Specifications of the device B and the device G.
Abstract
Although track maintenance is important, many operators of regional railway with limited financial resources are unable to conduct sufficient track inspections. In response to this problem, a track condition diagnosis system using car-body vibration sensors has been developed. In this study, a track condition monitoring system using a smartphone for general use has been developed. A technique for identifying train location using global navigation satellite system (GNSS) speed is proposed. The results of field testing shows that track condition diagnosis is possible using a smartphone-based monitoring system.
Keywords
- railway
- track
- condition monitoring
- wavelet transform
- Hilbert-Huang transform
- smartphone
1. Introduction
Railway track management helps support and guide wheels and is very important in terms of comfort and safety. Track management is usually carried out by track maintenance workers and usually uses track geometry cars. However, these methods are labor-intensive and costly, and many operators cannot carry out sufficient track inspections of regional railways due to limited financial resources. To address this problem, a track diagnosis system using an exclusive onboard sensing device has been developed [1, 2, 3, 4, 5].
Currently, track maintenance and management on railways is based on measured data of track displacement. However, track displacement measurement requires expensive equipment such as track inspection vehicles and measuring devices. A more economical management method is required, especially for regional railways. Other methods of track management exist in addition to displacement measurement, e.g., train vibration inspection using vehicle vibration measurement, but none of them constitutes a fundamental inspection method such as track displacement inspection owing to the low reproducibility of measurement data.
Meanwhile, recent advances in the practical use of IoT devices such as smartphones incorporating accelerometers based on microelectro mechanical systems (MEMS) have been reported. These devices can be used as simple and inexpensive vehicle vibration measurement devices. Many studies on track monitoring systems using vehicle vibration measurement with such IoT devices were conducted.
In this study, a track condition monitoring system for use on a smartphone was developed to reduce the cost of such a system. Using two types of IoT devices for business use and a commercially available smartphone, we took measurements of an actual car, compared the performance of both IoT devices, and diagnosed the condition of the track.
2. Literature review on track condition monitoring based on in-serve train vibration
2.1 Track condition monitoring from an in-service vehicle
Track maintenance and management on the railway is based on measured data of track displacement. However, track displacement measurement requires expensive equipment such as track inspection vehicles and measuring devices. A more economical management method is required, especially for regional railways.
Other methods of track management exist in addition to displacement measurement, e.g., train vibration inspection using in-service vehicle vibration measurement [6, 7]. Many studies on track monitoring systems using in-service vehicle vibration measurement with on-board sensing devices were conducted both in Japan and abroad.
2.2 Axle-box-mounted sensors
Chen
Sun
Chudzikiewicz
2.3 Bogie-mounted sensors
Some types of track faults were detected by measuring the acceleration of bogies. Weston
Malekjafarian
2.4 Car-body-mounted sensors
Tsunashima
Bai
A track condition monitoring based on the bogie and car-body acceleration measurements was presented and verified in Shang Hai metro Line 1 [17].
Balouchi presented a cab-based track monitoring system developed in the UK. They presented through comparison of vibration response from sites with known defects and outputs from Network Rail’s New Measurement Train (NMT). Good agreement was reported for track faults in relation to vertical and lateral alignment and dip faults [18].
2.5 Signal processing
To extract a signal on faulty tracks from measured vehicle vibration, several techniques using nonmodel-based and model-based method were proposed.
Tsunashima
A Kalman filter-based method to estimate the track geometry of Shinkansen tracks from car-body motions was proposed [19]. The proposed Kalman filter-based estimation technique was modified and applied for conventional railways [20].
Tsunashima proposed a classifier based on machine learning techniques for identifying track faults automatically from measured car-body vibration [3]. It is shown that the degradation of track can be classified in the feature space consists of car-body vibration RMS.
A new method for automatically classifying the type and degradation level of track faults using a convolutional neural network (CNN) by imaging car-body acceleration on a time-frequency plane by continuous wavelet transform [5].
2.6 Smartphones-based system
Chellaswamy
Rodríguez
Cong
Paixão
3. Constructing the track diagnosis system
3.1 Track irregularities and track faults
Major private railway companies and Japan Railways (JR) use track inspection vehicles to measure track displacement, and track management is based on such measurements. Track irregularities such as longitudinal level, alignment, gauge, cross level, and twist (depicted in Figure 1) should be controlled properly.
However, it is difficult for regional railway companies to introduce track inspection vehicles because of the cost. Moreover, manual inspection by track maintenance staff is inefficient and expensive.
3.2 Overview of track management with proposed system
Figure 2 depicts the track management method used in this study. A 3-axis accelerometer mounted on a smartphone was used to measure the vibration of the car-body, a 3-axis Gyro sensor was used to measure the angular acceleration, and a GNSS sensor was used to collect information about the position and traveling speed; all data are then transmitted to the server. By analyzing the transmitted data, the condition of the track can be diagnosed, and, based on the result, railway operators can prioritize track maintenance and work.
3.3 Measurement devices
A BL-02 IoT device for business use (hereafter referred to as Device B) and a commercial smartphone Galaxy S7-edge (hereafter referred to as Device G) were used for measurements. Figure 3 shows a photograph of these devices, and Table 1 details their specifications.
IoT Devices | Device B: BL-02 | Device G: GalaxyS7-edge |
---|---|---|
CPU | Cortex-A7 | Snapdragon820 |
OS | Android6.0 | Andorid6.0 |
Display | 2.8inch | 5.5inch |
Sensor | 3-axis accelerometer, 3-axis Gyro sensor, GNSS sensor | 3-axis accelerometer, 3-axis Gyro sensor, GNSS sensor |
Sampling frequency | 232 Hz | 417 Hz |
Size | 94 × 58 × 16 mm | 151 × 73 × 8.3 mm |
Weight | 102 grams | 158 grams |
Both devices were equipped with a 3-axis accelerometer, a 3-axis Gyro sensor, a GNSS sensor that can determine the location and traveling speed, and 4G internet, which is required for data transmission and reception.
At the time of measurement, data from Device B were measured at 232 Hz and data from Device G were measured at 417 Hz; both were down-sampled to 80 Hz at the time of acquisition from the server to reduce the amount of processing required for analysis.
Using these devices, we can measure and diagnose the vibration of the car-body. Considering convenience and GNSS reception environment, we installed the smartphone near the driver’s cab, as shown in Figure 4.
3.4 Identifying the areas of interest from the vibration measurements
Smartphones are able to acquire latitude and longitude information; however, location detection errors increase when methods such as map matching are not employed. Therefore, we adopted a method to calculate the milage using the GNSS speed, which was, in turn, calculated using the Doppler effect of the GNSS carrier wave.
4. Verification of measurement data
4.1 Verification of vibration and angular velocity data
Figure 5 shows the measurements from Devices B and G installed on an actual car on Regional Railway A (line length: 30.5 km, Stations:17, Max. speed: 85 km/h) in December 2021. The data from both devices are almost identical in phase and amplitude.
Figure 6 shows the power spectral density (PSD) of the vertical acceleration of the vehicle. The frequency characteristics of both devices were consistent, and we can conclude that they yield sufficient accuracy as onboard sensing devices.
4.2 Identification of train location
4.2.1 Comparison of GNSS speeds
Identifying the location of a train is important for track management. In this system, the location of a train,
The measured GNSS speed was shown in Figure 7. It should be noted that the measured GNSS speed was affected by multipath errors. Multipath is a major error source for GNSS receivers [25].
The location of a train can be estimated as shown in Figure 8 using the measured raw GNSS speed. The location data are affected by the number of satellite navigation systems supported by the device. Device B supports few satellite positioning systems and is not compatible with A-GPS; therefore, the number of satellites it receives information from differs to that of Device G, and it is considered to be more susceptible to multipath errors.
We used a correction process that used a median filter for the GNSS speeds affected by multipath errors. By performing median filter processing with a window size of 800 data for approximately 5 seconds, and taking into account the magnitude of the effect of the multipath errors, we were able to improve the rapid decrease in speed due to multipath errors of the GNSS speeds measured by Device B, as shown in Figure 9.
In addition, to evaluate the effect of the median filter processing on the accuracy of the location identification, we investigated the relationship between the vehicle location obtained by integrating the GNSS speeds and that obtained using the car-body vertical acceleration, as shown in Figure 10.
This figure shows that the position error is greatly improved between 450 and 550 m, which is a section that is particularly affected by multipath errors.
5. Track condition diagnosis using time-frequency analysis
5.1 Effect of track faults on time-frequency plane
5.1.1 Continuous wavelet transform (CWT)
The wavelet transform is well-known technique for analyzing nonstationary signals [26, 27]. A CWT gives simultaneous detection of the frequency and time characteristics for a nonstationary signals using a wavelet
The CWT is calculated using the mother wavelet
where
Eq. (3) translates a source signal
In this study, the
The CWT is subject to the uncertainty principle on time-frequency domain. In case of fault detection using CWT, if we are focusing on frequency related on the fault, the time when the fault occurred will be vague. If we are focusing on the time when the fault occurred, the frequency will be spread widely on the time-frequency plane.
5.1.2 Hilbert-Huang transform (HHT)
The Hilbert-Huang transform (HHT) has been proposed for analyzing nonlinear and nonstationary data by Huang
The HHT consists of two operations. The first operation is the empirical mode decomposition (EMD) and the second operation is Hilbert transform.
The EMD operation breaks time domain data into intrinsic oscillatory modes called intrinsic mode functions (IMFs). The second operation is the Hilbert transform. Instantaneous amplitude, instantaneous phase, and instantaneous frequency of the IMFs are obtained by the Hilbert transform.
An IMF must satisfy the following requirements: (1) the number of local extrema and the number of zero crossings must either equal or differ by at most one. (2) the mean value of the envelopes of local maxima and local minima is zero at any point.
For extracting IMFs from the original signal, the iterative sifting process is applied. Once the first IMF is calculated, it is subtracted from the original signal to obtain a residual value. The EMD operation is applied again to the residual. This process repeats until the residual no longer contains any oscillation modes.
The original signal,
where
Followed by the EMD operation, the analytical signal
where
where PV shows Cauchy principal value.
Instantaneous amplitude,
where
This data-driven method is highly adaptive. However, intrinsic mode functions (IMFs) obtained by EMD strongly depend on the data itself. Thus, a small change in the data will appear on different decomposition level.
5.2 Track condition diagnosis for regional railway lines
5.2.1 Regional railway A
Time-frequency analysis was performed on the measured data to identify and evaluate the detailed location and type of track fault. When a train runs on a track where a fault exists, characteristic vibration corresponding to the type of track fault occurs. Therefore, one could identify the type of track fault and location of its occurrence by analyzing the time-frequency plane of measured car-body vertical acceleration.
Figure 11 shows the time-frequency analyses, CWT and HHT of data measured in December 23, 2021 in Railway A (line length: 30.5 km, Stations:17, Max. speed: 85 km/h). The data used for this analysis are data measured using Device G.
It can be seen from Figure 11 that a high-frequency vibration appeared at 27000 m, which was caused by the joint depression [4]. Whereas a large vibration can be seen in low frequency in a 27,015–27,025 m section. This is caused by longitudinal-level track irregularities.
5.2.2 Regional railway B
Figures 12 and 13 show the time-frequency analyses, CWT and HHT of data measured in June and October 2022, respectively. In June 2022, vibrations due to longitudinal-level irregularity were detected at 1–2 Hz between 600 and 700 m but were no longer detected in October 2022 due to track irregularity correction. The data used for this analysis are data measured using Device G on Regional Railway B (line length: 6.4 km, Stations: 8, Max. speed: 40 km/h) in June 2022 and October 2022.
Figure 14 displays a photograph of the track section between 600 and 700 m in October 2022; the ballast was newly replenished, line maintenance work was carried out, and the longitudinal-level irregularity was eliminated. Thus, by performing time-frequency analysis using data measured by a smartphone, the type and location of track fault can be identified, and the effects of track irregularity correction can be confirmed.
6. Conclusion
In this study, we measured the car-body vibration of an in-service train using a smartphone and verified whether the track condition could be diagnosed. We were able to monitor and diagnose the track condition using both the IoT device for business use and the commercially available smartphone.
The accuracy of the GNSS speed, which is necessary to identify the location of the train, was reduced by the number of satellites received by the smartphone, that is, by the number of compatible satellite positioning systems. Therefore, when selecting a smartphone, the number of supported satellite positioning systems must be considered. In addition, we determined that the performance of devices susceptible to multipath errors can be improved by performing median filtering on the GNSS speed.
Time-frequency analysis of measured car-body acceleration obtained by a smartphone shows that proper diagnosis of track condition is possible using smartphone-based track condition monitoring system.
In the future, we plan to acquire data on a continuous basis and conduct track condition diagnosis.
Abbreviations
CWT | continuous wavelet transform |
HHT | Hilbert–Huang transform |
GNSS | global navigation satellite system |
MEMS | microelectro mechanical systems |
IoT | internet of things |
RMS | root mean square |
EMD | empirical mode decomposition |
References
- 1.
Mori H, Ohno H, Tsunashima H, Saito Y. Development of compact size Onboard device for condition monitoring of railway tracks. Journal of Mechanical Systems for Transportation and Logistics. 2013; 6 (2):142-149. DOI: 10.1299/jmtl.6.142 - 2.
Tsunashima H, Mori H, Ogino M, Asano A. In: Zboinski K, editor. Development of Track Condition Monitoring System Using Onboard Sensing Device. London, UK: Railway Research; IntechOpen; 2015. DOI: 10.5772/61077 - 3.
Tsunashima H. Condition monitoring of railway tracks from car-body vibration using a machine learning technique. Applied Sciences. 2019; 9 (13):2734. DOI: 10.3390/app9132734 - 4.
Tsunashima H, Hirose R. Condition monitoring of railway track from car-body vibration using time–frequency analysis. Vehicle System Dynamics. 2020; 60 (4):1170-1187. DOI: 10.1080/00423114.2020.1850808 - 5.
Tsunashima H, Takikawa M. Monitoring the condition of railway tracks using a convolutional neural network. In: Bulnes R, editor. Recent Advances in Wavelet Transforms and their Applications. London, UK: IntechOpen; 2022. DOI: 10.5772/intechopen.102672 - 6.
Weston P, Roberts C, Yeo G, Stewar E. Perspectives on railway track geometry condition monitoring from in-service railway vehicles. Vehicle System Dynamics. 2015; 53 (7):1063-1091. DOI: 10.1080/00423114.2015.1034730 - 7.
Kraft S, Causse J, Coudert F. Vehicle response based track geometry assessment using multi-body simulation. Vehicle System Dynamics. 2018; 56 (2):190-220. DOI: 10.1080/00423114.2017.1359418 - 8.
Chen X, Chai X, Cao X. The time-frequency analysis of the train axle box acceleration signals using empirical mode decomposition. Computer Modelling and New Technologies. 2014; 18 (7):356-360 - 9.
Karis T, Berg M, Stichel S, Li M, Thomas D, Dirks B. Correlation of track irregularities and vehicle responses based on measured data. Vehicle System Dynamics. 2018; 56 (6):967-981. DOI: 10.1080/00423114.2017.1403634 - 10.
Tsai HC, Wang CY, Huang NE, Kuo TW, Chieng WH. Railway track inspection based on the vibration response to a scheduled train and the Hilbert-Huang transform. Proceedings of the Institution of Mechanical Engineers, Part F: Journal of Rail and Rapid Transit. 2014; 229 (1):1-15 - 11.
Sun X, Fei Y, Shi J, Zaitian K, Yunlai Z. On-board detection of longitudinal track irregularity. IEEE Access. 2021; 9 :14025-11437. DOI: 10.1109/ACCESS.2021.3052099 - 12.
Chudzikiewicz A, Bogacz R, Kostrzewski M, Konowrocki R. Condition monitoring of railway track systems by using acceleration signals on wheelset axle-boxes. Transport. 2019; 33 (2):555-566. DOI: 10.3846/16484142.2017.134210 - 13.
Weston P, Ling C, Goodman C, Roberts C, Li P, Goodall R. Monitoring vertical track irregularity from in-service railway vehicles. Proceedings of the Institution of Mechanical Engineers, Part F: Journal of Rail and Rapid Transit. 2007; 221 :75-88. DOI: 10.1243/0954409JRRT65 - 14.
Weston P, Ling C, Goodman C, Roberts C, Li P, Goodall R. Monitoring lateral track irregularity from in-service railway vehicles. Proceedings of the Institution of Mechanical Engineers, Part F: Journal of Rail and Rapid Transit. 2007; 221 :89-100. DOI: 10.1243/0954409JRRT64 - 15.
Malekjafarian A, OBrien E, Quirke P, Bowe C. Railway track monitoring using train measurements: An experimental case study. Applied Sciences. 2019; 9 :4859. DOI: 10.3390/app9224859 - 16.
Bai L, Liu R, Li Q. Data-driven bias correction and defect diagnosis model for In-service vehicle acceleration measurements. Sensors. 2020; 20 :872. DOI: 10.3390/s20030872, 2020 - 17.
Wei X, Liu F, Jia L. Urban rail track condition monitoring based on in-service vehicle acceleration measurements. Measurement. 2016; 80 :217-228. DOI: 10.1016/j.measurement.2015.11.033 - 18.
Balouchi F, Bevan A, Formston R. Development of railway track condition monitoring from multi-train in-service vehicles. Vehicle System Dynamics. 2021; 59 (9):1397-1417. DOI: 10.1080/00423114.2020.1755045 - 19.
Tsunashima H, Naganuma Y, Kobayashi T. Track geometry estimation from car-body vibration. Vehicle System Dynamics. 2014; 52 (sup1):207-219. DOI: 10.1080/00423114.2014.889836 - 20.
Odashima M, Azami S, Naganuma Y, Mori H, Tsunashima H. Track geometry estimation of a conventional railway from car-body acceleration measurement. Mechanical Engineering Journal. 2017; 4 (1) JSME, Paper No.16-00498:1-12. DOI: 10.1299/mej.16-00498 - 21.
Chellaswamy C, Geetha TS, Vanathi A, Venkatachalam K. An IoT based rail track condition monitoring and derailment prevention system. International Journal of RF Technologies. 2020; 11 :81-107. DOI: 10.3233/RFT-190210 - 22.
Rodríguez A, Sanudo S, Miranda M, Gomez A, Benavente J. Smartphones and tablets applications in railways, ride comfort and track quality. Transition Zones Analysis. Measurement. 2021; 182 :1-12. DOI: 10.1016/j.measurement.2021.109644 - 23.
Cong J, Gao M, Miranda M, Wang Y, Chen R, Wang P. Subway rail transit monitoring by built-in sensor platform of smartphone. Frontiers of Information Technology & Electronic Engineering. 2020; 21 (8):1226-1238. ISSN 2095–9184 (print); ISSN 2095-9230 (online) - 24.
Paixão A, Fortunato E, Calçada R. Smartphone’s sensing capabilities for on-board railway track monitoring: Structural performance and geometrical degradation assessment. Advances in Civil Engineering. 2019; 2019 :1-13. Article ID 1729153. DOI: 10.1155/2019/1729153 - 25.
Hsu L. GNSS multipath detection using a machine learning approach. In: 2017 IEEE 20th International Conference on Intelligent Transportation Systems (ITSC). Nagoya, Japan: IEEE ITS Society; 2017 - 26.
Daubechies I. Ten lectures on wavelets. Society for Industrial and Applied Mathematics. Philadelphia. 1992. pp. 1-357. DOI: 10.1137/1.9781611970104 - 27.
Mallat S. A Wavelet Tour of Signal Processing. New York: Academic Press; 1998 - 28.
Le T. Use of the Morlet mother wavelet in the frequency-scale domain decomposition technique for the modal identification of ambient vibration responses. Mechanical Systems and Signal Processing. 2017; 95 :488-505. DOI: 10.1016/j.ymssp.2017.03.045 - 29.
Huang NE, Attoh-Okine NO. Hilbert-Huang Transforms in Engineering. Boca Raton, Florida, USA: CRC Press; 2005. ISBN 9780849334221