Exploration of Using FBG Sensor for Axle Counter in Railway Engineering

K.Y. LEE, K.K. LEE, S.L. HO Department of Electrical Engineering, The Hong Kong Polytechnic University Hung Hom, Hong Kong SAR, CHINA

Abstract: - Fibre Bragg Grating (FBG) sensors are finding increasing applications in smart structures. They can be embedded in civil structures such as bridges, buildings etc for the detection and monitoring of parameters such as strain, pressure and temperature. It allows assessments on the integrity of structures at various manufacturing and construction stages, and facilitates health management of many important structures in their subsequent operations. However, their application in railway engineering is rare. This paper explores the technical feasibility and potential usage of FBG sensor for axle counter application in railway engineering. In essence, an axle counter system is used for counting the number of train axles coming in and out of a section of rail track. Field tests have been carried out and concluded that FBG is having great potential for replacing conventional railway axle counter system, which is based on the principles of magnetic field disturbance technology as the wheels are passing by. FBG axle counter have lots of advantages over their conventional counterparts due to its electrical immunity to noise, multiplexing capability, compactness and more importantly, its relatively low cost of construction. Provided there is a weather proof and robust mounting methodology in the next phase, the FBG axle counter will become a strong competitor to conventional systems. Once the physical integrity of FBG systems is proven to be satisfactorily, one can extend its application to build a smart railway, particularly as part of a new generation of signaling systems.

Key-Words : - Fibre Bragg Grating, Axle Counter, Train Detection & Identification

1 Introduction applications is on the railway axle counter in Railway engineers are understandably very tracks. Due to its unique construction nature, the conservative in the use of any new technology cost of FBG axle counter is relatively low when for train operations. Only proven technology compared with those of conventional with good track record will be used. Any change technologies that are used to provide the same to the conventional equipment and technical and operational information. There is methodology must be brought-in under a a good future in using FBG axle counter to technically and commercially prudent manner. operate a smart railway service.

Fibre Bragg Grating (FBG) technology has been widely adopted in smart structure engineering 2 Fibre Bragg Grating Technology with good records. Its application on railway The fabrication of refractive index grating is engineering is limited at this stage, although reported by Hill et al. in 1978 [2]. Subsequently, there are applications in, for example, the Meltz et al. devise a method to controll the monitoring of the strain/stress on the train bogie fabrication of grating using UV laser [3]. frame composite structure [1]. An in-depth Basically, when a germanium-doped silica core review on the FBG characteristic revealed that it fibre is exposed to ultraviolet (UV) radiation can bring massive benefits to railway operation (with wavelength around 240 nm), it produces a by giving additional train operational permanent change in the refractive index of the information to rolling stock engineers and germanium-doped region, due to the signaling engineers in the monitoring and photosensitivity nature of the fibre and, using planning of the train traffic. One of its possible such an exposure, it is possible to obtain realize changes in the refractive index by factors as -3 2π 2π large as 10 in germanium-doped silica fibre. If 2 n = − λ eff Λ the fibre is exposed to a pair of interfering UV Β (2) beams as shown below, then in regions of constructive interference which correspond to or high UV intensity, the local refractive index will λ = 2 Λ n increase. Β eff (3)

where λB is called the Bragg wavelength – that is, the wavelength that satisfies the Bragg condition. The Bragg wavelength is dependent on the effective index as well as the grating period.

Fig 1. UV beams on optical fibre through a phase mask

At the same time, in regions of destructive interference, where the intensity of UV light is negligible, there is no index change. Therefore, an exposure to an interference pattern will result in a periodic refractive index modulation along Fig 2. Vector diagram the length of the fibre, the period of modulation being exactly equal to the spacing between the interference fringes.

3 Bragg Condition When light is made to propagate through an UV Reflected exposed fibre with a periodically modulated Signal refractive index, it is found that, under certain conditions, the propagating light beam can be strongly coupled to a mode propagating in the Fig 3. Reflected Signal satisfying Bragg condition backward direction. This happens when the Bragg condition is satisfied, i.e. the difference in the propagation constants of the two-coupled 4 Optical Interrogation System modes equals to the spatial frequency of the The function of the optical interrogation system grating, is to detect the wavelength shift in relation to the 2π external perturbation. Due to the characteristics βg − ( − βg ) = K = of FBG, the sensor possesses an intrinsic Λ (1) self-referencing capability as the sensed where, βg is the propagation constant of the information is directly encoded in accordance to forward propagating guided mode, - βg is that of wavelength, which is an absolute parameter and the backward propagating guided mode, and Λ is independent of the light power and loss in the transmission fiber. This feature is widely the spatial period of the grating. Figure 2 shows the corresponding vector diagram acknowledged as one of the most important advantages of the FBG sensors. wherein the propagating vectors and the grating vector satisfy the condition given by (1). If neff is The wavelength-encoded characteristic enables the effective index of the mode, then equation (1) the quasi-distributed sensing of measurands, such can be written as as temperature, strain, and pressure etc., to be arbitrarily located at any point along the fiber.

Based on the Bragg’s Conditions described before, it can be understood easily that by measuring the changes in λ, the changes in Λ

can be calculated by assuming that neff is relatively stable with respect to ambient changes. Since Λ is sensitive to both thermal expansion due to temperature changes and Fig 4. Interrogation System strain/stress, changes in Λ actually reflects strain/stress received and therefore the Each FBG of a quasi-distributed FBG sensor corresponding force/weight applied can be array has a unique reflected wavelength. The estimated by measuring changes in λ with wavelength interval between two neighboring suitable calibration. FBGs should be greater than the maximum wavelength shift of the individual FBG that Rail beam can actually be regarded as a simple results from external perturbations. In this way, structural beam experiencing compression, the magnitude of the measurant and at each neutral and extension stratum as shown in the arbitrary sensing position along the fiber can be diagram below. tracked simultaneously. In other words, there could be many FBG sensors engraved on a single 1. Compression optical fibre. This is obviously a very significant 2. Neutral 3. Extension advantage compared with conventional electrical sensors which require an independent circuit for 1 each sensor. 2 3 It can be shown that when force is added to the sensor, the response of the reflective wavelength Fig 6. Wheel weight on rail is different as shown below.

6 Site Test Arrangement Eight FBG sensors were fabricated and glued onto the rail beam as shown in Fig 7.

Fig 5. Reflective Wavelengths

5 FBG Axle Counter Exploration The weight of the train exerts significant force upon the rail to generate a high strain/stress on Fig 7. FBG sensor on Rail the rail beam. Such observation is very interesting for engineers wanting to exploit the The weight of the Electro-multiple Units intrinsic advantages of FBG sensors in railway (EMUs), i.e. the trains, would create both engineering, particularly the usage of FBG as tensile strain and stress on the rail beam as the development of a new generation of axle shown in Fig 6. The corresponding changes in Λ counters in railway engineering. thus resulted can be detected by attaching FBG sensors at positions 1 and 3 while on the neutral

plane, the FBG could be used to measure the Sensor ambient temperature of the rail beam. Practically, changes in the FBG grating Λ sensed would be presented as changes in λ and hence, according to the principle, the weight of the EMU will trigger a shape change in wavelength (Δλ). Once it is captured, it will Glue on the rail Finished represent the presence of an axle that is entering into a given section of the rail track. Fig 9. FBG sensors at Tai Wai Station

Eight FBG sensors were fabricated and glued Optical sensor interrogator and portable onto the rail beam at East Rail Tai Wai Station computer were used to record the data. The down line. The layout of the sensors are as optical fibre sensors were connected in series shown below in Fig 8. and fed into the input port of the interrogator as displayed below in Fig 10. The block diagram of the complete set up for measurement is shown in Fig 11. Data obtained were then transferred to the computer for recording purpose. Results are presented in terms of continuous waveform, showing the weight of

the train passing all the sensors. Fig 8. FBG sensors layout at Tai Wai railway station

The eight FBG sensors were located at the entrance and exit of the station. Their spacings were in line with the bogies and axles distances of the passenger coach. The wavelength of these sensors was shown in the table below-

FBG Wavelength Senso (nm) r 1 1536.7 2 1544.4 3 1562.3 Fig 10. FBG Sensor Interrogation Systems Display 4 1539.91 5 1546.81 Optical fiber connection 6 1549.72 between sensors and 7 1555.2 interrogator 8 1558.19 Fiber optic sensors Interrogator and installed on rails portable computer set

The range of lightwave’s wavelength of the Fig 11. Block diagram showing the complete set up for the

FBG sensor is set within 1530nm – 1570nm in measurement accordance to the specifications of the Optical

Interrogation System being used. The photos in Fig 9 indicate how the sensors 7 Field Data Analysis The 12-car EMU train running in East Rail is were glued onto the rail for testing purpose. composed of motor cars (M) and trailer cars (T) in the following configuration.

T-M-T-T-M-T-T-M-T-T-M-T

The basis configuration of the motor car and trailer car are as shown below in Fig 12-

Fig 14. Bending Moment Diagram

MLR-First Axle-S4 0.16 0.14 0.12 0.1 0.08 Compression Extension 0.06 0.04 0.02 0 -0.02 Change of wavelength (nm) -0.04 -0.06 Fig 15. Compression vs Extension Waveform

Fig 12. Train configuration 9 Axle Counter As each peak represents the presence of each All peaks of the strain obtained as shown below axle, by counting the number of peaks generated (Fig 13) for Sensors S2 and S3 are supposed to from each sensor, the FBG can perform the indicate the force of the vehicle wheels acting function of axle counter that are installed along on the rail through a transformation of strain the East Rail track in the Kowloon-Canton from the rail which are supported on ballasts. It Railway. From the waveform as shown in Fig is obvious that two distinguishable peak levels 16, one can see clearly that the number of axles are observed, one is for motor car and one is for passing through the sensor. As the waveforms trailer car, for which the latter is lighter. were taken for a 12-car train with 48 axles, thus 48 peaks were recorded. The importance is to MLR-S2 0.25 capture the peak created by the extension of the 0.2 0.15 rail while the inherent compression part of the 0.1 0.05 0 waveform is of no concern in the axle counting -0.05 manipulation algorithm. -0.1

MLR-S3 0.3

0.2

0.1

0 -0.1 Fig 13. Waveforms of two sensors

8 Offset Fig 16. Axle peaks The offset or the negative changes in Thus, it is possible to use FBG sensors as a good wavelength so observed at each peak is mainly alternative to the conventional axle counter as due to the compression seen by the sensor when shown below in Fig 17. the train wheels are approaching. The following The conventional axle counter is based on the bending moment diagram (Fig 14 and Fig 15) technology which detects a change in magnetic explains the situation. flux when an axle passes through. When the train passes the conventional axle counter, the magnetic flux will be changed as shown in Fig counter either. This is a very desirable 18. The track relay will control the evaluator characteristics of the proposed system as and also count the number of axles running on compared to the conventional method which is the rails. based on the magnetic flux change in the detection of wheels.

As fiber optic sensor system is totally immune from electromagnetic interference (EMI) throughout the entire process, that is, from signal generation to signal transmission to the interrogation system, there is no chance for any EMI disturbance in the system to cause any degrade in the integrity of the axle detector. This feature is in contrast to conventional strain gauge systems that suffer from EMI, particularly when electrical signals have to be transmitted via long cabling. Indeed, the problems associated with EMI is posing a great challenge to railway engineers, particularly for trains operating with a 25kV traction supply.

As the FGB axle counter is disturbance-free as compared with the conventional one in terms of its mode of detection and transmission, the Fig 17. Axle Counter in Tai Wai Station former is worthy of thorough investigation in a modern railway.

11 Virtual Track circuit using FBG Axle Counter FBG axle counter is wavelength-encoded, and gives an absolute measurement of axle Fig 18. Magnetic Flux Change once metal wheel is passing counting. This method avoids the problems with scale resetting and signal intensity variation that 10 Disturbance-Free FBG Axle are plaguing the intensity, phase and Counter vs Conventional One polarization modulated sensing systems. FBG axle counter have multiplexing ability. For the conventional method, if there is any Coupled with a high spatial resolution of this metallic part mounted on a vehicle running in kind of sensor, it makes the FBG system the best the vicinity of the sensor heads, the magnetic candidate in the building of virtual track circuits flux may be disturbed to result in the system on top of any moving block signaling system mis-counting the number of axle passes by. This that offers high frequency passenger trains will bring in disturbance to the system due to a service. By multiplexing sensors onto the same mis-match in the IN and OUT figures. With the optical fiber and distributing them over a FBG axle counter, even the actual weight of the distance of track of say, more than 10km, one vehicle wheel can be detected, hence there is no can sectionize the track into multi-sections of chance of any hanging metal on train causing any lengths, of any numbers and at any location, disturbance to the FBG sensor system. Any depending on the strategic needs as shown in change in shape of the wheel profile due to Fig. 19 below. Thanks for the low loss property normal wear and tear will not bring in any of the optical fiber and its immunity to EMI on uncertainty on the detection by the FBG axle long transmission. identify of each train type and by checking it, one can identify the type of trains entering and

Signal In Grating passing a specific section. This is of great use to

λ1 λ2 the operator for running hybrid train service and

Reflected Signal for running maintenance-engineering trains at mid-night, particularly if the manual system in Optical Fiber Gratings entering the train information is collapsed and it is necessary to re-identify what are the train Fig 19. Multiplexing FGB sensors types in the network for re-entering into the The virtual traction circuit is of great advantage system. since it can allow the train operator to locate the positioning of the trains and thus provide a 13 Train Speed Detection using standby system in running trains, particularly FBG axle counter for night engineering trains, which have train configurations that are totally different from any By placing FBG gratings suitable along a short of the standard passenger train length running in length of track, one can detect the timing when the daytime. the axle passes through them and one can calculate the train speeds as shown in Fig 21, which is another important data in checking the 12 Train Identification using running profile of the train along the entire FBG axle counter section of track.

Grating

λ1 λ2

distance

Speed = distance /time elasped Optical Fiber Gratings

Fig 21. Speed Detection

14 Conclusion Successful measurement results were obtained for further development of FBG axle counter in railway applications. The response signal obtained by the FBG sensors are positive and clear enough for further development. In essence, the FBG system could be tailor-made for the following railway applications including: a. Counting the numbers of axles in and out of a given section (track circuit Fig 20. Train Identification occupation) Through the interrogation system, it can be b. Train identification by tracking the observed that different types of trains passing differences of axle bases, weights & car through the sensors are associated with different numbers of various train stocks. patterns of waveforms as shown below in Fig 20. c. Speed detection As each train rake is configured with different In essence, a small piece of FBG sensor can vehicle weight, axle length and car numbers, generate lots of train running information with thus, the pattern can be treated as a bar code or good data integrity. It can be applied to monitor the train running pattern of the railway by building a virtual signal system on top of the conventional one as a standby system. A smart railway can be constructed based on FBG axle counter. Further analysis and measurements together with a full blown development on rail are necessary before the maturity of the FBG axle counter technology into practical usage in a highly conservative industry.

15 Acknowledgement The authors are indebted to Professor HY Tam of Hong Kong Polytechnic University for supporting and providing the testing facilities. Sincere thank is expressed to Kowloon-Canton Railway Corporation in allowing the test to be carried out at stations and supporting the publication of this paper.

References [1] Laurent MAURIN, Jonathan BOUSSOIR, Stephane ROUGEAULT, Michel BUGAUD, Pierre FERDINAND, Alain G. LANDROT, Y ves-Henri GRUNEVALD and Thierry CHAUVIN, FBG-based smart composite bogies for railway applications TuC12002 IEEE [2] K.O. Hill, Y. Fujii, D. C. Johnson, and B. S. Kawasaki, Photosensitivity in optical fibre waveguides: Application to reflection filter fabrication, Applied Physics Lett., 32, pp. 647 –649, 1978. [3] G. Meltz, W. W. Morey, and W. H. Glenn, Formation of Bragg gratings in optical fibres by a transverse holographic method, Optics Lett., 14, pp. 823 825, 1989