Home Search Collections Journals About Contact us My IOPscience

Multiplexed Bragg grating optical fiber sensors for damage evaluation in highway bridges

This article has been downloaded from IOPscience. Please scroll down to see the full text article.

1998 Smart Mater. Struct. 7 209

(http://iopscience.iop.org/0964-1726/7/2/008)

View the table of contents for this issue, or go to the journal homepage for more

Download details:
IP Address: 218.248.13.196
The article was downloaded on 30/07/2010 at 11:56

Please note that terms and conditions apply.

Smart Mater. Struct. 7 (1998) 209–216. Printed in the UK PII: S0964-1726(98)90264-6

Multiplexed Bragg grating optical

fiber sensors for damage evaluation
in highway bridges
R L Idriss†, M B Kodindouma†, A D Kersey‡ and M A Davis‡
† New Mexico State University, Las Cruces, NM 88003-8001, USA
‡ Optical Sciences Division, Code 5673, Naval Research Laboratory, Washington,
DC 20375, USA

Received 16 October 1996, accepted for publication 27 June 1997

Abstract. A multiplexed Bragg grating optical fiber monitoring system is designed

and integrated at the construction stage in an experimental full scale laboratory
bridge. The test bridge is a 40 ft span non-composite steel girder concrete deck
bridge. The network of sensors is used to measure the strain throughout the
bridge, with sensors bonded to the tension steel in the slab and attached to the
bottom flange of the girders. Resistive strain gages and Bragg grating sensors are
placed side by side to compare results. The strain data are obtained for the pristine
structure, then damage is introduced at midspan for an exterior girder. Several
levels of damage in the form of cuts in one of the girders are imposed with the final
cut resulting in a half depth fracture of the girder. The load path in the structure is
obtained using the built in sensor system.

1. Introduction (v) remote monitor critical structures, and provide a

A challenging task that faces the bridge engineer when
inspecting a deficient structure is damage assessment, i.e.,
quantifying the effect of damage on the structural integrity,
the overall performance and the load carrying capacity
of the bridge. From the bridge engineer’s standpoint,
the major concern with respect to bridge deterioration is
the form of deterioration that impairs the structural load
carrying capacity of the bridge. Cost-effective bridge
rehabilitation makes it necessary to identify the presence
of damage accurately and very early.
The objective of this multidisciplinary research project
is the development and implementation of a bridge
monitoring and evaluation system. Strain measurements at
critical locations in the structure, obtained from a bridge
monitoring system, can give a quantitative, simple and
straightforward means of evaluating an existing bridge, its
performance under load, and the effect of any damage on
its capacity and structural integrity. This could result in
better decision making, more efficient bridge management
and tremendous savings. This sensor system can provide
an objective tool to:
(i) assess the loading history
(ii) evaluate the effects of damages on the capacity and
performance of a bridge
(iii) assess the effectiveness of repairs and maintenance
programs
(iv) check the performance compared with the design
assumptions
warning when abnormal conditions occur.
2. Research approach
The multidisciplinary research study is an on-going testing
program, implemented in several phases; (a) large scale
component testing, (b) full scale bridge testing in the
laboratory and (c) full scale implementation in the field on
existing bridge structures. Phase (b) is the current phase of
the project, and will be covered in this paper.
In the current phase of the project, a multiplexed Bragg
grating optical fiber monitoring system is designed and
integrated at the construction stage in an experimental full
scale laboratory bridge. The network of sensors is used
to measure the strain throughout the bridge, with sensors
bonded to the tension steel in the slab, and attached to
the bottom flange of the girders. The strain data are
first obtained for the pristine structure, then damage is
introduced at midspan of one of the exterior girders. The
damage is in the form of torch cuts in the steel girder,
introduced in several stages with the final cut resulting in a
half depth fracture of the girder. The real time after fracture
response of the structure to damage is obtained using the
built in sensor system.
3. Bridge description
The structure is a full scale 40 ft (12.2 m) simple span, non-
composite steel girder bridge with a concrete deck (figures 1

0964-1726/98/020209+08$19.50

c 1998 IOP Publishing Ltd 209
R L Idriss et al
Figure 1. Elevation view and longitudinal sensors.
Figure 2. Midspan cross section.
and 2). The girders are W21 × 62 steel beams. The
deck is 11 ft (2.4 m) wide with a 6 in (15.2 cm) thick
reinforced concrete slab. The girders are spaced 4.50 ft
(1.4 m) center to center, and are welded to standard C12 ×
30 diaphragms over the abutments. The abutments consist
of two 600 × 2200 × 3200 (15.2 × 55.9 × 8.13 cm3 ) standard
highway New Jersey bounces. The girders bear directly on
two C8 × 11.5 channels over the supports.
4. Sensor layout
A dual system was used to monitor the bridge. Optical
fiber sensors and conventional foil gages were placed side
by side to compare results. Forty-eight optical fiber Bragg
grating sensors and 48 resistive gages were used to monitor
the bridge. The location of the sensors was selected based
on results from previous research [1, 2] and finite element
simulations of the intact and damaged bridge [3]. When
damage occurs in one of the girders in a multigirder steel
bridge, the structure responds by redistributing the load
along alternate paths.
The damaged girder carries less load. The load is
redistributed to the other girders via the secondary members
210
and the slab. The load redistribution to the intact girders
through the slab occurs in the transverse direction, and is
most pronounced at the vicinity of the damage. The closer
the girder to the damage, the more the load redistribution
to it [1].
To capture the after-fraction response of the bridge,
sensors were placed along the bottom flange of the girders
and embedded in the slab (figures 1, 2 and 3).
The sensors in the slab were placed in a grid pattern
in the longitudinal and transverse direction, bonded to the
tension steel. The longitudinal sensors were located over
the girders. This layout allowed a comparison of the
strain pattern in the slab and in the girder bottom flanges.
Although non-composite, due to the embedment of the top
flange in the slab, composite action between the deck and
the girders was expected in the elastic range.
5. Instrumentation
A dual system was used to monitor the bridge. The optical
fiber sensor system and a conventional data acquisition
system along with resistive gages were run simultaneously
to compare results.

Figure 3. Sensor layout. Plan views.
5.1. Fiber optic sensing system
Probably one of the most attractive features of optical fiber
sensors is their inherent ability to serve as both the sensing
element and the signal transmission medium, therefore
greatly simplifying the instrumentation of large structures.
Another advantage is the immunity of the optical fibers to
electric and electromagnetic interference.
The fiber Bragg grating (FBG) sensors used in the
research are intrinsic sensors, i.e., sensing occurs along the
fiber itself, while the light remains guided through the fiber.
The fiber optic intracore Bragg grating is a segment of an
optical fiber that has been internally modified by exposure
to UV light so that it reflects light at one wavelength.
This manufacturing process is fast and lends itself to mass
production, ultimately resulting in much lower cost for the
sensor.
The Bragg grating sensor is used for point sensing. The
sensor’s wavelength depends on the strain and temperature
imposed on the optical fiber at the location of the Bragg
grating. Therefore, as the strain changes the wavelength at
Damage evaluation in bridges
which the light is reflected shifts. One major advantage
of the FBG sensors is the ease with which several can
be multiplexed along a single optical fiber. A number
of discrete sensors designed to operate as point sensors
can be arranged in a network or array configuration, with
the individual sensor outputs. The sensor elements can be
tailored in terms of response and range to meet the specific
application [4].
The instrumentation system has the capability to
monitor 12 FBG sensors along each of five separate fibers
for a total of 60 sensor elements. The strain resolution of
this system was determined to be 0.95 microstrain. The
sensors and instrumentation used in this bridge research
were developed at the Naval Research Laboratory.
5.2. Conventional foil gage sensor system
The Megadac, a conventional, fully programmable, 56
channel data acquisition system developed by OPTIM
Electronics Corporation was used to acquire the data from
the resistive strain gages.
211
R L Idriss et al
Figure 4. Girder dead load strain during curing.
6. Bridge testing
6.1. Cure monitoring
The sensors were installed during the construction phase.
This allowed the structure to be continuously monitored
during the pouring and curing of the concrete. The strain
measured in the girders during this monitoring period is
plotted in figure 4, along with the temperature strain due
212
Figure 5. Girder behavior at cut 1.
to ambient temperature fluctuations. The girders were
unshored. At the end of curing the strain at midspan
measured in the bottom flange of the girders under dead
load was 350 microstrain for the middle girder, and 225
microstrain for the external girders.
6.2. Damage description
A half depth fracture was introduced near midspan in the
East girder, as shown in figure 3, at 210 (6.42 m) from the
north end. This fracture was inflicted using a series of five
sequential torch cuts in the web and flange of the girder.
The first cut (cut 1) initiated at mid-height of the girder, in
the web, and extended 500 (12.7 cm) down. In the second
level of damage (cut 2), the cut was extended to the top of
the bottom flange. In the third and fourth cuts (cut 3 and
cut 4) a quarter of the bottom flange was cut on each side.
In the last stage the flange was severed completely leaving
a half depth fracture at midspan of the girder (cut 5).
6.3. After-fracture response
The system was zeroed prior to each cut. Dead load strain
was recorded with both the optical fiber system and the

Figure 6. Girder behavior at cut 2.
Megadac as the cut was taking place. The dual systems
were left running, until the strain in the structure had
stabilized. Readings from both systems were within 3 to 5
microstrain.
6.3.1. Girder response.
6.3.1.1. Cut 1. Cut 1 initiated in the web at mid-height
of the girder, and extended 500 (12.7 cm) down. There
was a slight but definite change in the structure’s response
(figure 5). The damaged east girder shows a localized
increase in strain at gage C10 near the cut. The middle and
the west girder show a slight increase in strain of about 5
Damage evaluation in bridges
Figure 7. Girder behavior at cut 3.
microstrain.
6.3.1.2. Cut 2. At cut 2 the same pattern is observed as in
cut 1, but the change is more pronounced (figure 6). We
can see the large localized strain at C10 in the damaged
girder, and a larger increase in load in the middle girder.
6.3.1.3. Cut 3 and cut 4. The same pattern of load
213
R L Idriss et al
Figure 8. Girder behavior at cut 4.
redistribution is observed for cuts 3 and 4. Localized strain
increases at C10 in the damaged girder, and there is an
additional global redistribution of load to the other girders,
214
Figure 9. Girder behavior at cut 5.
mainly to the middle girder (figures 7 and 8).
6.3.1.5. Cut 5. The most drastic response occurred when
the bottom flange was completely severed at the final cut
(figure 9). Load redistribution occurred throughout the
structure.
In the intact bridge, dead load strain at midspan, in the
bottom flange was at about 350 microstrain for the middle
girder, and 225 microstrain for an exterior girder. When
fracture occurred, under dead load, the damaged girder
carried less load, with about 40% drop in strain at midspan.
The load was shed to the middle girder which picked up a
30% increase in strain at midspan. The west girder showed
little change with only a 5% increase in strain.
6.3.2. Slab response. A consistent pattern was also
observed in the slab. The transverse sensors at section C at

Figure 10. Slab response at cut 1.
the vicinity of the crack show the most pronounced change
in response, at every cut level.
At cut 1, the change in strain is very slight and most
pronounced at section C, where an increase in strain is noted
(figure 10).
At cut 5, the change in response is significant, as can be
seen in figure 11. The largest change occurs at the vicinity
of the crack (section C), showing a negative strain due to
the slab cantilevering from the west girder to the damaged
girder.
7. Conclusions
The test results showed the system to be a powerful bridge
diagnostic tool. When damage was introduced in the
bridge, the monitoring system recorded a definite change in
the structure’s response. It indicated damage had occurred,
along with the time and the location of the damage.
The time of occurrence corresponds to the time the
change in condition was recorded. The crack location could
be determined from the actual response of the structure.
When observing the change in the response of the bridge
Damage evaluation in bridges
Figure 11. Slab response at cut 5.
the slab indicated damage in the bridge at the vicinity of
cross section C. The girder response indicated damage in
the east girder where the loss in load occurred. The change
in response was most accentuated at the vicinity of the
crack. The larger the damage, the greater the response
from the system.
Acknowledgment
This project is a collaborative research project between
New Mexico State University and the Naval Research
Laboratory. It is supported by the National Science
Foundation grant no CMS-9457604, the New Mexico State
Highway and Transportation Department and the Federal
Highway Administration.
215
R L Idriss et al
References
[1] Idriss R L and White K R 1991 Secondary load paths in
bridge systems Transportation Research Record No 1290,
vol 1, 194–201
[2] Idriss R L, White K R, Woodward C B and Jauregui D V
1995 After fracture redundancy of a two girder steel
bridge Transportation Research Board, Fourth Int. Bridge
Conf. vol 2, pp 316–26
216
[3] Kodindouma M B 1996 Investigation of an optical fiber
monitoring system for highway bridges PhD Dissertation
New Mexico State University, Las Cruces
[4] Davis M A, Berkoff T A and Kersey A D 1995 Serially
configured matched filter interrogation technique for
Bragg grating arrays Proc. Smart Sensing, Processing and
Instrumentation Conf., SPIE 2444