Draft...Draft 2 24 Abstract: 25 Distributed fiber optic sensing (DFOS) is gaining increasing...
Transcript of Draft...Draft 2 24 Abstract: 25 Distributed fiber optic sensing (DFOS) is gaining increasing...
Draft
Quantifying Progressive Failure of Micro-Anchored Fiber Optic Cable–Sand Interface via High-Resolution Distributed
Strain Sensing
Journal: Canadian Geotechnical Journal
Manuscript ID cgj-2018-0651.R1
Manuscript Type: Article
Date Submitted by the Author: 21-Apr-2019
Complete List of Authors: Zhang, Cheng-Cheng; Nanjing University, Department of Geoengineering and Geoinformatics, School of Earth Sciences and Engineering; University of California, Berkeley, Department of Civil and Environmental EngineeringZhu, Hong-Hu; Nanjing University, School of Earth Sciences and EngineeringLiu, Suping; Nanjing University, Department of Geoengineering and Geoinformatics, School of Earth Sciences and EngineeringShi, Bin; Nanjing UniversityCheng, Gang; North China Institute of Science and Technology
Keyword: Micro-anchor, Progressive failure, Interfacial shear stress, Strain distribution, Distributed fiber optic sensing (DFOS)
Is the invited manuscript for consideration in a Special
Issue? :Not applicable (regular submission)
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
1
1
2 Quantifying Progressive Failure of Micro-Anchored Fiber Optic Cable–Sand Interface
3 via High-Resolution Distributed Strain Sensing
4
5 Cheng-Cheng Zhang1,2, Hong-Hu Zhu1, Su-Ping Liu1, Bin Shi1, and Gang Cheng3
6
7 1School of Earth Sciences and Engineering, Nanjing University, Nanjing, Jiangsu 210023,
8 China.
9 2Department of Civil and Environmental Engineering, University of California, Berkeley, CA
10 94720-1710, USA.
11 3North China Institute of Science and Technology, Beijing 101601, China.
12
13 Corresponding author: Prof. Hong-Hu Zhu, Ph.D.
14 Address: 362 Zhugongshan Bldg, 163 Xianlin Ave, Qixia District, Nanjing, Jiangsu 210023,
15 China.
16 E-mail: [email protected]
17 Telephone: +86 158-9599-6665
18
19
20 Revised manuscript submitted to Canadian Geotechnical Journal for re-review
21 and possible publication as an Article
22
23 21 April 2019
Page 1 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
2
24 Abstract:
25 Distributed fiber optic sensing (DFOS) is gaining increasing interest in geotechnical monitoring.
26 By using soil-embedded fiber optic cables, strain profiles as well as deformation patterns of
27 geotechnical infrastructures can be captured. Probing the fiber optic cable–soil interfacial
28 behavior is vital to the advancement of DFOS-based geotechnical monitoring and our
29 understanding of the soil–inclusion interaction mechanism. To this aim, laboratory pullout tests
30 were performed to investigate the progressive failure of the interface between
31 unanchored/micro-anchored cables and the surrounding sand. High-resolution strain profiles
32 recorded using Brillouin optical time-domain analysis (BOTDA) not only elucidated the
33 influence of anchorage on strain measurements, but also allowed the classical soil–inclusion
34 interaction problem to be studied in detail. Interfacial shear stresses calculated from step-like
35 strain profiles provided clear evidence of the contribution of each micro-anchor to the pullout
36 resistance. The cable–soil contact is a combination of overall bonding and point fixation
37 depending on the level of mobilized interfacial shear stress, and therefore the validity of
38 measured strains is correlated to a three-stage process of interface failure. This study also shows
39 that installing heat-shrink tubes on the fiber optic cable is a rapid, low-cost, effective approach
40 to make an anchored DFOS system for deformation monitoring of earth structures.
41
42 Keywords: Micro-anchor; Progressive failure; Interfacial shear stress; Strain distribution;
43 Distributed fiber optic sensing (DFOS)
Page 2 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
3
44 Introduction
45 Deformation measurements play an important role from evaluating the performance of massive
46 and complex geotechnical structures to validating and refining relevant analytical and
47 numerical models. Traditional instrumentation using strain gauges, linear variable displacement
48 transducers, inclinometers and extensometers only allow deformation to be recorded at discrete
49 locations or directions, and therefore localized features of the monitored geotechnical
50 infrastructures may not be identified in some cases (Pelecanos et al. 2018).
51 The distributed fiber optic sensing (DFOS) technology turns common telecommunication
52 fiber optic (FO) cables into sensing elements, which enables continuous strain measurements
53 at kilometre scales and beyond (Shi et al. 2003; Thévenaz 2010; Habel and Krebber 2011; Bao
54 and Chen 2012; Soga and Luo 2018). Recent advances in FO sensing have rendered this
55 technique better suited to geotechnical applications (Schenato 2017; Nöther and von der Mark
56 2019; Soga et al. 2019). FO cables can be readily integrated into geotechnical structures such
57 as ground anchors, piles, retaining walls, and geosynthetics, because of their geometrical
58 versatility. Recently, field and laboratory attempts have been made to monitor and evaluate the
59 performance of geotechnical structures using quasi-distributed fiber Bragg gratings (FBGs)
60 (e.g., Lee et al. 2004; Zhu et al. 2011; Huang et al. 2012; Zhang et al. 2014b; Borana et al. 2017;
61 Zhu et al. 2017) or Rayleigh and Brillouin scattering-based fully-distributed approaches (e.g.,
62 Shi et al. 2003; Klar et al. 2006, 2016; Mohamad et al. 2012; Lienhart 2015; Schenato et al.
63 2017; Ni et al. 2018; Zhang et al. 2018; Wheeler et al. 2019).
64 Recent trends aimed at monitoring soil deformation using embedded FO cables have
65 encouraged increased use of micro-anchors (Fig. 1; Iten et al. 2011; Hauswirth et al. 2012,
66 2014; ASTM F3079-14 2014; Zhu et al. 2014; Damiano et al. 2017). In ground-improvement
67 engineering, anchorages have already been utilized to enhance the pullout resistance of soil
68 reinforcements such as ground anchors (Pérez et al. 2017). For geogrids, the soil in front of
69 transverse ribs is subject to passive earth pressure; this effect, along with the solid geogrid
Page 3 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
4
70 surface–soil skin friction, results in the mobilized interaction mechanism (Koerner et al. 1989;
71 Bergado et al. 1993; Liu et al. 2009; Moraci et al. 2014). The passive resistance can be further
72 increased by adding anchorages to the transverse elements. One such geogrid named grid-
73 anchor has been introduced by Mosallanezhad et al. (2008) and Alamshahi and Hataf (2009).
74 This was made by attaching short anchors to one side of the geogrid and so greater bearing
75 resistance can be obtained. More recently, Sadat Taghavi and Mosallanezhad (2017) developed
76 an anchored geogrid by adding a set of steel equal angles to the geogrid; the pullout resistance
77 of the new reinforcing system can be increased by up to 65% compared to traditional ones. For
78 DFOS-based geotechnical monitoring, micro-anchors are expected to enhance the interlocking
79 effects between a FO cable and the surrounding soil and therefore extend the measurement
80 range of a DFOS system. Schenato et al. (2017) used the DFOS technology to measure strains
81 exerted on a corrugated FO cable induced by shallow landslide triggering; the evolution of the
82 landslide coincided with the phases identified during the progressive cable–soil failure.
83 Hauswirth et al. (2010) developed a system with three-dimensional micro-anchors attached to
84 a FO cable, and evaluated its performance through pullout and shearing tests. Lately, this
85 system was successfully applied to detect landslide movements in Switzerland (Hauswirth et
86 al. 2012) and ground deformation induced by tunneling in the UK (Hauswirth et al. 2014).
87 The accuracy of strains measured by a DFOS system with soil-embedded FO cables is
88 subject to the cable–soil interaction mechanism. Probing the cable–soil interfacial properties is
89 thus of great importance for the advancement of both the DFOS technology and our
90 understanding of the interaction mechanism between soil and an inclusion. As such, classical
91 pullout tests have been introduced to study the FO cable–soil interfacial behavior (Iten et al.
92 2009; Zhang et al. 2014a). The test procedure is similar to that of an earth reinforcement such
93 as a soil nail (Wang et al. 2017). The difference lies in that the FO cable is treated as both a soil
94 inclusion and a distributed strain sensor, and therefore no extra strain gauges are needed which
95 might potentially affect the FO cable–soil interfacial properties. In addition, the axial stiffness
Page 4 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
5
96 of a FO strain cable (equivalent to Young’s modulus multiplied by cross-sectional area) is
97 comparably low, typically in the range between 1–60 kN (Iten et al. 2011; Klar et al. 2016;
98 Schenato et al. 2017). This results in an evident progressive failure behavior of a smooth FO
99 cable in soil under low overburden pressures (Iten et al. 2009; Zhang et al. 2014a). However,
100 until now, the pullout performance of an anchored FO cable embedded in soil has not been
101 investigated in detail. While it is widely accepted that an unanchored soil inclusion has a linear
102 distribution of axial force (accordingly, a uniform distribution of interfacial shear stress) at the
103 critical state (Fig. 2a), these distributions have not been directly measured for an anchored
104 inclusion (Fig. 2b).
105 In this study, a rapid, low-cost, effective approach was first proposed to make an anchoring
106 system for DFOS-based geotechnical monitoring. Then, laboratory pullout tests were
107 conducted to investigate the micro-anchored FO cable–sand interfacial characteristics. High-
108 resolution strain profiles along the FO cable were captured during progressive interface failure
109 by using a Brillouin optical time-domain analysis (BOTDA) interrogator, which elucidated the
110 influence of anchorage on the measured FO strain data and allowed the soil–inclusion
111 interaction to be studied in detail. Finally, the implications of the experimental results for
112 DFOS-based geotechnical monitoring, as well as the limitations of the present study, were
113 discussed.
114
115 Principle of DFOS using BOTDA
116 DFOS is a technology that enables spatially continuous, long-distance, and near-real time
117 measurements along a FO cable. The FO cable itself is not only the transmission medium but
118 also the sensing element. When a light wave generated from a FO interrogator travels through
119 the core of an optical fiber, backscattered lights (Rayleigh, Raman, or Brillouin) are generated
120 at any point along the fiber (Hartog 2017).
Page 5 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
6
121 For Brillouin scattering (Fig. 3), the frequency shift of the backscattered light is
122 approximately linearly proportional to the axial strain or temperature exerted on the optical
123 fiber (Bao and Chen 2012):
124 (2)B BB B0 T
T
125 where and = Brillouin frequency shifts before and after a measurement, respectively; B0 B
126 = axial strain change; = temperature change; and and = coefficients T B / B / T
127 for strain and temperature changes, respectively. According to eq. (1), if the Brillouin frequency
128 shifts along the entire length of the fiber are detected, a continuous distribution of strain and
129 temperature with respect to distance can be determined. As the spontaneous Brillouin scattering
130 light is very weak, it is recommended to utilize BOTDA to measure strain and temperature
131 along the optical fiber. By deploying the pulse pump and probe at both ends of the fiber, the
132 frequency changes in stimulated Brillouin scattered light can be detected with high spatial
133 resolution and accuracy. Further information about the fundamentals of the BOTDA technology
134 may also be found in Thévenaz (2010), Schenato (2017), and Soga and Luo (2018), which is
135 beyond the scope of this paper.
136
137 Materials and Methods
138 Materials
139 The soil used in this experimental investigation was a clean, angular river sand. A sieve analysis
140 was carried out and the obtained grain size distribution curve is shown in Fig. 4. From this curve,
141 the values of and were 0.23 mm and 0.35 mm, respectively. The coefficient of 10D 50D
142 uniformity, , was 1.61, and the coefficient of curvature, , was 1.06. The sand was uC cC
143 classified as a poorly graded sand (SP) according to the Unified Soil Classification System
144 (ASTM D2487-17 2017). Following ASTM D4253-16 (2016) and ASTM D4254-16 (2016),
145 the maximum and minimum dry density, and , were determined to be 1.69 g/cm3 dmax dmin
146 and 1.21 g/cm3, respectively. By using direct shear tests (ASTM D6528-17 2017), the friction
Page 6 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
7
147 angle, , of the sand at a relative density, , of 23% was measured to be 32.3°. Detailed rD
148 physical and mechanical properties of the sand are summarized in Table 1.
149 FO strain sensing cables for monitoring soil deformation have to meet two main
150 requirements, including being robust enough to survive in harsh construction conditions and
151 ensuring efficient strain transfer from soil to the fiber core. A 2 mm-diameter tight-buffed
152 polyurethane-coated FO cable (Nanzee Sensing, Suzhou, China) was used as the strain cable
153 (Fig. 5a), whose performance has been tested in several field monitoring projects as well as
154 physical model tests by the authors’ research group. Detailed properties of this cable are
155 summarized in Table 2. The soft polyurethane coating lowers the Young’s modulus of the cable,
156 which however is still capable of withstanding moderate impacts during installation. Due to its
157 relatively low modulus, it can be easily prestrained and directly integrated into loose material
158 such as soil. Note that this cable was chosen for the current experimental study for that it has
159 shown good performance both in the field and laboratory. However, if the construction activity
160 involves heavy equipment, it is recommended to use cables of higher robustness such as steel
161 strand- or fiber reinforced polymer-buffered cables.
162 Two types of heat-shrink tube, made of polyethylene (PE) and polymethyl methacrylate
163 (PMMA), were used to make micro-anchors on the FO cable. These two tubes are commonly
164 used to protect a fused optical fiber; they will shrink radially to wrap tightly around the optical
165 fiber by heating with a hot air gun or a built-in heater of an optical fiber fusion splicer. The
166 tubes were shrunk onto the cable coating at a certain interval using a KL-500 fusion splicer
167 (Jilong, Nanjing, China), as shown in Fig. 5b. Two separate FO cables, each having one type
168 of tube, were prepared and tested independently. The average diameter of the micro-anchored
169 cables was calculated according to the anchor distribution, and the average Young’s modulus,
170 , of the anchored segments was estimated based on the cross-section geometry. Detailed aE
171 micro-anchor properties are listed in Table 3.
172
Page 7 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
8
173 FO Interrogator
174 The experimental investigation presented in this paper used a NeubreScope NBX-6050A
175 interrogator (Neubrex, Kobe, Japan). This instrument provides a minimum readout resolution
176 of 10 mm with a spatial resolution of 50 mm and a noise level of about 7.5 , allowing for
177 distributed measurements along a FO cable of 1 km. The spatial resolution, which is a key factor
178 affecting strain measurements, achieved using the BOTDA technology is higher than that of
179 the Brillouin optical time-domain reflectometry (BOTDR) technology (normally, 500–1000
180 mm; Shi et al. 2003; Pelecanos et al. 2018). Here, the minimum readout resolution (i.e., 10 mm)
181 and the highest spatial resolution (i.e., 50 mm) of this BOTDA interrogator were adopted to
182 obtain optimal measurements at a laboratory scale.
183
184 Test Setup and Procedure
185 A schematic drawing of the test setup is shown in Fig. 6. The test container was 1500 mm long,
186 200 mm wide, and 200 mm high, made of aluminum alloy walls, 15 mm thick, having a 60
187 mm-diameter hole in the middle of the front wall. The width and depth of the test container
188 were 100 times the cable diameter and therefore the boundary effect during testing was avoided.
189 The test bed was prepared using a sand raining technique. When the prepared sand reached
190 the mid-height of the test container, it was gently levelled out and the FO cable centered
191 carefully in the container. The depth of sand above and below the cable was 100 mm. No
192 additional confining pressures were applied on the surface of the sand bed. The density and
193 water content of the sand measured after testing were 1.32 g/cm3 and 1.92%, respectively. This
194 corresponded to a relative density, , of 23%, indicating that the sand sample was in a loose rD
195 state (Mitchell and Soga 2005). The embedded cable length was 1200 mm, whereas a 200 mm-
196 long free segment was left between the embedded cable and the front wall of the container to
197 minimize boundary effects. A special clamp having two rubber pieces acting as a cushion was
198 utilized to clamp the free cable segment (Fig. 7); the clamp was then connected to a tensile
Page 8 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
9
199 tester. The free segment at the other end ran from the top of the test container. Finally, the free
200 segments at both ends of the cable were connected to the FO interrogator.
201 Prior to the test, a zero measurement was performed using the FO interrogator to record
202 the initial strain distribution along the FO cable. During testing, the pullout displacement was
203 applied using an electric motor with a velocity of 0.1 mm/s and was later calculated according
204 to the motor speed and time elapsed; the pullout force was recorded using a force gauge with a
205 resolution of 0.1 N. At each displacement step of 1 mm, the displacement was maintained until
206 a strain measurement was performed using the FO interrogator.
207 Preliminary tests were carried out to ensure the suitability of the test setup and procedure.
208 Afterward, micro-anchors with two different diameters, i.e., Da = 3.6 mm and 6.0 mm, were
209 investigated. A smooth cable without micro-anchors was also tested for comparison.
210
211 Results
212 Characteristics of Pullout Force–Displacement Curves
213 Curves of pullout force, , versus pullout displacement, , are shown in Fig. 8. Notice that 0F 0u
214 the elongation of the free cable segment was deduced from the measured displacement. A
215 nonlinearity was observed from the curves before the peak pullout force was reached, due to
216 the partial failure of the cable–soil interface. This was expected because the axial stiffness of
217 the cable was comparatively low and therefore progressive failure was evident during testing
218 (Zhang et al. 2014a).
219 Moreover, there was only a slight difference between the slopes of the initial sections,
220 which was confirmed by the secant interfacial shear stiffness calculated at a displacement step
221 of 2 mm, with the average value being 198.7 kPa/m (Fig. 9). This indicated that the effect of
222 anchorage on the cable–soil interfacial shear stiffness was not obvious. On the contrary, its
223 influence on the peak pullout force and peak interfacial shear stress (ISS) was much more
224 pronounced. For the unanchored cable, the pullout force peaked at a displacement of
Page 9 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
10
225 approximately 3.82 mm and then reduced with increasing displacements. This illustrated that
226 the unanchored cable–soil interface exhibited a strain-softening behavior during pullout, which
227 agreed well with the results of Zhang et al. (2014a). In contrast, a strain-hardening behavior
228 was observed for the micro-anchored cables, which largely increased the peak ISS (Fig. 9). The
229 peak ISS was increased by 78.48% and 146.6% for the micro-anchored cables with Da = 3.6
230 mm and 6.0 mm, respectively, as compared to the unanchored cable. This demonstrated that
231 the use of micro-anchors can increase the interfacial bond between FO cable and soil and
232 therefore extend the measurement range of a DFOS system (Hauswirth et al. 2010; Iten et al.
233 2011; Damiano et al. 2017).
234
235 Behavior of Unanchored Cable–Sand Interface
236 Prior to data analysis, the monitored strains along the 200 mm-long free cable segment were
237 converted to tensile forces and compared with force gauge measurements (Fig. 10). There was
238 generally good agreement between the two forces, indicating the validity of the measured strain
239 data. Fig. 11 shows the obtained strain data (relative to the zero strain measurement) along the
240 FO cables for each displacement step. Because the strain profiles exhibited no waviness, no
241 signal processing procedure (e.g., filtering) was performed. In addition, temperature
242 compensation for the measured strains was not considered as the variation of temperature was
243 negligible during testing. For ease of comparison, the pullout force–displacement curves are
244 shown in the upper right corner. The displacement steps are denoted as closed circles for the
245 strain distributions presented and discussed.
246 For the unanchored FO cable, the strain propagated from the cable head toward the toe
247 under increasing displacement steps (Fig. 11a). This indicated that initially only a small portion
248 of ISS was mobilized to resist the pullout force; however, with the increase of the pullout force
249 the ISS increased gradually in terms of both magnitude and mobilized length. Once the peak
250 pullout force was reached (displacement step = 5 mm), the maximum strain began to decline,
Page 10 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
11
251 indicating the decrease of the average ISS. To examine the evolution of ISS in detail, the pullout
252 behavior of the unanchored FO cable in sand was further simulated using a pullout model
253 incorporating strain-softening of the cable–soil interface (Zhang et al. 2014a). The correlation
254 between pullout force and pullout displacement for five pullout phases (Phase I–V) can be
255 expressed as (Zhang et al. 2014a):
256 (3)
1 10
12
2 2 s max max 2 s max20
2 res 2 2 s 1 2 res
2 max2 2 200
2 1 2
2 2r res max max res
0r r 1 2 2
π tanh (Phase I)
π cot π π cot(1 ) (Phase II)sin
π tanπ tan (1 ) (Phase III)
ππ π ( )4 2 4
DG L u
DG L D D LGuL G
D LDG L GuFG
DLD E D EuL L G G G
res
(Phase IV)
π (Phase V)D L
257 where D, L, and E = diameter, length, and Young’s modulus of the unanchored FO cable,
258 respectively; and = lengths of softening and residual zones, respectively; and = sL rL 1G 2G
259 cable–soil interfacial shear stiffnesses corresponding to ascending and descending branches,
260 respectively; and = peak and residual cable–soil ISSs, respectively; and max res
261 . The expression of axial strain and ISS distributions and further details 4 / ( 1, 2)i iG ED i
262 of the model can be found in Zhang et al. (2014a).
263 The values of independent parameters for simulating the behavior of the unanchored FO
264 cable during pullout are as follows: D = 2 mm; L = 1,200 mm; E = 0.34 GPa; = 3.00 MPa/m; 1G
265 = 55.83 kPa/m; = 0.75 kPa; and = 0.40 kPa. The simulated results are shown in 2G max res
266 Figs. 12a–c. Note that here the pullout force was calculated according to strain readings along
267 the free cable segment. There was generally good agreement between the measured force–
268 displacement curve and the simulated curve, except for the final residual phase where the
269 measured force did not rigorously reach a constant value (Fig. 12a). Five pullout phases are
270 also marked in the figure according to the simulation results. Prior to reaching the final residual
Page 11 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
12
271 phase (Phase V), the percentage of displacement generated during each pullout phase (Phase I–
272 IV) was 3.04%, 37.18%, 39.15%, and 20.63%, respectively. The two transitional phases—
273 Phase II (elastic–softening phase) and Phase IV (softening–residual phase)—was significant,
274 further illustrating the progressive failure nature of the cable–soil interface. The distribution of
275 ISS along the FO cable was highly nonuniform (Fig. 12c). Initially, the ISS increased along the
276 whole length of the cable. Once the peak ISS was reached, the ISS began to decrease for the
277 segment close to the cable head, whereas that along the remaining segment continued to
278 increase. The ISS was fully mobilized when the peak pullout force was reached. Afterward, the
279 ISS along the whole length of the cable decreased toward the residual value ( = 0.40 kPa).res
280
281 Behavior of Anchored Cable–Sand Interface
282 The pattern of strain profile for the micro-anchored cables was different from that of the
283 unanchored one (Figs. 11b and c). The former exhibited a unique step-like pattern due to the
284 addition of micro-anchors. The inclination of the strain curve of unanchored segments was
285 smaller than that of anchored segments, indicating that the ISS along the unanchored segments
286 was comparatively low. Moreover, although the average ISS over the whole length of the FO
287 cable increased sharply with increasing displacement steps, the variation of ISS along
288 individual segments (unanchored or anchored) appeared to be less significant. To the authors’
289 knowledge, no analytical model is available in the literature to reproduce such behavior as well
290 as the strain profiles shown in Figs. 11b and c. Therefore, the ISS along each segment was
291 estimated using:
292 (4)seg seg d4 d
E Dx
293 where = axial strain gradient; and and = average Young’s modulus and d / dx segE segD
294 diameter of a particular segment, respectively. For the unanchored segments, linear fits were
295 used to obtain the slope of a curve; the average coefficients of determination (R2) were 0.96 and
296 0.83 for the anchored cable with Da = 3.6 mm and 6.0 mm, respectively. For each anchored
Page 12 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
13
297 segment, the slope was regarded as the strain difference between two neighboring unanchored
298 segments divided by anchor length. For clarity, the unanchored segments are marked as A, B,
299 C, D, and E from left to right, whereas the anchored segments are marked as a, b, c, d, and e
300 (Figs. 11b and c). Note that the unanchored segment at the cable toe was neglected considering
301 its short length, and therefore the last anchored segment e was not considered as well. Finally,
302 the overall average ISS was estimated using , where = average diameter of the 0 / πF DL D
303 FO cable.
304 The calculated ISSs are shown in Fig. 13. The overall average ISS over the whole cable
305 length increased gradually toward a constant value (Figs. 13a and d), consistent with the
306 correlation between the pullout force and the pullout displacement (Fig. 8). The unanchored
307 segments of the anchored FO cables exhibited strain-softening (cable with Da = 3.6 mm in
308 particular; Figs. 13b and e), which was similar to the behavior of the unanchored FO cable.
309 However, the average ISS along the unanchored segments decreased toward 0.46 kPa (or 0.53
310 kPa) for the anchored cable with Da = 3.6 mm (respectively, Da = 6.0 mm), which was larger
311 than the residual ISS along the unanchored cable ( = 0.40 kPa).res
312 Upon applying a pullout displacement, the ISS was simultaneously mobilized along the
313 unanchored segments (Figs. 13b and e) and the anchored segments (Figs. 13c and f). However,
314 as expected, the ISS along the unanchored segments was much smaller than that along the
315 anchored segments, demonstrating that the inclusion of micro-anchors led to the difference in
316 peak ISS (Fig. 9). Additionally, the average ISS along the unanchored segments peaked at a
317 displacement step of approximately 7 mm for the anchored cable, which was earlier than the
318 peak of average ISS along the anchored segments. The latter reached a peak value at the
319 displacement step being almost identical to that which resulted in the overall failure of the
320 anchored cables. The difference in average ISS between anchored and unanchored cable
321 segments of the micro-anchored FO cable was further characterized using their ratio, which is
322 defined as
Page 13 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
14
323 (5)b b a
a a b
E AkE A
324 where and = effective axial stiffnesses of the anchored and unanchored cable a aE A b bE A
325 segments, respectively; and and = average ISSs along the anchored and unanchored a b
326 cable segments, respectively. Fig. 14 shows the difference in average ISS between the anchored
327 and unanchored segments calculated using eq. (4). The ratio increased approximately linearly
328 with increasing displacement steps, further illustrating the contribution of anchored and
329 unanchored segments to the ISS during pullout. For the contribution made by each micro-
330 anchor, a close examination of Figs. 13c and f suggested that anchored segments close to the
331 cable head contributed, in general, more to the ISS than those close to the toe. However, the
332 ISS of anchored segment c became the largest after applying the displacement step of 9 mm
333 (Fig. 13f). This could have occurred if this particular micro-anchor had a larger diameter and a
334 rougher surface, as the property of each micro-anchor could not be strictly controlled during
335 heat shrinking.
336
337 Discussion
338 Implications for Geotechnical Monitoring
339 The proper deployment of FO strain sensing cables is one of the major concerns in DFOS-based
340 geotechnical monitoring (Damiano et al. 2017; Schenato 2017). This becomes extremely
341 difficult to deal with natural geomaterials (e.g., soils) as compared to man-made materials.
342 Accordingly, anchoring systems have been used by geotechnical practitioners to prevent the
343 relative slippage between FO cable and soil (Hauswirth et al. 2010, 2014; Iten et al. 2011; Zhu
344 et al. 2014; Damiano et al. 2017). Such a system has been preliminarily evaluated by using
345 classical pullout tests in this paper. The obtained results indicate that the mechanical coupling
346 between the strain sensing cable and the surrounding soil can be effectively enhanced (Fig. 9)
347 with relatively little effort, that is, by shrinking commercially available heat-shrink tubes onto
Page 14 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
15
348 the cable coating and therefore the equally spaced tubes can serve as a micro-anchorage system
349 (Fig. 5b).
350 FO strain cables are usually point fixed (Pelecanos et al. 2018) or overall bonded (Shi et
351 al. 2003) to a monitored object. Iten et al. (2011) proposed that a combination of overall bonding
352 and point fixation is possible by using a soil-embedded micro-anchored FO cable, which is
353 further confirmed by the development of ISS along the anchored and unanchored cable
354 segments calculated from the high-resolution strain profiles (Fig. 13). The results reveal that
355 the micro-anchored FO cable–soil contact is a combination of overall bonding and point
356 fixation depending on the level of mobilized ISS, and therefore the validity of FO strain data is
357 correlated to a three-stage process during the cable–soil interface failure (Fig. 15 and Table 4).
358 In Stage I, the cable is bonded along its entire length to the soil considering that neither the
359 unanchored nor the anchored segment reaches the peak ISS. Since deformation compatibility
360 between the cable and the soil is ensured, the strain measurements can truly reflect the soil
361 deformation. In Stage II, the peak ISS of the unanchored segment is reached and therefore the
362 anchors contribute a substantial part of the ISS. In this stage, the micro-anchored FO cable is
363 similar to a gauged strainmeter and so the measured strain is averaged over the gauge length,
364 namely, anchor spacing. Therefore, one has to carefully consider the anchor spacing so as not
365 to compromise the spatial resolution of the DFOS system (Schenato 2017). Stage III follows
366 when the anchors begin to fail provided that the cable itself has not yielded. The measured data
367 are completely unreliable in this stage. In this section, the reliability of strain data measured
368 from a micro-anchored FO cable has been linked to the progressive failure process of the cable–
369 soil interface, but for field monitoring how to distinguish when the interface has transitioned
370 from one stage to another remains to be investigated.
371
372 Limitations
Page 15 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
16
373 The quality of distributed FO strain data is intrinsically influenced by its distributed nature. The
374 measured strain is averaged over a certain length called spatial resolution, which has remarkably
375 increased during the past few years to several millimeters for the optical frequency-domain
376 reflectometry (OFDR) technology (Ni et al. 2018; Friedli et al. 2019). In this study a spatial
377 resolution of 50 mm and a readout resolution of 10 mm were achieved by using BOTDA.
378 Because this spatial resolution was two times the length of the micro-anchor (i.e., 25 mm, Fig.
379 5b), the quality of the obtained strain data along the anchored segments was therefore inevitably
380 influenced. Moreover, similar to a geogrid the passive earth resistance produced against the
381 micro-anchors should be an important component of the mobilized interaction mechanism
382 between the anchored cable and the surrounding soil. This effect was, however, not fully
383 considered in this preliminary study. Further research is needed to obtain a more accurate strain
384 profile using advanced technologies such as OFDR and, hence, a refined analysis of the passive
385 earth pressure effect can be performed.
386 Furthermore, although anchoring systems such as the one proposed in this study or by
387 other researchers (e.g., Hauswirth et al. 2010) can effectively improve the cable–soil
388 mechanical coupling, they do not allow lateral strains to be measured in a soil. In fact, it is a
389 limitation of the DFOS technology that only strains along the longitudinal axis of a FO cable
390 can be recorded. Therefore, the information a single FO cable can provide on the global strain
391 state in a soil at a specific spatial location is limited.
392
393 Conclusions
394 In this study, the interaction between unanchored/micro-anchored FO strain cables and sand
395 was investigated through laboratory pullout tests. High-resolution strain profiles along the FO
396 cables during cable–soil interface failure were recorded by using the BOTDA technology. The
397 strain profiles showed that both unanchored and micro-anchored FO cables exhibited
398 remarkable progressive failure behavior in dry, loose sand during pullout. The unanchored
Page 16 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
17
399 cable exhibited a strain-softening behavior, which was described by using an existing pullout
400 model whereby the evolution of ISS was obtained. The simulated pullout force–displacement
401 curves and strain distributions were generally in good agreement with measured curves. The
402 micro-anchored cables had unique step-like strain profiles. The average ISS along the anchored
403 segments were much higher than that along the unanchored segments; their ratio increased
404 approximately linearly with increasing displacement steps. Moreover, the anchored segments
405 close to the cable head contributed, in general, more to the ISS than those close to the toe.
406 Shrinking commercially available heat-shrink tubes onto the cable coating is a rapid, low-
407 cost approach to make an anchored DFOS system for geotechnical monitoring. The inclusion
408 of micro-anchors largely increased the peak ISS and hence the mechanical coupling between
409 FO cable and soil, without significantly affecting the interfacial shear stiffness. The micro-
410 anchored cable–soil contact is a combination of overall bonding and point fixation depending
411 on the level of ISS. The validity of FO strain data is correlated to a three-stage process during
412 the cable–soil interface failure: (1) the measured strain is valid when the cable is bonded along
413 its entire length to the soil; (2) the measured strain is an average value over the gauge length
414 when the unanchored segment fails and the cable is therefore point-fixed to the soil; and (3) the
415 measured strain becomes unreliable once an overall interface failure occurs.
416
417 Acknowledgments
418 This work was supported by the National Natural Science Foundation of China (NSFC) grants
419 41722209, 41672277, 41427801, and 41702347. C.-C. Zhang acknowledges a fellowship from
420 the China Scholarship Council grant 201706190165 for his academic visit to the University of
421 California, Berkeley. The authors thank D.-D. Chen, F. Li, Y.-F. Ni, and S.-H. Zeng (all of
422 Nanjing University) for laboratory assistance and Prof. K. Soga (the University of California,
423 Berkeley) for fruitful discussions.
424
Page 17 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
18
425 Notation list
426 The following symbols are used in this paper:
427 = cross-sectional area of fiber optic cableA
428 = curvature coefficient of soilcC
429 = uniformity coefficient of soiluC
430 = diameter of cableD
431 = effective grain size of soil10D
432 = average grain size of soil50D
433 = micro-anchor diameteraD
434 = relative density of sandrD
435 = diameter of a particular cable segmentsegD
436 = average cable diameterD
437 = Young’s modulus of cableE
438 = Young’s modulus of anchored cable segmentaE
439 = Young’s modulus of unanchored cable segmentbE
440 = Young’s modulus of a particular cable segmentsegE
441 = pullout force0F
442 = cable–soil interfacial shear stiffness corresponding to an ascending branch1G
443 = cable–soil interfacial shear stiffness corresponding to a descending branch2G
444 = specific gravity of soilsG
445 = embedded cable lengthL
446 = length of softening zonesL
447 = lengths of residual zonerL
448 = temperatureT
449 = pullout displacement0u
450 = pullout model coefficient
Page 18 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
19
451 = axial cable strain
452 = Brillouin frequency shift after a measurementB
453 = Brillouin frequency shift before a measurementB0
454 = cable–soil interfacial shear stress (ISS)
455 = peak ISSmax
456 = residual ISSres
457 = average ISS along anchored cable segmentsa
458 = average ISS along unanchored cable segmentsb
459 = friction angle of soil
460
461 References
462 Alamshahi, S., and Hataf, N. 2009. Bearing capacity of strip footings on sand slopes reinforced
463 with geogrid and grid-anchor. Geotextiles and Geomembranes, 27(3): 217–226.
464 doi:10.1016/j.geotexmem.2008.11.011.
465 ASTM D2487-17. 2017. Standard practice for classification of soils for engineering purposes
466 (Unified Soil Classification System). ASTM International, West Conshohocken, PA, USA.
467 doi:10.1520/D2487-17.
468 ASTM D4253-16. 2016. Standard test methods for maximum index density and unit weight of
469 soils using a vibratory table. ASTM International, West Conshohocken, PA, USA.
470 doi:10.1520/D4253-16.
471 ASTM D4254-16. 2016. Standard test methods for minimum index density and unit weight of
472 soils and calculation of relative density. ASTM International, West Conshohocken, PA,
473 USA. doi:10.1520/D4254-16.
474 ASTM D6528-17. 2017. Standard test method for consolidated undrained direct simple shear
475 testing of fine grain soils. ASTM International, West Conshohocken, PA, USA.
476 doi:10.1520/D6528-17.
Page 19 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
20
477 ASTM F3079-14. 2014. Standard practice for use of distributed optical fiber sensing systems
478 for monitoring the impact of ground movements during tunnel and utility construction on
479 existing underground utilities. ASTM International, West Conshohocken, PA, USA.
480 doi:10.1520/F3079-14.
481 Bao, X., and Chen, L. 2012. Recent progress in distributed fiber optic sensors. Sensors, 12(7):
482 8601–8639. doi:10.3390/s120708601.
483 Bergado, D.T., Chai, J.C., Abiera, H.O., Alfaro, M.C., and Balasubramaniam, A.S. 1993.
484 Interaction between cohesive-frictional soil and various grid reinforcements. Geotextiles
485 and Geomembranes, 12(4): 327–349. doi:10.1016/0266-1144(93)90008-C.
486 Borana, L., Yin, J.H., Singh, D.N., Shukla, S.K., and Pei, H.F. 2017. Influences of initial water
487 content and roughness on skin friction of piles using FBG technique. International Journal
488 of Geomechanics, 17(4): 04016097. doi:10.1061/(ASCE)GM.1943-5622.0000794.
489 Damiano, E., Avolio, B., Minardo, A., Olivares, L., Picarelli, L., and Zeni, L. 2017. A
490 laboratory study on the use of optical fibers for early detection of pre-failure slope
491 movements in shallow granular soil deposits. Geotechnical Testing Journal, 40(4): 529–
492 541. doi:10.1520/GTJ20160107.
493 Friedli, B., Pizzetti, L., Hauswirth, D., and Puzrin, A.M. 2019. Ground-buried fiber-optic
494 sensors for object identification. Journal of Geotechnical and Geoenvironmental
495 Engineering, 145(2): 04018109. doi:10.1061/(ASCE)GT.1943-5606.0002001.
496 Habel, W.R., and Krebber, K. 2011. Fiber-optic sensor applications in civil and geotechnical
497 engineering. Photonic Sensors, 1(3): 268–280. doi:10.1007/s13320-011-0011-x.
498 Hartog, A.H. 2017. An Introduction to Distributed Optical Fibre Sensors. In 1st edition. CRC
499 Press, Boca Raton, Florida, USA. doi:10.1201/9781315119014.
500 Hauswirth, D., Iten, M., and Puzrin, A.M. 2012. Detection of ground movements using soil-
501 embedded distributed fiber optic sensors. In Geotechnical and Geophysical Site
Page 20 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
21
502 Characterization 4. Edited by R.Q. Coutinho and P.W. Mayne. CRC Press, London, UK.
503 pp. 579–586. doi:10.1201/b13251-68.
504 Hauswirth, D., Iten, M., Richli, R., and Puzrin, A.M. 2010. Fibre optic cable and micro-anchor
505 pullout tests in sand. In Physical Modelling in Geotechnics, Two Volume Set: Proceedings
506 of the 7th International Conference on Physical Modelling in Geotechnics (ICPMG 2010),
507 28th June - 1st July, Zurich, Switzerland. Edited by S. Springman, J. Laue, and L. Seward.
508 CRC Press, FL, USA. pp. 337–342.
509 Hauswirth, D., Puzrin, A.M., Carrera, A., Standing, J.R., and Wan, M.S.P. 2014. Use of fibre-
510 optic sensors for simple assessment of ground surface displacements during tunnelling.
511 Geotechnique, 64(10): 837–842. doi:10.1680/geot.14.T.009.
512 Huang, A.-B., Lee, J.-T., Ho, Y.-T., Chiu, Y.-F., and Cheng, S.-Y. 2012. Stability monitoring
513 of rainfall-induced deep landslides through pore pressure profile measurements. Soils and
514 Foundations, 52(4): 737–747. doi:10.1016/j.sandf.2012.07.013.
515 Iten, M., Hauswirth, D., and Puzrin, A.M. 2011. Distributed fiber optic sensor development,
516 testing, and evaluation for geotechnical monitoring applications. In Proc. SPIE 7982,
517 Smart Sensor Phenomena, Technology, Networks, and Systems 2011. Edited by W. Ecke,
518 K.J. Peters, and T.E. Matikas. San Diego, California, USA. p. 798207.
519 doi:10.1117/12.881228.
520 Iten, M., Puzrin, A.M., Hauswirth, D., Foaleng-Mafang, S., Beugnot, J.-C., and Thévenaz, L.
521 2009. Study of a progressive failure in soil using BEDS. In Proc. SPIE 7503, 20th
522 International Conference on Optical Fibre Sensors. Edited by J.D.C. Jones. p. 75037S.
523 doi:10.1117/12.835115.
524 Klar, A., Bennett, P.J., Soga, K., Mair, R.J., Tester, P., Fernie, R., St John, H.D., and Torp-
525 Peterson, G. 2006. Distributed strain measurement for pile foundations. Proceedings of the
526 Institution of Civil Engineers - Geotechnical Engineering, 159(3): 135–144.
527 doi:10.1680/geng.2006.159.3.135.
Page 21 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
22
528 Klar, A., Uchida, S., and Levenberg, E. 2016. In situ profiling of soil stiffness parameters using
529 high-resolution fiber-optic distributed sensing. Journal of Geotechnical and
530 Geoenvironmental Engineering, 142(8): 04016032. doi:10.1061/(ASCE)GT.1943-
531 5606.0001493.
532 Koerner, R.M., Wayne, M.H., and Carroll, R.G. 1989. Analytic behavior of geogrid anchorage.
533 In Proc., Geosynthetics’ 89 Conf. IFAI, San Diego, California, USA.
534 Lee, W., Lee, W.-J., Lee, S.-B., and Salgado, R. 2004. Measurement of pile load transfer using
535 the Fiber Bragg Grating sensor system. Canadian Geotechnical Journal, 41(6): 1222–1232.
536 doi:10.1139/t04-059.
537 Lienhart, W. 2015. Case studies of high-sensitivity monitoring of natural and engineered slopes.
538 Journal of Rock Mechanics and Geotechnical Engineering, 7(4): 379–384.
539 doi:10.1016/j.jrmge.2015.04.002.
540 Liu, C.-N., Zornberg, J.G., Chen, T.-C., Ho, Y.-H., and Lin, B.-H. 2009. Behavior of geogrid-
541 sand interface in direct shear mode. Journal of Geotechnical and Geoenvironmental
542 Engineering, 135(12): 1863–1871. doi:10.1061/(ASCE)GT.1943-5606.0000150.
543 Mitchell, J.K., and Soga, K. 2005. Fundamentals of Soil Behavior. In 3rd edition. John Wiley
544 & Sons, Hoboken, New Jersey, USA.
545 Mohamad, H., Soga, K., Bennett, P.J., Mair, R.J., and Lim, C.S. 2012. Monitoring twin tunnel
546 interaction using distributed optical fiber strain measurements. Journal of Geotechnical
547 and Geoenvironmental Engineering, 138(8): 957–967. doi:10.1061/(ASCE)GT.1943-
548 5606.0000656.
549 Moraci, N., Cardile, G., Gioffrè, D., Mandaglio, M.C., Calvarano, L.S., and Carbone, L. 2014.
550 Soil geosynthetic interaction: Design parameters from experimental and theoretical
551 analysis. Transportation Infrastructure Geotechnology, 1(2): 165–227.
552 doi:10.1007/s40515-014-0007-2.
Page 22 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
23
553 Mosallanezhad, M., Hataf, N., and Ghahramani, A. 2008. Experimental study of bearing
554 capacity of granular soils, reinforced with innovative grid-anchor system. Geotechnical
555 and Geological Engineering, 26(3): 299–312. doi:10.1007/s10706-007-9166-z.
556 Ni, P., Moore, I.D., and Take, W.A. 2018. Distributed fibre optic sensing of strains on buried
557 full-scale PVC pipelines crossing a normal fault. Geotechnique, 68(1): 1–17.
558 doi:10.1680/jgeot.16.P.161.
559 Nöther, N., and von der Mark, S. 2019. Distributed Brillouin sensing for geotechnical
560 infrastructure: Capabilities and challenges. Geotechnical Engineering Journal of the
561 SEAGS & AGSSEA, 50(2): 8–12.
562 Pelecanos, L., Soga, K., Elshafie, M.Z.E.B., de Battista, N., Kechavarzi, C., Gue, C.Y., Ouyang,
563 Y., and Seo, H.-J. 2018. Distributed fiber optic sensing of axially loaded bored piles.
564 Journal of Geotechnical and Geoenvironmental Engineering, 144(3): 04017122.
565 doi:10.1061/(ASCE)GT.1943-5606.0001843.
566 Pérez, Z.A., Schiavon, J.A., Tsuha, C. de H.C., Dias, D., and Thorel, L. 2018. Numerical and
567 experimental study on influence of installation effects on behaviour of helical anchors in
568 very dense sand. Canadian Geotechnical Journal, 55(8): 1067–1080. doi:10.1139/cgj-
569 2017-0137.
570 Sadat Taghavi, S.H., and Mosallanezhad, M. 2017. Experimental analysis of large-scale pullout
571 tests conducted on polyester anchored geogrid reinforcement systems. Canadian
572 Geotechnical Journal, 54(5): 621–630. doi:10.1139/cgj-2016-0365.
573 Schenato, L. 2017. A review of distributed fibre optic sensors for geo-hydrological applications.
574 Applied Sciences, 7(9): 896. doi:10.3390/app7090896.
575 Schenato, L., Palmieri, L., Camporese, M., Bersan, S., Cola, S., Pasuto, A., Galtarossa, A.,
576 Salandin, P., and Simonini, P. 2017. Distributed optical fibre sensing for early detection
577 of shallow landslides triggering. Scientific Reports, 7(1): 14686. doi:10.1038/s41598-017-
578 12610-1.
Page 23 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
24
579 Shi, B., Xu, H., Chen, B., Zhang, D., Ding, Y., Cui, H., and Gao, J. 2003. A feasibility study
580 on the application of fiber-optic distributed sensors for strain measurement in the Taiwan
581 Strait Tunnel project. Marine Georesources & Geotechnology, 21(3–4): 333–343.
582 doi:10.1080/713773406.
583 Soga, K., Ewais, A., Fern, J., and Park, J. 2019. Advances in Geotechnical Sensors and
584 Monitoring. In Geotechnical Fundamentals for Addressing New World Challenges, 1st
585 edition. Edited by N. Lu and J.K. Mitchell. Springer International Publishing, Cham. pp.
586 29–65. doi:10.1007/978-3-030-06249-1.
587 Soga, K., and Luo, L. 2018. Distributed fiber optics sensors for civil engineering infrastructure
588 sensing. Journal of Structural Integrity and Maintenance, 3(1): 1–21.
589 doi:10.1080/24705314.2018.1426138.
590 Thévenaz, L. 2010. Brillouin distributed time-domain sensing in optical fibers: State of the art
591 and perspectives. Frontiers of Optoelectronics in China, 3(1): 13–21. doi:10.1007/s12200-
592 009-0086-9.
593 Wang, Q., Ye, X., Wang, S., Sloan, S.W., and Sheng, D. 2017. Experimental investigation of
594 compaction-grouted soil nails. Canadian Geotechnical Journal, 54(12): 1728–1738.
595 doi:10.1139/cgj-2017-0063.
596 Wheeler, L.N., Take, W.A., Hoult, N.A., and Le, H. 2019. Use of fiber optic sensing to measure
597 distributed rail strains and determine rail seat forces under a moving train. Canadian
598 Geotechnical Journal, 56(1): 1–13. doi:10.1139/cgj-2017-0163.
599 Zhang, C.-C., Shi, B., Gu, K., Liu, S.-P., Wu, J.-H., Zhang, S., Zhang, L., Jiang, H.-T., and
600 Wei, G.-Q. 2018. Vertically distributed sensing of deformation using fiber optic sensing.
601 Geophysical Research Letters, 45(21): 11,732–11,741. doi:10.1029/2018GL080428.
602 Zhang, C.-C., Zhu, H.-H., Shi, B., and She, J.-K. 2014a. Interfacial characterization of soil-
603 embedded optical fiber for ground deformation measurement. Smart Materials and
604 Structures, 23(9): 095022. doi:10.1088/0964-1726/23/9/095022.
Page 24 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
25
605 Zhang, C.-C., Zhu, H.-H., Xu, Q., Shi, B., and Mei, G.-X. 2014b. Time-dependent pullout
606 behavior of glass fiber reinforced polymer (GFRP) soil nail in sand. Canadian
607 Geotechnical Journal, 52(6): 671–681. doi:10.1139/cgj-2013-0381.
608 Zhu, H.-H., Shi, B., and Zhang, C.-C. 2017. FBG-based monitoring of geohazards: Current
609 status and trends. Sensors, 17(3), 452. doi:10.3390/s17030452.
610 Zhu, H.-H., Shi, B., Zhang, J., Yan, J.-F., and Zhang, C.-C. 2014. Distributed fiber optic
611 monitoring and stability analysis of a model slope under surcharge loading. Journal of
612 Mountain Science, 11(4): 979–989. doi:10.1007/s11629-013-2816-0.
613 Zhu, H.-H., Yin, J.-H., Yeung, A.T., and Jin, W. 2011. Field pullout testing and performance
614 evaluation of GFRP soil nails. Journal of Geotechnical and Geoenvironmental Engineering,
615 137(7): 633–642. doi:10.1061/(ASCE)GT.1943-5606.0000457.
Page 25 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
26
616 Table 1. Physical and mechanical properties of the sand.
Parameter Unit Value
Specific gravity sG — 2.67
Effective grain size 10D mm 0.23
Average grain size 50D mm 0.35
Coefficient of uniformity uC — 1.61
Coefficient of curvature cC — 1.06
Minimum dry density dmin g/cm3 1.21
Maximum dry density dmax g/cm3 1.69
Friction angle ° 32.3
617
Page 26 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
27
618 Table 2. Properties of the tight-buffed single-mode fiber optic (FO) strain cable.
Outer diameter
(mm)Coating
materialBuffer Coating
Average
Young’s
modulus
(GPa)
Unit
weight
(kg/km)
Tensile
strength
(MPa)
Effective
axial stiffness
(kN)a
Polyurethane 0.9 2.0 0.34 4.1 23.1 1.1
619 aDetermined under a tensile strain of 1%.
Page 27 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
28
620 Table 3. Properties of two types of micro-anchor.
MaterialLength
(mm)
Spacing
(mm)
Diameter
(mm)
Average Young’s modulus
of anchored segment
(GPa)a
Average
diameter of cable
(mm)b
PE tubec 25 250 3.6 0.78 2.17
PMMA
tubed25 250 6.0 2.70 2.42
621 aCalculated based on cross-section geometry.
622 bCalculated based on micro-anchor distribution.
623 cPE = polyethylene.
624 dPMMA = polymethyl methacrylate.
Page 28 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
29
625 Table 4. Validity of FO strain data related to a three-stage process of the cable–soil interface
626 failure.
StagePeak ISS of unanchored
segment reached or nota
Peak ISS of anchored
segment reached or not
Coupling
conditionFO strain data
I No NoOverall
bondingValid
II Yes NoPoint
fixationb
Averaged over
gauge lengthc
III Yes Yesd Decoupling Invalid
627 aISS = interfacial shear stress.
628 bThe micro-anchored FO cable is similar to a gauged strainmeter.
629 cThe gauge length is equivalent to anchor spacing.
630 dProvided that the cable itself has not yielded.
Page 29 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
30
631 Figure captions
632 Fig. 1. Schematic diagram of ground deformation sensing using a micro-anchored fiber optic
633 (FO) strain cable.
634 Fig. 2. (a) Distributions of axial force and interfacial shear stress (ISS) along an unanchored
635 soil inclusion during pullout at the critical state; and (b) hypothesized axial force and ISS
636 distributions for an anchored soil inclusion.
637 Fig. 3. Schematic of the principle of distributed fiber optic sensing (DFOS) using Brillouin
638 optical time-domain analysis (BOTDA).
639 Fig. 4. Grain size distribution curve of the soil.
640 Fig. 5. (a) Structure of the FO cable; and (b) layout of micro-anchors on the FO cable (unit:
641 mm).
642 Fig. 6. Schematic of the test setup (unit: mm; not to scale).
643 Fig. 7. Schematic of the clamp used to clamp the FO cable (unit: mm; not to scale).
644 Fig. 8. Experimental curves of pullout force versus pullout displacement under different micro-
645 anchor diameters.
646 Fig. 9. Influence of micro-anchor on the secant interfacial shear stiffness and peak ISS. The
647 secant interfacial shear stiffness was determined under a displacement step of 2 mm.
648 Fig. 10. Comparison between pullout forces measured by the force gauge and those calculated
649 using FO strain measurements.
650 Fig. 11. Strain profiles along the FO cables captured using BOTDA during the progressive
651 failure of the cable–soil interface: (a) unanchored FO cable; (b) micro-anchored FO cable (Da
652 = 3.6 mm); and (c) micro-anchored FO cable (Da = 6.0 mm).
653 Fig. 12. Simulated results for the unanchored FO cable: (a) pullout force–displacement curve;
654 (b) strain profiles; and (c) ISS profiles. Five pullout phases are marked in (a) according to the
655 simulation results.
Page 30 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
31
656 Fig. 13. Overall average ISS and ISS along unanchored and anchored segments for the micro-
657 anchored FO cable with: (a–c) Da = 3.6 mm; and (d–f) Da = 6.0 mm.
658 Fig. 14. Difference of average ISS between anchored and unanchored segments of the micro-
659 anchored FO cables.
660 Fig. 15. Validity of FO strain data as related to a three-stage process of the cable–soil interface
661 failure. The unanchored segment B and two neighboring anchored segments a and b for the
662 micro-anchored FO cable with Da = 6.0 mm are taken as an instance.
Page 31 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
Stable ground Unstable groundShear zone
Movement directionFO cable Micro-anchors
Fig. 1. Schematic diagram of ground deformation sensing using a micro-anchored fiber optic (FO) strain cable.
Page 32 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
(a)
(b)
ISS
Axial force
ISS
Axial force
Pulloutforce
Pulloutforce
Mobilized ISS
Mobilized ISS
Fig. 2. (a) Distributions of axial force and interfacial shear stress (ISS) along an unanchored soil inclusion during pullout at the critical state; and (b) hypothesized axial force and ISS distributions for an anchored soil inclusion.
Page 33 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
DraftBrillouin frequency
FO cable distance
Appliedstrain
Pumping pulse light
Brillouin backscattering
Probelight
Light intensity
Fig. 3. Schematic of the principle of distributed fiber optic sensing (DFOS) using Brillouin optical time-domain analysis (BOTDA).
Page 34 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
0
25
50
75
100
101
Grain size (mm)10010-1 10-2
Perc
enta
ge fi
ner b
y w
eigh
t (%
)
Fig. 4. Grain size distribution curve of the soil.
Page 35 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
(b)
250
1,200
250 250 250 50150
25
6.0
3.6
Da
(a)
Polyurethane coating
Optical fiber
Buffer
Fig. 5. (a) Structure of the FO cable; and (b) layout of micro-anchors on the FO cable (unit: mm).
Page 36 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft1,500
Soil
200 200
Test container
Clamp
FO cable
1,20010.2
Forcegauge
BOTDAinterrogator
Tensile tester
Computer
Fig. 6. Schematic of the test setup (unit: mm; not to scale).
Page 37 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
FO cable
Rubber pieces(69×20×2.2)
Aluminumpieces
(69×20×5.6)
Connected totensile tester
Bolts
Cross-sectional view
Plan view
Fig. 7. Schematic of the clamp used to clamp the FO cable (unit: mm; not to scale).
Page 38 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
DraftPullout displacement, u0 (mm)
Pullo
ut fo
rce,
F0 (
N)
UnanchoredAnchored (Da = 3.6 mm)Anchored (Da = 6.0 mm)
0
5
10
15
20
25
0 5 10 15 20 25
Fig. 8. Experimental curves of pullout force versus pullout displacement under different micro-anchor diameters.
Page 39 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
Seca
nt in
terfa
cial
shea
r stif
fnes
s (k
Pa/m
)Peak ISS (kPa)
Anchored(Da = 3.6 mm)
Unanchored Anchored(Da = 6.0 mm)
0
50
100
150
200
250
0
0.5
1
1.5
2
2.5Secant interfacial shear stiffnessPeak ISS
Fig. 9. Influence of micro-anchor on the secant interfacial shear stiffness and peak ISS. The secant interfacial shear stiffness was determined under a displacement step of 2 mm.
Page 40 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
0
5
10
15
20
0 5 10 15 20Calculated pullout force (N)
Mea
sure
d pu
llout
forc
e (N
)
UnanchoredAnchored (Da = 3.6 mm)Anchored (Da = 6.0 mm)
1:1 ratio
Fig. 10. Comparison between pullout forces measured by the force gauge and those calculated using FO strain measurements.
Page 41 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
0
1,000
2,000
3,000
4,000
5,000
2,000
4,000
6,000
8,000
10,000
0
3,000
6,000
9,000
12,000
15,000
Stra
in (μ
ε)St
rain
(με)
Stra
in (μ
ε)
0
7 11 1422
1 2 34 5 6
7 9
1 2 34 5 6
Disp. steps (mm)
7 9 1113 15 1719 21 24
1 2 34 5 6
0 200 400 600 800 1,000 1,200Distance from FO cable head (mm)
(a)
(b)
(c)
u⁰ (mm)
F⁰ (
N)
B
C
D
E
B
C
D
E
a b c d
a b c d
A
A
0
8
0 8 16 24
0
14
0 8 16 24
0
24
0 12 24
u⁰ (mm)
F⁰ (
N)
u⁰ (mm)
F⁰ (
N)
Fig. 11. Strain profiles along the FO cables captured using BOTDA during the progressive failure of the cable–soil interface: (a) unanchored FO cable; (b) micro-anchored FO cable (Da = 3.6 mm); and (c) micro-anchored FO cable (Da = 6.0 mm).
e
e
Page 42 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
0
1,000
2,000
3,000
4,000
5,000
0
2
4
6
0 5 10 15 20 25Pullout displacement (mm)
Pullo
ut fo
rce
(N)
MeasuredSimulated
Stra
in (μ
ε)
Distance from FO cable head (mm)(a) (b)
7 mm 9 mm
1 mm 2 mm3 mm 4 mm5 mm 6 mm
Applied
0
0.2
0.4
0.6
0.8
Sim
ulat
ed IS
S (k
Pa)
Distance from FO cable head (mm)(c)
0 200 400 600 800 1,000 1,200
0 200 400 600 800 1,000 1,200
Simulated
1 mm 2 mm 3 mm4 mm
5 mm
6 mm7 mm
8 mm
Fig. 12. Simulated results for the unanchored FO cable: (a) pullout force–displacement curve; (b) strain profiles; and (c) ISS profiles. Five pullout phases are marked in (a) according to the simulation results.
III III IV V
Page 43 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
0
0.5
1
1.5
2
2.5
-0.5
0
0.5
1
1.5
0
250
500
750
0 7 14 21 28 35Displacement steps (mm)
0
0.5
1
1.5
2
0
0.2
0.4
0.6
0.8
0
20
40
60
0 5 10 15 20 25
Aver
age
ISS
over
who
le c
able
leng
th (k
Pa) (a)
Displacement steps (mm)
ISS
alon
gun
anch
ored
seg
men
t (kP
a)IS
S al
ong
anch
ored
seg
men
t (kP
a)
(b)
(c)Av
erag
e IS
S ov
erw
hole
cab
le le
ngth
(kPa
)IS
S al
ong
unan
chor
ed s
egm
ent (
kPa)
ISS
alon
g an
chor
ed s
egm
ent (
kPa)
(d)
(e)
(f)
BA C D E Avg.
dba
c Avg.dba
c Avg.
BA C D E Avg.
Fig. 13. Overall average ISS and ISS along unanchored and anchored segments for the micro-anchored FO cable with: (a–c) Da = 3.6 mm; and (d–f) Da = 6.0 mm.
Residual ISS alongunanchored FO cable
Residual ISS alongunanchored FO cable
Page 44 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
0
5
10
15
0 5 10 15 20 25
y = 0.39x + 2.64R² = 0.95
y = 0.53x + 4.53R² = 0.97
Displacement steps (mm)
Diff
eren
ce o
f ave
rage
ISS
kAnchored (Da = 3.6 mm)Anchored (Da = 6.0 mm)
Fig. 14. Difference of average ISS between anchored and unanchored segments of the micro-anchored FO cables.
Page 45 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal
Draft
-0.5
0
0.5
1
1.5
0
250
500
750
0 7 14 21 28 35Displacement steps (mm)
Stage I Stage II Stage III
ISS
alon
g an
chor
ed s
egm
ent (
kPa)
ISS alongunanchored segm
ent (kPa)
Unanchored seg. B
Anchored seg. bAnchored seg. a
Fig. 15. Validity of FO strain data as related to a three-stage process of the cable–soil interface failure. The unanchored segment B and two neighboring anchored segments a and b for the micro-anchored FO cable with Da = 6.0 mm are taken as an instance.
Page 46 of 46
https://mc06.manuscriptcentral.com/cgj-pubs
Canadian Geotechnical Journal