Inspection for Voids in the Grout below the Protective Duct of an External Post-Tensioning Bridge Tendon Using a THz A-Scanner

Abstract

Grout voids in the tendons of a post-tensioning bridge reduce their strength. Grout voids are also severe flaws causing corrosion of the steel strands in the tendons. Detecting voids during construction and operation of the tendons is essential to prevent tendon failure, which is critical to bridge safety. This study presents a method for inspecting external tendons for voids in the grout below the protective duct pipe using terahertz electromagnetic waves. Due to low attenuation in the high-density polyethylene duct and the large reflectivity difference between the duct/grout and the duct/void interfaces, terahertz waves are suitable for detecting voids in the grout inside tendons. For this study, we developed a mobile frequency-domain terahertz A-scanner that can be used to measure terahertz A-scan data in real time. It is shown that the mobile terahertz A-scanner can be used to assess the area of the grout void in external bridge tendons.

1. Introduction

External tendons are widely used in the girders of post-tensioning bridges. Tendon failure can have a serious impact on the safety of a bridge system [1], so assessing the tendon condition is critical to ensuring bridge safety [2]. The most common tendon failure is caused by corrosion of the steel strands of the tendon. In particular, external tendons are more susceptible to corrosion than internal tendons embedded in hardened concrete because external tendons are more exposed to the surrounding environment, characterized by parameters such as humidity, salinity, and temperature [3]. Importantly, voids in the grout reduce the strength of the tendon system due to the lack of stress redistribution [4]. A more severe problem resulting from the voids is that they expose the steel strands to corrosive ions and moisture, accelerating the corrosion of the steel strands [5]. A grout void is a common defect caused by unskilled workmanship, low-quality grout material, sudden changes in the duct shape, and long-term grout sinking. Therefore, it is important to regularly inspect external tendons for voids and monitor changes in their shape and size, which requires a suitable inspection method.

In general, an external tendon in a post-tensioning system consists of several steel strands, cement grout, and a protective duct pipe surrounding them. Smooth high-density polyethylene (HDPE) is widely used as a duct in the tendon [6]. In recent decades, several methods have been proposed to inspect external tendons for voids in the grout. Some of these methods are based on mechanical waves, such as sounding inspection [7] and ultrasonic testing [8]. The sounding method is carried out by tapping an impactor on the outside of the duct, and a professional expert listens to the tapping sound to determine the presence of damage. This is a simple and fast method but is not suitable for detecting relatively small voids. For ultrasonic testing, ultrasonic probes are used to transmit and receive ultrasonic waves. Depending on the ultrasonic mode used and the arrangement of the probes, there are several techniques such as pulse-echo, through-transmission, linear array, guided waves, etc. It is essential to maintain coupling between the probe and the duct using a liquid couplant for ultrasonic transmission, which can make scanning for imaging and sizing of voids difficult. Other methods are based on electromagnetic waves, such as infrared thermography and radiography, which allow high-resolution imaging of voids [9]. In infrared thermography, voids act as thermal barriers and are detected by thermal imaging. Passive thermography using natural ambient heat is highly dependent on ambient temperature conditions and has limited detection capability, whereas active thermography requires artificial heat sources to transfer heat to tendons. Radiography using X-ray or gamma-ray radiation is a tool to provide high-detectability and high-resolution images of voids. However, radiography is often subject to heavy regulations due to concerns about radiation exposure during testing.

Terahertz (THz) electromagnetic waves in the frequency region of 0.1~10 THz can be used for nondestructive inspection of nonconductive materials owing to their transparency to THz waves [10,11]. This study presents a method based on THz waves to inspect external tendons for voids in the grout below the duct. An A-scanner using electromagnetic waves may be a promising tool for the real-time detection of voids between the duct and the grout in tendons. Electromagnetic waves characterized by low attenuation in the HDPE duct and a significant reflectivity difference between the duct/grout and duct/void interfaces are required to identify the presence of voids below the duct through A-scan measurements. THz waves exhibit the above characteristics so that a THz A-scanner can distinguish well between voids and cement grout in contact with the HDPE duct. In addition, THz waves are harmless to the human body below a certain power, unlike radiographic methods which can affect the human body despite providing high-resolution images of voids.

There are two kinds of THz A-scan methods: time-domain and frequency-domain methods. In time-domain THz A-scan methods, THz pulses generated by a femtosecond laser are used to directly acquire time-domain A-scan data while scanning the time delay between the THz pulse and the femtosecond optical sampling pulse [12]. In addition to the conventional mechanical delay line combined with lock-in detection, fast delay scanning methods using signal averaging have been demonstrated, such as asynchronous optical sampling, electronically controlled optical sampling, optical sampling by laser cavity tuning, and specifically designed mechanical delay tools [13,14,15,16,17]. Fast delay scanning methods have been used for high-speed THz spectroscopy or tomography with applications in real-time detection of explosives, high-speed nondestructive testing (NDT) of composites, etc [17,18]. The cost and size of a time-domain THz A-scanner can be reduced using a femtosecond fiber laser. The short duration or broad spectrum of THz pulses can provide high A-scan resolution. In general, there is a trade-off between A-scan rate and range, depending on time delay scanning methods.

In frequency-domain THz A-scan methods, A-scan data are obtained from the fast Fourier transform (FFT) of frequency-domain THz data acquired while the frequency of continuous-wave (CW) THz radiation is scanned [19,20,21,22,23,24]. The frequency-modulated CW (FMCW) THz method using electronic devices has the advantages of fast acquisition, small size, and long A-scan range, where the frequency modulation rate can be more than kHz and electronic devices are employed, including an oscillator with a narrow linewidth [21,23,24]. However, the FMCW THz method using electronic devices has the disadvantages of high cost and low A-scan resolution, due to the use of high-frequency electronic devices and a frequency modulation range of less than 100 GHz. The FMCW THz method using electronic devices was exploited for THz radar imaging applied to stand-off detection of concealed objects [23] and THz tomography applied to NDT of composites [24]. The advantages of the THz photomixing method are its low cost and compact size [25,26]. Also, it can have high A-scan resolution and long A-scan range due to its wide frequency sweep range of more than 1 THz and a narrow linewidth. However, it usually requires a relatively long frequency sweep time due to the slow tuning of the laser frequency. The THz photomixing method was used in the coherent homodyne detection scheme for THz spectroscopy and thickness measurement [20,22,27,28,29,30,31].

We previously developed a method for high-speed broadband frequency sweep of CW THz radiation [32]. Particularly, the use of a wavelength-swept (WS) laser and a distributed feedback laser diode (DFB-LD) in THz photomixing enabled THz frequency sweep with a THz sweep range and a kHz sweep repetition rate. However, instead of the WS fiber laser, a WS semiconductor laser is used in this study. Compared to the WS fiber laser with a coherence time of several tens of picoseconds, the WS semiconductor laser has the advantages of a long coherence time of about 700 ps and high frequency-sweep linearity [32,33]. In this study, we demonstrate that our mobile frequency-domain THz A-scanner, based on the high-speed broadband THz frequency sweep, is useful for nondestructive inspection of external tendons for voids below the duct.

2. Materials and Methods

The schematic of our THz A-scan measurement system is shown in Figure 1. The frequency of the CW THz radiation generated by THz photomixing corresponds to the optical frequency difference between the DFB-LD and the WS semiconductor laser. The high-speed broadband THz frequency sweep is implemented by the optical frequency sweep of the WS semiconductor laser. The DFB-LD is operated at a fixed wavelength of 1545 nm, and the output wavelength of the WS semiconductor laser is swept from 1544.5 to 1560 nm at a repetition rate of 18 kHz. The optical frequency step of the WS semiconductor laser can be calibrated using a fiber-coupled H¹³C¹⁴N gas cell [34]. Frequency-domain THz signals are measured in the frequency range of up to 1.88 THz at the same repetition rate using coherent homodyne detection. A longer A-scan range can be inspected due to the longer coherence time of the WS semiconductor laser [32,33]. The optical output of the WS semiconductor laser is amplified by an optical attenuator and an optical fiber amplifier and is then combined with that of the DFB-LD by a 3 dB fiber coupler. One of the two optical outputs of the coupler is input to a THz CW transmitter (THz-Tx) module and the other is input to a THz CW receiver (THz-Rx) module via a variable optical delay line. No polarization controller is used since all the optical components are connected using polarization-maintaining fibers.

The THz wave emitted by the THz-Tx is focused on a sample after passing through a thin-film THz beam splitter with a thickness of about a few micrometers [35]. The reflected THz wave from the sample is reflected by the beam splitter and focused onto the THz-Rx. The photocurrent output from the THz-Rx is amplified using a current preamplifier with a bandwidth of 14 MHz and then recorded using a 100 MS/s digitizer. The A-scan range is limited to about 400 ps by $\tau =f_d/\left(d{f}_{THz}/dt\right)$, where $\tau$ and $f_d$ are the time delay and the detection frequency limited by the 14 MHz bandwidth, respectively, and $d{f}_{THz}/dt$ is the sweep speed of the THz frequency, which is equal to 3.5 × 10⁴ THz/s at a sweep rate of 18 kHz [32]. The sweep start trigger signal output from the WS semiconductor laser triggers a waveform generator to provide the THz-Tx with a bias voltage modulated at half the repetition rate. At the same time, data traces are consecutively acquired at the repetition rate by the digitizer triggered by the sweep start trigger signal. The signal-to-noise ratio of THz data can be enhanced by subtracting noise traces acquired with the bias off from THz data traces carrying noise acquired with the bias on and averaging the resulting THz data traces. The A-scan data are obtained by FFT of the measured frequency-domain data in the frequency range of 0.5~1.88 THz [32]. When previously using the WS fiber laser, we measured the temporal variation of the optical frequency of the WS fiber laser (and thus that of the THz frequency) during the frequency sweep period and converted the time to the THz frequency using the temporal variation of the THz frequency to obtain frequency-domain THz data [32]. In addition, the frequency-domain THz data were interpolated at regularly spaced frequencies for their FFT. However, in the case of the WS semiconductor laser, the process is not needed due to its high frequency-sweep linearity.

In Figure 1, the yellow highlighted area illustrates our THz A-scan setup. Since the THz A-scan setup built on the optical table is not applicable to tendon inspection, a mobile THz A-scan unit is required for THz A-scan measurements on tendons. Thus, we designed a mobile THz A-scan unit with a tendon guide that fits tendons with a diameter of 110 mm. Figure 2 shows the top, side, and three-dimensional (3D) views of the designed mobile THz A-scan unit. For THz A-scan measurements on tendons, the THz A-scan unit can be moved manually in the longitudinal and circumferential directions of tendons with the help of the tendon guide.

Figure 3 shows photographs of the mobile THz A-scan unit fabricated according to the design. Mounts are made for the components including the THz-Tx, THz-Rx, off-axis parabolic mirrors, and thin-film THz beam splitter, along with a grooved bottom plate for precise positioning of each component, side plates, a cover, handles, and a white plastic tendon guide. All the parts were assembled to build the mobile THz A-scan unit. Optical and electrical cables pass through a hole in the side plate and a hose surrounding them is attached to the hole. The overall dimensions of the mobile THz A-scan unit are 396 × 275 × 215 mm (including the handles and tendon guide), and the weight of the unit is 8.0 kg. To complete the mobile THz A-scanner, the parts outside the yellow highlighted area in Figure 1 were put together into a movable THz A-scan controller.

3. Results and Discussion

The THz A-scanner has been tested to ensure that it is working properly. For this test, an aluminum mockup was made consisting of half a cylinder with the same diameter as that of the external tendons and a base plate (see Figure 3). The THz A-scanner was used to measure the THz A-scan signal reflected from the aluminum mockup. Figure 4 shows the measured frequency-domain THz data and the THz A-scan data obtained by FFT of the frequency-domain THz data. In Figure 4b, the peak at 10 ps represents the reflected signal from the surface of the aluminum mockup. The full width at half maximum (FWHM) of the peak is 1.75 ps, indicating that the A-scan resolution is 0.26 mm in free space. The A-scan resolution is limited by the measurable frequency bandwidth and could be enhanced by using a THz-Tx and THz-Rx with a wider bandwidth. The frequency band of 0.5~1.88 THz has a mean frequency of 0.88 THz when weighted with the amplitude. If the THz beam is assumed to be a Gaussian beam, the FWHM of the beam spot on the focal plane, that is, the lateral resolution, is estimated to be roughly 2.7 mm by $\sqrt{\ln 2}\left(4/\pi \right)\left(f\lambda /D\right)$, where $f$ is the focal length of the off-axis parabolic mirror, $\lambda$ is the wavelength corresponding to the mean frequency, and $D$ is the input beam diameter [36,37].

To test the performance of the THz A-scanner for detecting voids in external tendons, tendon specimens were constructed by inserting steel strands in an HDPE duct with outer and inner diameters of 110 and 100 mm, respectively, and injecting grout as a filler. As the grout hardened, voids were formed in the upper part of the duct interior. Photographs of one of the tendon specimens thus constructed are shown in Figure 5. Figure 5a,b shows the side and cross-sectional views of the tendon specimen, respectively. Figure 5b shows that there is a void between the duct and the grout at the top of the inner side of the duct.

The tendon specimen shown in Figure 5 was inspected using the THz A-scanner. Figure 6a shows the THz A-scan data measured in the void-free area of the tendon specimen. The first peak is the reflected signal from the duct surface, and the second peak is the reflected signal from the interface between the duct and the grout. The magnitude ratio of the second peak to the first peak is about 70%.

Figure 6b,c shows the THz A-scan data measured in the void area of the tendon specimen. In Figure 6b,c, the first peak is the reflected signal from the duct surface, and the second peak is the reflected signal from the interface between the duct and the void. The third peak in Figure 6c, which is not visible in Figure 6b, is the reflected signal from the interface between the void and the grout. The third peak may or may not be present because the reflected signal from the interface between the void and the grout is measured only when the interface is perpendicular to the THz wave propagation. In Figure 6c, the time delay of 14.7 ps between the second and third peaks indicates that the void is 2.2 mm deep. The magnitude ratio of the second peak to the first peak is about 30%, as shown in Figure 6b,c. Real-time THz A-scan measurements on the tendon specimen using the mobile THz A-scanner are demonstrated in Video S1. Indoor and outdoor THz A-scan measurements gave almost identical results, and all the results shown here were measured outdoors.

Reflectivity is determined by the refractive indices of two materials at their interface. It is presumed that the magnitude of the second peak in Figure 6b,c is much smaller than that of the second peak in Figure 6a because the interfacial reflectivity of the duct and the void is lower than that of the duct and the grout. Therefore, the presence and absence of voids can be clearly discerned by the magnitude of the second peak, indicating that the THz A-scanner is usable for inspecting external tendons for voids below the duct.

4. Conclusions

We have developed a mobile frequency-domain THz A-scanner that can be used to nondestructively inspect external tendons for voids below the duct. THz A-scan data can be rapidly measured with the THz A-scanner. It has been verified that the presence and absence of voids between the duct and the grout in external tendons can be clearly discerned using the THz A-scanner. The THz A-scanner can contribute to the safety of a post-tensioning bridge through nondestructive inspection for voids in external tendons. The present technique exhibits excellent sensitivity and spatial resolution in detecting voids below the duct. When combined with a C-scan automation equipment, the THz based inspection could possibly provide B-scan and C-scan images of the voids and then be more favorable for field applications than radiography, which is detrimental to the human body through radiation exposure.