Permanent Deformation and Temperature Monitoring of Subgrades Using Fiber Bragg Grating Sensing Technology

The hidden nature of subgrades makes the effective monitoring of their deformation very difficult. This paper addresses this issue by proposing the use of fiber Bragg grating (FBG) sensing technology. Here, an FBG is encapsulated within a monitoring tube formed from a polyvinyl chloride tube, and one end of the monitoring tube is fixed perpendicular to a concrete column, forming a cantilever beam monitoring system. The deformation is assessed according to the theoretical relationship between the horizontal strain on the FBG embedded in the monitoring tube and the vertical displacement of the cantilever beam. Then, the relationship between the variation in the wavelength of light reflected by the encapsulated FBG and the temperature and horizontal strain is obtained on this basis by calibration experiments. The monitoring tubes are buried at a proscribed depth below the top surface of the subgrade, which facilitates the monitoring of the deformation and temperature of the subgrade at different stages of construction through the collection of FBG wavelength data during different periods, such as after embedding the monitoring tubes, the completion of the test road surface, and during the period of operation. The proposed technology is verified by employing the system to monitor the instantaneous maximum deformation and permanent deformation of a subgrade under dynamic loads. The monitoring results demonstrate that the instantaneous maximum deformation values of the subgrade at 0.25m and 0.5m below the surface are 695.40 μm and 574.02μm, respectively, and the corresponding permanent deformation values are 53.00 μm and 41.54μm, respectively. The FBG sensor system is thereby verified to provide a reliable method for conducting long-term continuous, accurate, and efficient subgrade deformation and temperature monitoring.


Introduction
There are many factors that affect road service life [1]. Excessive settlement of highway subgrades is one of the main causes of roadway degradation, which can lead to serious driving safety risks [2][3][4]. However, conventional subgrade deformation monitoring methods are typically unable to provide timely and accurate information regarding the occurrence of subgrade deformation [5][6][7]. Accordingly, the development of effective subgrade deformation monitoring and early warning methods is essential for the prevention and control of highway degradation [8].
Fiber Bragg grating (FBG) sensing technology has developed rapidly over the past years [9][10][11] and represents an excellent alternative to conventional subgrade deformation monitoring methods owing to its many advantages such as corrosion resistance, electromagnetic interference resistance, good waterproof performance, and high measurement precision [12]. Here, an FBG sensor employs a periodic variation in the refractive index along a short length of an optical fiber to obtain an optical filter that reflects specific bandwidths of light in response to the transmission of broad-spectrum light through the optical fiber. The periodic variation in the refractive index of optical fibers can be either uniform or nonuniform [13], and because the refractive index varies along the length of an optical fiber, the specific bandwidth of light reflected by the FBG is sensitive to temperature and strain. This sensing technology has been applied gradually over recent years for conducting engineering health monitoring [14][15][16] and geotechnical engineering monitoring [17,18].
The specific designs and applications of FBG sensing technology have been the subject of intense development in recent years. For example, Guo et al. [28] applied a surfacemounted FBG strain sensor to successfully monitor the magnitude of the strain and its variation in an expressway via duct due to vehicle load pressure. Kesavan et al. [29] proposed an FBG sensor application for measuring the interfacial strain of reinforced concrete beams strengthened with carbon fiber-reinforced polymer (CFRP). An experimental application verified that the FBG sensor array effectively identified the beginning and expansion of CFRP separation from the concrete surface. An FBG humidity sensor has also been designed to monitor the corrosion rate of concrete sewer walls [30]. The FBG sensor was verified to provide high durability, good time response, and stability over an extended period in the corrosive hydrogen sulfide gas environment with high humidity. Accordingly, the FBG sensor was shown to have broad application prospects in harsh environments. Both long-gauge FBG and point FBG strain sensors were developed for monitoring the static and dynamic loads of a concrete railway bridge [31]. Experimental results demonstrated that both sensors were able to provide accurate strain measurements. A new type of mechanical sensor employing an FBG was proposed, and the design was demonstrated to improve the accuracy of damage detection and localization for civil engineering structures by amplifying the strain applied to the FBG sensor by a factor of about 36 [32]. Dyshlyuk et al. [33] presented the experimental application of an FBG-based measurement method based on optical timedomain reflectometry (OTDR) for monitoring strain in bent reinforced concrete beams. The proposed method was demonstrated to provide results that were consistent with those of direct spectral measurements. A vibration detection method based on FBG sensing technology was proposed to accurately measure the physical and mechanical properties of soil [34]. The data collected by the FBG sensor were demonstrated theoretically and experimentally to reflect vibration conditions clearly and quantitatively. Bellas and Voulgaridis [35] responded to the impact of geotechnical disasters on community housing by developing an advanced FBG-based method for monitoring the health of housing structures and serving as an early warning system. Nan et al. [36] contributed toward the prevention of subway tunnel collapse by applying ultraweak FBG sensing technology to detect the distributed vibrational response of the FBG positioned on the tunnel wall and the track bed, and the occurrence and characteristics of intrusion events simulated by the discrete and continuous pulses of an excavator were recognized under two loading attitudes. The above studies have demonstrated that FBG sensing technology can be implemented in a wide array of engineering applications under complex and harsh conditions. Accordingly, this technology has been applied for monitoring subgrade deformation [37,38]. However, a sufficiently exact relationship between the specific bandwidth of light reflected by the FBG and the vertical displacement of the subgrade was not established in these past studies, and the proposed methodology was unable to monitor subgrade temperature.
The present work addresses the shortcomings of past efforts seeking to apply FBG sensing technology toward subgrade deformation and temperature monitoring. Here, an FBG with a uniformly distributed variation in the index of refraction is encapsulated within a monitoring tube formed from a polyvinyl chloride (PVC) tube, and one end of the monitoring tube is fixed perpendicular to a concrete column, forming a cantilever beam monitoring system. The deformation is assessed according to the theoretical relationship between the horizontal strain on the FBG embedded in the monitoring tube and the vertical displacement of the cantilever beam. Then, the relationship between the variation in the wavelength of light reflected by the encapsulated FBG and the temperature and horizontal strain is obtained on this basis by calibration experiments. The monitoring tubes are buried at a proscribed depth below the top surface of the subgrade, which facilitates the monitoring of the deformation and temperature of the subgrade at different stages of construction through the collection of FBG wavelength data during different periods, such as after embedding the monitoring tubes, the completion of the test road surface, and during the period of operation. The proposed technology is verified by employing the system to monitor the instantaneous maximum deformation and permanent deformation of a subgrade under dynamic loads. The results verify that the deformation and temperature of the subgrade can be monitored in situ even after the completion of highway pavement construction, and each measurement can be completed in a few minutes.

Working Principles of FBG Sensors
2.1. Transmission Principle. The axial refractive index distribution of a uniform period FBG can be given as follows [39]: where n 0 is the refractive index of the optical fiber core, Δ n max is the maximum change in n 0 , Λ is the period length of the uniform grating, and z is the axial position along the optical fiber. The refractive index distribution is illustrated in Figure 1 [40]. According to coupled mode theory, the wavelengths of broad-spectrum light transmitted through an optical fiber  Journal of Sensors that meet the Bragg wavelength condition of the FBG will be reflected back to the incident end of the optical fiber, and the light of all other wavelengths will pass through freely [41]. This condition is illustrated in Figure 2.

Basic Sensing Principle.
Through coupling between the core mode of forward light transmission and the core mode of backward light transmission, the energy of the core mode of forward transmission is transferred to the core mode of backward transmission, forming the reflection of the incident wave [42]. The reflected wavelength of an FBG is given as follows: where n eff is the equivalent refractive index of the optical fiber core. The elastooptic effect of the fiber itself causes the value of n eff to change with changes in the strain state of the fiber, which changes the value of λ B reflected from the FBG [43]. As such, the value of λ B represents a measure of the strain state of the optical fiber.
If the influence of changes in the temperature is not taken into account, the change in λ B owing to tensile and compressive axial strain ε on the FBG can be expressed as follows: where P e is the effective photoelastic coefficient. Therefore, the axial strain can be calculated as follows:

Journal of Sensors
When accounting for changes in the temperature, the thermooptic effect and thermal expansion effect of FBG materials lead to a shift in λ B owing to a change ΔΛ in the value of Λ [43,44]. From equation (2), the value of Δλ B caused by a temperature change ΔT can be expressed as follows: Here, ∂n eff /∂T is the refractive index temperature coefficient of the fiber grating; ðΔn eff Þ ep is the elastooptic effect produced by the thermal expansion effect; ∂n eff /∂r is the waveguide effect produced by the thermal expansion effect, where r is the diameter of the optical fiber; and ∂Λ/∂T is the linear thermal expansion coefficient of the FBG. The right side of equation (5) can now be divided by the right side of equation (2) to obtain the following: where ∂n eff /∂T · 1/n eff is the thermooptic coefficient of the FBG, which is expressed by ζ; ΔΛ/ΔT · 1/Λ is the thermal expansion coefficient of the optical fiber, which is expressed by α; and 1/n eff ½ðΔn eff Þ ep + ∂n eff /∂r · Δr is the comprehensive action coefficient of the elastooptic and waveguide effects, which can be expressed by K H . Accordingly, equation (6) can be arranged as follows: which yields the following expression: However, the influence of the elastooptic and waveguide effects caused by thermal expansion on the temperature sensitivity coefficient is very weak. Therefore, K H can be ignored, and equation (8) can be simplified as   Journal of Sensors We can then combine equations (3) and (9) to obtain the change in λ B owing to both strain and temperature as follows: This can be further simplified using the definitions ð1 − P e Þ · λ B = K ε and ðζ + αÞ · λ B = K T , as follows: Here, K ε is denoted as the strain sensitivity coefficient, and K T is denoted as the temperature sensitivity coefficient.

FBG Monitoring Tube
Design. An FBG is fragile and vulnerable to damage, so it cannot be directly embedded in the subgrade. Therefore, an FBG must be packaged within a suitable carrier structure, which can then be embedded in the subgrade. The present work employs a circular PVC tube with a diameter of 10 cm as the FBG carrier. As shown in Figure 3(a), the PVC tube is first slotted in advance, and slotting is applied symmetrically to the other side of the tube as well. Then, a single FBG is fixed in the upper and lower slots of the PVC tube using 101 glue and then encased with NANDA 703 silicone rubber, as shown in Figure 3(b) for the upper FBG. The technical parameters of FBGs used in the paper are shown in Table 1. Figure 4 presents a cross-sectional schematic of the traffic lanes of a standard roadway. Past studies have demonstrated that the maximum dynamic stress generated by vehicle loading appears directly under the wheels over a specific depth range of the subgrade [45]. Moreover, according to traffic rules and standard driving habits, the slow lane is mainly used by large and heavy vehicles, which travel down the middle of the slow lane under normal circumstances. Therefore, these positions in the slow lane represent the points of maximum subgrade deformation. Accordingly, the present work adopts the cantilever beam monitoring system illustrated in Figure 5. Here, the base end of the cantilever beam is embedded within a concrete column, and two FBGs are packaged symmetrically in the upper and lower slots at the end of the monitoring tubes.
The positioning of the FBGs is based on a standard heavy truck wheelbase width of 2 m with consideration for the contact area between the tire and the ground. The standard width of tire contact with the ground is approximately 0.3 m based on the heavy vehicle tire most commonly used, an axle load of 100 kN, and tire pressure of 0.7 MPa. Accordingly, the distance between the inner and outer FBGs and the base end of the cantilever beam is 2.85 m and 3.15 m, respectively, and the FBGs are arranged symmetrically at the upper and lower positions of the monitoring tube to realize simultaneous monitoring of subgrade deformation and temperature. As shown in Figure 5, the centers of FBG1 and FBG2 are on the same plumb line, and the centers of FBG3 and FBG4 are on the same plumb line. This represents an improved configuration relative to previously published work [46], where each monitoring tube was equipped with 8 FBG monitoring points typically. However, the monitoring points close to the base end of the cantilever provided limited information due to the fixed constraint at the base end. Therefore, the FBGs are placed only directly under the vehicle track in the present work. Figure 6, past research has demonstrated that the vertical dynamic stress of a subgrade owing to automobile loading decreases linearly with increasing subgrade depth in the range of 0.0-0.6 m, and that the dynamic stress decreases much more slowly at depths greater than 0.6 m [47,48]. This linearity is employed in the present work by adopting two FBG monitoring tubes buried at 0.25 m and 0.50 m below the top surface of the roadbed, as illustrated in Figure 7, which includes the optical fiber configuration employed in conjunction with an FBG acquisition demodulator and data processing module.  5 Journal of Sensors FC/APC connector was welded to one end of the optical fiber using a fiber fusion splicer, which would be connected to the interrogator when conducting tests. The monitoring system has 4 arrays of 2 FBG sensors each, as Figure 7 shows with the 4 lines coming out from the monitoring tubes.

FBG Calibration.
Accurate monitoring of the subgrade deformation and temperature requires that we first calibrate the strain sensitivity coefficient K ε and temperature sensitivity coefficient K T of the FBGs employed in the proposed monitoring system. However, we first obtained the values of λ B for 8 representative FBGs installed within monitoring tubes in the absence of strain at a temperature of 21.3°C, and the results are listed in Table 2.
The monitoring tubes were so long that the FBGs on them were not easy to be calibrated after encapsulation, so 2 FBGs were selected for calibrating the strain sensitivity coefficient K ε and the temperature sensitivity K T in the laboratory. According to the previous calibration experiment results, there is little difference between the P e , ζ, and α values of the same model FBGs from the same manufacturer, so it was assumed that the P e , ζ, and α values of the 8 FBGs used on the monitoring tubes are equal to those of the 2 FBGs used in the calibration experiment.
3.3.1. Calibration of Strain Sensitivity Coefficient. Two FBGs were selected for calibrating the strain sensitivity coefficient K ε at a temperature of 21.3°C according to the experimental setup illustrated in Figure 8. Here, 60 cm sections of PVC monitoring tubes were applied, and FBGs were installed on the lower tube sides in an equivalent manner as described in Subsection 3.1. Center reflection wavelengths λ B of 1536.593 nm and 1561.129 nm were first obtained for these two FBGs in the absence of strain. In addition, resistance strain gauges were pasted on the upper tube sides directly above the FBGs. The two ends of the PVC tube were fixed, and a variable weight was applied at the midpoint of the tube. Then, strain and λ B data were collected simultaneously from the resistance strain gauge monitor and the grating   Journal of Sensors acquisition demodulator, respectively. The experimental resistance strain gauge setup provided an accuracy of 1 με and a measurement range of −19999 με to +38000 με. The relationships between λ B and the measured strain are presented in Figure 9. The calibration results indicate that the values of λ B for the FBGs change linearly with respect to strain over the investigated strain range, indicating that the FBGs respond uniformly over that range. The values of K ε were obtained for the two FBGs from the slopes of the lines fitted to the plotted data, which were 1.37 pm/με and 1.39 pm/με. According to the calibration results, the strain sensitivity coefficients K ε of the FBGs on the monitoring tube could be calculated by K ε = ð1 − P e Þ · λ B: The values of K ε obtained similarly for all 8 FBGs listed in Table 2 are listed in Table 3.

Temperature Sensitivity Calibration.
The two FBGs employed for strain sensitivity coefficient calibration were again employed for temperature sensitivity calibration over an experimental temperature range of −20°C to 30°C. A frequency conversion refrigerator was used to control the temperature between −20°C and 10°C, and the temperature   7 Journal of Sensors inside the refrigerator was monitored using a K-type thermocouple connected with a tt-k-24-sle thermocouple line to a temperature monitor located outside of the refrigerator. Testing at temperatures in the range of 10°C to 30°C was conducted in a laboratory equipped with a constant frequency air conditioner, and the temperature was again monitored with a K-type thermocouple. The change in the values of λ B for the FBGs with respect to temperature was monitored every 5°C in the absence of strain. The experimental results are shown in Figure 10. The temperature calibration results indicate that the values of λ B for the FBGs change linearly with respect to the temperature over the investigated temperature range, indicating that the FBGs respond uniformly over that range. The values of K T were obtained for the two FBGs from the slopes of the lines fitted to the plotted data, which were 43.94 pm/°C and 44.61 pm/°C. According to the calibration results, the temperature sensitivity coefficients K T of the FBGs on the monitoring tube could be calculated by K T = ð ζ + αÞ · λ B: The values of K T obtained similarly for all 8 FBGs listed in Table 2 are listed in Table 4.

Principle of Deformation Calculation
As discussed above, the horizontal strain on a monitoring tube can be determined according to the change Δλ B in the wavelength of the light reflected by the FBG. However, monitoring the deformation of the subgrade based on the cantilevered monitoring tubes illustrated in Figures 5 and 7 requires that the relationship between the horizontal strain and the vertical deflection of a monitoring tube be determined. This is evaluated according to the stress conditions acting on the axis of the cantilever beam, as illustrated in Figure 11. Here, the beam length between points A and B is l, x is the horizontal distance from the fixed end to any position along the cantilever beam, and q 1 and q 2 are uniformly distributed loads, respectively, on the upper and lower surfaces of the beam owing to a force F applied at point B.
The conditions illustrated in Figure 11 are further analyzed by considering the bending moments under load F and distributed loads q 1 and q 2 . First, the moment equation of the cantilever beam under load F is given as according to the moment diagram shown in Figure 12, where c 1 is the centroid of the bending moment diagram. The bending moments of the cantilever beam under q 1 and q 2 are, respectively, analyzed accordingly to the moment diagrams shown in Figures 13 and 14, where c 2 and c 3 are the respective centroids of the two bending moment diagrams, and the bending moment equations are given as follows: Accordingly, these bending moment diagrams represent quadratic parabolas.
Extending the above analysis, a unit force f is applied to the end of the cantilever beam, as shown in Figure 15(a), and the bending moment diagram is shown in Figure 15(b). Accordingly, the bending moment equation is given as follows: The ordinate values of the unit load bending moment diagram corresponding to the center of gravity of the bending moment of the cantilever under applied loads F, q 1 , and q 2 are, respectively, The vertical deformation of point B can be obtained using graph multiplication as follows: where the terms MðxÞ, M 1 ðxÞ, and M 2 ðxÞ are the bending moments of the cantilever beam under an external load (F, q 1 , q 2 ), which are defined as follows: Here, the horizontal strain ε is now a function of x, and y is the distance from a point on the monitoring tube along the radial direction to the beam centerline axis.
The analysis at point B is now extended to some arbitrary point C along the horizontal direction of the cantilever beam of radius R, as shown in Figure 16. Here, the bending moment equation under f is given as The relationship between the vertical displacement and the horizontal strain at point C is given as follows: The horizontal strain can be analyzed more clearly by considering a short segment of the cantilever beam of length dx at C in Figure 16 and expanding it as illustrated in Figure 17(a), where y is the distance between any longitudinal line segment MM ′ and the centerline axis OO ′ . This is applied to a case of pure beam bending through an arc of d θ, as illustrated in Figure 17(b), where the strain is given as and ρ is the radius of curvature of the centerline axis after bending. According to formulas (20) and (21), the deflection of point C under stress can be obtained as follows: which represents the collaborative deformation of the monitoring tube and soil under an applied force. Here, the minus sign (−) indicates that the deflection is downward, and the value of ρ can be determined from experimental measurement data.

Monitoring Tube Installation under Test
Roadway. The monitoring tubes were packaged according to the design in Figure 5 and installed in the subgrade soil underneath a test roadway section according to the layout scheme illustrated in Figure 7. The embedding process is shown in Figure 18.

Initial Data Acquisition.
The initial center wavelength of the reflected light from the 8 FBGs and the temperature of the subgrade was collected immediately after laying the two monitoring tubes. The test data were collected using a sm130 FBG demodulator (American Micron Optics, Inc.). The sm130 demodulator includes a built-in largebandwidth scanning laser light source. It also includes 4 channels with a wavelength range of 1510-1590 nm and an accuracy higher than 1 pm. Bluetooth wireless temperature measurement devices with low power consumption were embedded in the subgrade at depths corresponding to the monitoring tubes [49]. The collected data are listed in Table 5. The FBG monitoring system requires only the initial temperature from the wireless temperature sensors, and subsequent subgrade temperatures can be obtained by monitoring the change in the λ B values of the respective FBGs.   According to formula (11), the wavelength changes of FBG1 and FBG2 monitoring tube 1 in Figure 7 are as follows: where ΔT 1 and ΔT 2 are the subgrade temperature difference between the monitoring time and the initial time. Δλ B1 and Δλ B2 can be obtained by monitoring. It can be seen from Figure 17 that ε 1 = −ε 2 . Since the subgrade temperature change is small within 10 cm, ΔT 1 = ΔT 2 can be set. The strain and temperature changes of FBG1 and FBG2 are calculated as follows: Therefore, as long as the initial temperature of the monitoring point is known, the subgrade temperature at point ① in Figure 7 can be calculated when monitoring. According to the method, the temperature of points ②, ③, and ④ in Figure 7 can also be calculated.
Similarly, the center values of λ B obtained from the 8 FBGs were collected immediately after the pavement of the test roadway was completed. The collected data are listed in Table 6. Finally, the deformation of the subgrade at the monitoring points during the construction period was calculated based on the data collected before and after construction, and the subgrade temperature at the completion of roadway paving was also obtained. These results are listed in Table 7 along with vertical deformation and temperatures of the subgrade obtained after construction.

Subgrade Deformation Monitoring under Vehicle
Loading. Subgrade deformation monitoring of the test roadway under vehicle loading was conducted during the summer of the second year after the roadway was opened to traffic. First, the center wavelengths of the reflected light from the FBGs were collected, and the collection results are listed in Table 8. The permanent deformation of the subgrade at the monitoring points since the completion of the roadway   Tables 6 and 8, and the results are listed in Table 9 along with the calculated subgrade temperature at the time of monitoring.
We conducted dynamic subgrade deformation monitoring under vehicle loading using a heavy 4-axle truck traveling over the subgrade monitoring area at a speed of 10 km/h. The dynamic load monitoring test process is shown in Figure 19.    Figure 20. The maximum deformation, elastic deformation, and permanent deformation of each monitoring point under vehicle loading were determined from the collected data, and the results are listed in Table 10. It can be seen from Figure 20 that the FBG monitoring system accurately monitors the process by which the values of λ B change with respect to time under dynamic vehicle loading conditions, and the monitoring system can effectively detect the maximum deformation, elastic deformation, and permanent deformation of the subgrade from the time histories. It can be seen from Table 10 that the deformation values obtained by the FBGs increase as the distance between the FBGs on the cantilever beam and the anchor point in the concrete column increases. We then selected the larger value as the monitoring result. We also note that the maximum deformation, elastic deformation, and permanent deformation values of the subgrade decrease substantially with increasing distance from the subgrade surface.
The subgrade deformation was obtained by analyzing the data collected by the monitoring tube 1 and tube 2 under the load of vehicles such as a light vehicle, water truck, medium truck, transit mixer truck, and heavy truck during the fall of the second year after the roadway was opened to traffic. These results are shown in Figure 21. The subgrade deformation value was randomly monitored during the normal driving process of the vehicles; also, it was affected by many factors such as vehicle speed and cargo load.
It can be seen from Figure 21 that the deformation law of subgrade obtained by the two monitoring pipes under different loads is the same. The permanent deformation values of deep subgrade are smaller than those of shallow subgrade. It indicates that the permanent deformation of the subgrade decreases with the increase of the subgrade depth under the action of load. The permanent deformation of the subgrade is very little under the action of a light truck, but it increases obviously when the load increases. The order of magnitude and variation law of the subgrade deformation obtained by the monitoring method are similar to those obtained in the literatures [46,50].

Conclusions
The present work addressed the difficulty of monitoring the deformation of highway subgrades in situ using FBG sensing technology. An FBG was encapsulated within a monitoring tube formed from a PVC tube, and one end of the monitoring tube was fixed perpendicular to a concrete column, forming a cantilever beam monitoring sys-tem. The deformation was assessed according to the theoretical relationship between the horizontal strain on the FBG embedded in the monitoring tube and the vertical displacement of the cantilever beam. Then, the relationship between the variation in the wavelength of light reflected by the encapsulated FBG and the temperature and horizontal strain was obtained on this basis by calibration experiments. Two monitoring tubes were buried at depths of 25 cm and 50 cm below the subgrade surface of a test roadway, and the system was demonstrated to facilitate the monitoring of the deformation and temperature of the subgrade at different stages of construction through the collection of FBG wavelength data during different periods, which included after embedding the monitoring tubes, after the completion of the test roadway surface, and during a period of operation under dynamic vehicle loading. The proposed monitoring system was verified to effectively detect the maximum deformation, elastic deformation, and permanent deformation of the subgrade from the time histories of the center wavelengths of the light reflected by the FBGs under conditions of dynamic loading. Accordingly, we can conclude that the FBG monitoring system can realize long-term, accurate, and efficient monitoring of subgrade deformation as well as subgrade temperature in the range of −20°C to 30°C.

Data Availability
The data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
The authors declare that they have no conflicts of interest.