Claims:We Claims
1. The method of axle counting of railway vehicle, comprising the steps of:
• mounting of Fiber Bragg grating (FBG) sensors on the rail to measure strain;
• interrogating the light reflected by FBG sensors spaced apart from one another; and
• processing of available samples further comprises arranging of samples in consecutive windows; stabilizing the base line of the sensor dataset; filtering of local maxima; setting of threshold value and compare to get an accurate count.
2. The method of axle counting as claimed in claim1, where mounting of atleast one FBG sensor between the two sleepers of the rail.
3. The method of axle counting as claimed in claim2, wherein the distance between the two sensors is less than the wheelbase length of the train.
4. The method of axle counting as claimed in claim2, wherein the FBG sensors are placed on the web of the rail.
5. The method of axle counting as claimed in claim4, wherein FBG sensors are pasted on rail using glue and steel tape.
6. The method of axle counting as claimed in claim1, wherein arrangement of samples in consecutive windows is using sliding window mechanism.
7. The method of axle counting as claimed in claim6, wherein the steps of arranging the samples in consecutive windows using sliding window mechanism comprises
• dividing the data in three types of windows;
• first and second windows are overlapping windows;
• third window is intersection of first and second window; and
• add the peaks of first and second and subtract the intersection part to detect false peaks.
8. The method of axle counting as claimed in claim7, wherein window is having fixed sample size to be processed.
9. The method of axle counting as claimed in claim1, where in stabilization of the baseline of sensor dataset is using rolling window technique
10. The method of axle counting as claimed in claim9, where in the step of stabilizing the baseline of sensor dataset using rolling window technique comprises
• fixing of size of rolling window for samples of dataset;
• identifying the difference between maximum and minimum element in the window; and
• calculation of mean of the sample difference to get stabilised data.
11. The method of axle counting as claimed in claim10, wherein the size of rolling window is less than samples of dataset.
12. The method of axle counting as claimed in claim1, wherein filtration of local maxima is using peak detection technique.
13. The method of axle counting as claimed in claim1, wherein a setting of threshold is to obtain prominent peaks of samples.
14. The method of axle counting as claimed in claim13, wherein peaks of samples is more than predetermined threshold value increments the counter.
15. The method of axle counting as claimed in claim 1, wherein the steps of finding of direction comprises
• Identification of peaks from available sensors dataset;
• Comparison of sample number of peaks; and
• Counting of in-out axles.
16. The method of axle counting as claimed in claim 1, wherein the steps of classification of railway vehicle comprises
• Identification of peaks from available sensors dataset;
• Comparison of amplitude of dataset of engine and wheel; and
• Counting of axles across engine and wheel.
, Description:Field of the invention
The present invention relates to method for vehicle axle counting system for rail, and more particularly, to axle counting method using Fiber Bragg Grating (FBG) sensors mounted on a rail. The area of concern of this system is to count the number of axles of train, direction of train and the classification between train locomotive and its wagons.
Background
An Axle counter system plays a major role in railway signalling. An axle counter is a reliable and cost-effective device to monitor track sections. The device detects the passing of train between two points on a track. As train passes through the two points to detect the occupancy of track. The device can also detect the direction and speed of train.
Conventionally, vehicle axle counting is performed by Electromagnetic sensors (EMS) which needs power at sensing point and effected by electromagnetic interference that leads to miscounting.
The Application No. CN 101376392A, entitled “Vehicle axle counting method based on steel rail deformation/stress parameter” relates to method for counting vehicle axles based on steel-rail deformation or the stress. The application includes sensor either Fiber grating or electrical strain foil to be mounted on rail. This sensor senses strain or stress in rail during movement of train on rail. The sensor test point sets the threshold value to identify axle. But this method is more generic and uncertain.
The Application No. CN1676389B, entitled “System for monitoring railway tracks” relates to method for determining the characteristics of track. The system includes optical fiber, signal transmitter and signal analyser. The signal transmitter transmits optical signal to optical fiber, attached to a track. The optical analyser receives and analyses the changes in characteristics of optical fiber due to fiber Bragg grating in optical fibre. Hence, the method includes detecting and analysing the characteristics of optical fiber varies with the track. But the application is focussing on method to detect health of railway tracks not on the method of axle counting.
The Application No. WO2016/150670, entitled “Axle counting Method and Axle counting Device” relates to axle counting method for rail bound vehicles using two FBG sensors with Bragg wavelength and full width at half maximum. The method includes generation of temporal changes of the difference of two shear stresses at two sensor positions placed apart from one another. When this shear stress difference signal exceeds or falls below a predetermined limiting value then generate a wheel signal. But, this method of axle counting is more complex and using a pair of sensors for wheel detection.
The Application No. CN106494454B, entitled “Railway axle counting apparatus, a method, a system, and a signal processing device” relates to railway axle counter apparatus, method, system and a signal processing device. The invention provides an apparatus for railway axle counters comprising a broadband light source, a ring, a meter shaft unit, the optical fiber grating demodulation unit. The axle counting method includes receiving an electrical signal from the gauge axis railway axle counter device; processing of counter shaft electrical signal; then count determines the direction of train, train operation and speed of train. But this system and method is very complex and converting optical signal to electrical signal and processing electrical signal to count the number of axles.
Therefore, a stable method for real time counting of axles of railway vehicle with false peaks detection at the window boundaries of sensor data is highly desired.
Summary of invention
The present invention fulfils the foregoing needs by providing an optical fiber sensing based method for counting of vehicle axle in real time. Firstly, FBG sensor mounted on a rail sensing points. The FBG sensors are connected to an optoelectronic instrument using optical patch cords to interrogate reflected optical signals with a Bragg wavelength. When the train axle passes over the rail, the track gets strain and the FBG sensors cause a shift in Bragg wavelength.
A Data processing unit (DPU) processes the dataset received from optoelectronic instrument and stabilize baseline of the dataset using rolling window technique. It further includes filtering out of local maxima using peak detection technique and finding of peaks and detection of prominent peaks by setting of threshold value. When the peaks amplitude is more than predetermined threshold value, the counter increments one and considered as an axle passing.
It also includes comparing of sample number of peaks of both FBG sensors and count the number of IN-OUT axles to find direction of train.
Hence present invention provides a stable and real time solution to count the axles of railway vehicle through the strains on the rail measured using FBG sensors. The invention also detects direction and classify railway vehicle.
Brief description of figures
Exemplary embodiments of the present invention are fully explained with the description below and the accompanying figures, wherein:
Figure1: Illustrate System architecture.
Figure2: Flow chart for software development
Figure3: Sliding Window Mechanism
Figure4 (a,b,c): Example of false peak detection using overlapping window at edge of window
Figure5: FBG sensor S1 dataset with Bragg wavelength
Figure6: FBG data using Rolling window
Figure7: Local maxima of the plot
Figure8: Prominent peaks of available data samples
Figure9: Detection of direction of train
Figure10: Classification between locomotive and its wagons
Figure11: Experimentation of method of Axle count with Rajdhani Express (22961)
Figure12: Raw data of Rajdhani Express with engine and two coaches
Figure13: Data stabilization using Rolling window with engine and two coaches of Rajdhani Express
Figure14: Filtering peaks using Peak detection technique with engine and two coaches of Rajdhani Express
Detailed description of invention
The foregoing description of the embodiments, the various features, and advantageous details of the invention has been presented for the purpose of illustration. It is not intended to be exhaustive or to limit the invention to the precise form disclosed as many modifications and variations are possible in light of this disclosure for a person skilled in the art in view of the figures and description. The examples used herein are intended merely to facilitate an understanding of ways in which the embodiments herein may be practiced and to further enable those of skill in the art to practice the embodiments herein. Unless defined otherwise, all technical and scientific terms used herein have the same meanings as commonly understood by person skilled in the art.
The embodiments herein below provide a method for counting the axle of railway vehicle using FBG sensors in real time.
Fig 1 illustrates the system architecture includes FBG sensor 101, optoelectronic instrument 102, data processing unit (DPU) 103, Ethernet Switch 104, Wi-Fi Access Point 105.
The optical fiber with Fiber Bragg grating (FBG) sensors S1 and S2 101 as shown in fig.1 is most commonly used and broadly deployed optical sensors. The sensors reflect variation in strain when the trains wheel pass over the rail.
These sensors are pasted on railway track using special type of glue and special type of tape. Generally, steel tape is used for pasting.
FBG sensors (S1, S2) 101 are placed side by side on the web of the rail. Generally, each sensor is placed in between the two sleepers.
Further in fig1, an optoelectronic instrument 102 acts as a light source for optical fiber with FBG sensors 101 and receives, analyses the reflected light from FBG sensor.
FBG sensors (S1, S2)101 are connected to optoelectronic instrument 102 via optical patch cords.
When the train wheels pass over the rail, the strain on rails gets magnified.
Optoelectronic instrument 102 is connected to Data processing unit (DPU) 103 via USB port. The data acquisition from optoelectronic instrument 102 to DPU 103 is first in, first out (FIFO) method.
Additionally, to fig1, Data processing unit (DPU) 103 processes streams of data collected from optoelectronic instrument 102. It consists of complex techniques, software, database etc. to process the data received from optoelectronic instrument. DPU 103 stabilizes the baseline of received dataset to find peaks, filtering of local maxima and detection of prominent peaks by setting of threshold value.
When the peaks amplitude is more than threshold value, the counter increments one. The sensor records the pressed wheel condition occurs and considered as an axle passing.
Data Communication System (DCS) Consists of Ethernet Switch 104, Wi-Fi Access Point 105, Gateway for wired/wireless communication between DPU and the local/remote users.
Fig 2 shows the flowchart for software development. The process starts 201 with dataset collected by FBG sensors 202 by strains on the rail. Firstly, stabilizing the baseline of sensor dataset 203 using rolling window of fixed size; finding of local maxima 204 using peak detection technique; setting of threshold value 205 to get prominent peaks. As the peak amplitude is greater than the predetermined threshold value 206, the counter increments one and considered as an axle passing 207.
Further in fig.2, includes identification of peaks 208 from both FBG sensors (S1, S2) 101, and comparison of sample number of both peaks 209 to find direction 210 based on in-count and out-count of axles.
Fig.3 shows the sliding window mechanism for detection of false positives and false negatives at the window boundaries. The method is used to real time processing of FBG sensor data. The optoelectronic instrument sends the FBG sensor data to DPU in packets of N samples. Each sample is appended to a queue, from which data is sent in consecutive windows (W1, W2, W3,.......) of a fixed sample size to be processed. The pre-processed data of Window W1 is sent in overlapping window W2 and so on. In this way, the peaks that are missed at the edge of one window is identified by its overlapping window. Also, the intersection of two window samples (W12, W23, W34,.........) is subtracted from overlapping window samples (W1-W2, W2-W3, W3-W4) respectively to count the total number of peaks. This method also detects false peaks (positive and negative) at window boundaries.
For example, fig.4a shows plot of first window size of 10000 samples collected from sensor S1 and three peaks detected after axle counting method. Fig.4b shows plot of next consecutive 10000 samples collected from same sensor S1 and one peak detected after axle counting method. But, fig. 4c shows five peaks detected for same sample window of total size 20000 using peak detection technique. Hence sliding window mechanism detected one false negative/missed peak from a data of 20000 samples.
Fig.5, sensor S1 data samples with Bragg wavelength, when the train wheel passes over the rail.
Further, rolling windows technique is used to stabilize the baselines of the available sensor dataset. The method includes dataset of N samples, with rolling window of fixed size W (