GB2536452A

GB2536452A - Methods and systems for alerting a track operator to the likelihood of a fault in a track circuit

Info

Publication number: GB2536452A
Application number: GB1504483.7A
Authority: GB
Inventors: Peter Myers Andrew; Jordan Gary
Original assignee: Thales Holdings UK PLC
Current assignee: Thales Holdings UK PLC
Priority date: 2015-03-17
Filing date: 2015-03-17
Publication date: 2016-09-21
Anticipated expiration: 2035-03-17
Also published as: GB2536452B; GB201504483D0

Abstract

A method for alerting a track operator to the likelihood of a fault in a track circuit associated with a section of railway track. The method comprises receiving data representative of a measured track circuit response S501 as a train passes along the section of railway track and comparing the received data with a model response for the track circuit so as to obtain difference data S502 indicative of the difference between the measured response and the model response. The model response is computed by averaging track circuit responses measured for trains that have previously passed along the section of track. An alert that a fault has occurred or is likely to occur in the track circuit is issued dependent on analysis of the difference data S503. The alert may be issued in the form of an electronic communication such as an email, an aural sound or via a display screen S504.

Description

Methods and systems for alerting a track operator to the likelihood of a fault in a track circuit

FIELD

Embodiments described herein relate to methods and systems for alerting a track operator to the likelihood of a fault in a track circuit associated with a section of railway track.

BACKGROUND

Track circuits (TCs) are electrical devices used to monitor the absence or presence of trains on different sections of railway track. The output from these circuits is used to control signalling on the line and maintain a safe distance between trains travelling along the same line as one another.

The operation of a conventional track circuit can be explained by reference to Figures 1 and 2. As shown in Figure 1, the track circuit includes two circuits; a first circuit 101 and a second circuit 102. The first circuit 101 comprises a power source 103 and electrical connections which run along opposite rails 105 to a relay coil 107. The second circuit 102 includes its own power source 109 that supplies current to one of two indicators 111, 113, typically a red light and a green light. The second circuit includes a switch 115 that is operated by the relay coil and which determines which one of the indicators is powered at any one time.

As shown in Figure 1, in the absence of a train on the section of track, a current is supplied to the relay coil, causing the switch to adopt a first position and allowing current to flow through the first indicator. In this example, the first indicator is a green light, for showing approaching trains that the section of track is unoccupied.

Figure 2 shows the same section of track as a train passes through. Here, the train axle(s) 117 form an electrical connection between the opposite rails, effectively short circuiting the first circuit, such that current no longer flows to the relay coil. The switch 115 in the second circuit consequently switches to its second position, diverting current to the second indicator, in this case the red light 113. A driver of another train approaching the section of track will see the red light and so realise that the track ahead is occupied.

The condition of the track circuit may be assessed using a current monitor 121 to monitor the current flowing through the relay coil 107 as trains pass along the section of track. The current monitor may output data representative of the current passing through the relay coil, which may in turn be forwarded to a track operator, either directly, or over a communications network, for example. If the track circuit is functioning correctly, the current passing through the relay coil should follow a typical response curve, which will be reflected in the data output from the current monitor 121.

In certain circumstances, a track circuit may cease to function reliably. For example, during periods of rainfall, rain water may short-circuit the track circuit, preventing current from reaching the relay switch and so producing the impression that the section of track is currently occupied when there is, in fact, no train present. Conversely, in the event that leaves are present on the track, they may prevent the train from forming a consistent electrical contact with the rails and so failing to short circuit the track circuit. More generally, as components experience wear, the electrical contacts throughout the circuit may cease to stay firmly in contact, causing current to intermittently flow through the circuit and so affecting the reliability of the output signals.

At a certain point, the track circuit may fail entirely, that is it may reach a point at which there is no longer any chance of returning to normal, acceptable levels of performance without intervention from a maintenance crew. Once such failure has occurred, it will be essential to send a maintenance crew to the site as soon as possible.

In deciding whether or not to dispatch a maintenance crew, the risk of a "late call out" (i.e. calling out a maintenance crew after the point at which the track circuit has actually failed) must be balanced against the risk of an "unnecessary call out" i.e. one in which a crew is dispatched to a site that is not in danger of undergoing failure anytime soon. A late call out may be extremely costly, as trains will be prevented from travelling on the line pending the circuit's repair, or at the very least will have severe speed restrictions imposed on them; an unnecessary call out meanwhile will result in a waste of time and resources, as well as exposing maintenance crew to unnecessary risks, since there will always be danger present for men working on the line.

It is desirable, therefore, to limit the dispatch of maintenance crews to times when failure of the track circuit is deemed to be imminent. However, since each track circuit is located in a unique location and has its own characteristic response curve, it is difficult to identify precisely when a track circuit may be on the verge of failure. A train passing over a track circuit in a certain part of the rail network might produce a response curve whose profile indicates that the respective circuit is working within acceptable tolerances; however, were the exact same response obtained for another track circuit in a different part of the rail network, it might be judged as clear evidence that this other track circuit were on the point of failure.

Thus, there is a need to provide improved means for notifying track operators as to when track circuits may be on the point of failure.

SUMMARY

According to a first embodiment, there is provided a method for alerting a track operator to the likelihood of a fault in a track circuit associated with a section of railway track, the method comprising: receiving data representative of a measured track circuit response as a train passes along the section of railway track; comparing the received data with a model response for the track circuit so as to obtain difference data indicative of the difference between the measured response and the model response, the model response being computed by averaging track circuit responses measured for trains that have previously passed along the section of track; and dependent on the difference data, issuing an alert that a fault has occurred or is likely to occur in the track circuit.

The alert may be issued in the event that the difference between the measured track circuit response and the model response is above a threshold.

In some embodiments, the method comprises: receiving, for each one of a plurality of trains, data representative of a measured track circuit response as the respective train passes along the section of railway track; for each train, comparing the received data with the data representing the model response for the track circuit so as to obtain difference data indicative of the difference between the measured response and the model response; and dependent on the difference data obtained for each train, issuing an alert that a fault has occurred or is likely to occur in the track circuit.

The alert may be issued in the event that the difference between the measured track circuit response and the model response is above a threshold for a predetermined number of trains.

In some embodiments, for each train, the difference data is processed to obtain a single output value. The method may then further comprise: averaging the output values obtained for a first group of trains to obtain a first average, wherein the trains in the first group have each passed along the section of railway track during a first period of time; and averaging the output values obtained from a second group of trains to obtain a second average, wherein the second group of trains includes a larger number of trains than the first group and at least some of the trains in the second group are ones that passed along the section of railway track at an earlier point in time than any of the trains in the first group. The alert may be issued dependent on a comparison of the first average value and second average value.

In some embodiments, for each train, the received data is indicative of a change in current through a part of the track circuit as the train passes along the section of track. For each train, the received data may represent a measure of the current at different time points as the train passes along the section of track.

In some embodiments, for each train, the data representative of the measured track response is split into a plurality of time periods, each period reflecting a different phase in the train's passage through the track section. The data representative of the measured track response may be compared with the model response by comparing the data in each time period with a corresponding portion of the model response curve.

The alert may be issued by sending an electronic communication to the track operator. The electronic communication message may be sent as an email. In some embodiments, the alert is issued as an aural sound. In some embodiments, the alert is issued by updating a display on a track operator's or flight engineer's display screen.

According to a second embodiment, there is provided a computer readable storage medium comprising computer executable instructions that when executed by a computer will cause the computer to carry out a method according to the first embodiment.

BRIEF DESCRIPTION OF FIGURES

Embodiments of the invention will now be described by way of example with reference to the accompanying drawings in which: Figure 1 shows the configuration of a track circuit at a point in time in which there are no trains present on the section of track; Figure 2 shows the configuration of a track circuit at a point in time in which a train is present on the section of track; Figure 3 shows an example response curve for a track circuit as a train passes along the section of track: Figure 4 shows an example of a response curve for a train being superimposed on a model response curve for the track circuit; Figure 5 shows a flow chart of steps used in a method according to an embodiment; Figure 6 shows a flow chart of steps used in a method according to an embodiment; Figure 7 shows an example of how a model response curve may be generated for a track circuit in an embodiment; Figure 8 shows a flow chart of steps used in a method according to an embodiment; Figure 9 shows a flow chart of steps used in a method according to an embodiment; Figure 10 shows a flow chart of steps used in a method according to an embodiment; Figure 11 shows a flow chart of steps used in a method according to an embodiment; and Figure 12 shows an example of trends in the long term filter output and short term filter output according to an embodiment.

DETAILED DESCRIPTION

Figure 3 shows an example response curve for a track circuit such as that shown in Figure 1 as a train passes along the section of track. The response curve indicates the current passing through the relay coil and can be divided into 4 sections: falling, occupied level, rising edge and clear level. In the "dear" section, the current passing through the coil is at a maximum, reflecting the absence of a train on the section of track. In the "occupied" period, the current passing through the coil falls to a minimum, as the train axels short the first circuit between the rails (as shown in Figure 2, for example). The "rising" and "falling" sections reflect the transition between clear and occupied states of the track.

In general, assuming that the track circuit is functioning correctly, the response curves for successive trains will be broadly similar, each curve having the same U-shaped profile as the current is first blocked from flowing through the relay coil and then recovering as the train moves out of the track section and ceases to short the first circuit. The precise amount of current that is detected flowing through the coil will still vary between trains.

Embodiments described herein seek to monitor the response curves seen for successive trains in order to build up a model response curve for the track circuit. The model response curve provides an expected curve profile for subsequent trains passing along the track; that is, in the event that the track circuit is functioning correctly, a train passing along the track section will be expected to generate a response curve having a curve profile similar to that of the model response. A deviation from the model response may be an indication that the track circuit is no longer functioning correctly, having developed a fault, for example.

Figure 4 shows an example of a response curve for a train superimposed on a model response curve for the track circuit in question; here, the shaded area 401 illustrates the difference between the current levels at each time point in the two curves.

Figure 5 shows an example sequence of steps according to an embodiment. In a first step S501, data representative of the measured track circuit response are received. The data may comprise, for example, current readings from the relay coil switch, or more generally any data indicative of a flow of current through part of the track circuit. In step S502, the received data is compared with model response data for the track circuit in question, where the model response data represents an expected current profile for the track circuit. Based on the comparison, a decision is made as to whether issue an alert (step S503). In the event that the decision is positive, the alert is issued (step S504), thereby indicating the presence of a fault in the track circuit.

The alert may take one of a number of different forms; the alert may be issued as an aural alarm (e.g. a buzzer or siren), a visual alert displayed on a user's monitor screen (the user may be, for example, a flight engineer with responsibility for deciding whether or not to dispatch a maintenance crew to a particular site). The alert may take the form of a flashing light or a pop-up on the user's screen, for example. In some embodiments, issuing an alert may comprise displaying a message or text on the screen, or saving that message or text to a memory file together with other such messages, each one of which may then be displayed on the screen when the memory file is opened. The alert may also take the form of an electronic message being delivered to a person's personal account, such as by email. In the event that the decision in step S503 is negative, no alert is issued and steps S501 -S503 are repeated for the next train to pass along the track section.

Figure 6 shows an example of how the model response data may be computed in an embodiment. In step S601, data representative of the measured track circuit response for a train are received. As discussed above, the data may comprise, for example, current readings from the relay coil switch, or more generally any data indicative of a flow of current through part of the track circuit. Once received, the data is stored in memory. In step S602, a decision is made as to whether or not data has been received for a predetermined number of trains. In the event that data has yet to be received for the predetermined number of trains, no further action is taken until the next train passes along the track section, at which point steps S601 and S602 are repeated. Once data has been received for a predetermined number of trains, the model response data is computed by averaging the response curves obtained for each of those trains (Step S603). Thus, the model response data represents an average response profile for trains passing along the section of track in question.

In some embodiments, the method may comprise splitting the data in each response curve into four component parts that represent the falling phase, occupied phase, rising phase and clear phase of each curve. That is, the response curve for each respective train may be broken down into 4 data sets, where those data sets represent, respectively, the period during which the current signal is at its highest (clear), the period in which the current is below a threshold and the track section is deemed to be occupied, the period during which the current is falling from the clear state to the occupied state, and the period during which the current is rising from the occupied state to the clear state. Then, when computing the model response data, each one of the 4 data sets in each curve may be analysed independently. This process is illustrated pictorially in Figure 7, which shows response curves A, B and C obtained for 3 separate trains. The curve D in turn represents the model response curve obtained by averaging the data from the 3 trains A, B and C. As shown on the right hand side of Figure 7, each response curve is separated into 4 separate data sets that represent the falling, occupied, rising and clear parts of the curve profile, respectively.

The data sets for the falling component are averaged together to return a model response W for the falling part of the curve and the data sets for the occupied, rising and clear components are similarly averaged to return model responses X, Y and Z for those respective parts of the curve.

In the present embodiment, it will be understood that a predetermined number of trains must pass along the track before the steps of comparing received response data with model response data (step S502) and deciding whether or not to issue an alert (step S503) can be carried out; that is, it is only once a number of track response curves have been obtained from trains passing along the track section that it will be possible to generate the model response data, which can then be compared with response curves obtained for subsequent trains passing along the track.

Figure 8 shows an example of an embodiment in which an alert is issued in the event that the difference between the received data and model response data is above a threshold.

The threshold may be set manually and changed dependent on conditions on the track, the age of the track circuit, weather conditions etc. The difference may be taken as the difference between the received response curve and the model response curve at a single point on the time axis, or the difference summed over one of the four sections (clear, occupied, rising and falling) of the curve. The difference may also be taken as the sum difference across the entire curve, for example.

It will be understood that where the difference between the measured response curve and the model response curve is taken across a portion of the curve, rather than at a single point on the curve, the difference may be taken as the sum of the differences above and beneath the model response curve; that is, any deviation from the model response curve (be it above or beneath the curve) will make a positive contribution to the value that is output as the difference between the two curves. In alternative embodiments, the output value may be calculated by summing the difference between the two curves in portions of the curves for which the measured response exceeds the model response and subtracting the difference between the two curves for portions in which the measured response is smaller than the model response. In this way, in the event that the values in the measured response curve are consistently above the model response curve, the difference value that is output will be greater than a case in which the measured response curve runs close to the model response but rises above and below the model response at different times.

Figure 9 shows an example of an embodiment in which an alert is issued in the event that the difference data for a predetermined number of trains is observed to be above a threshold, across a certain period of time. In this embodiment, the frequency with which trains are seen to generate anomalous response curves (i.e. response curves for which the difference between the response curve and the model response curve is above a threshold) determines whether or not the alert is issued.

It will be understood that features of the embodiments shown in Figures 8 and 9 may be combined. For example, referring to step S903, in the event that the number of trains for which the difference value is above an upper limit has yet to reach the number required to issue the alert, but an individual response curve is measured having a particularly large difference value that exceeds a threshold (as in Figure 8), the alert may still be issued. Thus, the decision to issue an alert may be based on whether or not the number of trains for which the difference between the measured response curve and the model response curve is above a limit has reached a threshold and / or whether or a particularly anomalous response curve has been measured for an individual train.

Figure 10 shows an example of an embodiment in which difference data is compared with difference data obtained from earlier trains passing through the track section (step S503). In this embodiment, trends in the state of the track circuit can be observed. As trains pass along the section of track, the received data is compared with the model response and the difference between the curve and the model data calculated for each of the 4 parts of the curve. This gives rise to a difference plot for each section; these differences may in turn be amalgamated to give a single output for the current state of the track circuit i.e. a large difference error in any one section will result in a lower confidence that the track circuit is If the track circuit is functioning correctly, then the difference data obtained for successive trains should remain fairly constant (in other words, the difference between the profile curves obtained for successive trains should stay within certain limits). Conversely, if the circuit has a fault and is heading towards failure, this may be reflected in a growing number of anomalous readings, or alternatively, a trend towards larger or smaller current values being monitored in different parts of the response curves.

Figure 11 shows an example of how the comparison step of Figure 10 may be implemented in practice. As discussed above, the difference data associated with each train is processed to provide a single respective output value, which is stored in memory. A comparison of the values is made by filtering those values into long and short term filters, where those filters are implemented by averaging the values over different numbers of trains. For example, the short term filter may be used to generate a first value by averaging the values for the difference data over the last 10 trains to pass through the track section, whilst the long term filter is used to generate a second value by averaging the values for the difference data over the last 100 trains to pass through the track section. (It will be understood that the actual number of trains over which the data is averaged can vary for both the short term filter and long term filter; the prime consideration is that the number of trains that are averaged will be greater in the case of the long term filter than the short term filter). Should the behaviour of the track circuit change, the short term average will show the result sooner than the long term average. Hence, the change in behaviour can be detected by comparing the outputs of the two filters. In one example, an alert may be issued in the event that the difference between the short term average and the long term average rises above a threshold. An example of trends in long term filter output 1201 and short term filter output 1202 is shown in Figure 12.

Embodiments can help to provide more accurate predictions of failure of the track circuit, in the event that a fault in the circuit is developing over time. Embodiments offer better diagnosis capability, prediction capability for slow varying faults, automatic diagnosis capability and open up the possibility for automatic calibration.

While certain embodiments have been described, these embodiments have been presented by way of example only, and are not intended to limit the scope of the invention. Indeed, the novel methods, devices and systems described herein may be embodied in a variety of forms; furthermore, various omissions, substitutions and changes in the form of the methods and systems described herein may be made without departing from the spirit of the invention. The accompanying claims and their equivalents are intended to cover such forms or modifications as would fall within the scope and spirit of the inventions.

Claims

Claims 1. A method for alerting a track operator to the likelihood of a fault in a track circuit associated with a section of railway track, the method comprising: receiving data representative of a measured track circuit response as a train passes along the section of railway track; comparing the received data with a model response for the track circuit so as to obtain difference data indicative of the difference between the measured response and the model response, the model response being computed by averaging track circuit responses measured for trains that have previously passed along the section of track; and dependent on the difference data, issuing an alert that a fault has occurred or is likely to occur in the track circuit.
2. A method according to claim 1, wherein the alert is issued in the event that the difference between the measured track circuit response and the model response is above a threshold.
3. A method according to claim 1, comprising: receiving, for each one of a plurality of trains, data representative of a measured track circuit response as the respective train passes along the section of railway track; for each train, comparing the received data with the data representing the model response for the track circuit so as to obtain difference data indicative of the difference between the measured response and the model response; and dependent on the difference data obtained for each train, issuing an alert that a fault has occurred or is likely to occur in the track circuit.
4. A method according to claim 3, wherein the alert is issued in the event that the difference between the measured track circuit response and the model response is above a threshold for a predetermined number of trains.
5. A method according to claim 3 or 4, wherein for each train, the difference data is processed to obtain a single output value, and wherein the method comprises: averaging the output values obtained for a first group of trains to obtain a first average, wherein the trains in the first group have each passed along the section of railway track during a first period of time; and averaging the output values obtained from a second group of trains to obtain a second average, wherein the second group of trains includes a larger number of trains than the first group and at least some of the trains in the second group are ones that passed along the section of railway track at an earlier point in time than any of the trains in the first group; wherein the alert is issued dependent on a comparison of the first average value and second average value.
6. A method according to any one of the preceding claims, wherein for each train, the received data is indicative of a change in current through a part of the track circuit as the train passes along the section of track.
7. A method according to claim 6, wherein for each train, the received data represents a measure of the current at different time points as the train passes along the section of track.
8. A method according to claim 3, wherein for each train, the data representative of the measured track response is split into a plurality of time periods, each period reflecting a different phase in the train's passage through the track section; wherein the data representative of the measured track response is compared with the model response by comparing the data in each time period with a corresponding portion of the model response curve.
9. A method according to any one of the preceding claims, wherein the alert is issued by sending an electronic communication to the track operator.
10. A method according to claim 9, wherein the electronic communication message is sent as an email.
11. A method according to any one of claims 1 to 8, wherein the alert is issued as an aural sound.
12. A method according to any one of claims 1 to 8, wherein the alert is issued by updating a display on the track operators display screen.
13. A computer readable storage medium comprising computer executable instructions that when executed by a computer will cause the computer to carry out a method according to any one of the preceding claims.Amendments to the Claims have been filed as follows:-Claims 1. A method for alerting a track operator to the fact that a fault has occurred or is likely to occur in a track circuit associated with a section of railway track, the method comprising: receiving data representative of a measured track circuit response as a train passes along the section of railway track; comparing the received data with a model response for the track circuit so as to obtain difference data indicative of the difference between the measured response and the model response, the model response being computed by averaging track circuit responses measured for trains that have previously passed along the section of track; and dependent on the difference data, issuing an alert that a fault has occurred or is likely to occur in the track circuit.2. A method according to claim 1, wherein the alert is issued in the event that the difference between the measured track circuit response and the model response is above a threshold.3. A method according to claim 1, comprising: receiving, for each one of a plurality of trains, data representative of a measured track circuit response as the respective train passes along the section of railway track; 1-20 for each train, comparing the received data with the data representing the model response for the track circuit so as to obtain difference data indicative of the difference between the measured response and the model response; and dependent on the difference data obtained for each train, issuing an alert that a fault has occurred or is likely to occur in the track circuit.4. A method according to claim 3, wherein the alert is issued in the event that the difference between the measured track circuit response and the model response is above a threshold for a predetermined number of trains.5. A method according to claim 3 or 4, wherein for each train, the difference data is processed to obtain a single output value, and wherein the method comprises: averaging the output values obtained for a first group of trains to obtain a first average, wherein the trains in the first group have each passed along the section of railway track during a first period of time; and averaging the output values obtained from a second group of trains to obtain a second average, wherein the second group of trains includes a larger number of trains than the first group and at least some of the trains in the second group are ones that passed along the section of railway track at an earlier point in time than any of the trains in the first group; wherein the alert is issued dependent on a comparison of the first average value and second average value.6. A method according to any one of the preceding claims, wherein for each train, the received data is indicative of a change in current through a part of the track circuit as the train passes along the section of track.7. A method according to claim 6, wherein for each train, the received data represents a measure of the current at different time points as the train passes along the section of track.8. A method according to claim 3, wherein for each train, the data representative of the measured track response is split into a plurality of time periods, each period reflecting a different phase in the train's passage through the track section; wherein the data representative of the measured track response is compared with the model response by comparing the data in each time period with a corresponding portion of the model response.1-20 9. A method according to any one of the preceding claims, wherein the alert is issued by sending an electronic communication to the track operator.O10. A method according to claim 9, wherein the electronic communication message is sent as an email.11. A method according to any one of claims 1 to 8, wherein the alert is issued as an aural sound.12. A method according to any one of claims 1 to 8, wherein the alert is issued by updating a display on the track operator's display screen.13. A computer readable storage medium comprising computer executable instructions that when executed by a computer will cause the computer to carry out a method according to any one of the preceding claims.