Optical Fiber Technology 42 (2018) 97–104

Contents lists available at ScienceDirect

Optical Fiber Technology

journal homepage: www.elsevier.com/locate/yofte

Improving the durability of the optical fiber sensor based on strain transfer T
analysis
⁎
Huaping Wanga,c, Lizhong Jianga,b, Ping Xianga,b,d,
a
School of Civil Engineering, Central South University, Changsha 410075, Hunan, China
b
National Engineering Laboratory for High-speed Railway Construction, Changsha, China
c
Department of Civil and Environmental Engineering, The Hong Kong Polytechnic University, Hong Kong Special Administrative Region
d
Department of Architecture and Civil Engineering, City University of Hong Kong, Hong Kong Special Administrative Region

A R T I C LE I N FO A B S T R A C T

Keywords: To realize the reliable and long-term strain detection, the durability of optical fiber sensors has attracted more
Optical fiber sensor and more attention. The packaging technique has been considered as an effective method, which can enhance
Interfacial bonding properties the survival ratios of optical fiber sensors to resist the harsh construction and service environment in civil
Strain transfer analysis engineering. To monitor the internal strain of structures, the embedded installation is adopted. Due to the
Interfacial debonding failure
different material properties between host material and the protective layer, the monitored structure embedded
Error modification
with sensors can be regarded as a typical model containing inclusions. Interfacial characteristic between the
Application design
sensor and host material exists obviously, and the contacted interface is prone to debonding failure induced by
the large interfacial shear stress. To recognize the local interfacial debonding damage and extend the effective
life cycle of the embedded sensor, strain transfer analysis of a general three-layered sensing model is conducted
to investigate the failure mechanism. The perturbation of the embedded sensor on the local strain field of host
material is discussed. Based on the theoretical analysis, the distribution of the interfacial shear stress along the
sensing length is characterized and adopted for the diagnosis of local interfacial debonding, and the sensitive
parameters influencing the interfacial shear stress are also investigated. The research in this paper explores the
interfacial debonding failure mechanism of embedded sensors based on the strain transfer analysis and provides
theoretical basis for enhancing the interfacial bonding properties and improving the durability of embedded
optical fiber sensors.

1. Introduction For the brittle material properties of silica fiber, bare optical fiber is
weak to resist the shear or torsion force in structural construction and
The structural safety of civil infrastructures, ocean platforms and operation. Especially for the embedded case, the packaging technique is
aerospace structures has received increasing attention, because the the most critical factor to guarantee the survival and enhance the
failure of those important structures usually leads to large abundant of durability of optical fiber based sensors. However, the existence of the
casualties and economical loss. To characterize the structural perfor- protective layer introduces the intermedium between the sensing fiber
mance, structural health monitoring (SHM) technology has been re- and the monitored structure, which makes the strain measured by the
cognized as one of the most effective and intelligent measures sensor not entirely represent the actual strain of host material [2]. The
[18,19,1,23,22,10,7]. By the use of smart sensors and components, the error between the measured strain and the actual strain is attributed to
real-time, long-term and continuous information of the in-situ struc- the strain loss in the transferring path. To eliminate the strain transfer
tures can be provided for the damage identification, disaster forecasting error and improve the measurement accuracy of optical fiber based
and warming, and safety and life-time assessment [35,17,20,16,26]. sensors, strain transfer theory has been developed to establish the
Among these smart sensing elements, optical fiber based sensors are the quantitative relationship of strains between the host material and the
most popular in civil engineering for the unique advantages of high optical fiber [13,33,9,28].
sensitivity and precision, corrosion resistance, anti-electromagnetic Considerable attempts have contributed to studying the strain
interference, good stability, geometrical shape-versatility, absolute transfer mechanism of optical fiber sensors. The earliest research
measurement and convenient integration of network [25,34,32,12]. started from the 1990s, and the strain relationship between the

⁎
Corresponding author at: School of Civil Engineering, Central South University, Changsha 410075, Hunan, China.
E-mail addresses: [email protected], [email protected] (P. Xiang).

https://doi.org/10.1016/j.yofte.2018.02.004
Received 15 November 2017; Received in revised form 28 January 2018; Accepted 5 February 2018
1068-5200/ © 2018 Elsevier Inc. All rights reserved.
H. Wang et al. Optical Fiber Technology 42 (2018) 97–104

Fig. 1. The optical fiber based sensors used in practical engineering: (a) Steel packaged FBG sensor developed by HBM; (b) GFRP packaged FBG sensor developed by SCAIME; (c) Stainless
steel packaged FBG sensor developed by MOI (d) CFRP/GFRP packaged FBG sensor developed by TELL.

concrete and sensing fiber was studied with a polymer-to-glass modulus effective measurement. For this reason, the interfacial damage identi-
ratio of 1/200 given [21]. Host material with optical fiber sensor em- fication and measurement accuracy requires further investigation
bedded was then simplified to infinite elastic cylinder model, and through the strain transfer analysis between the packaged sensor and
plane-strain theory was adopted to explore the strain transfer me- host material.
chanism [24]. However, the simplified model in the two theories Given the analysis above, the possible failure modes of optical fiber
couldn’t be used to accurately determine the strain transfer relation- sensors available in the market for the embedment in structures are
ships for various host materials with non-elastic behavior. In 1998, a discussed by considering the strain transfer mechanism. Analysis on the
systematic strain transfer theory of a three-layered structure embedded influence of local interfacial debonding between the embedded sensor
with optical fiber sensor was established [3]. To analyze the effects of a and host material is studied theoretically. The perturbation of the em-
local interfacial slippage on the strain transfer ratio, a two-layered bedded sensor on the strain field of the structure is discussed. Based on
mechanical model consisted of host material and optical fiber was the strain transfer analysis, theoretical approach to diagnose the oc-
discussed [13]. For a multi-layered structure with various packaging currence of interfacial debonding and debonding length is proposed.
layers, the unified strain transfer formula was conducted [33]. The Furthermore, the sensitive parameters influencing the interfacial
improved strain transfer deduction of a three (multi)-layered sensing bonding properties are discussed and suggestions on the application
model by the use of simplified geometrical and physical functions was design of embedded sensors are provided for improving the durability
proposed [15]. Strain transfer analysis was also extended to special of the sensor.
cases for considering the viscoelastic material properties of the mon-
itored structure [30]. For surface-attached point optical fiber sensors,
the strain transfer mechanism and sensitivity of influencing parameters 2. Optical fiber based sensors with enhanced performance
was studied [27]. The strain transfer of surface-attached distributed
optical fiber sensors with one crack in host material was explored [6]. The optical fiber based sensors have been prevailed for decades, and
Besides, the dynamic strain transfer relationship of the sensing model optical fiber has been packaged with various materials to enhance the
under fatigue load was investigated [31]. In general, most of the ex- performance and the robustness of the sensors in practical engineering.
isting strain transfer theory mentioned above is based on the non- Available sensors used for the inside strain detection of concrete (or
destructive models, and limit consideration focuses on the damaged composite) structures in market majorly contains the following four
model. Besides, the current studies are confined to the strain transfer types presented in Fig. 1, which are separately developed by four in-
error modification. The strain transfer theory has yet been adopted to ternationally famous companies. Steel, glass fiber reinforced plastic
explore the failure mechanism and application design of optical fiber (GFRP), carbon fiber reinforced plastic (CFRP) and stainless steel have
based sensors. been separately used as the packaging materials to protect the sensing
In practical engineering, the premature failure of optical fiber based fiber. It can be noted that the strain of the structure is majorly trans-
sensors becomes a common phenomenon. Many installed sensors are ferred to the sensor by interfacial shear force. Therefore, the interfacial
out of service in 5 years, the life cycle of which is quite shorter than that bonding strength is particularly important to guarantee an effective
of the monitored structure. For embedded cases, it is usually difficult measurement. Besides of the embedded FBG sensor provided by HBM,
for rehabilitation or replacement, which threats the real-time perfor- the surface of the other three sensors separately developed by SCAIME,
mance monitoring of on-site structures. Therefore, the durability and MOI and TELL have been polished to increase the roughness. This
long-term performance of optical fiber based sensors draws consider- measure partially improves the interfacial bonding state and finally
able attention [29]. When the rehabilitation is not so convenient and benefits the long-term effective measurement of the embedded sensors.
the sensor is claimed to work normally, strain transfer error modifica- However, how to quantitatively assess the interfacial bonding strength
tion with the damaged cases considered becomes particularly sig- and scientifically enhance the roughness of the packaged sensor still
nificant. The interfacial debonding between the embedded sensor and requires reliable theoretical investigation. For this reason, the strain
host material is one of the most common failure modes that should be transfer theory is adopted to explore the interfacial failure mechanism
carefully studied, for it associates with the reliable use and measure- of the three-layered sensing model.
ment accuracy of optical fiber sensors in practical application. If the
local interfacial debonding exists in the embedded sensing model, the
strain transfer error modification is also demanded to ensure the

98
H. Wang et al. Optical Fiber Technology 42 (2018) 97–104

Fig. 2. Optical fiber sensor embedded in structure: (a) Testing model; (b) Stress state of optical fiber; (c) Stress state of protective layer; (d) Stress state of the model.

3. Strain transfer analysis 3.2. Theoretical derivation

3.1. Model description As shown in Fig. 2(b), the equilibrium of optical fiber sensing layer
gives
In order to obtain the internal strain of concretes, composite
dσf (x ) rf
structures and multi-layered pavements, optical fiber based sensors are · + τf (rf ,x ) = 0
dx 2 (1)
usually required to be embedded and the protective layer is im-
plemented to guarantee the survival. For this reason, the typical optical where axial stress of optical fiber is independent with variables r and θ .
fiber sensing model is usually composed of host material, protective As the slenderness ratio of optical fiber is large, the inner transverse
layer and optical fiber, as shown in Fig. 2(a). The optical fiber used in interaction can be ignored and the assumption is reasonable [4].
sensing field is often constituted of fiber core, cladding and coating. The Equilibrium equation can be established by considering the shearing
materials of fiber core and cladding are both silicon dioxides, and that action of the protective layer
of the coating is organic polymer. The three layers jointly bear the 2π 2π
deformation induced by external action. The interfacial debonding ∫0 τf (rf ,x ) rf dθdx = ∫0 τp (rp,x ) rp dθdx . (2)
between the fiber core, cladding and coating is not desired in practical
testing [14]. Otherwise, the brittle fiber core is very vulnerable to Since normal stress of the sensing fiber, σf , is independent of θ , both
the shear stresses τf and τp are independent of θ on the basis of Eq. (1).
breakage, which leads to the failure of the measurement. Some pub-
Then, Eq. (2) can be simplified as
lished researches stated that the coating absorbs a small part of strain
and induces the strain transfer error [3,15]. However, compared with rp
τf (rf ,x ) = ·τp (rp,x )
the practical engineering scale, the thickness of the coating (62.5 μm) is rf (3)
extremely small, and the induced measurement error can be neglected.
Therefore, the optical fiber is considered as one layer in the theoretical where τp (rp,x ) is the interfacial shear stress between the protective layer
and host material.
model, with radius equal to 125 μm [31].
A part of the monitored structure is selected, as shown in Fig. 2(d).
The embedded sensor accompanied with host material bears the
From the overall equilibrium along the axis x direction, it can be known
external effect. When local interfacial shear stress is up to the critical
that
shear strength, the interface is possible to suffer from local debonding.
For the embedded case, the local interfacial debonding between the σf (x )·r f2 + σm·(rm2 −rp2) = σm0·rm2 . (4)
sensor and the host material is difficult to be repaired, which can induce
strain transfer loss and influence the measurement accuracy [28]. In Because the shear strain of protective layer, γp (x ) , is determined by
practical situation, the embedded sensor is required to continuously displacements of optical fiber and host material [4,11], an equation can
monitor the strain of the monitored structure. To accurately interpret be generated
the true information of the structure, it is particularly important to um (x )−uf (x ) = γp (x )(rp−rf ). (5)
diagnose whether the local interfacial debonding between the em-
bedded sensor and the host material occurs. Therefore, study on strain By adopting the geometrical and physical equations, and taking the
transfer mechanism of the embedded sensing model is significant, derivation of Eq. (5) with respect to x , we can obtain
which can be adopted to identify the debonding damage and explore (rp−rf ) dτp (rp,x )
the influence of local interfacial debonding on the strain transfer error- εm−εf (x ) = · .
Gp dx (6)
modification principle. The work conducted in this section aims to
provide a corresponding error modification formula for the high-pre- Taking the derivation of Eq. (1) with respect to x , and replacing
cision and long-term strain detection of the embedded sensors in the dτf (rf ,x )/ dx by Eqs. (1) and (6) can be further expressed as
whole life cycle.
Ef rf (rp−rf ) d 2εf (x ) σ
Before the theoretical analysis, the following assumptions are made: · −εf (x ) = − m .
(1) the deformation of host material is transferred to the sensing fiber 2Gp dx 2 Em (7)
by the contacted interfaces; (2) the disturbing of packaged optical fiber
By using Eq. (4), the stress of host material, σm , can be replaced by
sensor on the far-end stress field of host material is ignored. The three-
far-end stress of host material, σm0 , and then Eq. (7) can be rewritten as
layered structure is affected by tensile stress σm0 from far end. The ra-
diuses of optical fiber, protective layer and host material are rf , rp and d 2εf (x ) σm0
−λ12 ·εf (x ) = −λ 02
rm , respectively. Sensing length of the sensor is 2L . dx 2 Em (8)
where σm0/ Em is the far-end strain of host material and denoted as εm0 ,
constants λ1 and λ 0 are used to replace the constant terms and expressed

99
H. Wang et al. Optical Fiber Technology 42 (2018) 97–104

as

2Gp ⎡ Ef r f2 ⎤ 2Gp rm2

λ12 = ⎢1+ 2 2 ⎥
, λ 02 = .
Ef rf (rp−rf )
⎣ Em (rm−rp ) ⎦ Ef rf (rp−rf )(rm2 −rp2) (9)
The general solution of Eq. (8) can be given by
λ 02
εf (x ) = Asinh(λ1 x ) + B cosh(λ1 x ) + εm0.
λ12 (10)
Since no axial force or constraint is applied on the two ends of
optical fiber, the axial stresses at x = ± L are zero. According to the
Hooke’s law, the strain of optical fiber at the two ends can be known as
εf (± L) = 0. (11)
By substituting Eq. (11) into Eq. (10), the relationship between the
strain of optical fiber and the far-end strain of host material can be
expressed as
λ 02 ⎡ cosh(λ1 x ) ⎤
εf (x ) = εm0· 1−
λ12 ⎢⎣ cosh(λ1 L) ⎥⎦ (12)
Fig. 3. Strain transfer ratio along half of the bonded length.
where Eq. (12) explains the strain transfer relationship of the three-
layered sensing model at nondestructive stage.
To inspect the accuracy of the proposed strain transfer function,
quantitative comparison is required. Values of correlated parameters
included in Eq. (12) follow Table 1. Comparison with the representative
research from Ansari and Yuan [3] is displayed in Fig. 3. General trend
of the two curves is that strain transfer ratio decreases from symmetric
center to the ends, which declares the correctness and availability of the
proposed strain transfer function in this article. Strain transfer ratio at
the center (origin of abscissa axis in Fig. 3) is about 98.7%. For the
existence of the protective layer, strain of host material cannot be to-
tally delivered to the optical fiber. That is to say, strain transfer coef-
ficient β at the center ( x = 0 ) should be smaller than 1, which makes
this result approval.

4. Strain perturbation induced by the embedment of optical fiber

sensor

The analysis provided in this section can be used to consider the

perturbation of the embedded sensor on strain field of the monitored
structure and instruct the design on the protective thickness of the
sensor in practical engineering. General solution of Eq. (7) with two
Fig. 4. Strain transfer ratios of εf / εm and εf / εm0 at different radiuses of the protective
strain variables of optical fiber and host material around the sensor can
layer.
be obtained and further solved by considering boundary condition Eq.
(11):
and (13), the ratios of εf / εm and εf / εm0 are calculated out, as displayed in
cosh(λ2 x ) ⎤ Fig. 4. Good overlapping of the two curves is observed, which de-
εf (x ) = εm·⎡1−
⎢
⎣ cosh(λ2 L) ⎥ ⎦ (13) monstrates that the embedded sensing fiber with protective layer
2Gp (rp = 0.25 mm) has little influence on the strain distribution of the
where λ 22 = .
Ef r f (rp − r f ) structure.
Eq. (13) indicates the strain relationship between optical fiber and However, to ensure the survival of the embedded sensor in civil
host material around the embedded sensor. By comparing Eq. (12) with engineering, the out radius of the protective layer can be up to 5 mm. In
Eq. (13), the influence of embedded optical fiber sensors on the dis- this case, the influence of the sensor on the strain field of the monitored
turbance of local strain field of the monitored structure can be in- structure (as shown in Fig. 4) cannot be ignored, and using Eq. (12) can
spected. Substituting the correlated parameters in Table 1 into Eqs. (12) much accurately reflect the actual strain of the host material. It also
should be noted that the low strain transfer ratio induced by the thick
Table 1 protective coatings means the strain measured by the optical fiber
Material and geometrical parameters of the three-layered sensing model.
sensor quite smaller than that of the host material for the high strain
Mechanical title Label Value Unit loss in the transferring path. In other words, the optical sensing element
should be very sensitive to the variation of the micro strain, and then
Young’s modulus of optical fiber Ef 7.2 × 10 10
N/m2 FBG sensor with wavelength resolution up to 1 pm can be adopted in
Radius of optical fiber rf 1.25 × 10−4 m
this case. Besides, in practical application, the embedded sensor should
Shear modulus of the protective layer Gp 2.25 × 106 N/m2
Out radius of the protective layer rp 2.5 × 10−4 m have ignorable influence on the structural integrity. Therefore, the
Young’s modulus of host material Em 2.2 × 109 N/m2 thickness of the protective layer is limited in an appropriate range.
Radius of host material rm 0.05 m
Half of the bonded length L 0.04 m

100
H. Wang et al. Optical Fiber Technology 42 (2018) 97–104

5. Influence of local interfacial debonding on the strain transfer

ratio

For the embedded case, the sensor and the host material will jointly
bear the external force. After working for a much long time or suffering
from sudden large deformation, the interface between the sensor and
host material is possible to be locally damaged, and it may weaken the
sensing performance of the sensor and lead to the failure of the testing.
To implement the effective measurement, the local interfacial de-
bonding need to be diagnosed and the influence of local interfacial
debonding between the sensor and host material on the strain transfer
ratio is discussed.
When the interfacial stress between host material and the protective Fig. 6. The interfacial debonding between host material and the protective layer.

layer reaches the critical state, local interfacial debonding may occur.
Thus, the effect of shear force should be taken into consideration. By εf (−L + Ld ) = εf (L−Ld ) = 0. (16)
substituting Eq. (12) into Eq. (1), the shear stress τp (rp,x ) can be in-
Substituting Eq. (16) into Eq. (10), the relationship between the
corporated with Eq. (3), and further expressed as
strain of optical fiber and far-end strain of host material can be ex-
Ef r f2 λ 02 σm0 sinh(λ1 x ) pressed as
τp (rp,x ) = .
2rp λ1 Em cosh(λ1 L) (14) λ 02 ⎡ cosh(λ1 x ) ⎤
εf (x ) = εm0 1−
By utilizing data from Table 1 and Eq. (14), distribution of the ratio λ12 ⎣⎢ cosh(λ1 Ld ) ⎦⎥ (17)
of shear stress to far-end stress along half of the bonded length can be where Eq. (17) is the strain transfer error correction formula with the
obtained as shown in Fig. 5. The shear stress shows a nonlinear increase influence of local interfacial debonding considered. In the proposed
by approaching to the end. That is to say, the debonding damage often model, if the maximum interfacial shear stress between host material
begins from the end. In the proposed model, the maximum shear stress and the sensor is 0.05 times of the far-end normal stress, namely
located at the bonded-length end and the value is τp = 0.064σm0 . Critical τpcr = 0.05σm0 , the critical effective bonded length can be calculated by
interfacial shear stress between host material and the protective layer using Eq. (15) as Ld = 0.036 m. Substituting Ld to Eq. (17), the strain
can be represented as τpcr , which is usually a constant. When the ma- transfer ratio can be figured out and the variation along the bonded
terials of the monitored structure and protective layer are selected length is displayed in Fig. 7.
[14,8], it can be further determined by experiments. It can be seen that the local debonding brings about the decrease of
When τp |x = L < τpcr , no debonding appears in the interface and Eq. strain transfer efficiency, and the strain transfer ratio will drop sig-
(12) can be used to modify the strain transfer error of the sensing nificantly with the increase of the debonding length. Therefore, it is
model. Otherwise if τp |x = L ⩾ τpcr , local debonding will occur in the in- important to prevent the local debonding at the interface between the
terface between the host material and the sensor [5]. By using Eq. (14), packaged sensor and the host material.
the critical debonding point Ld is solved out

2rp λ1 Em τpcr 6. Parametric studies

sinh(λ1 Ld ) = cosh(λ1 L)
r f2 λ 02 Ef σm0 (15)
Given the analysis above, it can be noted that the interfacial shear
where Ld is the debonding point when the host material is subjected to stress determines whether the local debonding damage occurs.
the action of the far-end stress σm0 . At this state, the effective sensing Therefore, to effectively control the interfacial bonding state between
length is Ld , and the debonding length is L−Ld , as shown in Fig. 6. the sensor and the host material, and ensure the reliable detection of
Boundary conditions are changed in comparison with no-debonding optical fiber sensors embedded in civil structures, the correlated ma-
status, which gives terial and geometrical parameters that influence the interfacial shear

Fig. 5. Ratio of interfacial shear stress and remote stress along half of the bonded length. Fig. 7. Comparison of strain transfer ratios with and without local interfacial debonding.

101
H. Wang et al. Optical Fiber Technology 42 (2018) 97–104

Fig. 8. Influence of material and geometrical parameters of the protective layer on the interfacial shear stress: (a) Radius; (b) Shear modulus.

stress should be carefully discussed, so to understand the interfacial distribution of the interfacial shear stress along the bonded length with
mechanical action. Since the properties of optical fiber are constant, the Gp increased separately from 2.5 × 108 N/m2 (acrylate), 1.2 × 109 N/
parametric studies focus on that of the protective layer and host ma- m2 (polyimide) to 2.5 × 109 N/m2 (ormocer) has been investigated.
terial. The sensitivity analysis is based on the case with parameters When Gp varied from 2.25 × 107 N/m2 to 2.25 × 108 N/m2, the inter-
listed in Table 1. facial shear stress approaching the bonded-length end experiences a
large increase, with value varied from 0.2σm0 to 0.65σm0, which may
6.1. Effect of radius and shear modulus of the protective layer on the lead to end interfacial debonding. Although the higher shear modulus
interfacial shear stress of the protective coating can bring about a smaller interfacial shear
stresses in the bonded-length range of 0–0.035 m, the interfacial shear
Fig. 8(a) shows the ratio of interfacial shear stress τp and far-end stress at the end of the boned length is quite larger, which can easily
stress σm0 decreased with the increase of radius of the protective layer. induce interfacial end debonding. A comprehensive consideration is
The ratio τp/σm0 at the end of the bonded length varies from 0.064, required to reasonably select the protective material to guarantee the
0.017 to 0.004, with the radius of the protective layer changed from distributed interfacial shear stresses along the bonded length in much
0.25 mm, 0.5 mm to 1 mm. A significant decrease of the interfacial low level. In general, it is better to select flexible protective material
shear stress is observed with rp increased from 0.25 mm to 0.5 mm. with relatively high bonding performance. To prevent the occurrence of
Generally, it indicates that the thinner protective layer can bring about local interfacial debonding, the shear modulus of the protective layer is
larger interfacial shear stress. To prevent the local interfacial de- suggested to be around 2.25 × 107 N/m2 for the proposed model. For
bonding, the interfacial shear stress should be as smaller as possible. example, the high- or low-density polyethylene with the shear modulus
Therefore, it is important to relatively increase the thickness of the approaching the suggested values can be selected as the protective
protective layer, since the thicker protective layer may induce larger material for the sensing fiber.
strain transfer error.
Fig. 8(b) demonstrates the evolution of the ratio of interfacial shear 6.2. Effect of radius and Young’s modulus of host material on the interfacial
stress and far-end stress with the increase of shear modulus of the shear stress
protective layer. The protective material discussed in the model is
hyper-elastic silicone rubber, with a shear modulus of 2.25 × 106 N/ The well-overlapped curves in Fig. 9(a) demonstrate that the change
m2. The silicone rubber is a kind of flexible transparent adhesive with of radius of host material has ignorable influence on the interfacial
good elasticity, which has been widely used as coating material and shear stress. Therefore, other parameters should be particularly con-
substrate. Since acrylate, polyimide and ormocer can also be used as the sidered in the design of the embedded sensor, so as to decrease the risk
protective coatings of the sensing fiber, the case study on the of the local interfacial debonding. Fig. 9(b) illustrates that the higher

Fig. 9. Influence of material and geometrical parameters of the host material on the interfacial shear stress: (a) Radius; (b) Young’s modulus.

102
H. Wang et al. Optical Fiber Technology 42 (2018) 97–104

(1) The perturbation of the embedded sensor on the strain field of the
host material can be assessed by the overlapped degree of Eqs. (12)
and (13), which can be used to determine the maximum radius of
the protective layer.
(2) Eq. (14) can be used to diagnose the occurrence of local interfacial
debonding between the sensor and host material, and the related
strain transfer error modification of the model with influence of
local interfacial debonding considered follows Eq. (17). The local
interfacial debonding decreases the strain transfer ratio and enough
interfacial bonding strength should be guaranteed, especially at the
bonded-length ends of the multi-layered sensing model.
(3) To prevent the local interfacial debonding between the embedded
sensor and host material and enhance the effective measurement of
the sensor, the thicker protective layer with relatively lower mod-
ulus and the longer gauge length are suggested in the application
design. When the embedded sensor is employed to detect the strain
of host material with low modulus, the interfacial debonding should
be carefully considered by the proposed strain transfer theory.

Acknowledgements
Fig. 10. Influence of bonded length on the interfacial shear stress.
The writers are greatly appreciating for editors and referees
spending time in attentively reading and checking this article. This
Young’s modulus of host material brings about smaller interfacial shear work is supported by National Science and Technology Support Project
stress. When Em changes from 2.2 × 108 N/m2 to 2.2 × 109 N/m2, a (Grant No. 2011BAK02B01), State 863 Project (Grant No.
sharp decrease is noticed, and the decreasing amplitude becomes 2014AA110401) and Key Laboratory Scientific Research Special Topics
smooth with Em increased from 2.2 × 109 N/m2 to 2.2 × 1010 N/m2. (No. DUT1LAB03). Special thanks are given to Prof. Zhi Zhou and Prof.
This conclusion indicates that when the sensor is employed to detect the Jinping Ou of Dalian University of Technology. The findings and opi-
strain of host material with low modulus, the interfacial debonding nions expressed in this article are only those of the authors and do not
should be particularly considered. necessarily reflect the views of the sponsors.

6.3. Effect of bonded length on the interfacial shear stress References

Fig. 10 shows the evolution of interfacial shear stress with the [1] F. Ansari, Sensing Issues in Civil Structural Health Monitoring, Springer, Dordrecht,
variation of the bonded length. Generally speaking, the longer bonded The Netherlands, 2005.
[2] F. Ansari, Practical implement of optical fiber sensors in civil structural health
length brings about relatively larger interfacial shear stress. When L
monitoring of bridges, J Intell Mater Syst Struct 18 (8) (2007) 879–889, http://dx.
increases from 0.02 m to 0.04 m, the interfacial shear stress at the doi.org/10.1177/1045389X06075760.
bonded-length end goes through a slow growth, and the maximum [3] F. Ansari, L.B. Yuan, Mechanics of bond and interface shear transfer in optical fiber
sensors, J Eng Mech 4 (1998) 385–394, http://dx.doi.org/10.1061/(ASCE)0733-
value increases from 0.055σm0 to 0.063σm0. When half of the bonded
9399(1998) 124:4(385).
length increases from 0.03 m to 0.04 m, the growth of the maximum [4] C.T. Chon, C.T. Sun, Stress distributions along a short fibre in fibre reinforced
interfacial shear stress becomes smooth. It means that the bonded plastics, J. Mater. Sci. 15 (1980) 931–938, http://dx.doi.org/10.1007/
length has relatively small influence on the maximum amplitude of the BF00552105.
[5] J.G. Dai, W.Y. Gao, J.G. Teng, Bond-slip model for frp laminates externally bonded
interfacial shear stresses. The increasing gradients of the three curves in to concrete at elevated temperature, J Compos Constr 17 (2) (2013) 217–228,
Fig. 10 indicate that the shorter bonded length can bring about a http://dx.doi.org/10.1061/(ASCE)CC.1943-5614.0000337.
sharper growth of the interfacial shear stresses. That is to say, the larger [6] X. Feng, J. Zhou, C.S. Sun, X.T. Zhang, F. Ansari, Theoretical and experimental
investigation into crack detection with BOTDA distributed fiber optic sensors, J Eng
normal stress of the host material can lead to higher interfacial shear Mech 139 (12) (2013) 1797–1807, http://dx.doi.org/10.1061/(ASCE)EM.1943-
stresses in the bonded-length interval, and local interfacial debonding 7889.0000622.
damage is prone to occurrence. The increasing gradients become gra- [7] D.M. Frangopol, M. Soliman, Life cycle of structural systems: recent achievements
and future directions, Struct Infrastruct Eng 12 (1) (2016) 1–20, http://dx.doi.org/
dual with the growth of L from 0.02 m to 0.04 m. Therefore, in practical 10.1080/15732479.2014.999794.
engineering, it is better to select relatively long bonded length to delay [8] W.Y. Gao, J.G. Dai, J.G. Teng, Analysis of mode II debonding behavior of fiber-
the occurrence of the local interfacial debonding when the other re- reinforced polymer-to-substrate bonded joints subjected to combined thermal and
mechanical loading, Eng Fract Mech 136 (2015) 241–264, http://dx.doi.org/10.
quirements are satisfied. For the given model, L equal to 0.03 m can be
1016/j.engfracmech.2015.02.002.
the best choice. [9] B. Glisic, Influence of the gauge length on the accuracy of long-gauge sensors em-
ployed in monitoring of prismatic beams, Meas Sci Technol 22 (3) (2011) 5206,
http://dx.doi.org/10.1088/0957-0233/22/3/035206.
7. Conclusions [10] B. Glisic, Y. Yao, Fiber optic method for health assessment of pipelines subjected to
earthquake-induced ground movement, Struct. Health Monit. 11 (6) (2012)
It is well known that the uncoordinated deformation induced by the 696–711, http://dx.doi.org/10.1177/1475921712455683.
[11] S.C. Her, C.Y. Huang, Effect of coating on the strain transfer of optical fiber sensors,
mismatch of material properties of each layer in a multi-layered Sensors 11 (7) (2011) 6926–6941, http://dx.doi.org/10.3390/s110706926.
structure is prone to interfacial debonding failure between the adjacent [12] D. Kinet, P. Megret, K.W. Goossen, L. Qiu, D. Heider, C. Caucheteur, Fiber Bragg
layers under the external loading. For the commonly used sensing grating sensors toward structural health monitoring in composites materials: chal-
lenges and solutions, Sensors 14 (2014) 7394–7419.
model, the interface between the monitored structure and the em-
[13] M. LeBlanc, Interaction mechanics of embedded single-ended optical fibre sensors
bedded sensor is also vulnerable to interfacial damage. To enhance the using novel in-situ measurement Techniques, University of Toronto, Canada, 1999
interfacial bonding properties and guarantee the relatively long-term (Thesis for the degree of Doctor).
effective measurement, theoretical approach based on the strain [14] C.K.Y. Leung, X.Y. Wang, N. Olson, Debonding and calibration shift of optical fiber
sensors in concrete, J Eng Mech 126 (3) (2000) 300–307, http://dx.doi.org/10.
transfer analysis is performed and the following conclusions can be 1061/(ASCE)0733-9399(2000) 126:3(300).
drawn from the study:

103
H. Wang et al. Optical Fiber Technology 42 (2018) 97–104

[15] H.N. Li, G.D. Zhou, L. Ren, D.S. Li, Strain transfer coefficient analysis for embedded K. Chah, H. Thienpont, F. Berghmans, Disbond monitoring in adhesive joints using
fiber Bragg grating sensors in different host materials, J Eng Mech 135 (12) (2009) shear stress optical fiber sensors, Smart Mater Struct 23 (2014) 1–10.
1343–1353, http://dx.doi.org/10.1061/(ASCE)0733-9399(2009) 135:12(1343). [27] K.T. Wan, C.K.Y. Leung, N.G. Olson, Investigation of the strain transfer for surface-
[16] G. Luyckx, E. Voet, N. Lammens, J. Degrieck, Strain measurements of composite attached optical fiber strain sensors, Smart Mater Struct 17 (2008) 0964–1726,
laminates with embedded fibre Bragg gratings: criticism and opportunities for re- http://dx.doi.org/10.1088/0964-1726/17/3/035037.
search, Sensors 11 (2011) 384–408. [28] H.P. Wang, Strain transfer of optical fiber under damage conditions and its appli-
[17] A. Mufti, Guidelines for Structural Health Monitoring. Design Manual No.2, ISIS cation in multi-layered pavements, Dalian University of Technology, Dalian, 2015
Canada, Manitoba, Canada, 2006. (Thesis for Doctoral Degree in Engineering).
[18] H. Murayama, K. Kageyama, H. Naruse, A. Shimada, K. Uzawa, Application of fiber- [29] H.P. Wang, L.Z. Jiang, P. Xiang, Priority design parameters of industrialized optical
optic distributed sensors to health monitoring for full-scale composite structures, J fiber sensors in civil engineering, Opt Laser Technol 100 (2018) 119–128.
Intell Mater Syst Struct 14 (2003) 3–11. [30] H.P. Wang, P. Xiang, Strain transfer analysis of optical fiber based sensors em-
[19] H. Murayama, K. Kageyama, T. Kamita, H. Igawa, Structural health monitoring of a bedded in an asphalt pavement structure, Meas Sci Technol 27 (7) (2016)
full-scale composite structure with fiber-optic sensors, Adv Compos Mater 11 (3) 75106–75117, http://dx.doi.org/10.1088/0957-0233/27/7/075106.
(2003) 287–297. [31] H.P. Wang, P. Xiang, X. Li, Theoretical analysis on strain transfer error of FBG
[20] H. Murayama, K. Kageyama, K. Uzawa, K. Ohara, H. Igawa, Strain monitoring of a sensors attached on steel structures subjected to fatigue load, Strain (2016), http://
single-lap joint with embedded fiber-optic distributed sensors, Struct. Health Monit. dx.doi.org/10.1111/str.12195 in press.
11 (3) (2011) 325–344. [32] H.P. Wang, W.Q. Liu, J.P. He, X.Y. Xing, D.D. Cao, X.P. Gao, X.W. Hao, H.W. Chen,
[21] A. Nanni, C.C. Yang, K. Pan, J.S. Wang, R.R.M. Jr, Fiber-optic sensors for concrete Z. Zhou, Functionality enhancement of industrialized optical fiber sensors and
strain/stress measurement, ACI Mater. J. 88 (3) (1991) 257–264. system developed for full-scale pavement monitoring, Sensors 14 (5) (2014)
[22] Y.Q. Ni, K.H. Lam, D. Zhu, Y. Wang, J.P. Lynch, K.H. Law, In-construction vibration 8829–8850, http://dx.doi.org/10.3390/s140508829.
monitoring of a super-tall structure using a long-range wireless sensing system, [33] Z. Zhou, Optical fiber smart Bragg grating sensors and intelligent monitoring sys-
Smart Struct. Syst. 7 (2011) 83–102. tems of civil infrastructures, Harbin Institute of Technology, Harbin, 2003
[23] J.P. Ou, H. Li, Structural health monitoring in mainland China: review and future (Dissertation for the Doctoral Degree in Engineering).
trends, Struct. Health Monit. 9 (3) (2010) 219–231, http://dx.doi.org/10.1177/ [34] Z. Zhou, W.Q. Liu, Y. Huang, H.P. Wang, J.P. He, M.H. Huang, J.P. Ou, Optical fiber
1475921710365269. Bragg grating sensor assembly for 3D strain monitoring and its case study in
[24] Y.E. Pak, Longitudinal shear transfer in fiber optic sensors, Smart Mater Struct 1 highway pavement, Mech Syst Sig Process 28 (2012) 36–49, http://dx.doi.org/10.
(1992) 57–62, http://dx.doi.org/10.1088/0964-1726/1/1/008. 1016/j.ymssp.2011.10.003.
[25] J.S. Sirkis, Unified approach to phase-strain-temperature models for smart structure [35] Z. Zhou, J.P. Ou, Development of FBG Sensors for Structural Health Monitoring in
interferometric optical fiber sensors: part 2, applications, Opt Eng 32 (4) (1993) Civil Infrastructures. Sensing Issues in Civil Structural Health Monitoring, Springer,
762–773, http://dx.doi.org/10.1117/12.61198. 2005 pp. 197–207.
[26] S. Sulejmani, C. Sonnenfeld, T. Geernaert, G. Luyckx, P. Mergo, W. Urbanczyk,

104