Steel structures of various forms are widely used in constructions [1
], however, with adverse factors such as corrosion [3
], dynamic loads [5
], and impacts [7
], steel structures are subject to damages [9
]. Carbon fiber reinforced polymer (CFRP) materials are widely used to retrofit or reinforce damaged steel structures [11
] due to their high strength-to-weight ratio, excellent corrosion resistance, and fast and easy installation. More recently, research on the application of CFRP materials for reinforcement of steel structures has received much attention [14
]. CFRP has been demonstrated as promising for strengthening steel structures [15
], especially for the bending components, which can be strengthened conveniently by bonding a CFRP sheet or plate to the tension face [16
Among the CFRP materials used for reinforcement of steel structures, CFRP sheets and CFRP plates are most commonly used. Meanwhile, compared with CFRP sheets, CFRP plates possess more advantages for strengthening damaged flexural components [17
]. Therefore, the application of CFRP plates for strengthening steel structures has attracted much attention. Deng et al. [18
] conducted an experimental and theoretical study on notched steel beams strengthened by bonding CFRP plates. Hosseini et al. [19
] introduced a prestressed unbonded reinforcement (PUR) system to strengthen existing fatigued steel members, and they also studied bond behavior and anchorage resistance of prestressed CFRP plates to steel substrate. Chen et al. [20
] investigated fatigue improvements of CFRP-plate-strengthened steel beams. Martinelli et al. [21
] investigated the bond behavior of CFRP plates epoxied to the steel substrate.
Studies show that the debonding between a steel structure and the reinforcing CFRP plate is one of the main failure modes [22
], which can affect the reinforcement effectiveness [23
] and even cause brittleness and sudden failure of strengthened structures [24
]. The debonding defects mainly occur in four areas [25
]: the CFRP plate, the steel/epoxy interface, the CFRP/epoxy interface, and the epoxy layer. The debonding damage is a typical failure in strengthening steel structures, urgently calling for an effective and nondestructive detection technique, which can help to detect damage timely so that effective measures can be taken to avoid serious consequences. Nondestructive detection techniques, such as acoustic emission technology [26
], ultrasonic inspection technology [27
], fiber optic sensing [28
] and X-ray inspection [30
], have been applied in structural damage detection. However, most of these conventional methods require complex equipment and algorithms, which may be difficult to deploy in some engineering applications. Therefore, a simple and effective nondestructive testing (NDT) method that can be applied in debonding detection between a steel structure and the reinforcing CFRP plate is necessary.
As a commonly used piezoceramic material, lead zirconate titanate (PZT) has been applied in structural health monitoring [31
] due to its strong piezoelectric effect [34
] and wide bandwidth [36
]. With both sensing and actuation functions, PZTs are often used in active sensing methods for structural health monitoring (SHM) and damage detection. For example, the active sensing method was used in characterizing concrete hydration [38
], monitoring circular reinforced concrete columns under seismic excitations [39
], and detecting damages in circular RC columns [40
]. With the PZT transducer, the electro-mechanical impedance (EMI)-based technique was applied to health-monitoring plate-like structures [41
], pin connection loosening [42
], and grout compactness of concrete-filled steel tubes [43
]. There are other applications of PZT transducers in SHM [44
As the propagation of stress wave across the bonding interface is sensitively correlated to the bonding condition, the PZT-enabled active sensing technique is also used in debonding detection and monitoring. Qin et al. [46
] studied the bond-slip detection between concrete beams and a steel plate using piezoceramic smart aggregates. Zeng et al. [47
] performed bond-slip detection between concrete and steel using shear mode smart aggregates. Xu et al. [48
] discovered that the active sensing method can monitor the bond slip between concrete structures and the glass fiber reinforced polymer (GFRP) bars accurately. Di et al. [49
] investigated the debonding process between the fiber reinforced polymer (FRP)/steel bars and the hosting structures using PZT probes and acoustic emission. Kong et al. [50
] developed an active sensing approach to monitor the cyclic crack of a reinforced concrete column in the process of simulated pseudodynamic loading.
However, few studies on debonding detection between a CFRP plate and steel beam using active sensing have been reported. In addition, the PZT-enabled active sensing methods often require the permanent installation of the transducers on or in the host structures [4
], which brings inconvenience to large-scale implementation. These factors motivate the authors to develop a new approach to detecting debonding between steel structures and the reinforcing CFRP plates. The main innovations of this paper are the development of the removable PZT transducers and the use of them in active sensing to detect the debonding damage between the steel beam and the reinforcing CFRP plate. To verify the feasibility of the proposed PZT-enabled active sensing method, two removable PZT-based transducers were fabricated, and each of them contained a PZT patch, two wires, and a strong magnet. For the testing specimen, a steel beam was bonded with a CFRP plate with epoxy for reinforcement. In this study, three different debonding areas (Area-A, Area-B, Area-C) were preset on the steel beam/CFRP plate interface of the testing specimen, and Area N was set as the healthy status without debonding. The experimental results of the debonding detection clearly show that the magnitudes of received signals decrease significantly from the case without debonding to the case with debonding, and the amount of decrease increases with the severity of the debonding damage, which demonstrates the effectiveness of the debonding detection between steel beam and reinforcing CFRP plate using the removable PZT transducers and the active sensing.
4. Experimental Results and Analysis
The results of Test 1 in the debonding detection experiment are shown in Figure 9
, where each curve represents the received signal by S2 in 0.5 s. To enable the comparison of the sensor signals without and with debonding, the received signals for the sensors in Area-N1, -N2, and -N3 are used as the baseline data and are compared to those in Area-A, -B and -C, respectively.
a shows the time domain signal received by S2 in Areas N1, A1, A2, and A3. The aptitude of received signal in Area-N1 is between −0.02 V and 0.02 V, and it is much larger than the that of received signals in Area-A1, which reveals that the signal amplitude received by S2 in a debonding area is much smaller than that in an area without a debonding, which means the stress wave propagation from CFRP plate to steel beam is sensitive to the bonding condition. The amplitudes of the time domain signals in Area-A1 and Area-A2 have similar values; therefore, it is not possible to differentiate their difference. The WPEI analysis was applied to calculate the energy of the received signal. As the black curve shows in the figure, the signal received in Area-A3 has a much smaller amplitude than those in Area-A1 and Area-A2, which means the received signal in the debonding area of 36 cm2
is smaller than those in the areas of 4 cm2
and 9 cm2
. Figure 9
b,c shows the similar variation trend of the received signals by S2.
The results of Test 2 in the debonding detection experiment are shown in Figure 10
. By comparing the time domain signal responses in Test 2 with Test 1, a similar conclusion can be drawn.
In addition, the WPEI was used to quantify the total energy of the signal received by S2 in both Test 1 and Test 2, as shown in Figure 11
. The Areas N1, N2, N3 have no debonding, and Area-A1, -B1, and -C1; Area-A2, -B2, and -C2; and Area-A3, -B3, and -C3 have debonding, whose areas are 4 cm2
, 16 cm2
, and 36 cm2
, respectively. It can be seen in Figure 11
a that the WPEIs are between 1000 and 1200 in areas without debonding, and they are much larger than the WPEIs in debonding areas. When the debonding appears (4 cm2
), the WPEIs are reduced significantly to between 400 and 600 since the debond damage dissipates the stress wave propagation. As the debonding area increases to 16 cm2
, the WPEIs decrease to between 200 and 400. When the debonding area further increases to 36 cm2
, the WPEIs are much less than 200. Figure 11
b shows that the trend of the WPEIs in Test 2 is almost the same as that in Test 1. Figure 12
shows the comparison of the WPEIs in Test 1 and Test 2. There is a slight difference between the amplitudes of energy indices in each area in Test 1 and Test 2 because of the different contact status between specimen and transducers. For example, the surface smoothness of steel beams in different areas can lead to a different effective contact area with the sensor. Nevertheless, the energy index trends of each test area are almost the same in both Test 1 and Test 2. The WPEIs illustrate that the PZT-based transducer can effectively detect the debonding between steel beam and reinforcing CFRP plate. Additionally, the WPEIs also reflect the dimension of the debonding: the less the WPEI value, the larger the debonding area.