Structure and Infrastructure Engineering

This article was downloaded by: [Southeast University] On: 31 December 2013, At: 06:06 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Structure and Infrastructure Engineering: Maintenance, Management, Life-Cycle Design and Performance Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/nsie20

Structural health monitoring of a steel stringer bridge with area sensing Jian Zhang

a b

b

b

b

, Wan Hong , Yongsheng Tang , Caiqian Yang , Gang Wu

a b

& Zhishen Wu

a

b a

Key Laboratory of C&PC Structures of the Ministry of Education, Southeast University , Nanjing , China b

International Institute for Urban Systems Engineering, Southeast University , Nanjing , China Published online: 22 Apr 2013.

To cite this article: Jian Zhang , Wan Hong , Yongsheng Tang , Caiqian Yang , Gang Wu & Zhishen Wu , Structure and Infrastructure Engineering (2013): Structural health monitoring of a steel stringer bridge with area sensing, Structure and Infrastructure Engineering: Maintenance, Management, Life-Cycle Design and Performance, DOI: 10.1080/15732479.2013.787103 To link to this article: http://dx.doi.org/10.1080/15732479.2013.787103

PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions

Structure and Infrastructure Engineering, 2013 http://dx.doi.org/10.1080/15732479.2013.787103

Structural health monitoring of a steel stringer bridge with area sensing Jian Zhanga,b1, Wan Hongb2, Yongsheng Tangb3, Caiqian Yangb4, Gang Wua,b5 and Zhishen Wua,b* a

Key Laboratory of C&PC Structures of the Ministry of Education, Southeast University, Nanjing, China; bInternational Institute for Urban Systems Engineering, Southeast University, Nanjing, China

Downloaded by [Southeast University] at 06:06 31 December 2013

(Received 20 July 2012; final version received 16 January 2013; accepted 28 January 2013) The Federal Highway Administration Long-term Bridge Performance Programme initiated an International Bridge Study by selecting a steel stringer bridge as a benchmark structure for structural health monitoring. As a part of this programme, the authors studied the application of the Long-Gauge Fibre Bragg Grating (LG-FBG) sensors on this bridge. This paper aims at illustrating the LG-FBG-related state-of-the-art technologies by taking the bridge as the test bed. (1) The concept of the LGFBG sensor for area sensing is presented. Most fibre optic sensors measure point strains for local monitoring. In contrast, the developed LG-FBG area sensor has a long gauge (e.g. 1 – 2 m), and it can be connected to each other to make a sensor array for distributed strain measuring; (2) spectral analyses of the macro-strain time histories are performed to identify structural frequencies, and the results are compared with those estimated from acceleration measurements; (3) the neutral axis position of the girder of the investigated bridge is estimated from the recorded macro-strain time histories, and the results are compared with those from static truck tests and (4) a modal macro-strain-based damage index is applied for damage detection of the steel stringer bridge. Keywords: long-gauge; area sensor; macro-strain; dynamic measurement; structural health monitoring

Introduction Recent collapses of bridges, for instance, the I-35 W Bridge over the Mississippi River, have received a tremendous amount of public attention, which emphasise the importance of health monitoring for civil infrastructures. To formulate and demonstrate best practice technologies in this field, the Federal Highway Administration Long-term Bridge Performance Programme initiated an International Bridge Study (IBS) by selecting a steel stringer bridge in northern New Jersey as a benchmark structure (Weidner, Prader, Moon, & Aktan, 2011). Over 10 groups worldwide attended this IBS to demonstrate and verify their technologies for structural health monitoring (SHM). The most widely way adopted for SHM is acceleration-based structural testing (AbdelGhaffar & Scanlan, 1985; ASCE, 2012; Brownjohn et al., 2009; Conte et al., 2008; Ko, Sun, & Ni, 2002; Pakzad & Fenves, 2009; Peeters & DeRoeck, 2001; Siringoringo & Fujino, 2008; Zhang et al., 2013). However, accelerationbased SHM still faces challenges. For instance, the identified structural modal frequencies (e.g. frequencies) are considered being insensitive to a specific local damage even near the accelerometer. Strain is more sensitive to local damages and has potential to be a good candidate for the measurement for damage detection. The fibre optic strain sensing technique,

*Corresponding author. Email: [email protected] q 2013 Taylor & Francis

under rapid development recently, presents a promising tool for strain measurement (Ansari, 2007). Owing to small size and light-weight, optical fibre can be either surface attached or embedded into structures without changing their original structural mechanical behaviours. Among the fibre optic sensors, the Fibre Bragg Grating (FBG)-based strain sensor is the most suitable with its special features of high precision level, stable sensing capacity and reliability. However, the fibre optic sensors and other traditional strain gauges fail to detect a local damage unless the sensor covers the damage region because they generally have short gauge lengths (around 1– 2 cm). The emerging brillouin optical time domain reflecting (BOTDR) sensing system and Pulse Pre-Pump (PPP) – Brillouin optical time-domain analysis (BOTDA) technologies make the distributed sensing with optic fibres be possible (Murayama, Kageyama, Ohara, Uzawa, & Kanai, 2008; Zhang & Wu, 2008). Unfortunately, distributing sensing still has not been widely applied in engineering practices because the current distributed sensing systems (BOTDR, PPP-BOTDA, etc.) have limited precision, space resolution and low speed. Due to the limitations of traditional point strain gauges and distributed sensing systems discussed above, a LongGauge FBG (LG-FBG) area sensor has been developed by the authors (Li & Wu, 2007; Wu & Zhang, 2011). It has a


2

Figure 1.

J. Zhang et al.

The investigated bridge: (a) side view, (b) girders and (c) stringers.

long-gauge length (e.g. 1 – 2 m) aiming at measuring not only local strain but also carrying out area sensing for global structural monitoring. As a part of the IBS described above, the LG-FBG area sensor was applied for SHM of the benchmark bridge. The first objective of this study is to verify the effectiveness of the developed LG-FBG sensors for area strain measurement. This part includes comparing the time histories from the developed LG-FBG sensor and those from traditional strain gauges and comparing the spectral analysis results from macrostrains and accelerations, respectively. The second objective is to verify the macro-strain based structural diagnose technologies, including the neutral axis position determination from the macro-strain time histories and the

1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8

Southbound

SPAN 1#

SPAN 2# Northbound

1.52m 2.44m 17.8m

The bridge configuration.

Bridge description The investigated bridge is a multi-girder steel stringer bridge constructed in 1984 (Figure 1(a)). It has four spans with a standard steel stringer design of girders (Figures 1(b) and 2), and each span is simply supported with pin and rocker bearings. The girders of the bridge have variable section properties and geometries. Their flange thickness varies from 2.5 to 6.3 cm. The bridge contains diagonal wind braces between girders (Figure 1(c)). The diaphragms

Test Area

0.24m

Figure 2.

structural damage detection by employing a modal macrostrain (MMS)-based damage index.

SPAN 3#

SPAN 4#

Structure and Infrastructure Engineering


are standard truss-type composed of four single angles connected to the girders with bolted connections and gusset plates. The deck of the bridge was cast using stay-in-place forms. Fatigue cracks are present at the location of wind bracing-to-girder connection, and several cracks were also found on the piers, especially the west pier between the first and second spans has a large vertical crack (about 0.2 mm). This bridge was selected for investigation in the IBS because it displays very common problems associated with approach settlement, bearing alignment, substantial vibrations and fatigue cracking (Weidner et al., 2011).

Concept of the LG-FBG area sensor The LG-FBG sensor developed by Li and Wu (2007) was used to monitor the benchmark bridge described in last section. By designing the FBG sensor with a long gauge and fixing its two ends (Figure 3), the in-tube fibre has the same mechanical behaviour and hence the strain transferred from the shift of Bragg centre wavelength represents the average strain (or say macro-strain) over the sensor gauge length. An improved packaging design has also been proposed to enhance the measuring sensitivity of LG-FBG sensors by utilising two materials with different modulus to package the optic fibre and to impose deformation within the gauge length, largely on the essential sensing part of the FBG (Figure 3). Proof tests performed on a cantilever beam illustrate that the package almost has no effects on the accuracy of the macro-strain measurements. Due to its long-gauge length, the developed LG-FBG strain-sensing technique has the merit to measure structural damage (e.g. cracks) in a large area. Furthermore, the long-gauge sensors can be connected in series to make an

(a)

Fixing end

Connector

Fiber sheath

3

FBG sensor array (Figure 3(c)), which is suitable for area macro-strain measuring.

Bridge test and data analyses Instrument plan and field test The second span on the southbound side (Figure 2) of the bridge was focused on in the field test because it traverses a park which allows for unrestricted underside access. This span has the skew on one side as shown in Figure 2. To monitor this span using the developed LG-FBG area sensors, the gauge length of the sensor was designed to be 1 m, and a total of 44 sensors were deployed on two girders of this span, as depicted in Figure 4. The first 10 sensors were mounted 25 and 50 cm up the web of the centre area of the third girder (Figure 5). Another 34 sensors were mounted on the top surface of the bottom flanges of girders 3 and 6 to measure their distributed strain responses under ambient excitations. All LG-FBG sensors were manufactured in the laboratory, including optic fibre pre-stressing, packaging, calibrating and heat sealing. Field test includes precisely determining the gauge location, removing the paint and corrosion, gluing the gauge in place with epoxy and running the cable along the girder to the designated data acquisition system. The epoxy became firm within a few minutes, making these gauges be effectively a permanent installation. The bond is typically very strong and the gauge can be used confidently over a long time, which facilitates long-term monitoring or future testing. The sm130 Optical Sensing Interrogator was used during the data acquisition, and the sampling rate for dynamic data collection was set to be 500 Hz.

FBG

Plastic tube

Gage length

Fixing end

Connector

(b) Sensing part

Connector

(c)

Connector

Figure 3.

Sensing part

Gage length (1-2m)

Packaged LG-FBG sensor: (a) schematic picture; (b) actual sensor and (c) sensor arrays.

Connector

4

J. Zhang et al. 9.91m 1

9.91m 2

3

4 Sensors

10 Sensors 9 Sensors

4 Sensors

9 Sensors

9.91m 4

4 Sensors

5

Girder 3

17.08m

1.5m

F6-F10 F5-F1 F15–F23

F24–F27

4 Sensors

Girder 6

LG-FBG Strain-gauge

F11–F14

37.64m Girder 3


1.5m F24-F27

F15-F23

F11-F14

33.15m Girder 6

Figure 4.

Instrument plan.

Macro-strain time histories During the ambient vibration test of the benchmark bridge by the authors, 30 and 60 datasets were recorded by the LG-FBG area sensors on 8 and 9 June 2011, respectively. Each dataset has the length around 10 min. To verify the accuracy of the developed sensors, the recorded strain time histories are compared with those from traditional strain gauges installed by Drexel University. The strain gauge used by Drexel University is the Hi-Tec weldable quarterbridge strain gauge, which has a 2 in. shim length and a 1 in. gauge length (Weidner et al., 2011). Figure 5 presents

Strain gauge Long gauge FBG

152.4

50.8

50.0 25.0 16

Figure 5.

12 50

Typical sensor layout on a section.

the typical longitudinal configuration of some Hi-Tec strain gauges that coincide with the LG-FBG sensors on Girders 3 and 6. These two kinds of sensors were set to collect data simultaneously during the ambient vibration test in order to compare with each other. The macro-strain time history at the line 2 of the girder 3 recorded by the LG-FBG sensor 24 (Figure 4) is plotted in Figure 6(a), in which the strain time history from the traditional gauge at the same time and location is also plotted for comparison. It is seen that the peaks of two datasets have very close magnitudes. A time window from 980 to 1030 s is plotted in Figure 6(a) to illustrate the details. It is clearly seen that the free-vibration responses excited by the traffic are successfully recorded by both kinds of sensors and they are almost the same. The strain time histories plotted in the figure were preprocessed with a wavelet filter for observation noise reduction. Similarly, Figure 6(b) shows the dynamic strains from the LG-FBG sensor and the Hi-Tec strain gauge mounted at the line 2 of the girder 6, which further illustrate that the developed sensors have the same capacity for dynamic strain measurement as the widely applied traditional strain gauges. When the structure measured has good conditions, for instance the benchmark bridge studied here, the LG-FBG sensor outputs the similar results at normal sensors. It is expected that the LG-FBG sensors will output different data with that from the traditional strain gauges when structural damages exist because the LG-FBG sensor is able to detect cracks within the long-gauge length (1 m in this study), while the traditional gauge can only measure the local strain (within 1 in length).

Micro Strain

(a) 75 60 45 30 15 0 –15 –30

Micro Strain

Structure and Infrastructure Engineering

500 1000 Time (Sec)

1500

75

60

60 Micro Strain

(b) 75 Micro Strain


0

45 30 15

LG-FBG (SEU.) Strain gauge (Drexel U.)

990

1000 1010 Time (Sec)

1020

–15 0

500 1000 Time (Sec)

1500

45 30 15 –15 1260

1265 1270 Time (Sec)

1275

Strain time history comparison: (a) line 2 of girder 3 and (b) line 2 of girder 6.

Macro-strain modal analyses Other than directly measuring local strains, the developed area sensor also has the merit to output structural global properties. Structural modal parameters, such as frequencies, reflect structural global features, which are generally identified from recorded accelerations. Traditional strain responses are local measurements which are sensitive to local structural conditions like corrosion. In contrast, the LG-FBG area sensor measures the averaged strain of the structural element within the long-gauge length, thus it is potential to output structural global features through modal analyses. As shown in Figure 7, a force is input at a node, and structural responses are recorded by both accelerometers and LG-FBG area sensors. Similar to the derivation of the acceleration frequency response function (FRF), the FRF of dynamic macro-strain is derived to be (Li & Wu, 2007),

1 ðvÞ m 1 r H mp ðvÞ ¼ Pp ðvÞ ¼

hm =Lm wpr ðwir 2 wjr Þ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi; 2 Mr v2r 2 v 2 þð2jr vr vÞ2

Accelerometers

εm

lth

f jth

ith

Lm

pth

N-DOF beam structural model Strain and acceleration FRF.

ð1Þ

where v is the frequency line, 1m is the macro-strain of the element m, Pp is the force at the node p, r H 1mp denotes macro-strain FRF at the mode r, hm is the beam cross section height, Lm is the element length, M r is the modal mass, vr is the rth structural frequency, jr is the the rth damping ratio and w is the mode shape. It is seen that the macro-strain FRF has a close relationship with the well-known acceleration FRF. Therefore, structural frequencies and damping ratios can also be identified from the poles of the macro-strain FRF. Moreover, the macro-strain FRF more closely resembles a displacement FRF than a velocity or acceleration one, which leads to

FBG sensors

Figure 7.

1030

0

0

Figure 6.

75 60 45 30 15 0 –15 –30 980

5

J. Zhang et al. 0.15 FBG 40 FBG 24 FFT magnitude

FBG 31 0.10

0.05

0.00

0

5

10 Frequency (Hz)

15

20

10 Frequency (Hz)

15

20

10 Frequency (Hz)

15

20

× 10–4 1.8 A9 A19 FFT Magnitude

that the macro-strain FRF provides a more sensitive indicator of structural frequencies at low modes, especially when the resonant frequency of the structure is small. This feature is valuable for the on-line monitoring of flexible structures, such as long-span bridges or high-rise buildings. Structural modal parameters of the benchmark bridge were estimated by model analyses of the dynamic macrostrain data recorded during the ambient vibration test. By performing Fast Fourier Transform (FFT), the macrostrain responses were transformed from the time domain to the frequency domain as shown in Figure 8(a). It is seen that spectral peaks in the first several modes are clear. Spectral analyses of the recorded accelerations during the ambient test were also performed as illustrated in Figure 8(b) for comparison. In addition, a multi-reference impact test was also performed by Drexel University using a newly developed drop hammer to excite the bridge. The Complex Mode Indicator Function (CMIF) plots (Catbas, Brown, & Aktan, 2004) as shown in Figure 8(c) further verify the correctness of the identified modal parameters from the macro-strain time histories. It is seen that all these three spectral figures in Figure 8 indicate consistent structural frequencies. By performing spectral analyses of the macro-strain time histories from all 44 LG-FBG sensors, structural frequencies in the first eight modes were identified and their averaged values and standard deviations are summarised in Table 1. The identified frequencies from the CMIF method are also provided in Table 1 for comparison. It is seen that the largest error between two methods is less than 5%.

A24

1.2

1.6

0

0

5

10–6

10–7

MMS-based neutral axis position determination This and next sections will present how to use the recorded dynamic macro-strains for structural diagnosis. Neutral axis position is an important factor for structural design and safety assessment. Static truck tests are generally used to determine the neutral axis position of a beam member (e.g. bridge girder) by installing at least two strain gauges at different height of a cross section. However, static truck tests cost much and require the bridge to be closed during the testing. Moreover, the truck location may affect the neutral axis position due to the nonlinear connection between the deck and girders which will be discussed later. Therefore, more convenient and reliable methods are needed in engineering practices to estimate the neutral axis position of key structural member for structural safety assessment. Methods have been proposed in the literature to utilise dynamic strain for neutral axis position determination (Jeffrey, Brena, & Civjan, 2009; MacLeod, 2010), in which statistical values like the maximum or mean values of the time histories are used to predict the strain profile of a given cross section. Unfortunately, due

Log(|SV|)


6

10–8

10–9

0

5

Figure 8. Strain and acceleration spectra analyses: (a) macrostrain, (b) acceleration and (c) CMIF plot of the impact test data.

to the estimated neutral axis positions from those methods are not stable. In this study, a more robust neutral axis position identification method is employed using a MMS concept. The instrument configuration consists of three gauge lines at the different heights of the centre area of the girder, three was designed for neural axis position determination. Each line includes five LG-FBG sensors. For instance, the sensors 19, 1 and 10 were installed on the topside of the

Structure and Infrastructure Engineering Table 1.

Identified frequency comparison. Macro-strain

Mode

Frequency (Hz)

Standard deviation

Acceleration Frequency (Hz)

Difference (%)

2.85 3.16 3.71 5.16 9.23 11.86 14.44 15.30

0.033 0.148 0.139 0.466 0.399 0.340 0.404 0.185

2.99 3.01 3.87 5.03 9.24 11.71 14.78 15.45

2 4.68 4.98 2 4.13 2.58 2 0.11 1.28 2 2.30 2 0.97

bottom flange, 25 cm up the web and 50 cm up the web of a section, respectively. In the static test, these three gauges provide strain data to plot the strain profile of that cross section. It is known that two gauges at different heights of a given cross section are enough to predict a linear strain profile and determine the neutral axis position. The purpose of mounting the third gauge in a section is to check the linearity of strain profile along the section and the compatibility between the deck and girders. Combination of concrete deck and steel girder makes a composite section. The neutral axis of bending locates in the upper half of the member, and small strain responses are expected to be in that region, thus there is no gauge installed in the upper half of the girder. As shown in Figure 8(a), the peak magnitudes of the spectral curves of the macro-strain time histories are defined as MMS. For instance, we define the peak value corresponding to the natural frequency in the spectral curve to be MMS1. From the dynamic macro-strains measured by three sensors at different heights of a given section, three MMS1 values were identified and they were used to plot the MMS profile, from which the neutral axis position was estimated. Figure 9(a) depicts the MMS profile of the section with sensors 19, 1 and 10, in which three points in each line denote three MMS1 values from a 10 min-long dataset recorded by those three sensors. It is (b) 200

150

150 Height (cm)

(a) 200

Height (cm)


1 2 3 4 5 6 7 8

100

seen that the neutral axis positions determined from all datasets are very consistent even the datasets are recorded at different time, which illustrates that the MMS-based neutral axis estimation method is robust. Similarly, Figure 9(b) presents the MMS profile of the section with sensors 8, 4, 18, which also illustrate very stable neutral axis position estimation results. The neutral axis position is also investigated by Drexel University through performing static truck tests (Figure 10). Two traditional strain gauges were mounted at the top of the bottom flange and 50.4 cm up the web in a typical strain gauge layout configuration (Weidner et al., 2011). Three static load cases by using six trucks with full loads, three trucks with full loads and three empty trucks were considered, and the loading locations are 1/4, 1/2 and 3/4 positions of the southbound span 2, respectively. The strain profile of a given section is plotted by employing the static strain measurements. For instance, Figure 11 shows the predicted strain profile of the cross section at the line 2 of the girder 3 from the static truck test. It is seen that when the static force location is fixed, the neutral axis position estimated from the strain profile is stable even when the static load is different. However, the estimated neutral axis position varies when the static truck load locations are different. This is probably because the property of the deck-girder connection has been changed with the varying truck

100 50

50 0 –0.8 –0.4 0 0.4 Modal Strain

Figure 9.

7

MMS profile.

0.8

0 –0.8 –0.4 0 0.4 Modal Strain

0.8

Figure 10.

Static truck tests performed by Drexel University.

8

J. Zhang et al. Load at 3/4

Load at 1/4

Load at 1/2

180

180

180

120

120

60

0 –50

0

50

100

60

150

Micro Strain


Figure 11.

0 –100

0

100 200

Micro Strain

60

0 –100

0

100 200

Micro Strain

Strain profile from static truck tests.

locations (Weidner et al., 2011). The estimated neutral axis position of the cross section at the centre of the gird 3 from both the static truck tests and dynamic macro-strain measurements is summarised in Table 2. It can be observed that the results from two methods are comparable, but it is obvious that the method using the dynamic macro-strain time histories has the merits that it does not require bridge closure during the test and it is much more convenient.

MMS for structural damage detection Other than the neutral axis position determination, the recorded dynamic macro-strain responses are also useful for structural damage detection. A basic idea is performing the regression analysis of the MMS extracted from dynamic strain measurements under different environmental and operational conditions (Adewuyi & Wu, 2011). For a simple beam under static loading, it is known that the stress or strain at any section solely depends on the magnitude of the moment. For a given loading configuration, the ratio of the strain between two measurement locations is independent on the load magnitude. When damages, such as cracks and corrosions, occur on a section, the ratio of strain between this section and the reference section will change. In other words, the change in the strain ratio between two sections indicates the stiffness reduction of the corresponding section. Therefore, the strain ratio between a reference Table 2.

Height (cm)

120

Height (cm)

Height (cm)

6 Full Trucks 3 Full Trucks 3 Empty Trucks

Neutral axis position estimation.

Ambient test

Average height (mm) 123.9

Static test with the load 3/4 span 1/2 span 1/4 span

120.1 122.0 126.2

Standard deviation 4.62 – – –

sensor and a target sensor has potential to serve as an index for structural damage detection. The strain ratio in the static case discussed above can be extended to the ambient vibration case by employing the MMS concept described in last section. By using the MMS of a macro-strain time history to replace the strain measurement in the static case, modal macro-strain ratio (MMSR) between a target section and a reference section of a certain condition should be also a constant, in which the used MMS can be for a particular mode or combination of several modes. The procedure to use the MMSR for damage detection of the steel stringer bridge is presented below. First, the dynamic macro-strains are collected from the LG-FBG sensors mounted on the reference and target locations in the longitudinal direction of the girder during the ambient vibration test. Second, FFT is performed to each macro-strain time history to get the magnitude of the spectral peak (i.e. MMS) corresponding to the structural natural frequency. Thirdly, a point is plotted by taking the MMS values of the reference and target sections as the x and y axes values, respectively. A dataset with a time period (e.g. 10 min) produces such a point, thus repeatedly processing a number of datasets which produces the same number of points. Ideally these points should lie on a line for an intact beam member. For the field test data with observation noises, a line is regressed from these points even when they are not exactly on the same line. For instance, Figure 12(a) illustrates the regression line by taking the sensors 6 and 10 as the reference and target sensors respectively. Thirty datasets recorded on 8 June 2011 were used to produce a regression line, and another 30 datasets recorded on 9 June 2011 were used to produce another regression, as illustrated by red and blue lines respectively in Figure 12(a). It can be seen that these two regression lines almost superpose, which illustrate that no damage occurs between these two tests. Similar investigations using the MMS values in the higher modes are also performed in this study. For instance, Figures 13 and 14

Structure and Infrastructure Engineering 0.8 0.7

9

0.8

0.3

The first test The first test The second test The second test

0.1 0

Figure 12.

0.2

0.4 Sensor 6

0.6

Sensor 13

Sensor 10

0.6 0.5

0.4 0.2 0

0.8

0

0.2

0.4 Sensor 6

0.6

0.8

MMSR in the first mode.

0.25

Sensor 13

Sensor 10

0.20

0.15 0.10

0.15 0.10

0.05

0.05

0.00 0.00

0.05

0.10 0.15 Sensor 6

0.20

0.05

0.10 0.15 Sensor 6

0.20

0.25

MMSR in the third mode.

0.04

0.04

0.03

0.03

0.02 0.01 0

Figure 14.

0.00 0.00

0.25

Sensor 13

Figure 13.

Sensor 10


0.25 0.20

0.02 0.01

0

0.01

0.02 Sensor 6

0.03

0.04

0

0

0.01

0.02 Sensor 6

0.03

0.04

MMSR in the fifth mode.

illustrate the regressions results using MMS values in the 3rd and the 5th modes. All the points are still located around the regression line even the varying operational and environmental condition caused some variation in the extracted features.

Conclusions At a part of the IBS programme, SHM of a steel stringer bridge has been performed by employing innovative LG-FBG sensors. The following conclusions are drawn: (1) The LG-FBG area sensor developed by the authors

has the merits to measure the averaged strain of a large area by designing a long-gauge length (e.g. 1 m in this study). Moreover, a sensor network has been realised by connecting the LG-FBG sensors, thus the developed sensors are able to carry out area sensing of the whole structure as least the most important areas at performed in this study. (2) The developed LG-FBG sensors and traditional strain gauges were employed simultaneously in the ambient vibration test of the steel string bridge. The comparison of the strain time histories from these two kinds of sensors illustrated that the developed sensor has a competitive resolution.


10

J. Zhang et al.

(3) Other than local strain measurement, the LG-FBG area sensor has the merit to output global structural properties. Spectral analyses of the recorded macrostrain time histories clearly estimated structural frequencies in the first eight modes in this study, and they are comparable with those from the spectral analyses of recorded accelerations. (4) The recorded macro-strain time histories were also employed to determinate the neutral axis position through a MMS concept. Comparison between the estimated results from the developed method and those from the static truck test illustrated that the neutral axis positions estimated from the LG-FBG sensors are stable, while the results from the static test are affected by the position of the trucks. (5) The MMS-based strain ratio was successfully verified to be an excellent damage detection index. The MMS ratios calculated from the macro-strains of a reference section and a target section keep to be a constant if the structure is intact, and it would shift once the damage occurs. Acknowledgements This work was sponsored by the National Science Foundation of China (51108076) and the Supporting Programme of the Twelve Five-year Plan for Science & Technology Research of China (2011BAK02B03). The second author would like to acknowledge the support from the Scientific Research Foundation of Graduate School of Southeast University and the Priority Academic Programme Development of Jiangsu Higher Education Institutions. We would particularly like to acknowledge the support from Mr Jeff Weidner during the field test and useful conversations with Drs F. Moon and A.E. Aktan from Drexel University.

Notes 1. 2. 3. 4. 5.

Email: [email protected] Email: [email protected] Email: [email protected] Email: [email protected] Email: [email protected]

References Abdel-Ghaffar, A.M., & Scanlan, R.H. (1985). Ambient vibration studies of golden gate bridge: I suspended structure. Journal of Engineering Mechanics, 111, 463– 482. Adewuyi, A.P., & Wu, Z.S. (2011). Modal macro-strain flexibility methods for damage localization in flexural structures using long-gauge FBG sensors. Structural Control and Health Monitoring, 18, 341– 360. Ansari, F. (2012). Practical implementation of optical fiber sensors in civil structural health monitoring. Journal of Intelligent Material Systems and Structures, 18(8), 879–889. ASCE (2012). Structural identification of constructed facilities: Approaches, methods and technologies for effective practice

of St-Id, A state-of-the-art report. ASCE SEI Committee on Structural Identification of Constructed Systems. Brownjohn, J.M.W., Magalha˜es, F., Caetano, E., Cunha, A., Au, I., & Lam, P. (2009). Dynamic testing of the Humber Suspension Bridge (pp. 93 – 102). International Conference of Experimental Vibration Analysis for Civil Engineering Structures, October, Wroclaw, Poland. Catbas, F.N., Brown, D.L., & Aktan, A.E. (2004). Parameter estimation for multiple-input multiple-output modal analysis of large structures. Journal of Engineering Mechanics, 130, 921– 930. Conte, J.P., He, X., Moaveni, B., Masri, S.F., Caffrey, J.P., Wahbeh, M., Tasbihgoo, F., Whang, D.H., & Elgamal, A. (2008). Dynamic testing of Alfred Zampa Memorial Bridge. Journal of Structural Engineering, 134, 1006– 1015. Jeffrey, A., Brena, S.F., & Civjan, S.A. (2009). Evaluation of bridge performance and rating through nondestructive load testing, Final Report, University of Massachusetts Amherst, Amherst-01003-MA. Ko, J.M., Sun, Z.G., & Ni, Y.Q. (2002). Multi-stage identification scheme for detecting damage in cable-stayed Kap Shui Mun Bridge. Engineering Structures, 24, 857– 868. Li, S.Z., & Wu, Z.S. (2007). Development of distributed longgauge fiber optic sensing system for structural health monitoring. Structural Health Monitoring, 6, 133– 143. MacLeod, A.B. (2010). Structural health monitoring of the traffic bridge in Saskatoon using strain gauges, Master of Science Thesis, University of Saskatchewan, Canada. Murayama, H., Kageyama, K., Ohara, K., Uzawa, K., & Kanai, M. (2008, June 15 –20). Novel measurement system with optic fiber sensor for strain distribution in welded tubular joints. Proceedings of the ASME 27th International Conference on Offshore Mechanics and Arctic Engineering, Estoril, Portugal. Pakzad, S.N., & Fenves, G.L. (2009). Statistical analysis of vibration modes of a suspension bridge using spatially dense wireless sensor network. Journal of Structural Engineering, 135, 863– 872. Peeters, B., & Roeck, G.D. (2001). One-year monitoring of the Z24-Bridge: Environmental effects versus damage events. Earthquake Engineering and Structural Dynamics, 30, 149– 171. Siringoringo, D.M., & Fujino, Y. (2008). System identification of suspension bridge from ambient vibration response. Engineering Structures, 30, 462– 477. Weidner, J., Prader, J., Moon, F., & Aktan, E. (2011). Static, crawl, ambient and forced vibration testing (Technical report). Philadelphia: Drexel University. Wu, Z.S., & Zhang, J. (2011). Advancement of long-gauge fiber optic sensors towards structural health monitoring, The 6th International Workshop on Advanced Smart Materials and Smart Structures Technology (ANCRiSST2011), 25 – 26 July, Dalian, China. Zhang, J., Prader, J., Moon, F., Grimmelsman, K.A., Aktan, E., & Sayama, A. (2013). Experimental vibration analysis for structural identification of a long span suspension bridge. ASCE Journal of Engineering Mechanics Published online 16 February 2012, DOI: 10.1061/(ASCE)EM.1943-7889. 0000416. Zhang, H., & Wu, Z.S. (2008). Performance evaluation of BOTDR-based distributed fiber optic sensors for crack monitoring. Structural Health Monitoring, 7, 143– 156.

Structure and Infrastructure Engineering

Structure and Infrastructure Engineering

Suggest Documents