a form in the first 24 h after placement. A continuous wavelet transform was used to compute the travel time of the propagating surface waves. The cylindrical compressive strength and penetration resistance tests were also performed during the test period. Four mixtures and five curing temperatures were used for the specimens. The surface wave velocity was confirmed to be applicable to estimating the concrete strength at a very early age in wall-like elements. An empirical formula is proposed for evaluating the early-age compressive strength of concrete considering the 95% prediction intervals. Keywords: ultrasonic sensors; surface wave velocity; early-age concrete; compressive strength

1. Introduction The setting and hardening of concrete are important processes during construction work and influence the form removal time. Recent advances in design and construction technology have expanded the construction market to super-high-rise buildings and long-span bridges, which require the placement of huge quantities of concrete within the shortest time possible. Accordingly, appropriate timing for form removal is important to reduce the construction period and costs. Concrete specifications in Korea prescribe removing the form when the concrete compressive strength reaches 5–8 MPa. For example, the Standard Concrete Specification [1], Standard Specification for Temporary Works [2], and Expressway Construction Guide Specification [3] each specify a form removal concrete strength of 5 MPa. The Manual of Concrete Practice [4] specifies a value of 8 MPa for high-strength concrete with a design strength of 40 MPa or more. On the other hand, ACI 318-14 [5], ACI 347-14 [6], and EM 1110-1-2009 [7] all state that the designer should directly decide the removal time of forms considering that “the concrete exposed by form removal shall have sufficient strength not to be damaged by deflection or twisting during removal operation.” To monitor the early-age strength of concrete, various nondestructive evaluation methods such as active sensing methods using embedded piezoelectric transducers [8–10] and electro-mechanical impedance methods [11–13] have been investigated. Ultrasonic wave velocity methods are widely Sensors 2017, 17, 1817; doi:10.3390/s17081817

www.mdpi.com/journal/sensors

Sensors 2017, 17, 1817 Sensors 2017, 17, 1817

2 of 15 2 of 15

impedance methods [11–13] have been investigated. Ultrasonic wave velocity methods are widely used to monitor the solidification of concrete because the material strength influences the used to monitor the solidification of concrete because the material strength influences the propagation propagation velocity of such waves [14]. The P-wave velocity measurement is most commonly used velocity of such waves [14]. The P-wave velocity measurement is most commonly used to evaluate the to evaluate the in situ concrete strength in a structure. This technique is prescribed in ASTM C597-16 in situ concrete strength in a structure. This technique is prescribed in ASTM C597-16 [15], provided [15], provided that opposite sides of the structures are accessible. However, Figure 1a shows the that opposite sides of the structures are accessible. However, Figure 1a shows the practical limitations practical limitations of measuring the P-waves in concrete elements. The disturbance to the passing of measuring the P-waves in concrete elements. The disturbance to the passing body waves by the body waves by the aggregate and reinforcing elements in the concrete structure make it difficult to aggregate and reinforcing elements in the concrete structure make it difficult to access opposite surfaces access opposite surfaces of some concrete structures, such as pylons and towers. Figure 1b shows the of some concrete structures, such as pylons and towers. Figure 1b shows the propagation of a surface propagation of a surface wave in concrete. Surface wave-based methods have been investigated wave in concrete. Surface wave-based methods have been investigated because they only need access because they only need access to one side [16,17]. Research on the relationship between the surface to one side [16,17]. Research on the relationship between the surface wave velocity and concrete wave velocity and concrete strength has focused on hardened concrete [18–21]. This study strength has focused on hardened concrete [18–21]. This study investigates the feasibility of using investigates the feasibility of using surface wave velocities to predict the strength of in situ concrete surface wave velocities to predict the strength of in situ concrete inside the form at a very early stage inside the form at a very early stage and the correlation between the surface wave velocity and and the correlation between the surface wave velocity and concrete strength. concrete strength.

(a)

(b)

Figure 1. Schematic description of wave propagation in concrete: (a) body-propagating wave and (b) Figure 1. Schematic description of wave propagation in concrete: (a) body-propagating wave and surface-propagating wave. (b) surface-propagating wave.

2. Finite Element (FE) Simulation of Surface Wave Propagation 2. Finite Element (FE) Simulation of Surface Wave Propagation 2.1. Objective Objective 2.1. Stress waves when pressure is induced by ultrasonic sensorssensors attached to a concrete Stress wavesare aregenerated generated when pressure is induced by ultrasonic attached to a surface. Rayleigh waves have been adopted by many researchers for nondestructive testing concrete surface. Rayleigh waves have been adopted by many researchers for nondestructive testing purposes [18–21]. [18–21]. When When generated generated in in concrete, concrete, Rayleigh Rayleigh waves waves propagate propagate along along the the free free surface surface of of purposes the concrete. In this study, surface waves that propagate along the interface between concrete and the concrete. In this study, surface waves that propagate along the interface between concrete and formwork are are needed needed to to determine determine the the in in situ situ concrete concrete strength strength at at an an early early curing curing stage. stage. Ultrasonic Ultrasonic formwork sensors were installed on the side of the concrete wall with an acrylic mounting panel in the sensors were installed on the side of the concrete wall with an acrylic mounting panel in the formwork formwork to generate and record the surface waves. Figure 2 shows the wall-side installation of the to generate and record the surface waves. Figure 2 shows the wall-side installation of the ultrasonic ultrasonicThe module. The mounting was to fix theofposition of theinsensors contact with the module. mounting panel waspanel used to fixused the position the sensors contactinwith the concrete concrete from the plastic state. Surface waves were generated from one of the ultrasonic sensors and from the plastic state. Surface waves were generated from one of the ultrasonic sensors and traveled traveled along the interface between the concrete and mounting panel. The waves are, in fact, along the interface between the concrete and mounting panel. The waves are, in fact, Stoneley waves Stoneleyalong waves along the interface between two solids, which differwaves from traveling Rayleigh near waves guided theguided interface between two solids, which differ from Rayleigh a traveling near a solid surface. Accordingly, the surface wave velocity with the mounting panel should solid surface. Accordingly, the surface wave velocity with the mounting panel should be compared be compared withwave the Rayleigh velocity. with a FE simulation. with the Rayleigh velocity.wave This was doneThis withwas a FEdone simulation.

Sensors 2017, 17, 1817

3 of 15

Sensors 2017, 17, 1817

3 of 15

Sensors 2017, 17, 1817

3 of 15

(a)

(b)

Figure 2. Preparation of the steel mold for surface wave velocity measurement: (a) mounting the

(b) Figure 2. Preparation of(a)the steel mold for surface wave velocity measurement: (a) mounting the acrylic panel within the steel mold before the installation of the ultrasonic transducer and (b) the steel acrylic panel within the steel mold before the installation of the ultrasonic transducer and (b) the steel Figure 2. Preparation of the steelultrasonic mold fortransducer. surface wave velocity measurement: (a) mounting the mold after the installation of the moldacrylic after panel the installation of the ultrasonic transducer. within the steel mold before the installation of the ultrasonic transducer and (b) the steel mold after the installation 2.2. Finite Element Modeling of the ultrasonic transducer.

2.2. Finite Element Modeling

To compare surface wave velocity with the Rayleigh wave velocity, two types of FE models 2.2. Finite Element the Modeling were developed depending on whether wasRayleigh installed. wave Figurevelocity, 3a showstwo the types configuration of To compare the surface wave velocityacrylic with the of FE models To compare the surface wave velocity with the Rayleigh wave velocity, two types of FE models the FE model with an acrylic mounting panel, which reflects the experimental condition of this study. were developed depending on whether acrylic was installed. Figure 3a shows the configuration of were developed depending onleaving whether was installed. Figure 3a showsthe theinterface configuration of In model this case, thean surface wave theacrylic transducer was transferred through between the FE with acrylic mounting panel, which reflects the experimental condition of this study. the FE model with an acrylic mounting panel, which reflects the experimental condition of this study. thecase, concrete and acrylic Figure 3b shows the model without through acrylic, where Rayleigh waves the In this the surface wavepanel. leaving the transducer was transferred the interface between In this case, the surfacewere wavetransferred leaving thealong transducer was transferred the interface between leaving the transducer the concrete with a free through surface. ABAQUS/Explicit was concrete and acrylic panel. Figure 3b shows the model without acrylic, where Rayleigh waves leaving the concrete acrylic panel. Figure 3b isshows the model without acrylic, Rayleigh waves adopted for and the FE simulation. CAX4R a four-node, axisymmetric solidwhere element with reduced the transducer were transferred along the concrete with a free surface. ABAQUS/Explicit was adopted leaving the transducer the concrete a free surface. ABAQUS/Explicit was integration and was were used transferred to model along the concrete and with acrylic panel. CINAX4 is a four-node, for the FE simulation. CAX4R is a four-node, axisymmetric solid element with reduced integration adopted for the FE simulation. CAX4R is a four-node, axisymmetric solid element with reduced axisymmetric infinite element and was used for modeling the far edge of the concrete zone. The and was used toand model the concrete and acrylic panel. CINAX4 ispanel. atofour-node, infinite integration was used to and model concrete and acrylic CINAX4 is the a four-node, element size was set to 1.25 mm the the integration time step was set 17 µs, basedaxisymmetric on suggestion axisymmetric infinite element and was used for modeling the far edge of the concrete zone. The element and was for modeling the far edge the concrete zone. were The element size set to by Moser et al.used [22]. During the measurement, twoof ultrasonic transducers placed 100 mmwas apart size set to side 1.25 time mm and thewas integration step was setthe to 17 µs, based on suggestion 1.25 element mm andsame thewas integration step setSensors to 17time µs,were based on suggestion bythe Moser et al. [22]. on the surface of the specimen. modeled as single nodes. The vertical by Moser et al.at [22]. two spaced ultrasonic transducers were placed mm apart displacement theDuring location ofmeasurement, the other sensor, mm apart from loading position, was During the measurement, twothe ultrasonic transducers were100 placed 100 mm apart on 100 the same surface on the same surface side of the specimen. Sensors were modeled as single nodes. The vertical investigated. The nominal frequency of the ultrasonic waves in FE simulation is 50 kHz. The time side of the specimen. Sensors were modeled as single nodes. The vertical displacement at the location displacement at the location of the other sensor, spaced 100as mm from loading position, tests. was history nominal frequency of the waves areloading the same in apart surface measurement of the otherand sensor, spaced 100 mm apart from position, was wave investigated. The nominal investigated. Thethe nominal frequency of the ultrasonic waves in FE for simulation is 50 kHz. The time Figure 4 shows loading signal normalized by its peak value the simulation. The vertical frequency of the ultrasonic waves in FE simulation is 50 kHz. The time history and nominal frequency history and nominal frequency theapart waves arethe theloading same as in surface wave measurement displacement at the location 100of mm from point was investigated. The signaltests. was of the waves are thethe same as insignal surface wave measurement tests. Figure 4simulation. shows theThe loading signal Figure 4 shows loading normalized by its peak value for the vertical processed with the continuous wavelet transform to compute the travel time of the propagating normalized by its peak value for the simulation. The vertical displacement at the location 100 mm displacement surface waves.at the location 100 mm apart from the loading point was investigated. The signal was apartprocessed from the with loading was investigated. The signal was processed continuous wavelet the point continuous wavelet transform to compute the travelwith timethe of the propagating transform compute the travel time of the propagating surface waves. surfacetowaves.

(a)

(b)

Figure 3. Configuration of finite element (FE) modeling: (a) with acrylic and (b) without acrylic.

(a)

(b)

Figure 3. Configuration of finite element (FE) modeling: (a) with acrylic and (b) without acrylic.

Figure 3. Configuration of finite element (FE) modeling: (a) with acrylic and (b) without acrylic.

Sensors 2017, 17, 1817

4 of 15

Sensors 2017, 17, 1817

4 of 15

Sensors 2017, 17, 1817

4 of 15

Figure 4. Excitation signal measured in the experimental study.

Figure 4. Excitation signal measured in the experimental study. 2.3. Signal Processing

2.3. Signal Processing

The surface wave velocity is determined by using the arrival time of surface waves between two ultrasonic There areisseveral methodsby forusing determining the arrival such aswaves using the first The surface sensors. wave velocity determined the arrival time time, of surface between two peak in the time domain [19], using the cross-correlation function [23], and using the continuous Figure 4. Excitation signal measured in the experimental study. time, such as using the ultrasonic sensors. There are several methods for determining the arrival wavelet transform (CWT) [18]. The CWT allows a time-frequency analysis of a signal [24]. In contrast first peak in the time domain [19], using the cross-correlation function [23], and using the continuous to2.3. theSignal well-known Fourier transform, the wavelet-transformed signal is shown in both the frequency Processing wavelet and transform (CWT)The [18]. The transformation CWT allows aistime-frequency analysis signal [24]. In time domains. wavelet presented in Equation (1). Itof is a a very efficient toolcontrast The surface wavetransform, velocity is determined by using the arrivalsignal time ofissurface waves between two to the well-known Fourier the wavelet-transformed shown both the frequency for calculating the arrival time of Rayleigh waves because the energy waves from the in incident loading ultrasonic sensors. There are several methods for determining the arrival time, such as using the first predominantly converted into Rayleigh waves (67%) rather than P-waves S-waves (26%) tool for and timeare domains. The wavelet transformation is presented in Equation (1). (7%) It is or a very efficient peak in the time domain [19], using the cross-correlation function [23], and using the continuous [25]. The CWT has been reported to be effective for the signal processing of Rayleigh waves [26,27]. calculating the arrival time of Rayleigh because the energy analysis waves from the incident loading are wavelet transform (CWT) [18]. The waves CWT allows a time-frequency of a signal [24]. In contrast There are numerous wavelet functions, often called the mother wavelet. In this study, a Morlet predominantly converted into Rayleigh waves (67%) rather than P-waves (7%) or S-waves (26%) [25]. to the well-known Fourier transform, the wavelet-transformed signal is shown in both the frequency wavelet is adopted, which is a locally, periodic sinusoidal wave windowed by a Gaussian envelope: domains. Theto wavelet transformation presented in Equation (1). It is a very efficient tool There The CWTand hastime been reported be effective for theissignal processing of Rayleigh waves [26,27]. 1 for calculating thefunctions, arrival timeoften of Rayleigh waves because the energy waves the incident loading are numerous wavelet called the mother wavelet. In thisfrom study, a Morlet wavelet is , (1) predominantly converted into Rayleigh√waves (67%) rather than P-waves (7%) or S-waves (26%) adopted,are which is a locally, periodic sinusoidal wave windowed by a Gaussian envelope: [25]. The CWT has been reported to be effective for the signal processing of Rayleigh waves [26,27].   Z +called There are numerous wavelet functions, often the ∞ cosmother (2) 1 exp t5− b wavelet. In this study, a Morlet 2 √ b, a) =periodic g(t)ψ wave windowed dt (1) wavelet is adopted, which iswa(locally, sinusoidal by a Gaussian envelope:

a

−∞

a

Here, is the loading signal, is 1the wavelet function, a is a scale parameter, and b is a , wavelet (1)  function.  shift parameter. Equation (2) is the Morlet It is widely used in the time-frequency √ t2 analysis of elastic waves and shows in ψ(good t) = performance exp − costime (2) (5t)localization [28,29]. Figure 5 shows 2 the contour plot of the Morlet wavelet for the the measured time history of surface wave motion and cos 5 exp (2) propagating waves. 2 function, a is a scale parameter, and b is a Here, g(t) is the loading signal, ψ(t) is the wavelet

shift parameter. Equation (2)loading is the signal, Morlet wavelet It is widely used in the time-frequency Here, is the is the function. wavelet function, a is a scale parameter, and b is a parameter. Equation (2) is the Morlet wavelet function. It is localization widely used in[28,29]. the time-frequency analysis shift of elastic waves and shows good performance in time Figure 5 shows analysis of elastic waves and shows good performance in contour time localization Figure 5 showsfor the the measured time history of surface wave motion and the plot of [28,29]. the Morlet wavelet the measured time history of surface wave motion and the contour plot of the Morlet wavelet for the propagating waves. propagating waves.

(a)

(b)

Figure 5. Measured time history of surface wave motion and the contour plot of the Morlet wavelet for the propagating waves. (a) Measured signal; (b) Travel time calculation with the Morlet wavelet.

(a)

(b)

Figure 5. Measured time history of surface wave motion and the contour plot of the Morlet wavelet

Figure 5. Measured time history of surface wave motion and the contour plot of the Morlet wavelet for for the propagating waves. (a) Measured signal; (b) Travel time calculation with the Morlet wavelet. the propagating waves. (a) Measured signal; (b) Travel time calculation with the Morlet wavelet.

Sensors 2017, 17, 1817

5 of 15

2.4. Surface Wave Velocity between the Concrete and Acrylic Layer Table 1 presents wave velocities calculated from FE simulations. Case 1 shows the estimated velocities for the surface and Rayleigh waves. The elastic modulus and density of concrete are assumed to be 6.23 GPa and 2400 kg/m3 , respectively. Poisson’s ratio of concrete is assumed to be 0.3 to reflect the characteristics of early-age concrete. Rayleigh wave velocity along the free surface is estimated to be 857.29 m/s, whereas the surface wave velocity between the concrete and acrylic layer is 883.96 m/s. The difference is less than 3.2%. This indicates that the acrylic panel, used as a mounting device for the ultrasonic transducer, has a negligible effect on the velocity. Table 1. Comparison of wave velocity from FE simulation. Case Case 1: effect of acrylic layer

Case 2: effect of concrete density

Case 3: effect of Poisson’s ratio

Specification

Start Time (µs)

Arrival Time (µs)

Velocity (m/s)

Difference (%)

Rayleigh wave (without acrylic)

3.2

119.846

857.29

100

Surface wave (with acrylic)

3.2

116.327

883.96

103.1

Density (kg/m3 )

2200 2250 2300 2350 2400

3.2 3.2 3.2 3.2 3.2

114.992 116.292 117.577 118.550 119.805

894.52 884.24 874.30 866.93 857.60

100 98.9 97.7 96.9 95.9

Poisson’s ratio

0.16 0.20 0.28 0.30 0.35

3.2 3.2 3.2 3.2 3.2

116.327 117.364 119.454 119.805 121.134

883.96 879.05 875.93 872.31 860.55

100 99.4 99.1 98.7 97.4

Case 2 shows the surface wave velocities when considering differences in concrete density due to material uncertainty. The concrete density ranges from 2200 to 2400 kg/m3 . The elastic modulus and Poisson’s ratio are assumed to be 6.23 GPa and 0.3, respectively. The velocity with a density of 2200 kg/m3 is 894.52 m/s. The higher the density is, the slower the waves propagate, as shown in Table 1. For a concrete density change of 2200–2400 kg/m3 , the change in velocity is less than 5%. Case 3 shows the surface wave velocities considering differences in Poisson’s ratio due to material uncertainty. Poisson’s ratio ranges from 0.16 to 0.35. The elastic modulus and density of concrete are assumed to be 6.23 GPa and 2400 kg/m3 , respectively. The surface wave velocity with a Poisson’s ratio of 0.35 is 847.93 m/s. The lower the Poisson’s ratio is, the faster the waves propagate. For a Poisson’s ratio change of 0.16–0.35, the change in velocity is less than 5%. 3. Measurement of Surface Waves 3.1. Objective The objective of the surface wave measurement is to evaluate the feasibility of using the surface wave velocity to predict the strength of concrete inside a form at a very early stage. A series of surface wave measurements on in situ concrete were conducted within the first 24 h after pouring by using the ultrasonic sensors. The cylindrical compressive strength test and the penetration resistance test were also performed as the surface wave velocity was measured. 3.2. Experimental Setup and Materials The time-of-flight of the propagating surface waves between the concrete and mounting acrylic panel was used to derive the surface wave velocity for predicting the compressive strength of concrete

Sensors 2017, 17, 1817

6 of 15

Sensors 2017, 17, 1817

6 of 15

inside the form. Figure 6 shows the schematic setup for the surface wave measurement: a pulser and receiver (Ultracon-3030, MKC Korea, Seoul,Seoul, Korea)Korea) for generating a 600aor 1200 V signal and pulser and receiver (Ultracon-3030, MKC Korea, for generating 600 or 1200 V signal for measuring the the received signal in the range ofof 1–10 and for measuring received signal in the range 1–10MHz, MHz,aapiezoelectric piezoelectric ultrasonic ultrasonic transducer transducer pair (CT-1010, MKC Korea, Seoul, Korea) with a nominal frequency of 50 kHz, and a wireless pair (CT-1010, MKC Korea, Seoul, Korea) with a nominal frequency of 50 kHz, and a wireless data data acquisition unit unit (MKDQ-710, (MKDQ-710, MKC MKC Korea, Korea, Seoul, Seoul, Korea) Korea) to to transfer transfer the the measured measured signal signal to to the the PC PC acquisition wirelessly. Theultrasonic ultrasonicsensors sensors were installed same the specimen 100apart. mm wirelessly. The were installed on on thethe same wallwall side side of theofspecimen 100 mm apart. The pulser sent a short, high-voltage signal to a transmitter to cause it to vibrate at its resonant The pulser sent a short, high-voltage signal to a transmitter to cause it to vibrate at its resonant frequency. The The surface surface wave wave leaving leaving the the transmitter transmitter was was transferred transferred through through the the surface surface of of the the frequency. specimen and arrived at the receiver. The signal was received by another transducer attached to the specimen and arrived at the receiver. The signal was received by another transducer attached to the same surface and the velocity was calculated using the arrival time. The CWT was used to calculate same surface and the velocity was calculated using the arrival time. The CWT was used to calculate the arrival arrival time. time. the

Figure 6. Schematic of the surface wave measurement setup. Figure 6. Schematic of the surface wave measurement setup.

The mixing ratio of concrete can significantly affect the development of the concrete strength ratio of concrete can were significantly affect the development of the concrete over The time.mixing Four mixture proportions used and designated as MIX1–MIX4, given in strength Table 2. over time. mixture proportions were used and designated as MIX1–MIX4, given instrength, Table 2. Because theFour curing temperature can also significantly affect the development of concrete Because the curing temperature can also significantly affect the development of concrete strength, specimens T35, T30, T20, T15, and T05 were prepared at curing temperatures of 35, 30, 20, 15, and T35, T30, T20, temperatures T15, and T05 were prepared curing temperatures of 35, 30, 20,from 15, and 5 ◦ C, 5specimens °C, respectively. These reflect concreteatconstruction in different seasons summer respectively. reflect concrete in 39.2%. different seasons summer to to winter. TheThese watertemperatures to cement ratio (W/C) rangedconstruction from 35.4% to The sand tofrom aggregate ratio winter. The water to cement ratio (W/C) ranged from 35.4% to 39.2%. The sand to aggregate ratio (S/A) ranged from 42% to 46%. MIX1 and MIX2 were used in the construction of the main tower of (S/A) ranged from in 42% to 46%. MIX1 and MIX2 were usedcement in the construction of theuses main tower of Yi Yi Sun-shin Bridge Korea. MIX1 uses ordinary Portland (OPC) and MIX2 blast furnace Sun-shin Bridge in Korea. MIX1 uses ordinary Portland cement (OPC) and MIX2 uses blast furnace slag cement (SC). MIX3 contains fly ash (FA). It was used in the construction of the main tower of slag cement (SC). MIX3 contains fly ash (FA). It was compressive used in the construction of the main tower Seohae Bridge in Korea. MIX1–MIX3 have a design strength of 40 MPa. MIX4 hasof a Seohae strength Bridge inofKorea. have a design compressive strength of 40 MPa. MIX4(HPC). has a design 80 MPaMIX1–MIX3 to reflect the recent developments of high performance concrete design2 strength MPa to reflect recent developments of high performance concrete (HPC). Table presents of the80details of the mixthe proportions. Table 2 presents the details of the mix proportions.

Sensors 2017, 17, 1817

7 of 15

Table 2. Mixture proportions per cubic meter for the surface wave velocity measurement, cylindrical compressive strength test, and penetration resistance test. Type

Index

W/C (%)

Unit Weight (kg/m3 )

S/A (%) Fly Ash

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

OPC

MIX1-T35 MIX1-T30 MIX1-T20 MIX1-T15 MIX1-T05

SC

MIX2-T35 MIX2-T30 MIX2-T20 MIX2-T15 MIX2-T05

FA

MIX3-T35 MIX3-T30 MIX3-T20 MIX3-T15 MIX3-T05

HPC

MIX4-T35 MIX4-T30 MIX4-T20 MIX4-T15 MIX4-T05

35.4

35.4

38

39.2

46

46

42

45

-

-

104

-

Super-Plasticizer

4.75

4.75

7

4.2

Cement

475

475

419

420

Sand

760

760

672

674.3

Curing Temperature (◦ C)

Design Strength (MPa)

935

35 30 20 15 5

40

935

35 30 20 15 5

40

932

35 30 20 15 5

40

827.9

35 30 20 15 5

80

Gravel

Sensors 2017, 17, 1817

8 of 15

3.3. Experimental Procedure Sensors 2017, 17, 1817

8 of 15

Twenty concrete specimens were prepared to cover the various mixture proportions and curing 3.3. Experimental Procedure temperatures for the surface wave velocity measurement. Freshly mixed concrete was placed in a Twenty concreteformwork specimens were to cover the various and steel mold to simulate withprepared dimensions of 400 mm × mixture 300 mmproportions × 300 mm ascuring shown in temperatures for the surface wave velocity measurement. Freshly mixed concrete was placed in a Figure 7a. The surface wave measurements were conducted on the side of a wall. Two ultrasonic steel mold to simulate formwork with dimensions of 400 mm × 300 mm × 300 mm as shown in Figure sensors were placed 100 mm apart. To fix the sensor positions, an acrylic plate with a thickness of 7a. The surface wave measurements were conducted on the side of a wall. Two ultrasonic sensors 1 mm wasplaced mounted onapart. the steel mold. The two ultrasonic transducers then of inserted into an were 100 mm To fix the sensor positions, an acrylic plate with were a thickness 1 mm was acrylic module. After placement, the specimens were kept under isothermal conditions as shown mounted on the steel mold. The two ultrasonic transducers were then inserted into an acrylic module. in Figure 7b placement, for 24 h, during which time experiments were conditions conducted. After the specimens werethe kept under isothermal as shown in Figure 7b for 24 Along with thetime surface wave velocity the penetration resistance and cylindrical h, during which the experiments weremeasurement, conducted. Along with thetests surface wave velocity measurement, thewith penetration resistance andKS cylindrical compressive strength were conducted in accordance KS F 2436 [30] and F 2403 [31], compressive strength tests were conducted in accordance with KS F 2436 [30] and KS F 2403 [31], was respectively. Figure 7c,d shows the two tests, respectively. The penetration resistance test respectively. Figure 7c,d shows the two tests, respectively. The penetration resistance test performed to determine the setting time of concrete and to investigate the validity of using thewas surface theconcrete. setting time of concrete and to investigate the validitytest of using surface to waveperformed velocity to fordetermine early-age The cylindrical compressive strength wasthe conducted wave velocity for early-age concrete. The cylindrical compressive strength test was conducted to estimate strength of in situ concrete. These two tests were performed at intervals of 30 min to 1 h estimate strength of in situ concrete. These two tests were performed at intervals of 30 min to 1 h while the surface wave velocity was measured. Each test was repeated three times and the results while the surface wave velocity was measured. Each test was repeated three times and the results werewere averaged. averaged.

(a)

(b)

(c)

(d)

Figure 7. Experiments for in situ concrete at an early age: (a) specimen for the surface wave velocity

Figure 7. Experiments for in situ concrete at an early age: (a) specimen for the surface wave velocity measurement, (b) specimen in an isothermal control machine, (c) specimen for the penetration measurement, (b) specimen in an isothermal control machine, (c) specimen for the penetration resistance resistance test, and (d) specimen for the cylindrical compressive strength test. test, and (d) specimen for the cylindrical compressive strength test.

4. Experimental Results

4. Experimental Results 4.1. Surface Wave Velocity for Monitoring Early-Age Concrete

4.1. Surface Wave Velocity for Monitoring Early-Age Concrete

Figure 8 shows the development of the surface wave velocity with the curing age. To investigate the feasibility of using the surface wave velocity towave monitor early-age concrete, the initial and final Figure 8 shows the development of the surface velocity with the curing age. To investigate setting times were superposed. According to KS F 2436 [30], the initial and final settings of concrete the feasibility of using the surface wave velocity to monitor early-age concrete, the initial and final are times determined when the penetration resistance a sieved sample reaches 3.5 and 28 setting were superposed. According to KS of F 2436 [30],mortar the initial and final settings ofMPa, concrete

are determined when the penetration resistance of a sieved mortar sample reaches 3.5 and 28 MPa,

Sensors 2017, 17, 1817

9 of 15

Sensors 2017, 17, 1817

9 of 15

respectively. These penetration resistance values correspond to two particular practical points, which respectively. These penetration resistance values correspond to two particular practical points, which are loosely defined as the limit of handling and the beginning of mechanical strength development, are loosely defined as the limit of handling and the beginning of mechanical strength development, respectively. The initial and final setting times determined by the penetration resistance test for the respectively. The initial and final setting times determined by the penetration resistance test for the 20 concrete specimens of this study 3. 20 concrete specimens of this studyare arepresented presented in in Table Table 3.

(a)

(b)

(c)

(d)

Figure 8. Surface wave velocities with the initial and final setting times: (a) MIX1, (b) MIX2, (c) MIX3,

Figure 8. Surface wave velocities with the initial and final setting times: (a) MIX1, (b) MIX2, (c) MIX3, and (d) MIX4. and (d) MIX4. Table 3. Initial and final setting times determined by the penetration resistance test for 20 concrete

Tablespecimens. 3. Initial and final setting times determined by the penetration resistance test for 20 concrete specimens.

Required Time Required Time (h)(h) Initial Setting Final Setting Initial Setting Final Setting MIX1-T35 3.7 5.3 MIX1-T35 3.7 5.3 MIX1-T30 4.6 6.1 MIX1-T20 8.8 MIX1-T30 4.6 6.4 6.1 MIX1-T15 7.3 10.2 MIX1-T20 6.4 8.8 MIX1-T05 11.2 16.6 MIX1-T15 7.3 10.2 MIX2-T35 5.0 6.5 MIX1-T05 11.2 16.6 MIX2-T30 4.9 7.3 MIX2-T20 10.9 MIX2-T35 5.0 7.4 6.5 MIX2-T15 13.0 MIX2-T30 4.9 9.2 7.3 MIX2-T05 12.7 19.4 MIX2-T20 7.4 10.9 Type Type

Type Type MIX3-T35 MIX3-T35 MIX3-T30 MIX3-T20 MIX3-T30 MIX3-T15 MIX3-T20 MIX3-T05 MIX3-T15 MIX4-T35 MIX3-T05 MIX4-T30 MIX4-T20 MIX4-T35 MIX4-T15 MIX4-T30 MIX4-T05 MIX4-T20

Required TimeTime (h) (h) Required Initial Setting Final Setting Initial Setting Final Setting 3.9 5.8 4.2 3.9 6.3 5.8 5.0 4.2 7.8 6.3 6.6 5.0 10.2 7.8 8.1 14.7 6.6 10.2 3.6 6.9 4.0 8.1 7.2 14.7 5.4 3.6 9.0 6.9 4.8 4.0 10.1 7.2 8.1 16.7 5.4 9.0

MIX2-T15 9.2 that at a very13.0 MIX4-T15 10.1 Observations showed early age there was considerable4.8 scatter and no well-defined trend in the computed velocities under 500 m/s. However, the plots in Figure 8 show an initial MIX2-T05 12.7 19.4 MIX4-T05 8.1 16.7 sharp increase in the velocity with age in the first few hours followed by asymptotic leveling after the casting. At the time of initial setting, the velocity development curves were already well-defined for Observations showedAfter that at agesetting, there was scatter and no well-defined all concrete mixtures. thea very time early of final the considerable velocity development curves showed

trend in the computed velocities under 500 m/s. However, the plots in Figure 8 show an initial sharp increase in the velocity with age in the first few hours followed by asymptotic leveling after the casting. At the time of initial setting, the velocity development curves were already well-defined for all

Sensors 2017, 17, 1817

10 of 15

concrete mixtures. After the time of final setting, the velocity development curves showed asymptotic Sensors 2017, 17, 1817 10 of 15 leveling. Thus, the surface wave velocity appears to be suitable for monitoring the strength gain in youngasymptotic concrete. A consistent wavewave velocity can appears be obtained concrete after a few leveling. Thus,surface the surface velocity to be from suitable for monitoring thehours of casting and such measurements are sensitive to the developing strength up until the final setting strength gain in young concrete. A consistent surface wave velocity can be obtained from concrete Sensors 2017, 17, 1817 10 of 15 time. Thus, the surface wave velocity is indeed suitable for monitoring the hardening process of after a few hours of casting and such measurements are sensitive to the developing strength up until very asymptotic leveling. Thus,the thesurface surface wave velocity appears to befor suitable for monitoring the final setting time. Thus, wave velocity is indeed suitable monitoring the hardening youngthe concrete. strengthofgain youngconcrete concrete. A consistent surface wave velocity can be obtained from concrete process veryinyoung after a few hours of casting and suchStrength measurements are sensitive to the developing strength up until 4.2. Correlation between the Compressive and Surface Wave Velocity 4.2. Correlation the Compressive and Surface Wavesuitable Velocityfor monitoring the hardening the final settingbetween time. Thus, the surfaceStrength wave velocity is indeed Figure shows the measurement results for all specimens. It shows the development of the process9 of very young concrete Figure 9 shows the measurement results for all specimens. It shows the development of the concrete strength during the first 24 h and the corresponding surface wave velocities. The surface concrete strength during the first 24 h and the corresponding surface wave velocities. The surface 4.2. Correlation between the Compressive Strength as andthe Surface Wave Velocity wave velocity increased with concrete strength specimen hardened with age. During the 24 h, wave velocity increased with concrete strength as the specimen hardened with age. During the 24 h, the strength of the9ofconcrete upup to about and surface wave velocity increased to Figure shows thedeveloped measurement results for10 all specimens. shows wave the development of the the strength the concrete developed to about 10MPa MPa and the theItsurface velocity increased aroundto 2000 m/s. concrete strength during the first 24 h and the corresponding surface wave velocities. The surface around 2000 m/s. wave velocity increased with concrete strength as the specimen hardened with age. During the 24 h, the strength of the concrete developed up to about 10 MPa and the surface wave velocity increased to around 2000 m/s.

Figure 9. Relationship between the compressive strength of concrete and surface wave velocity; the

Figuremeasurement 9. Relationship the compressive strength concreteconcrete. and surface wave velocity; the data between of fresh concrete and the existing data ofofhardened measurement data of fresh concrete and the existing data of hardened concrete. To further examine the relationship betweenstrength the surface wave velocity andwave concrete strength Figure 9. Relationship between the compressive of concrete and surface velocity; the at an early age, the measurement data for hardened concrete from previous studies [18–20], as measurement data of fresh concrete and the existing of hardened To further examine the relationship between thedata surface waveconcrete. velocity and concreteshown strength at in Figure 10, are superposed in Figure 9. Shin et al. [18] considered concrete at an age of 2–28 days. an early age, the measurement data for hardened concrete from previous studies [18–20], as at shown To further examine the arelationship thedays. surface wave and concrete strength Popovics et al. [19] adopted curing age between of 24 h–28 Gallo andvelocity Popovics [20] provided surface in Figure 10, are superposed in Figure 9. Shin et al. [18] considered concrete at an age of 2–28 an early age, thefor measurement for14, hardened concrete from previous [18–20], as shown days. wave velocities concrete ageddata for 7, and 28 days. Figure 9 shows thatstudies the relationship between Popovics et al. 10, [19]are adopted aand curing agewave ofShin 24velocity h–28 Gallo and Popovics surface in Figure superposed insurface Figure 9. et al. days. [18] considered concrete at an[20] ageh provided of days. the compressive strength of early-age concrete within 24 is 2–28 consistent Popovics al. [19] adopted a curing ageand of 2428 h–28 days. Gallo9and Popovics [20] provided wave velocities for concrete aged for 7, 14, days. Figure shows that the relationship between with that et of hardened concrete obtained from previous studies using Rayleigh waves. Thissurface result wave velocities forthe concrete aged for wave 7, 14, and 28 days. 9 shows that the relationship demonstrates that use ofsurface surface waves, propagating between concrete and a mounting the compressive strength and velocity ofFigure early-age concrete within 24 h between isacrylic consistent the compressive strength and surface wave velocity of early-age concrete within 24 h is consistent panel, is feasible for the evaluation of the compressive strength of early-age concrete. with that of hardened concrete obtained from previous studies using Rayleigh waves. This result with that of hardened concrete obtained from previous studies using Rayleigh waves. This result demonstrates that the use of surface waves, propagating between concrete and a mounting acrylic demonstrates that the use of surface waves, propagating between concrete and a mounting acrylic panel, is feasible for the evaluation of the compressive strength of early-age concrete. panel, is feasible for the evaluation of the compressive strength of early-age concrete.

Figure 10. Existing data for hardened concrete.

Figure 10. Existing data for hardened concrete.

Figure 10. Existing data for hardened concrete.

Sensors 2017, 17, 1817

11 of 15

Sensors 2017, 17, 1817

11 of 15

To and surface surface wave wave velocity, velocity, To find find the the relationship relationship between between the the early-age early-age concrete concrete strength strength and theoretical and empirical formulas were investigated. For the theoretical approach, the modulus theoretical and empirical formulas were investigated. For the theoretical approach, the modulus of of elasticity E was set as proportional to the square of the pulse velocity V . According to ACI 318-14 [5] R elasticity E was set as proportional to the square of the pulse velocity . According to ACI 318-14 [5] and and KHBDC KHBDC (Korea (Korea Highway Highway Bridge Bridge Design Design Code, Code, Limit Limit State State Design) Design) [32], [32], compressive compressive strength strength is is proportional to the square of the modulus of elasticity and third power of the modulus of elasticity, proportional to the square of the modulus of elasticity and third power of the modulus of elasticity, respectively. respectively. Accordingly, Accordingly, the the compressive compressive strength strength was was represented represented as as aa power power function function with with the the surface wave velocity V as given in Equation (3), which was adopted by early researchers [33,34]. surface wave velocity R as given in Equation (3), which was adopted by early researchers [33,34]. However, However, this this equation equation is is not not supported supported adequately adequately by by experimental experimental results results because because it it assumes assumes that that the material is homogeneous and linearly elastic, which concrete is not. An empirical formula has been the material is homogeneous and linearly elastic, which concrete is not. An empirical formula has developed by statistical meansmeans [18,35–38] and the frequently used form given in Equation (4). Here, been developed by statistical [18,35–38] and the frequently usedisform is given in Equation (4). fHere, strength (MPa), V is the surface wave velocity (km/s), and the constants a and b c is the compressive R is the compressive strength (MPa), is the surface wave velocity (km/s), and the are parameters determined by the least squares method to fit the measurement data in this study. constants a and b are parameters determined by the least squares method to fit the measurement data

in this study.

f c = a(VR )b

(3) (3) f c = aebVR (4) (4) Figure 11 shows the fitted graphs superposed with measurement data. The data were fitted 11 (3) shows the fitted superposed with measurement data.(4). TheCompressive data were fitted with with Figure Equation in Figure 11a,graphs while Figure 11b was fitted with Equation strength Equation with (3) inthe Figure 11a,wave while Figurein11b was fitted Equation (4). Compressive increased surface velocity each case. Thewith coefficient of determination (R2 )strength is 0.746 increased with the0.9083 surface velocity in eachthe case. The coefficient ) is 0.746 in Figure 11a and inwave Figure 11b. Because concrete strength of fordetermination form removal(is less than in MPa Figure 11a and 0.9083 in Figure the concrete strength forgraph form removal is less than 10 [1–7], the measured values11b. needBecause to be compared with the fitted in the low-strength 10 MPaFigure [1–7], 12 theshows measured values needwith to be compared with graph axis. in theThe low-strength region. the fitted graph a logarithmic scalethe forfitted the vertical measured region. Figure 12 shows the fitted graph with a logarithmic scale for the vertical axis. The measured data fit well with the empirical formula. Therefore, the following empirical formula is proposed for datarelationship fit well with the empirical formula. Therefore, the following empirical formula is proposed for the between very early compressive strength and surface wave velocity: the relationship between very early compressive strength and surface wave velocity: R f c = 0.0098e3.412V (5) (5) 0.0098 . Equation is determined calculating coefficients andbbof of the the exponential function Equation (5) is(5)determined by by calculating thethe coefficients a aand function presented presented in in Equation Equation (4) (4) using using aa nonlinear nonlinear regression regressionmethod. method.

(a)

(b)

Figure 11. 11. Compressive velocity: experimental data fitted by by (a) Figure Compressive strength strengthof ofconcrete concretevs. vs.surface surfacewave wave velocity: experimental data fitted thethe power function of Equation (3)(3) with a =a0.2813 and b =b5.3814 and (b)(b) thethe exponential function of (a) power function of Equation with = 0.2813 and = 5.3814 and exponential function Equation (4) with a = 0.0098 and b = 3.412. of Equation (4) with a = 0.0098 and b = 3.412.

Sensors 2017, 17, 1817 Sensors 2017, 17, 1817

12 of 15 12 of 15

Sensors 2017, 17, 1817

12 of 15

(a)

(b)

Figure Figure 12. 12. Compressive Compressive strength strength of of concrete concrete vs. vs. surface surface wave wave velocity velocity on on aa logarithmic logarithmic scale scale for for the the vertical axis: experimental data fitted by (a) the power function of Equation (3) with a = 0.2813 and vertical axis: experimental data fitted by (a) the power function of Equation (3) with a = 0.2813 and (a) (b) b=5.3814 0.0098 3.412. b = 5.3814and and(b) (b)the theexponential exponentialfunction functionofofEquation Equation(4) (4)with witha a= = 0.0098and andb =b = 3.412. Figure 12. Compressive strength of concrete vs. surface wave velocity on a logarithmic scale for the vertical axis: experimental fitted by (a)Compressive the power function of Equation (3)the with a = 0.2813 andVelocity 4.3. Proposed Empirical Formula todata Evaluate the Strength by Using Surface Wave 4.3. Proposed Empirical Formula to Evaluate the Compressive Strength by Using the Surface Wave Velocity b=5.3814 and (b) the exponential function of Equation (4) with a = 0.0098 and b = 3.412. Figure 13 shows the data of compressive strength and surface wave velocity from MIX1 with the Figure 13 shows the data of compressive strength and surface wave velocity from MIX1 with proposed formula of Equation (5).toIn general, the recorded compressive was higher than the 4.3. Proposed Empirical Evaluate Compressive Strength by Usingstrength the Surface Wave Velocity the proposed formula ofFormula Equation (5). Inthe general, the recorded compressive strength was higher value calculated by the formula at 30 and 35 °C and lower at 5 °C. This indicates that the curing ◦ Csurface shows theby data compressive waveat velocity from indicates MIX1 withthat the the than the Figure value 13 calculated theofformula at 30strength and 35and and lower 5 ◦ C. This temperature has an effect on the relationship between the early-age compressive strength and surface proposed formula of Equation (5).on In the general, the recorded compressive strength was higher than the curing temperature has an effect relationship between the early-age compressive strength wavevalue velocity. To consider the influence of the curing temperature on the proposed formula, the calculated by the formula at 30 and 35 °C and lower at 5 °C. This indicates that the curing and surface wave velocity. To consider the influence of the curing temperature on the proposed equation was modified to on include the coefficient the of early-age temperature, , as strength in Equation (6). The temperature has anwas effect the relationship compressive and surface formula, the equation modified to includebetween the coefficient of temperature, k, as in Equation (6). The measurement data To forconsider each mixture were divided into three subgroups. A consisted wave velocity. the influence of the curing temperature on theGroup proposed formula, of thedata measurement data for each mixture were divided into three subgroups. Group A consisted of data equation was modifiedbetween to include the coefficient of temperature, , as in Equation The measured at temperatures 30 and 35◦°C, to reflect summer construction. Group (6). B consisted measured at temperatures between 30 and 35 C, to reflect summer construction. Group B consisted of measurement each mixturebetween were divided into 20 three A consisted of data of data measureddata at for temperatures 15 and °Csubgroups. and wasGroup referred to as room-cured datameasured measuredatattemperatures temperatures between 15 and 20to◦ C and summer was referred to as room-cured specimens. between 30 and 35 °C, reflect construction. Group B consisted specimens. Group C consisted of at 5 °C to reflect winter construction. ◦ Cdata Group C consisted of data at 5 to reflect winter construction. of data measured at temperatures between 15 and 20 °C and was referred to as room-cured . (6) specimens. Group C consisted of data at 5 °C0.0098 to reflect winter construction. 3.412VR f c = 0.0098ke (6)

0.0098

.

(6)

Figure 13. Relationship between and measurement measurement data at different curing Figure 13. Relationship betweenthe theproposed proposed formula formula and data at different curing Figure 13. Relationship between the proposed formula and measurement data at different curing temperatures (MIX1). temperatures (MIX1). temperatures (MIX1).

Figure 14 shows line bestfitfitwith with aa 95% 95% prediction group A with MIX1. Figure 14 shows thethe line ofofbest predictioninterval intervalforfor group A with MIX1. Considering the field applicability, the coefficient of temperature in Equation (6) was determined Figure 14 shows the line of best fit with a 95% prediction interval for group A with MIX1. Considering the field applicability, the coefficient of temperature in Equation (6) was determined with the lower bound limit. The coefficient of temperature for temperature groups A, B, and C for Considering thebound field applicability, the coefficient of temperature in Equation (6) was determined with the lower limit. The coefficient of temperature for temperature groups A, B, and Cwith for four mixture this study of aretemperature listed in Tablefor 4. temperature groups A, B, and C for four the lower boundproportions limit. The in coefficient

four mixture proportions in this study are listed in Table 4. mixture proportions in this study are listed in Table 4.

Sensors 2017, 17, 1817

13 of 15

Sensors 2017, 17, 1817

13 of 15

Figure 14. Fitted formula with the 95% prediction bound (MIX1). Figure 14. Fitted formula with the 95% prediction bound (MIX1). Table 4. Temperature coefficient, , for mixtures 1–4. Table 4. Temperature coefficient, k, for mixtures 1–4.

Type Type

MIX1 MIX1

Temperature (°C)

◦ Temperature 30–35( C)

MIX2

R2

Type Type

30–35 15–20 15–205 5

1.2394 1.0955 0.874 0.901 1.0955 0.52677 0.901 0.934 0.52677 0.934

MIX3

30–35 15–20 15–20 5 5

1.1981 0.83825 0.939 0.95 0.83825 0.95 0.40681 0.961 0.961 0.40681

MIX4

30–35

MIX2

K

2 K1.2394 R0.874

1.1981

MIX3

0.939 MIX4

Temperature (°C) ◦ Temperature 30–35 ( C) 30–35 15–20 15–20 5 5 30–35 30–35 15–20 15–20 55

K K 2.4437 2.4437 1.1207 1.1207 0.64678 0.64678 2.3002 2.3002 1.0433 1.0433 0.41227 0.41227

R2 R2 0.925 0.925 0.567 0.567 0.982 0.982 0.846 0.846 0.943 0.943 0.975 0.975

5. Conclusions 5. Conclusions In this study, the surface wave velocity measured with ultrasonic sensors was verified to be In this study, the surface wave velocity measured with ultrasonic sensors was verified to be applicable to the monitoring of the solidification of concrete inside the form and an empirical formula applicable to the monitoring of the solidification of concrete inside the form and an empirical formula was developed for evaluating early-age concrete strength according to the surface wave velocity. A was developed for evaluating early-age concrete strength according to the surface wave velocity. A pair pair of ultrasonic sensors was installed on the same side prior to placement. The sensors were used of ultrasonic sensors was installed on the same side prior to placement. The sensors were used to to measure the surface wave velocity of concrete. An acrylic panel was used to fix the position of the measure the surface wave velocity of concrete. An acrylic panel was used to fix the position of the sensors and to increase the applicability on structures such as walls and towers. An FE simulation sensors and to increase the applicability on structures such as walls and towers. An FE simulation was was conducted to compare the surface wave velocity with the Rayleigh wave velocity and the conducted to compare the surface wave velocity with the Rayleigh wave velocity and the difference difference between them was less than 5%, even when material uncertainties were considered. The between them was less than 5%, even when material uncertainties were considered. The surface surface wave velocity for early-age concrete inside a form were comparable with the measured wave velocity for early-age concrete inside a form were comparable with the measured Rayleigh Rayleigh wave velocity for hardened concrete from previous studies. A series of experiments wave velocity for hardened concrete from previous studies. A series of experiments including surface including surface wave velocity measurement, a cylindrical compressive strength test, and a wave velocity measurement, a cylindrical compressive strength test, and a penetration resistance test penetration resistance test were conducted on in situ concrete during the first 24 h after placement. were conducted on in situ concrete during the first 24 h after placement. The penetration resistance The penetration resistance test showed that a consistent surface wave velocity can be obtained prior test showed that a consistent surface wave velocity can be obtained prior to the initial setting and to the initial setting and the measurement was sensitive to the developing strength up until the final the measurement was sensitive to the developing strength up until the final setting time. Thus, the setting time. Thus, the surface wave velocity is suitable for monitoring the hardening process of very surface wave velocity is suitable for monitoring the hardening process of very young concrete. Based young concrete. Based on the relationship between the measured surface wave velocity and the on the relationship between the measured surface wave velocity and the corresponding cylindrical corresponding cylindrical compressive strength, an empirical exponential function was developed. compressive strength, an empirical exponential function was developed. The formula was obtained The formula was obtained using all measurement data in this study. Then, a temperature coefficient, using all measurement data in this study. Then, a temperature coefficient, to account for the effect of to account for the effect of construction in different seasons, was added to the formula considering construction in different seasons, was added to the formula considering the 95% prediction bound. the 95% prediction bound. The proposed formula can be used with the measured surface wave velocity for concrete inside a form to predict the compressive strength of concrete. The method can help with determining the appropriate form removal time for curing concrete, thereby contributing to the reduction of a

Sensors 2017, 17, 1817

14 of 15

The proposed formula can be used with the measured surface wave velocity for concrete inside a form to predict the compressive strength of concrete. The method can help with determining the appropriate form removal time for curing concrete, thereby contributing to the reduction of a construction period. The method has been developed based on the mix proportion of concrete used in this study and, therefore, further research is needed to investigate the applicability of the method for various mix proportions of concrete. Acknowledgments: This research was supported by a Grant (13SCIPA02) from the Smart Civil Infrastructure Research Program funded by the Ministry of Land, Infrastructure and Transport (MOLIT) of the Korean government and the Korea Agency for Infrastructure Technology Advancement (KAIA). Author Contributions: Hyejin Yoon conceived the main idea and prepared this article as the first author; Hyejin Yoon and Hee Seok Kim performed the experiments; Hyejin Yoon and Jun Won Kang contributed finite element simulation; Hyejin Yoon and Jun Won Kang wrote the paper. All of the work in this paper has been done under the supervision of Young Jin Kim and Hyun-Moo Koh. Conflicts of Interest: The authors declare no conflict of interest.

References 1. 2. 3. 4. 5. 6. 7. 8. 9.

10. 11. 12.

13. 14. 15. 16. 17.

Ministry of Land, Transport and Maritime Affairs. Standard Concrete Specification; Ministry of Land, Transport and Maritime Affairs: Seoul, Korea, 2009. Ministry of Land, Infrastructure and Transport. Standard Specification for Temporary Works; Ministry of Land, Infrastructure and Transport: Seoul, Korea, 2014. Korea Expressway Corporation Expressway Construction Guide Specification; Korea Expressway Corporation: Seoul, Korea, 2009. Korea Concrete Institute. Manual of Concrete Practice; Kimoondang Publishing Co.: Seoul, Korea, 2000. Building Code Requirement for Structural Concrete and Commentary; ACI Committee 318; American Concrete Institute: Farmington Hills, MI, USA, 2014. Guide to Formwork for Concrete; ACI Committee 347; American Concrete Institute: Farmington Hills, MI, USA, 2014. US Army Corps of Engineers. EM 1110-1-2009: Engineering and Design—Architectural Concrete; US Army Corps of Engineers: Washington, DC, USA, 1997. Gu, H.; Song, G.; Dhonde, H.; Mo, Y.L.; Yan, S. Concrete early-age strength monitoring using embedded piezoelectric transducers. Smart Mate. Struct. 2006, 15, 1837. [CrossRef] Dumoulin, C.; Karaiskos, G.; Carette, J.; Staquet, S.; Deraemaeker, A. Monitoring of the ultrasonic P-wave velocity in early-age concrete with embedded piezoelectric transducers. Smart Mater. Struct. 2012, 21. [CrossRef] Kong, Q.; Hou, S.; Ji, Q.; Mo, Y.L.; Song, G. Very early age concrete hydration characterization monitoring using piezoceramic based smart aggregates. Smart Mater. Struct. 2013, 22. [CrossRef] Shin, S.W.; Qureshi, A.R.; Lee, J.Y.; Yun, C.B. Piezoelectric sensor based nondestructive active monitoring of strength gain in concrete. Smart Mater. Struct. 2008, 17. [CrossRef] Shin, S.W.; Oh, T.K. Application of electro-mechanical impedance sensing technique for online monitoring of strength development in concrete using smart PZT patches. Constr. Build. Mater. 2009, 23, 1185–1188. [CrossRef] Tawie, R.; Lee, H.K. Monitoring the strength development in concrete by EMI sensing technique. Constr. Build. Mater. 2010, 24. [CrossRef] Reinhardt, H.W.; Grosse, C.U. Advanced Testing of Cement-Based Materials during Setting and Hardening. Final Report of RILEM TC 185-ATC; RILEM Publications: Paris, France, 2005. Standard Test Method for Pulse Velocity through Concrete; ASTM Standard C597; ASTM Committee C09.64; ASTM International: West Conshohocken, PA, USA, 2016. Kim, J.H.; Kwak, H. Rayleigh wave velocity computation using principal wavelet-component analysis. NDT E Int. 2011, 44, 47–56. [CrossRef] Popovics, J.S.; Abraham, O. Surface wave techniques for evaluation of concrete structures. In Non-Destructive Evaluation of Reinforced Concrete Structures; Woodhead Publishing Ltd.: Washington DC, USA, 2010; Volume 2, pp. 441–465.

Sensors 2017, 17, 1817

18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35.

36.

37. 38.

15 of 15

Shin, S.W.; Yun, C.B.; Popovics, J.S.; Kim, J.H. Improved Rayleigh wave velocity measurement for nondestructive early-age concrete monitoring. Res. Nondestr. Eval. 2007, 18, 45–68. [CrossRef] Popovics, J.; Song, W.; Achenbach, J.; Lee, J.; Andre, R. One-sided stress wave velocity measurement in concrete. J. Eng. Mech. 1998, 124, 1346–1353. [CrossRef] Gallo, G.E.; Popovics, J.S. The use of surface waves to estimate in situ concrete strength. J. Adv. Concr. Technol. 2005, 3, 355–362. [CrossRef] Lim, Y.Y.; Kwong, K.Z.; Liew, W.Y.H.; Soh, C.K. Non-destructive concrete strength evaluation using smart piezoelectric transducer—A comparative study. Smart Mater. Struct. 2016, 25. [CrossRef] Moser, F.; Jacobs, L.J.; Qu, J. Modeling elastic wave propagation in waveguides with the finite element method. NDT E Int. 1999, 32, 225–234. [CrossRef] Wu, T.-T.; Fang, J.-S.; Liu, G.-Y.; Kuo, M.-K. Determination of elastic constants of a concrete specimen using transient elastic waves. J. Acoust. Soc. Am. 1995, 98, 2142–2148. [CrossRef] Daubechies, I. The wavelet transform, time-frequency localization and signal analysis. IEEE Trans. Inf. Theory 1990, 36, 961–1005. [CrossRef] Miller, G.; Pursey, H. On the partition of energy between elastic waves in a semi-infinite solid. Proc. R. Soc. Lond. Ser. A Math. Phys. Sci. 1955, 233, 55–69. [CrossRef] Kishimoto, K.; Inoue, H.; Hamada, M.; Shibuya, T. Time frequency analysis of dispersive waves by means of wavelet transform. J. Appl. Mech. 1995, 62, 841–846. [CrossRef] Inoue, H.; Kishimoto, K.; Shibuya, T. Experimental wavelet analysis of flexural waves in beams. Exp. Mech. 1996, 36, 212–217. [CrossRef] Kim, Y.Y.; Kim, E.H. A new damage detection method based on a wavelet transform. In Proceedings of the 18th IMAC, San Antonio, TX, USA, 7–10 February 2000. McGetrick, P.J.; Kim, C.W. A Parametric Study of a Drive by Bridge Inspection System Based on the Morlet Wavelet. Key Eng. Mater. 2013, 569–570, 262–269. [CrossRef] Testing Method for Time of Setting of Concrete Mixture by Penetration Resistance; Korean Standards KS F 2436; Korean Agency for Technology and Standards (KATS): Seoul, Korea, 2007. Korean Standards KS F 2403. Standard Test Method for Making and Curing Concrete Specimens; Korean Agency for Technology and Standards (KATS): Seoul, Korea, 2014. Ministry of Land, Transport and Maritime Affairs. Korean Highway Bridge. Design Code (Limit State Design); Korea Road and Transportation Association: Seoul, Korea, 2015. Kheder, G.F. A two stage procedure for assessment of in situ concrete strength using combined non-destructive testing. Mater. Struct. 1999, 32, 410–417. [CrossRef] Pessiki, S.P.; Carino, N.J. Setting time and concrete strength using the impact-echo method. ACI Mater. J. 1989, 85, 389–399. Turgut, P. Evaluation of the ultrasonic pulse velocity data coming on the field. In Proceedings of the 4th International Conference on NDE in Relation to Structural Integrity for Nuclear and Pressurized Components, London, UK, 6–8 December 2004. Nash’t, I.H.; Saeed, H.A.; Sadoon, A.A. Finding a unified relationship between crushing concrete strength and non-destructive tests. In Proceedings of the 3rd MENDT—Middle East Nondestructive Testing Conference & Exhibition, Manama, Bahrain, 27–30 November 2005. Trtnik, G.; Kavˇciˇc, F.; Turk, G. Prediction of concrete strength using ultrasonic pulse velocity and artificial neural networks. Ultrasonics 2009, 49, 53–60. [CrossRef] [PubMed] Lee, H.K.; Yim, H.; Lee, K.M. Velocity-strength relationship of concrete by impact-echo method. Mater. J. 2003, 100, 49–54. © 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/).