DFOS Applications to Geo-Engineering Monitoring

Optical fiber sensing technology has developed rapidly since the 1980s with the development of the optical fiber and fiber optical communication technology. It is a new type of sensing technology that uses light as a carrier and optical fiber as a medium to sense and transmit external signals (measurands). Distributed fiber optical sensors (DFOS) can continuously measure the external physical parameters distributed along the geometric path of the optical fiber. Meanwhile, the spatial distribution and change information of the measured physical parameters over time can be obtained. This technology has unmatched advantages over traditional point-wise and electrical measurement monitoring technologies. This paper summarizes the state-of-the-art research of the application of the distributed optical fiber sensing technology in geo-engineering in the past 10 years, mainly including the advantages of DFOS, the challenges in geo-engineering monitoring, related fundamental theoretical issues, sensing performance of the optical sensing cables, distributed optical fiber monitoring system for geo-engineering, and applications of optical fiber sensing technology in geo-engineering.


Introduction
China is one of the countries with the most severe geohazards and the most threatened population in the world. Due to complex geological conditions and frequent tectonic activities, hazards such as rockfall, landslides, debris flows, the ground collapses, ground subsidence, and ground fissures are widely distributed and highly concealed, sudden and destructive, and are difficult to prevent. In recent years, affected by extreme weather, earthquakes, and engineering constructions, geohazards occurred frequently, causing serious losses to people's lives and properties. According to the Ministry of Natural Resources of China, the direct economic loss from geohazards in 2019 alone reached 2.77 billion yuan. On the other hand, in large-scale foundation engineering constructions, various geotechnical engineering problems emerged, such as foundation pit failure, tunnel deformation, and leakage, diaphragm wall collapse, and differential settlement. These have brought many hidden dangers to engineering safety, greatly increased the construction and operation costs, prolonged the project cycle, and caused a lot of life and property losses. Therefore, preventing and controlling geohazards and solving all kinds of geotechnical engineering problems is the central task for geo-engineering researchers and practitioners and a major national demand, and monitoring is a prerequisite for this work.
Geohazards and geotechnical engineering are featured with large scale, multi-field effects, complex influencing factors, strong concealment, spanning multiple regions, harsh environments, high real-time monitoring requirements, and long monitoring periods. Because of these characteristics, the current point-wise electrical sensing technologies are still difficult to meet the needs of hazard prevention and mitigation, and have brought huge challenges to hazard warning, forecasting, and prevention.
The fiber optic sensing technology, which has developed rapidly with the development of optical fiber and fiber optic communication technology since 1980s, has the advantages of distributed and long-distance measurement, anti-interference, and corrosion resistance. Therefore, it has been increasingly used in geohazard and geotechnical engineering monitoring [1]. In the early days, the quasi-distributed fiber Bragg grating (FBG) technology and the fully distributed optical time-domain reflectometry (OTDR) technology were mainly used. In the past ten years, with the development of fully distributed fiber optic sensing technologies such as Raman optical time-domain reflectometry (ROTDR), Brillouin optical time-domain reflectometry (BOTDR), Brillouin optical time-domain analysis (BOTDA), and Brillouin optical frequency-domain analysis (BOFDA), the fully distributed monitoring technology for geo-engineering has also developed rapidly. Iten and Puzrin [2] buried fiber optic cables in an asphalt road passing through a creeping landslide and localized the landslide boundary by monitoring the deformation of the asphalt with BOTDA. Habel et al. [3] developed a GeoStab sensor based on the FBG technology and accurately detected the position of the sliding surface of a slope in an open coal pit. Soga's group [4][5][6][7][8][9] applied the BOTDR technology to foundation pile testing and tunnel deformation monitoring. Researchers [10][11][12] from the University of Campania "Luigi Vanvitelli" (formerly the Second University of Naples) in Italy applied BOTDA to laboratory landslide model tests, revealing the failure mechanism of pyroclastic soil slopes under rainfall infiltration. Based on the FBG sensing technology, Huang's group [13][14][15] has successively developed various types of fiber optic sensors suitable for slope monitoring, such as pressuremeters, pore pressure gauges, displacement gauges, and axial force gauges, which have been successfully applied to field landslide monitoring. Yin's group [16][17][18][19] developed FBG-based sensing rods, inclinometers, displacement meters, etc., and successfully applied them to GFRP soil nail pullout tests, dam model tests, debris flow, and slope monitoring. In Chinese mainland, with the rapid development of infrastructure construction and the urgent need for monitoring technology, a large number of research and application results related to geo-engineering fiber optic monitoring emerged. The authors' group [20] has carried out a lot of research on distributed monitoring technology of strain, temperature, and deformation in geo-engineering, and achieved many important results. In the past five years, research on the distributed sensing of water content and seepage of rock and soil mass has also made breakthroughs [21][22][23]. Liu et al. [24] used the BOTDR technology to conduct long-term monitoring of anti-slide piles of a highway slope. Chai et al. [25] applied the BOTDA to three-dimensional model tests of coal mining.
Combined with the research results of the authors' group, this review makes a general summary of research on fiber optic monitoring in geo-engineering in the past ten years.

Characteristics of geological body and monitoring requirements
Different from man-made structures such as reinforced concrete structures, steel structures, and synthetic material structures, geological bodies are the products of natural history and are a multiphase system consisting of solid, gas, and liquid. The rock mass is hard and its structure is complex. On the other side, the soft soil not only has the characteristics of porousness and low strength but also is constantly affected by natural and human engineering activities. From the perspective of geo-engineering monitoring, geological bodies have the following characteristics and put forward higher requirements for the monitoring technology.
(1) Complex structure and large spatial variability After a long process of geological evolution, geological bodies have a very complex structure, large spatial variability, and high uncertainty. Therefore, the monitoring technology should have the function of distributed monitoring to prevent missing key factors of geological bodies.
(2) Large scale, long distance, and great depth The scale, scope, and depth of geological bodies related to various projects are much larger than the structures. Such as large-scale slopes with a volume of tens of millions of cubic meters, river embankments with a length of tens or even hundreds of kilometers, and mines with an influence depth reaching thousands of meters. Therefore, to understand the dynamic process of such a large geological body, a long-distance and well-covered monitoring system is necessary to obtain sufficient information on various parameters of the geological body.
(3) Weak penetration and strong concealment It is hard to go to the sky and even harder to enter the ground. The geological body is composed of the rock mass, soil, and fluid, which is not easy to penetrate for the detection signal. Meanwhile, underground lithology and environmental factors are complex and difficult to grasp intuitively. This poses a huge challenge for the installation of monitoring components.

(4) Multi-field effects and complicated influencing factors
Since the surface of the crust is at the junction of the lithosphere, hydrosphere, atmosphere, and biosphere, it is bound to be coupled by multiple fields such as stress, temperature, moisture, water flow, and chemicals. This leads to the demand for multi-fields and multi-parameters monitoring function for the monitoring technology.
(5) Irregular shape and diverse geological environment The shapes of natural geological bodies and engineering rock-soil bodies are generally irregular, forming a variety of landforms. The irregularity of the shape causes difficulties to the installation of monitoring components, data processing, and mechanism analysis. The geological environment is complex and changeable, such as high mountains and deep valleys, high and low temperatures, and shallow and deep subsurface parts. This brings great challenges to the installation and protection of the geological monitoring system, as well as the reliability and durability of the system.

Shortcoming of conventional monitoring technologies
It is not easy to obtain multi-field and multi-parameter of geological bodies and the geotechnical engineering structures accurately, timely, and on a large scale. It requires an advanced monitoring system and advanced theories and methods.
At present, conventional monitoring technologies for geological and geotechnical engineering mainly use vibrating wire sensors and resistive sensors in point monitoring mode. Figure 1 summarizes the commonly used contact-type monitoring technologies for rock and soil deformation. The limited monitoring points of these technologies may lead to the problems of missed detection and blind spots. They are not suitable for the large-scale, long-distance, and great-depth monitoring in rocks and soils. Meanwhile, sensors are often subjected to harsh environments such as high temperature, low temperature, high pressure, and high humidity, which are great threats to the sensors and lead to low survival rates, corrosion, and poor durability. Sensors based on the principle of electromagnetic induction are susceptible to electromagnetic interference. They may have a long-term zero drift, which causes poor monitoring accuracy and stability. Generally, the deficiencies of these monitoring technologies obstruct a better understanding of the dynamic process of geological disasters and geotechnical engineering problems and therefore weaken the ability of long-term monitoring, prediction, and early warning. It is necessary to develop advanced geological and geotechnical engineering monitoring technologies to provide strong support to the theoretical research and engineering applications in geological disasters and geotechnical engineering.

Advantages of DFOS and its challenges to geo-engineering monitoring
In the optical fiber sensing technology, the distributed fiber optical sensors (DFOS) is very suitable for geo-engineering long-distance, long-period, and distributed monitoring because of its advantages such as distributed measurement capability, anti-electromagnetic interference, corrosion resistance, flexiblity, and ease to implement network monitoring. In DFOS techniques, the sensing fiber is structured into a one-dimensional, two-dimensional, or three-dimensional network according to a certain topology. It is like a sensory neural network, which is planted in a "dead" geological body, forming a geo-aware system (Fig. 1) and monitoring the relevant changes of measurement object in longitude, plane, and three dimensions. This technique overcomes the disadvantages of the traditional point monitoring method and improves the success ratio and effectiveness. Therefore, DFOS is particularly useful for overall strain and temperature monitoring in large geo-engineering such as uneven land subsidence in long-distance tunnel, ground crack distribution and deformation on a large scale, oil and gas pipeline leakage, dam and embankment leakage, and slope stability. The characteristics and deficiencies of distributed fiber optic sensing techniques in geo-engineering are listed in Table 1.
As shown in Table 1, the DFOS technology is a collection of various sensing techniques with different measurement variations and principles, and each technique has its characteristics and shortcomings. Therefore, suitable technology with a specific monitoring scheme should be selected, according to different monitoring objects and requirements. However, as mentioned in Section 1, because of the complex structure of the geological body, the multi-cracking properties of the rock mass, the multi-pore structure of soft soil, and the random behavior of rock and soil in multiple fields, the application of DFOS in geo-engineering monitoring faces many challenges. (1) Theoretical challenges. To apply the fiber optic sensing technology to geo-engineering monitoring, three basic theoretical problems must be solved. (a) Standards for evaluating the performance of the bare fiber encapsulated in optical cables should be established to provide a theoretical basis for the development and correction of sensing cables. (b) Sensing optical cables must be coupled with rock and soil to be tested for accurate measurement, hence a criterion for evaluating the coupling is required. (c) It is necessary to establish the corresponding characterization model between the basic quantities measured by DFOS and other measured parameters to realize the multi-parameters monitoring of geological engineering with DFOS.
(2) Technical challenges. According to the requirements of geo-engineering monitoring, three technical bottlenecks of sensing cables must be solved. (a) How to develop a sensing fiber optic cable that is both robust and sensitive? (b) The measuring scope of ordinary optical fibers generally does not exceed 2%, and how to develop sensing cables to meet the monitoring requirements of large deformation? (c) To estimate the moisture, seepage, and micro-seismic monitoring of geological bodies, special sensitivity enhanced fiber should be developed.
(3) Application challenges. Besides theoretical and technical challenges, the application of fiber optic sensing techniques to geo-engineering monitoring requires the corresponding DFOS monitoring system and construction method to install and maintain optical cables for different monitoring quantities.
In recent years, many researchers have carried out fruitful research and achieved a series of the breakthrough. Thus, geo-engineering DFOS monitoring techniques are greatly promoted, and the application field is extended.

Strain transfer analysis
Bare fibers less than 0.25 mm in diameter are too fragile to be used directly as sensing elements for geoengineering monitoring; they should be well coated and jacketed to survive in the harsh geologic environment (Fig. 2) [26]. However, the sensitivity of these cables will inevitably be affected by the multi-buffers. Strain transfer analysis is therefore important to retrieve the actual strain present in the monitored soil or rock. Most existing strain transfer models have assumed that the strain of host material is evenly distributed [27,28]. We performed elaborate matrix-to-fiber strain transfer analysis considering uneven matrix strain distributions and matrix-fiber interface decoupling (Fig. 3) [29], which is summarized here. In the following descriptions, r and D represent the radius and diameter, respectively; τ and ε represent the shear stress and normal strain, respectively; G and E represent the shear modulus and Young's modulus, respectively; the subscripts m, p, and g, denote the matrix, protective layer, and fiber core, respectively.  We first considered a parabolic matrix strain distribution ( Fig. 4): where m ε is the matrix strain; L is the half of the fiber embedment length; 0 ε is the maximum strain at the midpoint; x is the horizontal axis. Based on the shear-lag theory, we derived the strain transfer ratio as where k and n have the following forms, respectively: We then investigated the situation where the matrix-fiber interface decoupling occurs during the monitoring process. Figure 5 shows the assumed distribution of interface shear stress mp τ which can be divided into three zones. In the first zone adjacent to the fiber head (0≤l≤L move ), the matrix and fiber decouple and mp τ equals to the residual interface shear strength res τ . In the second zone (L move <l≤L s ), mp τ is linearly distributed, decreasing from the peak interface shear strength max τ to zero. In the third zone (l>L s ), mp τ is not mobilized and the monitored strain is equivalent to the matrix strain.
and move L and s L can respectively be derived as 2 move max res If max τ and res τ are determined, the lengths of the first and second zones can be calculated.

Cable-soil coupling evaluation
When the sensing cables are installed in the ground for monitoring, the coupling between the cable and the soil directly affects the accuracy of fiber-optic measurements, among which the deformation coupling is the most critical [1,[30][31][32][33][34][35][36].

Shear strength criterion for interface bond
A shear strength-based criterion was proposed for determining the bond state of a cable-soil interface [32,37] and was further developed by Zhang et al. [38]. Laboratory evidence showed that the cable-soil interface failed progressively, exhibiting strain-softening, or strain-hardening behavior depending on the soil properties [32,33]. Hence, by comparing the interface shear stress with the peak or residual shear strength ( max where max c and max δ are the peak cable-soil interface cohesion and friction angle, respectively; res c and res δ are the residual cable-soil interface cohesion and friction angle, respectively; h σ is the confining pressure normal to the cable axis. On the other hand, the cable-soil interface shear stress c τ may be determined by using K is the lateral earth pressure coefficient, and γ is the bulk unit weight of the overlying ground. For ε , a small cr h is required to guarantee a good cable-soil interface bond which is desirable in engineering practice.

Microanchor failure criterion
The cable-soil interface is likely to fail under low confining pressures. To address this problem, researchers have proposed to install microanchors on the cable coating to increase the cable-soil interface shear strength [12,[39][40][41][42]. In this regard, the deformation coupling between a microanchored cable and surrounding soil can also be evaluated by using the shear strength criterion [43,44]. Alternatively, the microanchored cable-soil deformation coupling may be assessed by using the geotechnical bearing capacity theory (Fig. 6) [45]. The force a F applied on a microanchor (interaction force between the microanchor and soil) can be derived based on the measured strain: where c ε Δ is the difference in strain measured by the two adjacent unanchored cable segments. For the three microanchor types shown in Fig. 6, the ultimate anchor-soil interaction force ar F can be derived from the bearing capacity theory as ( ) If the microanchor has not reached its capacity (i.e., ar a F F < ), it is considered that the cable couples to the surrounding soil and the strain measurements are reliable. Fig. 7 Three typical fiber-optic cable microanchors: (a) disc, (b) cylinder, and (c) spindle.

Cable-soil coupling index
To make the cable-soil coupling evaluation operational, laboratory pullout tests can be conducted where a fiber-optic cable is gradually pulled out from a confined soil under increasing pullout displacements and the strain distributions along the cable are simultaneously measured. We devised such a displacement-controlled pullout apparatus as shown in Fig. 8 [46]. The apparatus consists of (1) a pressure chamber where a confining pressure can be applied to a specimen up to megapascals; (2) a tensile tester for applying pullout forces/ displacements to a fiber-optic cable embedded in the specimen; (3)  In a pullout test, a series of strain profiles can be obtained. For a given strain profile, the strain propagation length d ε can be determined as the farthest propagation length of strain from the cable head toward the toe. Further, the maximum strain propagation length for all the strain profiles under a certain confining pressure is denoted by max d ε . Based on this concept, we proposed a cable-soil coupling index c s ζ − to quantitatively evaluate the cable-soil coupling, which is given by where 0 L is the length of the cable embedded in the specimen. According to this index, the cable-soil coupling condition can be classified; Table 2 shows such an example.

Transformation representation models for multi-parameter sensing
The basic physical quantities directly measured by DFOS are limited, mainly including temperature, strain, and vibration. To monitor other physical quantities (e.g., water content, seepage, density, and humanity) from measurements of the basic ones, a transformation representation model must be established between them to achieve multi-parameter geoengineering monitoring.
The actively heated fiber-optic (AHFO) method can estimate water content and dry density using temperature rise (T t ). In this method, a cable or alundum tube can be heated for a specific time, and T t can be detected by the fiber-optic inside. Because of the sensor's sufficiently large length to diameter ratio, T t can be given by the solution of the line heat source: where T t is the temperature rise ( ), a ℃ lso known as the temperature characteristic value, T(t) is the temperature of the heat source ( ) corresponding to ℃ the heating time t, T 0 is the initial ambient temperature ( ), ℃ Q is the heating power per unit length (W⋅m -1 ), λ is the thermal conductivity of soil (W⋅m -1 ⋅K -1 ), R is the thermal resistance per unit length (m⋅K⋅W -1 ) between the sensor and the soil wall, K is the thermal diffusivity (m 2 ⋅s -1 ) of the soil, a is the outer diameter (m) of the sensor, and c = 1.781 1 = exp (γ) where γ is the Euler-Mascheroni constant (= 0.577 2).
The thermal conductivity is the main factor determining the temperature rise. The thermal diffusion coefficient has a small contribution to the temperature rise. Equation (17)  , and p QR = .

Water content
Firstly, we proposed an empirical function that can be used to quantify the relationship between the temperature characteristic value (T t ) and the soil moisture (w) [ Fig. 9(a)] [47]: where A, B, C, D, and E are fitting parameters, and F is the threshold mass water content. According to Tarnawski and Leong [48], the threshold volume water content of fine-grained soil is generally 0.05 m 3 /m 3 -0.10 m 3 /m 3 . Considering the complex parameters and the intangible threshold in (19), a semi-empirical model is put forward based on the Côté and Konrad model [49]: where K e is the Kersten number, which is a function of the saturation S r , λ dry is the dry thermal conductivity (W⋅m -1 ⋅k -1 ), and λ sat is the saturated thermal conductivity (W⋅m -1 ⋅k -1 ). K e is given by where κ is an empirical parameter, which is related to the soil type. S r can be expressed as / where θ is the volume water content (m 3 /m 3 ), n is the void ratio, ρ d is the dry density (Mg/m 3 ), and ρ w is the density of water (Mg/m 3 ). According to (18), (20), (21), and (22), the relationship between T t and w can be obtained:  (23) is suitable for unfrozen soil. When the soil temperature is below zero, the water in the soil solidifies into ice, and (23) is still applicable. However, it should be noted that the thermal conductivity of frozen soil (λ satf ) varies with the different states of water in frozen soil [ Fig. 9(b)]. Therefore, the values of fitting parameters A, B, and C in (23) will be different. For unfrozen soil and frozen soil, the expressions of saturated thermal conductivity λ satu and λ satf are as follows: where λ s is the thermal conductivity of soil particles, λ w is the thermal conductivity of water, n is the soil porosity, λ i is the thermal conductivity of ice and its value is 2.24 W⋅m -1 ⋅K -1 , and ϕ u is the volume fraction of unfrozen water in the soil. The empirical parameter κ is different in both unfrozen soil and frozen soil. Therefore, the frozen soil calibration formula can be determined by the formula of unfrozen soil and the relationship between the calibration fitting parameters.
The parameters A u , B u , and C u are obtained by calibration test of unfrozen soil. For frozen soil, λ satf and κ can be determined by using A u , B u , C u , and measured (w, T t ) before and after freezing, to further determine the fitting parameters A f , B f , and C f of frozen soil, thus obtaining the calibration formula of frozen soil and realizing the water content measurement in frozen soil.

Dry density
Besides the water content, the thermal conductivity is also affected by dry density which can be described by the Singh and Devid model [ where d is a constant related to soil type. Combining (18) and (26), an exponential function can be obtained to describe the relationship between T t and ρ d :

Humidity
The humidity measurement is realized by the ORMOCER ® -coated FBG (Fig. 11) [51]. When the ambient humidity changes, the volume of ORMOCER ® will change, resulting in the axial straining of the FBG. Since the ambient temperature may change in the process of humidity measurement, an uncoated FBG is employed for temperature compensation. Based on the above analysis, the formula for determining the humidity and temperature of the OR-coated FBG humidity sensor is where T Δ is the temperature variation, RH K is the humidity sensitivity coefficient, RH Δ is the relative humidity variation, T1 K and T2 K are the temperature sensitivity coefficients of the OR-coated FBG and uncoated FBG sensing elements, 1 A mathematical matrix of the OR-coated FBG humidity sensor used to calculate humidity and temperature is presented as follows:

R&D of robust sensing cables
As the "sensing nerve" in the monitoring of geological engineering, the optical sensing cables are both sensors and signal transmission medium, which should be both strong and sensitive to meet the requirements of multi-range and multi-parameter monitoring of geological engineering. For the DFOS monitoring of geological engineering, it is necessary to optimize the structure of the optical sensor cables according to the monitoring requirements, conduct performance testing and parameter calibrations through laboratory and field tests, and propose the installation and on-site protection methods. Figure 12 shows five types of strain optical sensing cables developed by the authors' group.

Structure and type of strain optical sensing cable
(1) Thin sheathed strain optical sensing cable. It has the characteristics of a thin-layer sheath and good strain transmission, but the strength is low, as shown in Fig. 12(a). It is mainly used to package high-strength sensing optical cables, distributed pressure sensors, distributed stress sensors, etc.
(2) Low elastic modulus strain optical sensing cable. It has a soft tight jacket with low elastic modulus and good coupling with low-modulus materials as shown in Fig. 12(b). It is mainly used for strain monitoring in indoor and outdoor soil model tests.
(3) High strength strain optical sensing cable. It adopts metal stranded wire tightly wrapped structure, which has the characteristics of high strength, high elastic modulus, and good straightness, as shown in Fig. 12(c). It can be used for strain and deformation monitoring of rock mass and concrete structures.
(4) High stability strain optical sensing cable. It adopts integrated packaging with high-temperature resistance and low creep materials, which has the characteristics of high-temperature resistance and high long-term stability, as shown in Fig. 12(d). It is mainly used for monitoring the creep and long-term deformation of surrounding rocks.
(5) Ribbon strain optical sensing cable. It is encapsulated into a ribbon-type cable with metal sheets, fiber cloth, etc. for strain sensing, which is easy to be installed on the surface by welding or pasting, as shown in Fig. 12(e). It is mainly used for the monitoring of structures such as pipelines, tunnels, and steel pipe piles.
The sensing cables for distributed temperature sensing (DTS) and distributed vibration sensing (DAS) were designed for applications [20,52].

Fatigue characteristics
Under the condition of long-term tension, the materials of the optical cables will show recoverable elastic deformation and unrecoverable plastic deformation. There will be relative displacement between the various buffers of the optical cable. The fatigue characteristics of the optical sensing cable directly affect its long-term stability. The authors' group developed a device to evaluate the fatigue characteristics of the sensing optical cable according to the strain decay curves after low cyclic tension. The cable strain decayed exponentially, and the decay process involved quick decay, slow decay, and stabilization stages, as shown in Fig. 13. The sensing cables with different jackets differed in the amount and rate of decay. For the same cable, the strain decay increases with the initial strain and tensile frequency. As the unstrained period increased, the initial strain levels of the strain decay curves approached those of the first cyclic elongation. The proposed cyclic fatigue testing method can be used to identify the fatigue characteristics of distributed strain sensing cables with different jackets and can serve as a reference in assessments of cable performance during long-term use [53].  Fig. 13 Strain decay curve of the tested PU-jacketed sensing cable at a tensile frequency of 3 Hz.

Temperature sensing performance
To compare the temperature sensing performances of different cables, the authors' team conducted a comparison test in a walk-in constant temperature and humidity laboratory by using BOTDR [54]. Figure 14 shows the relationship curve between the Brillouin frequency shift and temperature of different optical sensing cables. It can be seen that the temperature coefficients of the optical cables with variant jackets are different. The temperature coefficient of some optical cables will change at a certain temperature. Therefore, when the optical sensing cables are used in high-temperature environments, the high-temperature-resistant cables should be developed, or the test results should be calibrated and corrected.

Multi-range deformation measurement
Because the measurable strain range of quartz optical fiber is generally less than ±2.0% to realize distributed monitoring of large deformation of rock and soil, a specific encapsulation of the sensing cable and the installation method should be adopted to achieve multi-range deformation monitoring according to the characteristics of rock and soil deformation. In summary, there should be three solutions [20]: (1) For the deformation monitoring with a measuring range less than 0.5%, it can be realized by using FBGs and optical sensing cables with fully contact installation, with which the FBG sensors and the sensing cables are directly attached on the surface or embedded in the measured object, as shown in Fig. 15(a). (2) For the deformation monitoring with a measuring range between 0.5% and 2%, the optical sensing cable fixed by points can be used to achieve homogeneous deformation, as shown in Fig. 15(b). The core is fixed with the jacket by points, at which the sensing cables are fixed with the measured object. Beyond the fixed point, the fiber core is free and is loosely sleeved. In this way, different strain ranges can be achieved by adjusting the distance between the fixed points.
(3) For the deformation monitoring with a measuring range between 2% and 50%, it can be realized by encapsulating a multi-point displacement meter with a spring or other deformation conversion modes, as shown in Fig. 15(c).

Actively heated fiber-optic cable
According to the principle of heat exchange, when using the DTS technology to monitor water and seepage fields in geological engineering, the actively heated fiber-optic cables are the basic sensing element. Figure 16 shows the structure of two sensing cables, the carbon-fiber heated cable (CFHC) and the metal-net heated cable (MNHC) [20,22]. CFHC consists of a multimode optical fiber, carbon-fibers, and a cable jacket. CFHC has large resistance, low required voltage, and strong corrosion resistance. CFHC is suitable for short-distance (<500 m) distributed monitoring of soil moisture. The structure of MNHC is similar to that of CFHC. However, MNHC is heated by the metal net. MNHC has the characteristics of low resistance, and long-distance and low-temperature value, which is more suitable for long-distance and large-area soil moisture monitoring.
Because the optimal spatial resolution of DTS is 1 meter, for moisture monitoring with high spatial resolution, the heated cables can be wrapped around a tube or a bar to form a heated sensing tube. Figure  17 shows the structure and the photo of the carbon fiber heated sensing tube (CFHST for short) wrapped by CFHC [22].

FOS-based geo-engineering monitoring systems
To apply the DFOS technology to geo-engineering monitoring, besides the development of special fiber optic sensors and sensing cables, another important task is to develop advanced monitoring systems. A distributed fiber optic monitoring system for geo-engineering is generally composed of the data acquisition subsystem, the data transmission subsystem, the data analysis subsystem, the decision-making subsystem, etc. Each subsystem is interrelated and indispensable, and each part is an organic part of the whole system. Due to the low degree of standardization of the fiber optic sensors, different types of sensors generally require specific data acquisition subsystems. Thus, once the sensing technology is selected, the corresponding data acquisition and processing technique is determined. That means the optimal placement of the sensors and the signal processing and analysis part are two key issues of a fiber optic monitoring system. In engineering practices, the monitoring information of a certain object obtained by fiber optic sensors is sent to the monitoring center through the data acquisition and transmission subsystems for data processing and assessment. On this basis, the stability state of the object under monitoring is evaluated. If the detected key parameters exceed the predetermined thresholds, the relevant management agents will be informed timely through the short message (SMS), E-mail, and other means, to take corresponding emergency measures to avoid significant loss of personnel and property.
As shown in Fig. 18, there are four subsystems in a distributed fiber optic monitoring system, i.e., the data acquisition subsystem, the data transmission subsystem, the data analysis subsystem, and the decision-making subsystem [55]. The basic functions of each subsystem are: (1) The establishment of the data acquisition subsystem mainly involves the selection of various fiber optic sensors and the sensing cables, the selection of suitable modulation methods, and then the determination of the topology of fiber optic sensors. On this basis, the sensor layout modes are selected (for example, the surface adhered or internal embedded type). Finally, how to realize the modulation signal demodulation and how to establish the field monitoring and control station will be considered. The terminal sensors are mainly distributed fiber optic sensors with different functions (including quasi-distributed), which are used for monitoring deformation field, temperature field, and seepage field and stress field of rock and soil mass.
The DFOS technology can be categorized into intensity modulation, phase modulation (interferometric), wavelength modulation, frequency modulation, and polarization modulation. Correspondingly, there are several demodulation technologies. Different types of modulation and demodulation instruments are combined to construct a DFOS monitoring system. By using the virtual instrument technology, the demodulator is automatically controlled, parameters are set and data are collected and stored by the controlling PC. Data acquisition should be remote, networked, and automated, and be combined with the database technology to realize local storage of data [56]. Fig. 18 Illustration of the geoengineering monitoring system.
(2) The data transmission subsystem connects the data acquisition subsystem and the data analysis subsystem, including the communication and data exchange between the Internet or wireless network and the remote terminal processor, and therefore the storage structure and mode of massive real-time data will be set. The data transmission subsystem mainly consists of three parts: the first part is the transmission from the signal modem to the sensor elements; the second part is the transmission from the signal modem to the data analysis subsystem; the third part is to quickly transmit the monitoring analysis results to the clients. The former is generally realized by wire transmission. However, with the continuous improvement of smart sensor technologies, the use of wireless transmission will be more and more preferred. In recent years, the employment of wireless, Internet, and communication satellites to transmit signals has been very popular. Since the DFOS systems for geo-engineering monitoring are often installed in very harsh local environments, it is hard for people to reach, so the use of wireless transmission should be the trend. Some systems already have a data analysis system installed in the signal modem, in which case the second part of the transmission is not required. Finally, there are many ways to quickly convey the monitoring analysis results to users, among which it is an inevitable trend to send the monitoring results and early warning signals to mobile phones, computer terminals, and public service platforms through Internet.
(3) The data analysis subsystem is the core part of the monitoring system, which is based on software technologies, such as geographic information system (GIS) and database. This subsystem is in charge of mass monitoring data processing (including information query, contrast, and extract.), management, validity analysis, selection of monitoring parameters, construction of the database, graphical presentation, etc.
With the improvement of the monitoring technologies, the diversification of monitoring information has been achieved. The management of monitoring data is not only the storage, query, and display of monitoring data, but interactive three-dimensional management combined with the spatial geography. Therefore, the management of data must enable the remote processing of updated data acquisition and direct visualization and can simulate all kinds of emergencies combined with virtual reality, to distinguish various events for the influence of engineering safety. The realization of these goals greatly increases the geoengineering monitoring level. Especially for the massive fully distributed monitoring data, they often need to be processed, including denoising, smoothing, coordinating, and subtracting. More importantly, it is necessary to identify, capture, and display abnormal points and areas of the monitoring data, and to get evaluation conclusions or issue prediction and early warning of the objects under monitoring, which all require the data analysis subsystem to complete. In recent years, a series of "soft computing methods", including fuzzy logic, artificial neural network, genetic algorithm, wavelet analysis, machine learning, and other computational intelligence technologies, have been developed rapidly. That makes it possible to conduct large-scale computing reasoning and parallel inference and provides a solid theoretical basis for health diagnosis in geoengineering.
(4) The decision-making subsystem has the functions of report output, safety evaluation, and the corresponding pre-disaster warning and forecasting.
According to the results of monitoring and analysis, users can make corresponding decisions and give timely warning to the hidden dangers of engineering safety.
In the decision-making process, the combination of traditional and modern methods is needed, such as mathematical statistics, numerical analysis, expert system, rough set theory, extension engineering theory, and dynamic fingerprint technique. In this way, the stability or health conditions of the object to be measured can be evaluated with high reliability. In addition, comprehensive analyses and parametric inversion of the calculation model can be conducted. Combined with the geological information, field observations, abnormal signals, and other related data, the forecasting of potential hazards can be realized [56].
Finally, appropriate emergency treatment measures should be taken, such as setting up warning areas, soil reinforcement, unloading, and evacuation, to minimize the risk of the loss of human life and properties.
From the point of view of engineering safety monitoring demand and technology development, distributed monitoring is the development trend of engineering monitoring.
In the future, combined with the GIS and database technology, the wireless data transmission technology enables the setup of monitoring networks at the engineering level, city level or regional level, and even nationwide level. In these monitoring networks, monitoring information center, fiber optic sensing network, and wired and wireless transmission serve as the core, the monitoring optical fiber network main entity, and the carrier, respectively. By this means, the normal operation of geo-engineering infrastructures can be wired and wireless transmission for the media monitoring system will run to the health of the geological and geotechnical engineering, thus providing a reliable guaranteed.

Applications to geotechnical engineering monitoring
In recent years, the optical fiber sensing technology has been widely used in the field of geo-engineering monitoring and has been rapidly developed with its unique advantages. In the detection and monitoring of foundation engineering, especially the detection of pile foundations, distributed optical fiber sensing technologies such as BOTDR/A and BOFDA can be used to perform fully distributed detection of the strain, force, and flexure status of the pile body [4,[57][58][59]. In tunnel and underground engineering structural health monitoring, DFOS has prominent advantages, unveiled in a series of research [6,8,[60][61][62][63][64][65]. In the deformation and stability monitoring of supporting structures such as foundation pits and retaining walls, the practice has proved that the distributed optical fiber sensing technology is feasible and effective, and has obvious advantages [66][67][68]. Besides, the distributed optical fiber sensing technology shows broad application prospects in monitoring the deformation and stability of artificial soil filling, excavation of slopes, highway and railway sub-grades, and river embankments [19,20,[69][70][71].
This section briefly introduces the applications of the distributed optical fiber sensing technology in engineering of piles and tunnels.

DFOS test system of the foundation pile
The DFOS test system of the foundation pile includes four subsystems: pile measurement, data collection, transmission and storage, and data processing and analysis. As shown in Fig. 19, the force and deformation characteristics of the pile are sensed by the optical sensing cable laid on the pile. The distributed information such as temperature, stress, and strain of the pile is obtained by the data collection system. The information is connected to the network through the data transmission line, and it is stored on the local hard disk or network hard disk. Finally, the results are obtained by the data processing and analysis subsystem.
For various types of foundation piles, the installation and layout methods of the sensing optical cables are different, as shown in Fig. 20. For precast piles, the sensing optical cables need to be installed in various parts of the precast pile body by a special laying process to ensure that the strain of the sensing optical cables is consistent with the deformation of the pile. For cast-in-place piles, just tie the metal base optical cable to the steel cage. For steel pipe piles, the metal base ribbon optical cable or carbon fiber composite base optical cable can be installed on the surface of the steel pipe by welding and cementing. The distributed optical fiber sensing technology can be applied in static load tests of various pile foundations, pile driving monitoring, and horizontal deflection monitoring of anti-slide piles. When the pile body is deformed under force, the built-in sensing optical cable will deform in coordination with the pile body. Then, the distributed strain sensing technology can test the entire strain distribution of the pile body. This method can test the fine distribution law of pile body strain in the whole section of the pile body and analyze the force behavior of the pile body (axial force, side friction resistance, pile compression, etc.) and the pile quality (elastic modulus, etc.), measure the bearing capacity of the pile foundation, and provide data reference for the optimization design of the pile foundation. Figure 21 shows the test results of the BOTDA of the pretensioned spun high strength concrete (PHC) piles under the horizontal load.

Research on the optical fiber monitoring technology in tunnel
Tunnel construction is generally divided into three types: open-cut, cover-cut, and shield methods.
The main monitoring contents of the open-cut tunnel include anti-slide pile monitoring, anchor cable and anchorage system monitoring, top beam and waist beam monitoring, main supporting structure monitoring, main body side displacement monitoring, ground surface deformation monitoring, and tunnel deformation monitoring. The optical fiber can be implanted into the soil around the anti-slide pile, top beam, waist beam, anchor cable, and tunnel, and connect each part in series to form a system. When any part is deformed, the strain of the optical fiber can be obtained accordingly, so the DFOS technology is used to comprehensively monitor the deformation of the supporting system and the surrounding soil. For some key points, the FBG sensing technology can be used to measure in real time, as shown in Fig. 22  The monitoring of the cover-cut tunnel mainly includes the stress distribution of the supporting pile, the deformation of the roof, the overall stability of the tunnel structure, and the monitoring of the surrounding soil deformation. The overall layout scheme of cable is shown in Fig. 22(b).
The shield tunnel monitoring mainly includes the uneven settlement of the tunnel, the convergence of the ring structure, the deformation of the segment structure, and the expansion and contraction of the segment joints. The layout of the cable is shown in Fig. 22(c). The opening, closing, and dislocation deformation of segment joints and structural seams are the main sources of deformation of shield tunnels, which need to be monitored. Generally, two methods are adopted: the first method is to fix the strain sensing optical cable on the two sides of the contact seam or the tube piece. When the joint is opened or closed, the strain sensing optical cable will stretch or return. And the optical cable will undergo strain changes, and the integral of the strain and the fixed-point distance is the change in the width of the joint. The second method is to use the FBG displacement sensors, which are installed on both sides of the joint to realize the joint opening and closing deformation monitoring. Figure 23 shows the BOFDA monitoring results of the expansion and contraction of the joints of the shield tunnel segments.

Applications to geo-hazards monitoring
Monitoring plays a key role in the risk assessment, prediction, and early warning of the geological hazard. In recent years, the optical fiber sensing technology has shown its advantages in geological hazard monitoring and early warning [67,72,73]. The monitoring technology has been successfully applied to monitoring geological hazards such as landslides, ground collapse, ground subsidence, and ground fissures [21,[74][75][76][77][78][79][80][81]. In geological hazard monitoring, deformation of rock and soil mass and water distribution are two very important monitoring contents. Therefore, this section briefly introduces two kinds of optical fiber monitoring technology that are suitable for monitoring land subsidence deformation and water content distribution.

Borehole full-section optical fiber monitoring technology
The borehole full-section optical fiber monitoring is to lay fiber-optic sensing cables or sensors in the borehole to form a distributed (including quasi-distributed and fully distributed) multi-fields and multi-parameters optical fiber comprehensive monitoring system. This monitoring technology can monitor quantitatively the distribution of borehole deformation, groundwater level, temperature, water content, pore water pressure, and so on, to get the change rule of multi-parameters of borehole full section geological body.
The system consists of four parts: (1) monitoring module, (2) signal modulation and demodulation module, (3) signal transmission and data analysis module, and (4) evaluation module. The monitoring module mainly includes the ground drilling, the layout and installation of sensing optical cable and sensor, and the establishment of monitoring station; the signal modulation and demodulation module mainly includes a modem, field data acquisition, and storage. Signal transmission and data analysis module mainly involves the communication, data exchange, and data processing analysis between the Internet or wireless network and remote terminal processor; the evaluation module mainly includes report output, multi-field potential evaluation, and early warning prediction, as shown in Fig. 24 To demonstrate the application of the borehole full section optical fiber monitoring system, this section takes the optical fiber monitoring of Suzhou Shengze land subsidence as an example to introduce the monitoring efficacy of this technology.
The thickness of the quaternary stratum in the Suzhou area is about 180 m. The site of the test borehole is selected in the sports ground of Shengze middle school in Suzhou. The diameter of the monitoring hole is 129 mm and the depth is 200 m. The optical fiber monitoring system of the whole section of the borehole is shown in Fig. 25. The initial value was measured after the consolidation of the borehole backfill. The monitoring cycle was once every 3 months to 4 months.

Optical fiber monitoring technology of moisture content
The distribution and degree of water content in the soil are the significant factors affecting the stability of the soil. Geological hazards such as landslide, debris flow, collapse, land subsidence, and geotechnical engineering problems (including foundation pit instability, retaining wall collapse, and tunnel leakage) are closely related to the soil moisture field. Therefore, the monitoring of the water content in soil is of great significance for mastering the engineering properties of soil, preventing geological hazards, and solving various geotechnical engineering problems.
In recent years, the distributed optical fiber temperature measurement technology based on the active heating optical fiber (AHFO) was used to measure the temperature of soil with different moisture contents, establish calibration curve, and then calculate the moisture content of soil according to the measured temperature change value. This method has been successfully applied to a large number of laboratory tests [23,[82][83][84][85]. Furthermore, some scholars have also tried to apply this method to the monitoring of soil moisture content in situ, and some achievements had been made [47,[86][87][88]. Distributed optical fiber monitoring of the water content in the soil is realized by measuring the size and distribution of soil moisture content. The authors' team has developed two kinds of optical fiber sensors (cables) with active heating function based on the FBG and DTS temperature measurement technology. The special optical fiber sensors (cables) are implanted into the soil to obtain the time history curves of active heating and temperature rise in the soil with different moisture contents and determine the temperature characteristic value (T t ) of the soil. Furthermore, by establishing the relationship between T t and volume moisture content in the soil, the size and distribution of water content in the soil were measured to realize the distributed monitoring of water content in the soil.
In practical application, the heated FBG sensor or DTS sensing optical cable was installed in the measured soil by layered embedding, inserting, or drilling. Especially, in the way of borehole embedment, it was necessary to backfill the earth material in the borehole and ensure that the optical fiber measuring tube was tightly packed with the surrounding soil and no space was left. The soil moisture field monitoring system based on the FBG and DTS was generally composed of heating, sensing, data processing, transmission, and analysis subsystems. Figure 27 shows a portable heating DTS moisture meter developed by Suzhou Nanzhi Sensing Technology Co., Ltd.  Figure 28 shows the variation of soil profile moisture content measured by CFHST during dewatering of a foundation pit. It can be seen that the groundwater level has been falling continuously during the whole test process, from 9.81 m at the beginning to 14.08 m at the end. The moisture content of the topsoil layer was affected by weather, and the influence depth was about 4 m. The test results were very clear. At present, this technology has been widely used in the loess water field and Yangtze River Dike Seepage projects.

Prospects and conclusions
With the continuous development of engineering activities to the deeper underground space, the engineering geological and hydrogeological conditions encountered are becoming more and more complex, and the geo-environment is becoming harsher and harsher. Therefore, there is an increasing requirement for robust sensing technologies and systems. In the past 10 years, the development of the distributed fiber optic sensing technology is very fast, and a series of breakthroughs have been made in terms of geotechnical sensing theories and technologies. A much wider range of applications has been observed at the same time. Many monitoring problems that are difficult to be solved by conventional methods have been solved. The widely used point-type geotechnical instrumentation technology is being replaced by the distributed sensing systems. Simultaneously, the electric sensors are being replaced by the fiber optic sensors. These achievements are reflected in the following aspects.
At the theoretical level, the criterion to evaluate the performance of sensing optical cables has been established, together with that of the deformation compatibility of the cable-soil interface. The calculation methods of critical confining pressure and critical depth of soil layer are proposed. The ultimate capacity of micro-anchors on the optical cable is proposed. In addition, several parameters have been introduced to describe the cable-soil bonding behavior. The characterization models of temperature characteristic value and soil moisture, water content, and seepage are established, which enables the distributed optical fiber in-situ monitoring of water field of rock and soil.
At the technological level, the secondary dismantling and synthetic cable is proposed. By modifying the fiber packaging material and optimizing the fiber structure, a series of robust sensitive-preserving sensing cables are developed. The encapsulation principles of fixed-point homogenization and the equivalent conversion are proposed and the sensing fiber with a different measuring range is developed. The carbon fiber reinforced polymer (CFRP) sensitized optical cable with internal heating function and the functional optical cable with smart and sensitive coating layer are invented, which provide powerful sensing elements for humidity and moisture content monitoring of rock and soil mass.
At the level of the monitoring system, the multi-parameter optical fiber monitoring system for the whole section of the borehole is established to improve the monitoring efficiency and reduce the overall cost. The optical fiber in-situ monitoring system of the soil moisture field is developed to realize the in-situ monitoring of large-area soil water field. The distributed fiber optic testing and monitoring systems for foundation piles and tunnels have been developed, which solve the problem of accurate and long-distance monitoring in geotechnical engineering.
However, despite the rapid development of the distributed fiber optic sensing technology in the past decade, it is still far from meeting the monitoring requirements of geo-engineering and needs further improvement in the data acquisition instrument for DFOS, the development of high temperature and anti-aging sensing optical fiber cable, and the installation technologies of sensing optical fiber cable. In recent years, there have been some new distributed fiber optic sensing technologies, such as the distributed fiber optic acoustic sensor technology (distributed acoustic sensing, DAS) and the weak grating array sensing technology. These technologies provide a new technical method and developing direction of geological engineering monitoring. In 2019, the world's top journals Science and Nature reported that based on the DAS technology, the undersea "dark fiber" was used to monitor earthquake, fault activity, ocean-solid earth interaction, and other research fields [89,90]. Therefore, there is every reason to believe that with the development of the DFOS technology, geoengineering optical fiber monitoring will usher in another golden decade.