1.

INTRODUCTION

Fiber Bragg Grating (FBG) sensors are increasingly being used for real-time structural health monitoring. The sensors are created inside optical fibres; usually standard telecommunication fibres (single mode SM 9/125 standard) are used, which allows reducing costs for wiring and operation, because standard tool for operating with the optical fibres can be used. The market for FBG sensor is increasing with the time, and FBG technology can be used to replace other sensing technology in an increasing number of applications. The advantages are many: the very low signal loss of the optical fibres allow positioning the monitoring system far away from the monitored structure; a number of sensors can be placed on a single fibre, reducing the number of cables needed for building a network of sensors; FBG sensors are immune to electromagnetic interferences and almost insensitive to corrosion, so that can be used in harsh environments and near power lines of electronic devices; there is no risk of electrical discharge from an optical fibre, so that FBG sensors are safe to use in places where fire hazard is a concern. Because of the small dimensions of optical fibres, FBG sensors can even be embedded inside the materials of the structures to be monitored, either polymeric composites or metallic alloys.^1–3 The insensibility to electromagnetic interferences made possible using FBG sensors for monitoring in extreme environments, such as in particle physics experiments.^4–6 Because of the resistance to corrosion, FBG sensors have been proposed and used for monitoring boats and structure in marine environment.^{7, 8} FBG sensors exhibit a good long term stability and zero drift, that makes this technology a preferred choice for monitoring civil engineering structures^9–13 and cultural heritage structures.^14–16 The FBG sensor written on the fibre is sensitive to strain and temperature, so that can be directly used as a strain gage or a temperature sensor. By applying interfaces designed on purpose, FBG technology can be used to monitor also accelerations, pressure, displacements and vibrations. In particular, metal coatings can be applied to improve the sensitivity toward temperature.^{17, 18}

Pipelines are a type of infrastructure that is particularly suited for exploiting the advantages of FBG technology, since pipelines can be placed in corrosive environments, needs very long arrays of sensors, shall be continuously monitored for long time, and the monitoring system shall not pose an additional hazard in case pipelines are used for flammable liquid and gases. Pipelines are vulnerable to operational and environmental damages (corrosion, fatigue), geohazard risks (earthquakes, soil movements, landslides, floods), and man-made damages such as vandalisms and illegal spilling. FBG sensors have been already proposed and tested for monitoring strain, pressure, leakage and movements of pipelines.^19–25

A system for detecting pipeline damages and in particular vandalism and oil stealing, based on the monitoring of vibrations, has been tested in laboratory and in an outdoor experiment on a section of pipeline.^{26, 27} The vibration sensor proposed consists of a FBG sensors glued on a cantilevered interface specifically designed for the amplification of the vibration. In the previous works, the device was able to detect vibrations induced by drilling a pipe at different distances from the monitored point. In this paper the device is tested on a pipe containing pressurized oil, and is compared with other FBG sensors for assessing the improvement in vibration monitoring.

2.

EXPERIMENTAL SETUP

The test has been performed at the SITEC (Sustainable Innovation Technology) laboratory, in the Casaccia Research Centre of ENEA (the Italian National Agency for New Technologies, Energy and Sustainable Economic Development). The lab is equipped with two high performance seismic tables for three axial seismic tests of structures. The larger, 6 degrees-of-freedom, 4 m x 4 m table, can test structures up to 10 ton mass and 3g acceleration applied at the Center of Gravity at 1 m from the base table. The FBG sensors have been applied on a pipe that feeds the pressurized fluid to the hydraulic actuators that move the table. The pipe is mechanically isolated from the actuators and from the seismic table. Three sensors have been applied to the pipe (Fig. 1): the vibration sensor, a copper coated FBG sensor and a bare FBG sensor. The copper coated and the bare FBG sensors have been glued to the pipe, while the vibration sensor is magnetic and is simply positioned on the steel pipe and can be easily removed. The copper coating on the second sensor was applied by the DICMA department of Sapienza University of Rome.

Figure 1.

Positions of the FBG sensors on the monitored pipe. (A) is the position of the vibration sensor; (B) is the copper coated FBG sensor; (C) is the bare FBG sensor.

The sensors were fixed with the axis parallel to the pipe axis because the main target of the test was a comparison in the sensitivity to the vibration and not a measure of the deformation of the pipe under pressure. The FBG sensors were monitored using a Micron Optic sm130 interrogation system operated at 1 kHz scan frequency. The pipe has a diameter of 16 cm and feeds high pressure hydraulic fluid to the actuators of the table. The pressure inside the pipe is 50 atm when the table is not operated, and reaches 200 atm during seismic tests. The temperature of the fluid starts from 30-35°C and reaches a maximum of 40°C.

The sensors were monitored during a test of a structure. A typical seismic test is performed as follows:

The resonance frequencies recorded after the first and the second sine-sweep tests are then compared to find damages from the simulated earthquake. The reference frame is such as that axis X and Y are in the horizontal plane, the axis Z is the vertical direction, perpendicular to X and Y. However, the sensors are not placed on the table but on the feeding pipe, so the vibration recorded by the FBG sensors are not the frequencies of the table but are due mainly to the vibration from the servohydraulic sistems and the servovalves.

3.

PRELIMINARY DATA ANALYSIS

In Fig.2 are reported the data read during the sine-sweep along the X-axis. The time domain plot shows that there is a wavelength shift on all the three sensors during the test. The shift is of 50 pm on the bare FBG and 70 pm on the vibration (“cantilever”) sensor. In fact, FBG sensor have a thermal sensitivity of about 0.1°/pm. Such a shift is compatible with a thermal wavelength shift from temperature increase of 5-7°C, that is in the operational range expected from the system. Indeed, all the three sensors are recording not only the thermal effects on the Bragg grating but also the thermal expansion of the material on which are glued, that can amplify or lower the shift due to temperature. The two sensors directly glued on the pipe recorded a lower shift than the vibration sensor: probably the interface of the vibration sensor amplify the thermal effect because of the material used for its manufacturing. Such a shift was observed an all the sine-sweep tests.

Figure 2.

Data recorded during a sine-sweep test along the X axis of the table: time domain.

Fig. 3 shows the frequency domain response of the FBG sensors during the same sine sweep test. The spectrum of the vibration sensor shows a strong response at 71.7 Hz, which was not recorded by the copper coated sensor. The bare FBG sensor recorded both the 71.7 Hz frequency and another peak at 173.3 Hz. Since the plots of the data in the time domain of the sine-sweep tests along Y and Z axis are very similar to the one of the test on X axis, in Fig. 4 and Fig. 5 are reported only the frequency domain responses.

Figure 3.

Data recorded during a sine-sweep test along the X axis of the table: frequency domain.

Figure 4.

Data recorded during a sine-sweep test along the X axis of the table: frequency domain.

Figure 5.

Data recorded during a sine-sweep test along the Z axis of the table: frequency domain.

In the Y-axis sine-sweep tests, both the coated FBG sensor and the bare sensor show two frequencies at 33.7 Hz and 71.7 Hz, while the vibration sensor shows only the peak at 71.7 Hz, but with a stronger amplitude.

The data from the simulated earthquake are reported in Fig. 6 (time domain) and Fig. 7 (frequency domain). In this last case, the frequency response of the bare FGB sensor shows two peaks at 29.0 Hz and 33.7 Hz, plus the peaks at 71.7 Hz and 173.3 Hz. The coated sensors shows only the peak at 29.0 Hz, and the vibration sensor only the peak at 71.7 Hz, with a larger amplitude than the bare sensor.

Figure 6.

Data recorded during a simulated earthquake test: time domain.

Figure 7.

Data recorded during a simulated earthquake test: frequency domain.