Application of Optical Fiber Sensor in Safety Monitoring of Railway Perimeter

Zheng Qian

（Wuhan Optics Valley Opto-interconnecting Network Technology Co.,Ltd,Wuhan 430000）

Abstract:Railway exhibits the characteristics of geographical dispersion,long distance along the line,complex and changeable natural environment,and its perimeter safety is usually difficult to be guaranteed.It has important theoretical and application value to build an efficient safety monitoring system to accurately identify the security risks of railway perimeter.This paper introduces the current mainstream perimeter security technology,analyzes the working principle of M-Z interference sensor,designs a railway optical fiber perimeter security monitoring system,and focuses on multichannel data acquisition and processing and intrusion vibration data extraction.The results show that the overall sensing accuracy of the system can reach 92.2%,which can meet the practical application requirements.

Key words:M-Z interference;Optical fiber vibration sensor;Railway perimeter

The natural environment around railways is complex and changeable,and is vulnerable to invasion caused by various natural disasters such as landslide,collapse,debris flow,earthquake and so on.The traditional railway perimeter safety protection mostly adopts human inspection,which can not meet the increasingly higher demands of railway perimeter safety[1].Therefore,the safety protection of high-speed railway perimeter has become an important issue to be addressed.

Optical fiber vibration sensor has the characteristics of long sensing distance,high sensitivity and simple structure,which is very suitable for the safety monitoring needs of large-scale infrastructure such as railway.However,in order to capture a large range of railway perimeter vibration space-time information and meet the needs of railway perimeter safety monitoring engineering,optical fiber vibration sensor is faced with a large amount of data concurrency,complex and diverse intrusion vibration modes,longterm stable operation and other technical problems[2].Therefore,in order to improve the safety level of railway perimeter,this paper discusses the application of optical fiber sensor in railway perimeter safety monitoring.

1 Current Mainstream Technology

Since the beginning of the new century,a variety of security technologies have emerged and been applied widely.At this stage,the perimeter security technologies used in the project are mainly divided into the following categories[3]:

1.1 Active Infrared Detection

The infrared detection system is mainly composed of infrared transmitter,infrared receiver and terminal signal processor.When there is a barrier between the transmitter and the receiver toprevent the receiver from receiving infrared light,the light intensity received by the receiver will change,and the system will output an alarm signal when it reaches the set alarm threshold.

1.2 Shielding Microwave Detection

The sheltered microwave detection system applies the principle of electromagnetic disturbance to detect the invading target.The system consists of microwave transmitter,microwave receiver and signal processing unit.When the system is working,the transmitter sends the processed modulated beam to the microwave receiver directly.When there is an intrusion object,the normal propagation of the microwave is changed,resulting in the change of the signal received by the receiver.The signal processing unit generates an intrusion alarm through the change of the signal.

1.3 Optical Fiber Sensing Detection

Optical fiber perimeter detection technology uses optical fiber as a sensor to sense the surrounding environment,and uses signal processing system to recognize the intrusion signal pattern,and then outputs the intrusion alarm.The optical signal generated by the laser is transmitted through the optical cable,and two interference optical signals are generated through the coupler.The signal can be used as the vibration sensing carrier for vibration detection,so as to realize the protection function and early warning.When the strength of the illegal intrusion signal is higher than the predetermined index,the alarm signal is generated.Table 1 makes a comparison of the characteristics of the above perimeter security technologies.

Tab.1 Characteristic Comparison of perimeter security systems

2 Working Principle of M‑Z Interference Sensor

M-Z interferometric optical fiber sensor is based on the principle of optical phase modulation for vibration detection.The sensor consists of laser source,two couplers and a photo-detector.The light source generates two interference beams through the coupler,and the two interference beams interfere after entering the coupler through the sensing arm and the reference arm respectively.If there is no vibration,the two coherent beams do not change relatively,and the detector can not detect the change of light intensity after the interference of coupler.Once the sensing arm and the reference arm vibrate,the core diameter,refractive index and length of the optical fiber at the vibration point of the sensing arm and the reference arm will change accordingly.Due to different optical fiber parameters,the phase of the two coherent beams will change relatively.After the interference in the coupler 2,the light intensity will change.The phase difference is obtained by detecting the coherent light intensity through the photoelectric converter.The optical path of M-Z dry sensor is shown in Figure 1.

Fig.1 Optical path of M-Z dry sensor

3 Design of Railway Optical Fiber Perimeter Safety Monitoring System

According to the characteristics of optical fiber perimeter alarm and the particularity of railway perimeter environment,this paper designs the system from three levels:multichannel acquisition processing layer,optimization processing layer and decision classification layer according to the general signal processing flow.

3.1 Multichannel Acquisition and Processing Layer

The railway distance between two adjacent stations is long,therefore,multiple defense areas are divided for data collection.The real-time monitoring of multi-defense areas produces a large amount of data,which requires the system to process a large amount of data in parallel and real-time.The data processing of a single defense area includes photoelectric conversion,analog amplification and filtering,A/D conversion,digital filtering,and the extraction,storage and transmission of vibration data.

3.2 Optimized Treatment Layer

It is difficult for computer to process a large number of data in multi-defense area in real time.FPGA high-speed parallel processing is used to process multichannel data.Some features of the original vibration data and vibration signal are extracted.Hardware and software are combined to obtain and process data,the intrusion vibration data and information are provided for the decision classification layer.

3.3 Decision Classification Level

Railway intrusion behavior are complex,which makes it difficult to train the patterns and for backup data,this layer mainly completes the pattern learning through a small amount of intrusion pattern training and environment adaptation,and uses the training template to quickly classify the vibration signal,so as to output the correct alarm.

The overall architecture of the system is shown in Figure 2.

Fig.2 Overall structure of the system

4 Multi‑channel Data Acquisition and Processing

Facing the demand of multichannel data parallel processing,the system uses FPGA as the main control chip of the acquisition card to realize multichannel data parallel pre-processing,which can effectively reduce the calculation amount of the upper computer algorithm and improve the real-time performance of the system.The system pre-processing acquisition card is mainly composed of photoelectric conversion module,FPGA logic control module and USB3.0 data transmission module.After amplifying and filtering the original electrical signal,the data can be collected by multichannel ad chip in real time without breakpoint,and the digital signal can be obtained by controlling a/ D chip with FPGA.Due to the different railway boundary environment,the analog filter can not completely filter out the noise,which is not conducive to the data processing.FPGA has a excellent speed in data parallel processing.The data filtering and event extraction are implemented in FPGA,which speeds up the data processing and promote real-time performance.The system uses two SDRAM for data ping-pong operation to ensure the integrity of data storage and the safety and reliability of data storage.Through the design of the interface between PFGA and USB3.0,it can complete the high-speed and reliable transmission of large amount of data.

5 Extraction of Intrusion Vibration Data

Because of the large fluctuation and non periodicity of the intrusion vibration signal,it provides a favorable condition for the extraction of the intrusion vibration event.When judging the event and extracting the intrusion signal data,we should accurately identify the intrusion vibration,ensure the integrity of an intrusion vibration when extracting the data,and try to ensure the simple operation of intrusion vibration judgment.In the extraction of vibration events,the commonly used methods include:short-time zero crossing,short-time energy and auto-correlation.Due to the randomness of optical fiber vibration signal,short-time energy is used to detect and extract the intrusion vibration data.

When there is no external vibration,the amplitude fluctuation of the optical fiber data is very small,and the short-time energy of the signal is small,while the energy is large when there is vibration.Therefore,the real-time short-time energy of the signal can be used to monitor the optical fiber vibration.The magnitude of the vibration amplitude can be expressed by the short-time energy of the signal.Noise signal energy is smaller than vibration signal energy,and vibration signal energy is the superposition of environmental vibrationand noise energy.In fact,for the vibration signal with high signalto-noise ratio,the short-term energy of vibration signal and noise signal is obviously different,thus the energy can be used to judge the environmental noise and vibration signal.

Tab.2 Extraction of vibration events

6 System Test and Application

The construction of perimeter alarm system environment mainly includes external optical fiber sensing and internal signal processing.The external optical fiber sensor is mainly used to detect the vibration,and the vibration signal is transmitted to the signal processing part through the sensitive sensing of the sensing optical cable.The signal processing part processes and stores the original vibration signal,and then outputs the alarm.

When the system extracts data,it extracts data in silent state,knocking state,crossing state,wind and rain state.In the actual construction process,due to the difference of fusion and optical path of optical cable,four defense areas are used for data acquisition.The energy threshold of the system is set to 2 times of the energy in the silent state.Each defense area collects 100 times of data,and 400 times of field data are collected in each vibration state of knocking state,crossing state and wind rain state.The results of event extraction are shown in Table 2.

The above table shows that no event is extracted in the silent state,but the number of event extraction in the knocking and crossing state is more than the actual situation.According to the actual waveform analysis,the reason for the large number of event extraction is that the fence will produce a small amplitude of shaking after severe knocking,resulting in a large number of event extraction times.Overall,the perception accuracy is about 97.5% in the knocking state,about 99% in the over state,about 80% in the wind and rain state,and about 92.2% in the comprehensive perception accuracy.

7 Conclusion

Taking the railway perimeter safety as the research subject,this paper studies the application of optical fiber sensor in the railway perimeter safety monitoring by using the characteristics of M-Z optical fiber,such as long sensing distance,high sensitivity and simple structure,combined with the safety monitoring requirements of railway linear distribution,aiming at improving the safety of railway perimeter.The research shows that the sensing accuracy is about 97.5% in the knocking state,about 1.0% in the crossing state,and 80% even in the rain and wind state.Therefore,the railway perimeter safety monitoring system proposed in this paper can accurately perceive the safety status of the railway perimeter,and plays an important role in the railway security work.

Reference: