On using the DFOS for damage identification in civil engineering structures

The application of distributed fiber optic sensing (DFOS) technologies to the monitoring of civil engineering structures has been widely developed in recent years. In particular, fiber optic sensing (FOS) using the Rayleigh backscattering phenomenon of the laser frequency spectrum has been used to continuously measure strain along the fiber optic sensor with high spatial resolution. However, the practical implementation of fiber optic sensorsin the field is still a challenge. These include, but are not limited to, the bonding between fiber sensor cables and the surrounding medium, the practical way in which fiber optic sensors can be installed in concrete structures, and the analysis of crack width using the strain distribution along the fiber. Compared to classical discrete sensors, the DFOS has a great potential to detect, localize and quantify crack formation even in the early stages of microcracking. This paper discusses the potential and challenges of using distributed fiber optic sensors for crack detection in concrete structures. In this sense, different types of mechanical tests on concrete structures using different fiber optic sensors, and different bonding techniques are presented to explore the possibilities of crack detection in concrete structures and its potential for practical structural health monitoring (SHM) systems of civil structures.


Introduction
Structural Health Monitoring (SHM) systems involve the use of sensors and other monitoring systems to continuously assess the condition of a structure.Today, SHM systems are essential for the maintenance and management of critical infrastructure such as bridges, tunnels, and buildings [1].SHM relies on the use of sensors to continuously determine the condition of a structure by monitoring various structural parameters, such as thermally induced strain and structural deformation.A wide range of sensors can be used in SHM systems, including but not limited to electrical strain gauges, accelerometers, and temperature sensors.
In addition to traditional sensor technologies, Distributed Fiber Optic Sensing (DFOS) is a promising technology for SHM that can provide continuous, real-time monitoring of temperature and strain along the length of fiber optic cables, even over long distances.DFOS, based on the acquisition of the Rayleigh backscattering of the laser frequency spectrum, allows the continuous measurement of temperature and strain along the fiber optic sensor with high spatial resolution.The application of DFOS has shown high efficiency for damage detection, especially in laboratory experiments [2].
The quality of the measurement results is strongly influenced by the type of sensor cable and the bonding or embedding technique.For surface bonding with a structural adhesive, the usual conditions of surface treatment and cleanliness must be ensured.For embedding in a structure, the bonding to the surrounding matrix and the exact position must be guaranteed.
The influence of various parameters on laboratory scale elements has been studied by many researchers.It has been shown that cyanoacrylate adhesive is suitable for localized bonding of fiber optic cables to the surface of concrete structures due to its low strain loss [3].However, the concrete surface must be well prepared before the fiber sensors are applied.The effect of the fiber type and the coating layers on the losses in the strain measurements between the matrix surface and the core of the fiber was investigated.They show that polyimide-coated fibers are more sensitive than other types of fibers used in the experiments [4].Other researchers have investigated the

Abstract
The application of distributed fiber optic sensing (DFOS) technologies to the monitoring of civil engineering structures has been widely developed in recent years.In particular, fiber optic sensing (FOS) using the Rayleigh backscattering phenomenon of the laser frequency spectrum has been used to continuously measure strain along the fiber optic sensor with high spatial resolution.However, the practical implementation of fiber optic sensorsin the field is still a challenge.These include, but are not limited to, the bonding between fiber sensor cables and the surrounding medium, the practical way in which fiber optic sensors can be installed in concrete structures, and the analysis of crack width using the strain distribution along the fiber.Compared to classical discrete sensors, the DFOS has a great potential to detect, localize and quantify crack formation even in the early stages of microcracking.This paper discusses the potential and challenges of using distributed fiber optic sensors for crack detection in concrete structures.In this sense, different types of mechanical tests on concrete structures using different fiber optic sensors, and different bonding techniques are presented to explore the possibilities of crack detection in concrete structures and its potential for practical structural health monitoring (SHM) systems of civil structures.ability to detect, locate and quantify cracks using DFOS.In this sense, algorithms have been developed to analyze the crack width based on the measured strain distribution [5,6].In this paper, we investigate different fiber optic sensors, sensor cables and different adhesives for surface bonding of DFOS, as well as the influence of different sensor positions and embedding in concrete structures on the resulting sensor data.

DFOS for damage identification
The influence of different fiber optic sensors using different structural adhesives for surface bonding of the DFOS on the possibilities of spatially resolved crack identification and crack propagation was tested.In addition, the efficiency of the DFOS in detecting crack propagation and quantifying crack width was realized.For this purpose, two experimental programs were conducted by testing reinforced concrete elements in bending and tension.

3-Point Bending Test -DFOS bonded at Concrete Surfaces
A 3-point bending test was conducted on four beams with 1200 x 150 x 100 mm using an incremental load at the midspan of 900 mm, as shown in Figure 1.The materials used were as follows: Concrete cover of 10 mm. Reinforcement type B500B: 2T8 at the bottom layer, and 2T8 at the top layer. Stirrubs of ∅6@100mm.
The beams were equipped with two types of fibre optic sensors with different coating layers: the TPE-coated (Hytrel) and the polyamide-coated fibres of 900 µm and 145 µm, respectively.The fibre optic sensors were bonded to the concrete surface in 4 x 600 mm measuring strips as follows:  Strip I: on the side of the beam at the compression zone (top).


Strip II: on the side of the beam at the tension zone (bottom).


Strip III and IV: at the bottom side of the beam.
To verify the influence of structural adhesives on the measurements of DFOS, the two types of fibre were combined with two adhesive materials of different modulus of elasticity of 240-280 MPa (Et_MW=260 MPa) for the "soft" adhesive on the basis pf polyurethane and 2000-2500 MPa (Et_MW=2250 MPa) for the "hard" adhesive material, respectively.Both adhesives are suitable for bonding of concrete, with short pot times (< 5 min) and viscosities of 6000 to 8500 mPa.s.
To validate the results of the DFOS measurements, the beams were each equipped with two 5 cm long electrical strain gauges (DMS) in the tension zone (bottom).At selected cracks, the crack width was measured using two special inductive displacement transducers (LVDTs) as shown in Figure 2.  The fiber optic sensors were prepared by splicing a FOS cable with a connector on the single-mode optical fiber.
The fiber end was prepared so that no light could be backscattered from the fractured surface of the fiber end.
The strain along the optical fiber sensors during the mechanical test was measured with a coherent frequency domain reflectometer (ODiSI B, LUNA Innovations Incorporated, USAs).The ODiSI allows a spatial resolution of 0.65 mm.The sampling rate was 1 Hz.Temperatures were kept constant during the experiments (23±2°C).
The concrete beam was loaded at midspan for 597 s corresponding to 16 kN.The load was then held constant to allow time for the crack width measuring device to be attached.At the time of 2050 s, the beam was continuously reloaded to failure at 37 kN.The loading regime is shown in Figure 5.Some of the results are shown in Figure 3 and Figure 4 for the stains for the Hytrel-coated fiber and the polyimide-coated fiber with the "soft" adhesive material.
The contours in Figure 3 (bottom-left) represent the change in the time-position history of the strains in the four measurement strips, i.e.Strip I to IV, from left to right.The x-axis represents the coordinates along the fibre in mm, while the y-axis represents the time in seconds.
To better highlight the results over the duration of the experiment, the strains are plotted in Figure 3 (top) at certain time intervals, i.e., 335, 1961, and 3018 s, as shown by the blue, magenta, and red curves, respectively.The peaks indicate the concentration of strains that coincide with crack propagation in the concrete, which means that DFOS follows the same behavior as rebar in the vicinity of cracks.In addition to the increase in strain values in the peaks with increasing applied load, it is evident that the strain values also increased around the crack tip due to The negative strain values in Strip I indicate the behavior of the concrete in the compression zone at the top of the beam.This shows the possibility to measure compressive strain by using DFOS in addition to the tensile strains in Strips II through IV.It is noteworthy that DFOS can also detect the microcracks that have developed in the compression zone, see for example the red curve in Figure 3 (top) at Strip I.The development of strains over the experiment duration is shown in Figure 3 (bottom-right) at the position intervals 1725, 3487, and 4648 mm represented with black, magenta and red curves, respectively.The strain-time curves show that the measurements accurately follow the time history of the loads applied to the beam in both the loading and unloading phases, as shown in the load-time curve in Figure 5 (bottom).The measurements of the polyimide-coated fiber with the "soft" adhesive are shown in Figure 4 using the same layout as the graph for the Hytrel-coated fiber in Figure 3.The polyamide fiber follows the same trend described above for the Hytrel coated fiber in the tension and compression zones of the beam, see for example the contours in Figure 4 (bottom left) and the strain curves in Figure 4 (top).
However, the strain values measured by the polyimide fiber are higher than those measured by the Hytrel-coated fiber.The lower strain values are due to the softer Hytrel coating (thermoplastic elastomer) and the greater distance to the surface caused by the 6 times larger diameter of the Hytrel-coated fiber.For example, see the strain values in Figure 4 (top and bottom right).It is also noteworthy that the sensitivity of the polyimide fiber is higher, especially at high strain values, as shown by the red curve in Figure 4 (top).However, if the strain gradient of two consecutive measuring points (especially in the crack transition regions) on the fiber is too high, the evaluation algorithm of the OdiSI fails and does not record any detectable useful values.
The DFOS was able to detect the high local strains that coincide with crack initiation at the very beginning of crack formation, even before they are visible to the observer.In the case of the soft adhesive Hytrel-coated fiber, the first structural crack was detected at 354 s, corresponding to a load of 9 kN.This is illustrated by the blue curve in Figure 5 (bottom) with strain-time curves at a fixed sensor position, where the DFOS measurements increased dramatically and began to deviate from the DMS measurements.However, it was not possible to visually detect a crack at this stage of the test, where the first crack was visually detected at 597 s, corresponding to a load of 16 kN.The same behavior can be seen from the red and black curves representing the DFOS measurements at the same location of the DMS.The first cracks at the DMS location were detected by the DFOS at 450 seconds.However, both the DFOS and DMS agree with the load time history shown in Figure 5 up to the formation of cracks in the vicinity of the DMS.Beyond this point, only the DFOS was able to follow the load time history.
The types of adhesive materials used in the experiments did not have a significant effect on the DFOS measurements.Figure 6 shows that the "soft" adhesive can detect cracks slightly before the "hard" adhesive.However, all selected combinations of fibres and adhesive materials efficiently localized the crack, as shown by comparing the DFOS measurements with the visual inspection of cracks R.01 to R.07 along the tested beam (see Figure 7).The most promising finding is that the DFOS also detects the microcracks smaller than 0.04 mm (the first crack seen visually) at the beginning of the test, see the magenta, red, and blue curves in Figure 7.The DFOS efficiently measured the strains up to the end of the test with 37 kN, which is 0.25 mm of the maximum measured crack width.

DFOS embedded into Concrete Elements
Investigations of DFOS sensor cables embedded in concrete were also carried out to test the practicality of the sensor application and the potential for crack detection.Three concrete cylinders of 100 mm diameter and 800 mm length, each with a type B500B, 16 mm steel bar, were prepared as shown in Figure 8.The cylinders were made of concrete Class C35/40.The material properties and dimensions were carefully chosen to ensure the formation of multiple cracks along the length of the specimen.
Tensile tests were performed by pulling on the steel bar with a universal testing machine Tiratest 28100 Tira GmbH using a 100 kN load cell.The displacement of the concrete was measured with externally mounted inductive displacement sensors.Tensile loads were continuously applied to the rebars until the yield strain reached approximately the 90 kN load.The force-displacement diagram is shown in Figure 12.
Strains in concrete and steel reinforcement were measured using two types of DFOS: 1) the EpsilonSensor, shown in Figure 9, made of single mode optical fiber surrounded by glass fiber/polyester system with 3 mm diameter from SHM System [7], 2) the Hytrel-coated fiber with a diameter of 900 µm, as shown in Figure 10.The Hytrelcoated fiber was bonded directly to the outer surface of the steel bars using the "soft" adhesive.The specimens were symmetrically equipped with two embedded Epsi-lonSensors, as shown in Figure 11.Strains in the Epsi-lonSensors and Hytrel-coated fiber were measured using a four-channel coherent frequency domain reflectometer (ODiSI 6104, LUNA Innovations Incorporated, USA).In parallel with the tensile tests, a camera system was used to evaluate the strain and crack evolution on the concrete surface based on Digital Image Correlation (DIC).
In addition to locating the cracks, the DFOS can accurately determine the time of each crack initiation, which is not possible with the global measurement of force and displacement, such as the classical load-time regime shown in Figure 12, where only the global crack event can be seen.In addition, the locations of the cracks detected by DFOS are exactly the same as those detected by digital image correlation using a camera system, as shown in Figure 14 (top).However, the cracks were visible to the naked eye shortly after the first strain peaks in the DFOS measurement during the test.Thus, the embedded DFOS can detect the time and location of crack events in the specimens tested.
The evolution of the strains along the concrete body over time is shown in the 3D plot, i.e. strain as a function of time and sensor position in Figure 13 for the Epsilon sensor.The two embedded EpsilonSensors behave symmetrically and show the same trend of the concrete and identical crack width, each peak representing a crack in the concrete body.Coordinates, time, and strain are represented in DFOS by the x, y, and z axes, respectively.To better illustrate the distribution of DFOS strains around the cracks, the 2D plot, i.e., strains as a function of sensor position, is shown in Figure 14            On the other hand, the Hytrel fibre bonded to the rebar does not smoothly represent the behavior of the rebar during the test period, as shown in Figure 17.However, this is due to the very small amount of adhesive material around the Hytrel fibre, about 1 mm in width.For a better representation of the reinforcement strains, it is recommended that the DFOS fibre be bonded to the rebar within a tiny notch filled with adhesive material.

Conclusion
This paper presents the possibilities of using DFOS for crack detection in concrete structures by using different DFOS sensor cables and coherent frequency domain reflectometry.FOS in combination with c-OFDR enables crack detection in concrete structures.The selection of DFOS cables and their application was essential for crack detection.DFOS bonded to concrete surfaces were suitable for detecting cracks propagating in concrete structures.However, the coating layers of the DFOS affect the measured strain values due to the loss of measurements through the different layers of the coating.In addition, the adhesive materials used in the tests were found to have less influence on the measured strains.In the case of the DFOS embedded in the concrete matrix, sensor cables were suitable for crack detection and were easier to apply compared to the surface DFOS.However, both the Hytrel fibre bonded to the surface of the concrete structure and the EpsilonSensor embedded in the concrete body were suitable for crack identification.It was evident, the Epsi-lonSensor has a better bond to the concrete and was, therefore, more suitable for crack quantification.Despite its high efficiency, further research is needed before the DFOS technique can be applied to real structures, such as studying the performance of DFOS under cyclic loading cases.In addition, a practical solution will be developed to deal with the very large amount of data generated by such measurement systems, especially in the case of high sampling frequencies, long fibres, and long measurement periods.

Acknowledgement
This work was carried out with the support of the Carl Zeiss Foundation as part of the "Beton 2.0" research project, which the authors would like to acknowledge.We would also like to thank the MFPA Weimar for supporting this research with their extensive experience in DFOS systems

Figure 1 3 -
Figure 1 3-Point bending test: the schematic layout (Top); the experiment arrangements at the laboratory (Bottom).

Figure 2 A
Figure 2 A photo of the bottom side of the concrete beam used in the 3-point bending test.

FOS
slip between the fiber and the surrounding concrete matrix.

Figure 3
Figure 3The stains measured along the DFOS of the Hytrel-coated fiber using the soft adhesive material with a modulus of elasticity of 260 Mpa.

Figure 4
Figure 4The stains measured along the DFOS of the Polyamide-coated fiber using the soft adhesive material with a modulus of elasticity of 260 Mpa.

Figure 5
Figure 5 Load time-history of the DFOS Hytrel-coated fiber using the "soft" adhesive material (top); Comparison of the measurements of DMS to the DFOS (bottom).

Figure 6
Figure6 Strains at the measuring strip III, for the "hard" and "soft" adhesive materials in the case of Hytrel-coated fiber.
(bottom).The x-axis represents the sensor position, i.e. the coordinates along the concrete body.The y-axis represents the time of the test in seconds.It can be seen that six cracks have propagated in the concrete body.In Figures15 and 16, the strains measured within the embedded EpsilonSensor are plotted at specific time intervals during the test duration to show the strain distribution along the fiber.Each peak along the plotted curves represents a crack tip.The distribution of strains around the cracks in each curve increased with increasing load due to slip between the fiber and the surrounding concrete matrix.

Figure 8
Figure 8The layout of the Cylinder tension test.

Figure 9 A
Figure 9 A photo of the EpsilonSensor fibre embeded into the concrete matrix.to the steel bars.

Figure 10 A
Figure 10 A photo of the Hytrel fibre bonded to the steel bars.

Figure 11
Figure 11 the general arrangement of DFOS within the cross-section of specimens.

Figure 12
Figure 12 Load-displacement curve of the tension test of the reinforced concrete cylinder.

Figure 13
Figure133D-plot of strain versus time and sensor position along the concrete sample over the test duration using EpsilonSensor (strain values indicated with colour in the scale bar).

Figure 14
Figure 14 2D contours illustrate the distribution of strains within the DFOS around the cracks (bottom); Crack detection using DIC (top).

Figure 15
Figure 15The strains measured along the EpsilonSensor DFOS fibre at specific time intervals using Channel 01 with correspondence crack widths in grey boxes.

Figure 16
Figure 16The strains measured along the EpsilonSensor DFOS fibre at specific time intervals using Channel 02 with correspondence crack widths in grey boxes.

Figure 17
Figure 17The strains measured along the Hytrel DFOS fibre at specific time intervals using Channel 03.The EpsilonSensor shows a better bond to the surrounding concrete when compared to the behavior of the bonded . The authors would like to thank Dr.-Ing.A. Flohr, T. Lehmphul of the F.A. Finger Institute for Building Materials