Full text
1 INTRODUCTION 1.1 Context As it is the case in many European countries, there is in Portugal a large number of bridges with stone arches. Such bridges are traditionally old and were not designed for the loads and traffic that are presently crossing them. Moreover, several bridges of this type show clear signs of structural degradation and lack of maintenance that may lead into situations where safety is not assured. Therefore, the possibility of monitoring and analyzing the structural behavior of a real bridge with stone arches represents an important contribution for the knowledge of this type of structures. The work presented here refers to the development and implementation of a fiber Bragg grating based sensing network for structural monitoring of a new stone bridge over Vizela River in Vila Fria, Felgueiras, Portugal, which was build to replace an old bridge with serious safety fragilities (Figure 1). Figure 1. The Vila Fria (a) old and (b) new stone masonry bridge. Implementation of a fiber Bragg grating sensor network for structural monitoring of a new stone bridge L. A. Ferreira, F. M. Araújo INESC Porto, Campus da FEUP, R. Dr. Roberto Frias, 378, 4200-465 Porto, Portugal FiberSensing, R. Vasconcelos Costa, 277, 4470-640, Portugal C. Barbosa, N. Costa FiberSensing, R. Vasconcelos Costa, 277, 4470-640, Portugal A. Arêde, A. Costa, P. Costa Fac. Eng. Universidade do Porto, R. Dr. Roberto Frias, s/n, 4200-465 Porto, Portugal ABSTRACT: The implementation of a large sensing network based on fiber optic sensors for the structural health monitoring of a new five arch stone bridge with a total span of 60 m over Vizela River (Felgueiras, Portugal) is presented. In addition to the implemented instrumentation with conventional sensors, four different types of fiber Bragg grating based sensors were specially developed for measuring temperature, strain and displacement in several critical points of the structure. In total, 85 Bragg sensors were installed and more than 800 m of fiber cables were deployed along the bridge. The fiber optic sensing network installed on the bridge was entirely monitored by using a single measurement unit designed for collecting and archiving the signals from all the fiber Bragg grating sensors, and also to allow remote access to the important data through a standard GPRS connection.
1.2 Why fiber optics? In general, modeling structures like the Vila Fria bridge brings several issues on how to define parameters to calibrate the mathematical models of the structure and on how to be certain about the results predicted by those models. Therefore, the possibility of monitoring an arch stone bridge from the beginning of its construction is of particular interest because it allows the installation of a large number of sensors according to a pre-established plan and accessing points impossible to achieve in structures that are already built. With the large amount of information that can be generated, improved calibration and models validation can be obtained. The fact that the bridge can then be monitored for a long period also allows the evaluation of its structural behavior and its relation to the usually observed pathologies in this kind of structures. Nevertheless, the establishment of a large sensing network for monitoring multiple parameters that are structurally relevant is not always easy when using conventional electric technology. In fact, conventional sensors suffer from high EMI/RFI sensitivity, environment induced drift and require individual electrical stimulus that makes large scale structural health monitoring highly complex, particularly in hazard-environments common in Civil Engineering monitoring applications. Fiber optic sensors, and in particular, fiber Bragg gratings1, constitute a particularly appropriate and competitive alternative for structural health monitoring applications2,3. In fact, fiber Bragg grating sensors add to the long recognized advantages of fiber optic sensors (e.g., immunity to EMI/RFI, remote monitoring, small size and weight, electrical isolation, intrinsically safe operation, high sensitivity, long-term reliability) the inherent multiplexing capability and the ability to provide absolute measurements without the need for referencing4. Fiber Bragg grating technology is therefore becoming the natural substitute for the conventional sensing technologies by easing health monitoring of large structures during construction, load tests and long-term service5. 1.3 Fiber optic monitoring project The monitoring of the Vila Fria stone masonry bridge pursued two main objectives: the establishment and calibration of analytical modeling techniques suitable for the numerical simulation of similar constructions; the evaluation and systematization of the main features of this type of structures in terms of their structural behavior and their relation with the most frequent structural pathologies. To achieve these goals, a vast instrumentation project was put forward to transform the bridge into a “live laboratory”, which included, in addition to the installation of several conventional sensors described in reference 6, the deployment of a large fiber sensing network for the measurement of temperature (28 sensors), displacement (48 sensors) and strain (9 sensors). Some of the sensors used in the bridge were specifically developed by FiberSensing for this project, as it is the case of the LPDS – Linear Position and Displacement Sensor, and the Long Gauge. Most temperature sensors were used for referentiation of both fiber optic and conventional sensors. The sensors were deployed in a tree network configuration with 15 branches. To interrogate this sensing network, a BraggMeter measurement unit containing an optical switch with 16 channels was adopted. In total, more than 800 m of fiber optic cables were installed in the bridge. Figure 2 shows a simplified scheme of the fiber optic sensing network architecture. Technical Cabinet Measurement Unit BraggMeter Fibre Optic Switch 1x16 (Built-in) Remote Transmission Module Optical cables to sensing network branches 1 16 Typical sensing branch (up to 10 sensors) LPDS Temp. sensor Strain sensor Long Gage sensor Optical connector Figure 2. Implemented fiber optic sensing network.
2 THE FIBER OPTIC SENSING NETWORK COMPONENTS 2.1 LPDSs – Linear Position and Displacement Sensors To measure relative displacements between some stone blocks of the bridge, a Linear Position and Displacement Sensor based in a fiber Bragg grating was developed. The working principle of the sensor is quite common; it is based on the transduction of the displacement in force applied to the Bragg grating through the use of a division spring7,8. The fiber Bragg grating works therefore as a dynamometer. Nevertheless, the mechanical construction of this sensor is highly demanding, as can be seen from the CAD drawing in Figure 3. In fact, in order to obtain linear behavior over the desired measurement range and avoid hysteretic behavior, it is extremely important to eliminate any internal sources of friction. Also, to allow the sensor to be used in series and fixed to the structure as a conventional LVDT, it is necessary to design it in such a way that the input and output fibers enter and exit the device outside the longitudinal axis, which implies bending the fiber near the dynamometer moving parts. Adding to this, the need to make the sensor robust and hermetic places many constrains in the selection of materials as well as in their processing. Figure 3. CAD drawing of the LPDS. Figure 4 shows the final aspect of the developed sensor and Table 1 summarizes its characteristics. In Figure 5 it is shown a typical calibration cycle. It can be seen that the response of the sensor is linear and no hysteretic behavior is observable. Figure 4. Final aspect of the LPDS. 0 2 4 6 8 10121416182022 0 250 500 750 1000 1250 1500 1750 Δλ = 76.0 Δd + 8.4 R2=1.000 Wavelength shift - Δλ (pm) Displacement - Δd (mm) Figure 5. Typical calibration cycle for the LPDS.
Table 1. Main specifications of the LPDS. Optical central wavelength C band (1530 nm to 1570 nm) spectral width (FWHM) < 0.2 nm reflectivity > 75% insertion loss < 0.1 dB side lobe suppression > 10 dB sensitivity 76 pm/mm maximum measurement range ±12.5 mm Mechanical packaging stainless steel dimensions 12 mm x 18 mm x 205 mm (max.) distance between anchoring points 170 ± 5 mm (adjustable) weight 200 g Environmental operation temperature -20 ºC to 80 ºC relative humidity < 90% at 40 ºC (no condensation) Inputs / Outputs buffer Ø 900 μm length 1 m connectors FC/PC On the bridge, 44 LPDSs were used to measure the opening and closing of the joints between selected stone blocks of the face of the arches – in a standard Z configuration, Figure 6 – and also in the intrados of the arches – in the longitudinal and transverse directions. In each Z set, and also in each intrados set, a temperature sensor (described below) was used to compensate for temperature cross-effects, accordingly to Table 2. For aesthetic reasons, special cavities were formed in the stone blocks to hide the sensors, as can be seen in Figure 6. All the fiber optic connections were made from inside during the bridge construction or immersed in the mortar used to fill the joints. Figure 6. Installation of the LPDSs: (a) Z configuration; (b) detail of the stone housing. Table 2. Association between sets of LPDSs and temperature sensors. Arches face (Z) D01, D02, D03 T1 D04, D05, D06 T2 D07, D08, D09 T3 D10, D11, D12 T4 D13, D14, D15 T5 D16, D17, D18 T6 D19, D20, D21 T7 D22, D23, D24 T8 D25, D26, D27 T9 D31, D32, D33 T10 D40, D41, D42 T11 Intrados D46, D47, D48 T24 D49, D50, D51, D52 T26 D53, D54, D55, D56 T28
2.2 Temperature sensors Temperature sensors used in the bridge are shown in Figure 7. Table 3 summarizes their characteristics and in Figure 8, a calibration line is presented. As mentioned previously, they were mainly used to compensate for temperature cross effects, not only for the fiber optic sensors but also for other type of sensors used in the bridge. Table 4 lists the temperature sensors that were used in combination with conventional differential fluid pressure sensors for level measurements (level sensors). Figure 7. Temperature sensor. Table 3. Main specifications of the temperature sensor. Optical central wavelength C band (1530 nm to 1570 nm) spectral width (FWHM) < 0.2 nm reflectivity > 75% insertion loss < 0.1 dB side lobe suppression > 10 dB sensitivity 29.2 pm/ºC maximum measurement range -10 ºC to 80 ºC Mechanical packaging stainless steel dimensions Ø 6.3 mm x 56.5 mm weight 100 g Environmental operation temperature -20 ºC to 80 ºC relative humidity < 90% at 40 ºC (no condensation) Inputs / Outputs buffer Ø 900 μm length 1 m connectors FC/PC 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 -250 0 250 500 750 1000 1250 1500 1750 Δλ = 29.2 T - 808.7 R2=1.000 Wavelength shift - Δλ (pm) Temperature - T (ºC) Figure 8. Calibration curve for the temperature sensor. Table 4. Temperature sensors associated with non-fiber optic pressure sensors. Level sensors T12 T13 T14 T15 T16 T17 T18 T19 T20 T21
2.3 Strain sensors Figure 9 shows an example of the sensors installed in the bridge for strain measurement. These sensors were produced by embedding pre-tensioned fiber Bragg gratings in composite materials, which provides a device that is very robust and easy to apply in field. Table 5 summarizes the main characteristics of this sensor. Nine sensors were installed in the arches of the bridge, mainly for the control of tension in the extrados and intrados, as indicated in Table 6. Figure 10 shows sets of two sensors being installed along two orthogonal directions. Figure 9. Composite strain sensor. Table 5. Main specifications of the composite strain sensor. Optical central wavelength C band (1530 nm to 1570 nm) spectral width (FWHM) < 0.2 nm Reflectivity > 75% insertion loss < 1 dB side lobe suppression > 10 dB Sensitivity 1.2 pm/με maximum measurement range ± 2 mε Mechanical Packaging stainless steel Dimensions 20 mm x 100 mm x 0.9 mm Weight 25 g Environmental operation temperature -20 ºC to 80 ºC relative humidity < 90% at 40 ºC (no condensation) Inputs / Outputs Buffer Ø 3 mm; Ø 900 μm Length 1 m Connectors FC/PC Table 6. Composite strain sensors. Arches E02E_L, E11I_L, E12I_T E02I_L, E11I_T, E12E_L E04I_L, E12I_L, E12E_T D40, D41, D42 Figure 10. Installation of the strain composite sensors: (a) extrados; (b) intrados. 2.4 Long Gauge sensors The long gauge sensor was also specially developed for this project. The objective was to measure small relative displacements between the bridge timpani walls that are separated by ~4.7 m. Basically, the displacement is measured through the measurement of deformation in an Invar
wire with uniform cross-section (∅ 1.27 mm) that is fixed to each one of the timpani walls under an initial tension. To allow the installation of this sensor inside the soil, it was necessary to develop a suitable protection for the wire, as well as a mechanism for the application of tension in situ. The CAD drawing of the complete system shown in Figure 11 illustrates how those elements were build. The wire is protected by a set of three tubes; two of them, with l m length each, are threaded to the parts to be fixed to the stone blocks; a longer tube, 4 m long, is used outside these ones, and can be moved to allow later access to the wire fixing points and the tensioning mechanism, which is based on a secondary thread. Before the installation, a few tests were implemented in order to evaluate not only the sensor performance, but also the procedure of its installation. The results of one of those tests is shown in Figure 12, where it can be seen the agreement between values measured with the fiber Bragg grating based Long Gauge and those measured with a conventional LVDT. Figure 11. CAD drawing of the Long Gauge sensor. 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 0 1 2 3 4 5 6 7 8 9 10 11 12 LVDT FBG Displacement - Δd (mm) Time - t (min) Figure 12. Laboratory testing of the Long Gauge. Figure 13 shows the in situ installation of the Long Gauge sensors, where it can be seen the metallic parts used for fixing the wire to the stone blocks, the tensioning mechanism, the Invar wire, and the hermetic stainless steel tube that protects all the elements (Figure 13 (a)). Figure 13. Installation of the Long Gauge sensors: (a) detail of the tension mechanism; (b) anchoring and protection.
As an additional mean of protection, a secondary larger tube was used, thus ensuring that the entire system is immune to bending and vertical movements due to self weight and traffic loads (Figure 13 (b)). Table 7 lists the four Long Gauges installed in the bridge together with the associated compensation temperatures sensors. Table 7. Association between sets of Long Gauge sensors and temperature sensors. Timpani walls D61 T22 D62 T23 D63 T25 D64 T27 2.5 Measurement unit The sensing network installed in the Vila Fria bridge was interrogated by using the BraggMeter measurement unit shown in Figure 14, which has an internal built-in optical switch with 16 channels. The unit has a sampling rate of 1 S/s (for all the sensors in one channel) and the switching time is less than one second, which means that measurement data relative to all the sensors in the network can be collected every minute or less. Table 8 lists the most relevant characteristics of this measurement unit. A GPRS transmission module was connected to the unit in order to provide remote access to gathered data. Figure 14. BraggMeter measurement unit. Table 8. Main specifications of the BraggMeter used in the Vila Fria bridge. Wavelength Measurement operating range C band (1530 nm to 1570 nm) resolution 0.5 pm absolute accuracy ±2.0 pm repeatability ±1.0 pm sensors per fiber 12 (for meas. ranges of ~3 nm per sensor) acquisition rate 1 S/s Laser Source optical output power 10 dBm (max) linewidth 150 kHz optical isolation 60 dB Environmental operation temperature 10 ºC to 40 ºC relative humidity < 90% at 40 ºC (no condensation) Inputs / Outputs optical connectors FC/PC instrument interface touch screen, Ethernet, USB Datalogger Functions sampling, archiving, transmission, Excel™ compatible files Mechanical dimensions 450 mm x 180 mm x 370 mm weight 15 kg Power Requirements voltage 100-240 V power 300 W frequency 50-60 Hz
Optical Switch channels 16 insertion loss 0.5 dB crosstalk - 80 dB back reflection - 62 dB polarization dependent loss 0.02 dB repeatability 0.02 dB switching time 25 ms internal optical fiber singlemode, 8/125 μm 3 DISTRIBUTION OF THE SENSING NETWORK ALONG THE BRIDGE 3.1 Architecture of the network The instrumentation project is of great importance for maximizing the monitoring efficiency. The location of all the equipments and sensors must be carefully studied in order to obtain the maximum information with optimized resources. Thus the architecture must be defined taking into account singular spots that are particularly important in terms of structural behavior. Figure 15 shows a plan of the fiber optical monitoring project of the Vila Fria bridge. Plan (extrados) Plan (intrados) D1-3;T1 D4-6;T2 D7-9;T3 D10-13;T4 D14-15;T5 D16-17;T6 D19-21;T7 D22-24;T8 D25-27;T9 D31-33;T10 D40-42;T11 Downstream View D48-47 D46;T24 D51-50 D49;T26 D52 D53 D55-56 D54:T28 D61;T22 D62;T23 D63;T24 D64;T25 T13 T12 T15 T14 T17 T16 T19 T18 T20 T21 E12E E11I E2I E12I E2E Legend: Strain sensors Temperature sensors Long gauge sensors Displacement sensors in 'Z' configuration Displacement sensors Figure 15. Monitoring plan of the Vila Fria bridge (fiber optic sensors). Displacement sensors were installed between stone blocks where displacements are expected to be maximal. The LPDSs applied along the arches in a Z configuration are meant to register relative movements both in longitudinal and transverse directions. Realizing how stone blocks move and rotate makes it better to understand the global behavior of such structures. Displacement sensors were also placed underneath the arch to monitor block interfaces. They are both in longitudinal and transverse positions along the bridge axis and transversely on the joints between the first row of blocks and the others approximately bellow the border between timpani walls and filling material.