The present invention provides an apparatus for monitoring railway track. The present invention further provides a method for monitoring railway track.

Rail track maintenance is a critical part of maintaining traffic density and speed in the rail network. Unforeseen track maintenance, frequent track maintenance or incorrect track maintenance can all conspire to reduce safety and capacity of the network, and can all be caused by a lack of in-depth information about key parameters describing the state of a piece of track. Information that drives track maintenance operations typically comes from special measurement trains. These are expensive pieces of equipment that measure, amongst other parameters, track geometry. They run at limited speeds and follow individually defined routes, so inevitably disrupt passenger services. The information from the measurement train is generally interpreted very strictly, driving maintenance operations that are chosen to correct geometry defects, but that might not correct long term underlying faults that reduce stability of the track or track support.

One of the parameters that is not measured by measurement trains is track stiffness. This is the resilience of the track as a train passes over it, and typically results in a vertical movement of the track of 2-3mm as a train passes over the track. This resilience determines how much energy is put pack into the train wheel, which affects wheel wear, and how much noise is dissipated into the ground, which affects ballast stability and local populations living near the track. Low track stiffness can also contribute to rail breaks. Track stiffness is, therefore, an important track property and is usually measured by pushing a mass down onto the track and measuring the deflection, which is a very time-consuming test, or can be measured by observing track movement from trackside equipment.

There is therefore a need in the art for an apparatus for, and a method of, monitoring track stiffness which can provide an improved measurement as compared to known apparatus and methods.

There is also a need in the art for an apparatus for, and a method of, monitoring track condition, in particular track stiffness, in-service, preferably in real-time.

There is a further need in the art for an apparatus for, and a method of, monitoring track stiffness which can monitor track condition to provide highly accurate measurement and analysis during an in-service test, preferably in real-time. The present invention at least partially aims to meet one or more of these needs.

Accordingly, the present invention provides an for monitoring railway track, the apparatus comprising a wireless sensor node fitted to an axle assembly of a railway vehicle, the wireless sensor node comprising a vibration energy harvester for converting mechanical energy from vibration in the axle assembly into electrical energy, a sensor for measuring a parameter, and a wireless transmitter for wirelessly transmitting the measured parameter or data associated therewith, and the apparatus further comprising a processor for processing the measured parameter to produce processed data, wherein the sensor is mounted to the axle assembly and is arranged to measure vibration in the axle assembly over a period of time to produce a vibration-time signal which varies with a periodicity corresponding to a spacing of sleepers along the railway track, and the processor is arranged to process the vibration-time signal to determine a track stiffness parameter from the measured vibration, wherein the processor comprises a periodicity module which is arranged to process the vibration-time signal to determine the presence of periodic alternating first and second signal portions respectively corresponding to railway track over a sleeper and railway track between adjacent sleepers.

The present invention further provides a method of monitoring railway track, the method comprising the steps of:

a. providing a wireless sensor node fitted to an axle assembly of a railway vehicle, the wireless sensor node comprising a vibration energy harvester for converting mechanical energy from vibration in the axle assembly into electrical energy, a sensor for measuring a parameter, wherein the sensor is mounted to the axle assembly, and a wireless transmitter for wirelessly transmitting the measured parameter or data associated therewith;

b. while the railway vehicle is in motion, the vibration energy harvester receiving input vibration energy which is converted into electrical energy to power the wireless transmitter;

c. while the railway vehicle is in motion, measuring, using the sensor, vibration in the axle assembly over a period of time to produce a vibration-time signal which varies with a periodicity corresponding to a spacing of sleepers along the railway track;

d. while the railway vehicle is in motion, wirelessly transmitting the vibration-time signal or data associated therewith using the wireless transmitter; and e. processing the vibration-time signal to determine a track stiffness parameter from the measured vibration, wherein the vibration-time signal is processed to determine the presence of periodic alternating first and second signal portions respectively corresponding to railway track over a sleeper and railway track between adjacent sleepers.

Preferred features of the apparatus and method of the present invention are defined in the respective dependent claims.

The apparatus and method of the preferred embodiments of the present invention solve the problem of providing an in-service track condition measurement, optionally in real-time, which can be utilized in a protocol for measuring track stiffness, in particular track stiffness between adjacent sleepers.

The apparatus and method of the preferred embodiments of the present invention are predicated on the finding by the present inventors that the railway track exhibits enhanced deflection between railway track sleepers. The sleepers are arranged periodically along the track, at accurately spaced locations. The vibration exhibited by the axle can be measured and analysed to determine a corresponding periodicity resulting from the vibration sensor passing regularly and intermittently over track portions that are between sleepers and exhibit enhanced deflection as compared to track mounted above sleepers. This invention solves this problem of accurate, efficient and effective measurement of track condition, in particular track stiffness, by measuring sleeper deflection from equipment mounted on trains in service, exploiting the periodicity of sleepers. The measurement and analysis may also accommodate the different axle loads on a passenger train corresponding to trailer/motorised axles and heavier or lighter coach bodies.

In preferred embodiments of the present invention, the processor is provided in a data server and analytics engine, which receives vibration data from trains and processes the data to extract track stiffness information. Typically, the invention is put into practice using a fleet of instrumented trains, with axle box vibration monitors and the facility to apply location and timestamp data before the vibration data is transmitted from the train. The preferred embodiments of the present invention utilise vibration energy harvester powered wireless sensor nodes (WSNs), fitted to all axle ends of the train. The wireless sensor nodes can typically be configured to measure, in addition to other parameters, the component of vibration generated by sleeper periodicity, since the rail bends between each sleeper as a result of the mass of the passing train urging the track downwardly as a result of the track having freedom of vertical movement between adjacent sleepers.

The bending deflection of the rail between each sleeper depends partly on track stiffness, and since this is the only parameter that varies during the course of a journey, monitoring the variation of this rail deflection, and aggregating the information across all sensors, which can improve the signal-to-noise ratio and location accuracy, can provides a measure of track stiffness across the track network.

Although wireless sensor nodes (WSNs) are known in the art for use in measuring axle vibration, the present invention employs the innovation of using a component of vibration generated in the wheel-rail interaction that is sensitive to track stiffness, and that can be extracted from noise caused by wheel and track condition. Tracking the periodic vibration provides additional sensitivity and specificity, and has not previously been disclosed as a method for measuring track stiffness. Other known methods for measuring track stiffness rely on individual, lengthy direct measurement that require specific equipment.

The preferred embodiments of the present invention provides a particular advantage by using signal processing using measurement of real time vibration data which is coherent with sleeper spacing vibration on in-service trains.

In the preferred embodiments of the present invention, historical data can be analysed and provided to generate a reference signal for comparison with each cycle of vibration, for example by multiplying the reference with the signal measured. In some embodiments, the output of that reference signal analysis can be integrated over time for an enhanced accuracy of measurement which is averaged over a given length of track. Additionally, reference data associated with individual sleepers can be utilised to investigate and identify individual problems with specifically identified sleepers.

In the preferred embodiments of the present invention, vibration produced by track movement between sleepers, which appears in the vibration spectrum as a steady speed dependent component, can be measured and tracked using Fast Fourier Transform (FFT) analytical techniques, or alternatively analytical techniques other than the Fast Fourier Transform, which is computationally intensive and slow, since the analysis effectively averages signals over many sleepers. In some embodiments of the present invention, a technique similar to phase sensitive detection may be used, in which historical vibration data can be examined to generate a reference vibration, for example an in-phase sine wave, for comparison with the speed/sleeper vibration component at each sleeper location, thus providing a measure of individual sleeper support. It is known for example that hanging sleepers are another cause of rail breakage or failure. In each case, close examination of vibration measured at the wheel axle, i.e. at the wheel hub, in particular the component of vibration dependent on sleeper periodicity, is examined closely for changes dues to variations in track stiffness. Other signal processing techniques, digital or analogue, could be used to monitor small variations in a narrowband signal in real time.

Embodiments of the present invention will now be described, by way of example only, with reference to the accompanying drawings, in which:

Figure 1 is a schematic end view of an apparatus for monitoring railway track according to an embodiment of the present invention;

Figure 2 is a schematic view of the processing system in the apparatus of Figure 1;

Figure 3 schematically illustrates an operation mode of the processor to process input vibration from the output of the sensor in the apparatus of Figure 1;

Figure 4 schematically illustrates another operation mode of the processor to process input vibration from the output of the sensor in the apparatus of Figure 1 ;

Figure 5 schematically illustrates another operation mode of the processor to process input vibration from the output of the sensor in the apparatus of Figure 1;

Figure 6 schematically illustrates another operation mode of the processor to process input vibration from the output of the sensor in the apparatus of Figure 1 ; and

Figure 7 schematically illustrates another operation mode of the processor to process input vibration from the output of the sensor in the apparatus of Figure 1.

Referring to Figures 1 and 2, there is shown an apparatus 2 for monitoring railway track 4. The apparatus 2 comprises a wireless sensor node 6 fitted to an axle assembly 8 of a railway vehicle 10. The railway vehicle 10 may be a locomotive, a passenger carriage or a freight car or truck.

The wireless sensor node 6 comprises a vibration energy harvester 12 for converting mechanical energy from vibration in the axle assembly 8 into electrical energy. A sensor 14 is provided for measuring a parameter, and the sensor 14 is mounted to an axle box assembly 30 at an end 32 of the axle assembly 8. A wireless transmitter 16 is provided for wirelessly transmitting the measured parameter or data associated therewith to a remote location for further processing and/or analysis; the remote location may be within the railway vehicle 10, or within a locomotive or other vehicle of a train which includes the railway vehicle 10. Typically, each axle assembly 8 within a train is provided with a monitoring apparatus as described herein and the apparatus 2 comprises a plurality of the wireless sensor nodes 6, each wireless sensor node 6 being fitted to a respective axle assembly 8 of the railway vehicle 10.

The apparatus 2 further comprises a processor 18 for processing the measured parameter to produce processed data. In the illustrated embodiment, the processor 18 is integral with the wireless sensor node 6, and the wireless transmitter 16 is arranged wirelessly to transmit the processed data. However, in alternative embodiments, the processor 18 is remote from the wireless sensor node 6, and the wireless transmitter 16 is arranged wirelessly to transmit the measured data which is then remotely processed by the processer 18 to produce the processed data.

The sensor 14 is mounted to the axle assembly 8 and is arranged to measure vibration in the axle assembly 8 over a period of time to produce a vibration-time signal which varies with a periodicity corresponding to the spacing of sleepers 20 along the railway track 4.

The processor 18 is arranged to process the vibration-time signal to determine a track condition parameter, which is a track stiffness parameter, and optionally additionally a sleeper condition parameter, from the measured vibration.

In the preferred embodiments of the present invention, the apparatus 2 further includes a speed correlation module 48 in the processor 18 to correlate the vibration-time signal against a measured speed of the railway vehicle 8. The time period between sensing adjacent sleepers is speed dependent, and the correlation can provide a signal which is independent of speed.

In the preferred embodiments of the present invention, the processor 18 comprises a periodicity module 22 which is arranged to process the vibration-time signal to determine the presence of periodic alternating first and second signal portions respectively corresponding to railway track over a sleeper and railway track between adjacent sleepers. The processor 18 further comprises a track condition analyser module 24 which is arranged to determine the track condition parameter from analysis of the second signal portions. The processor 18 is adapted to determine a track stiffness parameter as the track condition parameter, and optionally a sleeper condition parameter as the track condition parameter. In the preferred embodiments of the present invention, the apparatus 2 further comprises a time stamp module 26 which associates a time stamp parameter with the vibration-time signal and a geographic location module 28 which associates a geographic location parameter with the vibration-time signal. Therefore any given vibration-time signal, which can be used to determine a track stiffness parameter, and optionally a sleeper condition parameter, can be indexed or tagged by a timestamp and/or geographic location.

In the preferred embodiments of the present invention, the processor 18 includes a baseline noise signal module 34 which is adapted to store a predetermined baseline noise vibration signal. This signal can represent background noise from wheel and axle motion, and motion of the wheelset, including the axle and opposed wheels fitted thereto, over the track. A baseline comparator module 36 compares a current vibration signal against the predetermined baseline noise vibration signal and a calibrated analyser module 38 determines a calibrated parameter of the current vibration-time signal from the comparison, the calibrated parameter comprising part of the processed data. These modules enable noise to be filtered from the periodic signal so that the vibration-time signal can more readily be analysed to distinguish a track stiffness parameter, and optionally a sleeper condition parameter.

In the preferred embodiments of the present invention, the processor 18 includes a reference signal module 40 which is adapted to store a predetermined reference vibration-time signal which varies with a periodicity corresponding to the spacing of sleepers along the railway track, a reference comparator module 42 to compare a current vibration-time signal against the predetermined reference vibration-time signal and a reference analyser module 44 to determine a referenced parameter of the current vibration-time signal from the comparison, the referenced parameter comprising part of the processed data. These modules enable previous measurements of the railway track 4 to be stored and then used as reference data, including periodic vibrations resulting from the previously detected differences in track stiffness of railway track 4 above a sleeper 20 and railway track 4 between sleepers 20. In the preferred embodiments of the present invention, the wireless sensor node 6 is adapted to be operated continuously over a monitoring period thereby continuously to measure the vibration-time signal and continuously to compare the current axle vibration-time signal at any given time against the predetermined reference vibration-time signal. In the preferred embodiments of the present invention, the processor 18 includes an integration module 46 which combines a plurality of the vibration- time signals from previous monitoring operations to provide the predetermined reference vibration-time signal. The predetermined reference vibration-time signal is typically associated with a known length of railway track 4 and/or a known sleeper 20 along a length of railway track 4. The current periodic signal can be compared against the reference data so that the vibration-time signal can more readily be analysed to distinguish a track stiffness parameter, and optionally a sleeper condition parameter.

The apparatus 2 is used in a method of monitoring railway track 4. In the method, a wireless sensor node 6 as described above is fitted to the axle assembly 8 and preferably each axle assembly 8 within a train is provided with the monitoring apparatus 2 as described herein and the apparatus 2 comprises a plurality of the wireless sensor nodes 6, each wireless sensor node 6 being fitted to a respective axle assembly 8 of the railway vehicle 10.

While the railway vehicle 10 is in motion, the vibration energy harvester 12 receives input vibration energy which is converted into electrical energy to power the wireless transmitter 16. When the processor 18 is integrated into the wireless sensor node 6 the vibration energy harvester 12 can provide the electrical energy to operate the processor 18. The vibration energy harvester 12 can provide the electrical energy to operate any other powered components of the wireless sensor node 6.

Also while the railway vehicle 10 is in motion, the sensor 14 is used to measure, vibration in the axle assembly 8 over a period of time to produce a vibration-time signal which varies with a periodicity corresponding to the spacing of sleepers 20 along the railway track 4.

While the railway vehicle 10 is in motion, the vibration-time signal or data associated therewith is wirelessly transmitted using the wireless transmitter 16, as described above.

The measured vibration-time signal is processed by the processor 18, which is either integral to, or remote from, or has components that are integral to or remote from, the wireless sensor node 6, to produce processed data which includes a track condition parameter, and thereby determines a track condition parameter from the measured vibration. Typically, the processing step comprises the sub-step of correlating the vibration-time signal against a measured speed of the railway vehicle. As described above, the time period between sensing adjacent sleepers 20 is speed dependent, and the correlation can provide a signal which is independent of speed.

When the processed data is processed by the processor 18 which is integral with the wireless sensor node 6, the processed data is wirelessly transmitted. The processing by the processor 18 is preferably carried out in real-time simultaneously with the measuring and transmitting steps to measure the vibration-time signal and transmit the vibration-time signal or data associated therewith.

In the processing step, the vibration-time signal is processed to determine the presence of periodic alternating first and second signal portions respectively corresponding to railway track 4 over a sleeper 20 and railway track 4 between adjacent sleepers 20. The track condition parameter may be determined from analysis of the second signal portions. In the processing step, a track stiffness parameter is determined as the track condition parameter, and optionally a sleeper condition parameter is also determined as the track condition parameter.

In the preferred embodiments of the present invention, a time stamp parameter is associated with the vibration-time signal and/or a geographic location parameter is associated with the vibration-time signal. As described above, this provides the advantage that any given vibration- time signal, which can be used to determine a track stiffness parameter, and optionally a sleeper condition parameter, can be indexed or tagged by a timestamp and/or geographic location.

In the preferred embodiments of the present invention, the processing step includes the sub- steps of (i) storing predetermined baseline noise vibration signal, (ii) comparing a current vibration signal against the predetermined baseline noise vibration signal, and (iii) determining a calibrated parameter of the current vibration-time signal from the comparison, the calibrated parameter comprising part of the track condition parameter. As described above, this provides the advantage that noise can be filtered from the periodic signal so that the vibration-time signal can more readily be analysed to distinguish a track stiffness parameter, and optionally a sleeper condition parameter.

In the preferred embodiments of the present invention, the processing step (e) includes the sub- steps of (I) storing a predetermined reference vibration-time signal which varies with a periodicity corresponding to the spacing of sleepers along the railway track, (II) comparing a current vibration-time signal against the predetermined reference vibration-time signal, (III) determining a referenced parameter of the current vibration-time signal from the comparison, the referenced parameter comprising part of the track condition parameter. Typically, the wireless sensor node 6 is operated continuously over a monitoring period thereby continuously to measure the vibration-time signal and in step (II) the current axle vibration-time signal at any given time is continuously compared against the predetermined reference vibration-time signal. Typically, in the method there is a step of (A) integrally combining a plurality of the vibration-time signals from previous monitoring operations to provide the predetermined reference vibration-time signal. The predetermined reference vibration-time signal may be associated with a known length of railway track 4 and/or a known sleeper 20 along a length of railway track 4. As described above, the use of reference data provides the advantage that current periodic signal can be compared against the reference data so that the vibration-time signal can more readily be analysed to distinguish a track stiffness parameter, and optionally a sleeper condition parameter.

Figure 3 schematically illustrates an operation mode of the processor to process input vibration from the output of the sensor in the apparatus of Figure 1. The vibration input from sensor 14 is provided as current vibration data to the reference comparator module 42 of the processor 18. A dataset of historical track vibration data is provided to the reference signal module 40, which provides a sleeper spacing derived speed dependent reference signal to the reference comparator module 42. The reference comparator module 42 compares the current vibration signal against the historical vibration data and the output is passed to the reference analyser module 44 to determine a referenced calibrated parameter of the current vibration-time signal.

Figure 4 schematically illustrates an operation mode of the processor to process input vibration from the output of the sensor in the apparatus of Figure 1. The vibration input from sensor 14 is provided as current vibration data to the periodicity module 22 which processes the signal to determine the presence of periodic first and second signal portions as described above. The output is sent to the track condition analyser 24 which determines the track condition parameter which is output as a signal to the reference analyser module 44.

Figure 5 schematically illustrates an operation mode of the processor to process input vibration from the output of the sensor in the apparatus of Figure 1. The vibration input from sensor 14 is provided as current vibration data to the speed correlation module 48 of the processor 18. A measured speed signal is provided by a speed sensor (not shown) to the speed correlation module 48 which correlates the input vibration-time signal against the measured speed. The baseline noise signal module 34 may optionally output a baseline noise vibration signal to the sped correlation module to calibrate the correlated signal against speed and a baseline noise. The speed correlation module 48 outputs a speed correlated signal to the reference analyser module 44.

Figure 6 schematically illustrates an operation mode of the processor to process input vibration from the output of the sensor in the apparatus of Figure 1. Current vibration data to be processed is sent to the reference analyser module 44. The baseline noise signal module 34 receives baseline noise vibration. The outputs of the reference analyser module 44 and the baseline noise signal module 34 are received by the baseline comparator module which compares the current vibration data against the predetermined baseline noise vibration signal and the output is sent to the calibrated analyser module which outputs a data signal calibrated, and thereby compensated, against baseline noise.

Figure 7 schematically illustrates an operation mode of the processor to process input vibration from the output of the sensor in the apparatus of Figure 1. Plural vibration data inputs from different sensors 14 or different vibration measurements, each measuring at a common location, are provided to the integration module 46 of the processor 18. The integration module 46 outputs vibration data which corresponds to the common location and is noise-reduced as a result of cumulatively processing plural signals from the same location.

Various modifications to the preferred embodiments of the present invention will be apparent to those skilled in the art.