METHOD AND ARRANGEMENT FOR FAULT DETECTION IN A SWITCH SYSTEM

Abstract

For fault detection in a switch system including multiple switches, for a respective switch its operating data are acquired and transmitted to a central control. For a respective first switch of the switch system, a deviation of an operating behavior of the first switch from a set behavior is determined by the central control on the basis of its operating data. If there is a deviation, operating data of the first switch are compared with operating data of other, second switches of the switch system and, depending on that, a second switch with similar operating data is selected.

Description

CROSS-REFERENCE TO RELATED APPLICATION

This application claims priority to European application No. 18170074.1, having a filing date of Apr. 30, 2018 the entire contents both of which are hereby incorporated by reference.

FIELD OF TECHNOLOGY

The following relates to a method and arrangement for fault detection in a switch system.

BACKGROUND

For track-bound modes of transport or means of transport, switches are generally used for changing over from one track to another. This is the case for example for rail-bound means of transport such as railroads and streetcars and also for switchbacks, trolley coaches or pneumatic mail systems.

The switches are central elements of the transport route, which can greatly disrupt an operating sequence if there are malfunctions. It is consequently desirable that malfunctions and other faults in the switch system can be detected as precisely as possible and preferably can be predicted. An identification of a respective cause of a fault would likewise be very advantageous.

It is known to monitor switches and their operating mechanisms by means of assigned sensors and to signal a fault when there are deviations in operating behavior from a set behavior. However, predicting faults and/or determining their causes by means of the known methodology involves a relatively great effort.

SUMMARY

An aspect relates to a method and an arrangement for fault detection in a switch system that allow more detailed fault indications and/or require less effort.

For fault detection in a switch system that is distributed over at least one track route and comprises multiple switches, operating data of the switch concerned are respectively acquired for the switches of the switch system and transmitted to a central control. The term switch system should be understood here as also referring in particular to any desired group of multiple switches. For a respective first switch of the switch system,

• • a deviation of an operating behavior of the first switch from a set behavior is determined by the central control (CTL) on the basis of the operating data of the first switch,
  • if there is a deviation, operating data of the first switch are compared with operating data of other, second switches of the switch system,
  • depending on the results of the comparison, a second switch with similar operating data is selected,
  • on the basis of the operating data of the selected second switch, a fault indication for the first switch is derived, and
  • the fault indication for the first switch is output.

For carrying out the method according to the embodiment of the invention, an arrangement for fault detection in a switch system, a computer program product (non-transitory computer readable storage medium having instructions, which when executed by a processor, perform actions) and a computer-readable storage medium are provided.

The method according to the embodiment of the invention and the arrangement according to the embodiment of the invention can be carried out or implemented in particular by means of one or more processors, application-specific integrated circuits (ASICs), digital signal processors (DSPs, and or so-called “field programmable gate arrays” (FPGAs).

The embodiment of the invention can advantageously use the fact that switches arranged on the same track route are often passed by the same vehicles, and are therefore subjected to similar loading. By analogy with this, switches that are located in spatial proximity are generally exposed to similar ambient conditions, in particular similar weathering conditions. Behaviors, and in particular malfunctional behaviors of switches of a switch system are therefore often correlated with one another. Consequently, a known behavior or malfunctional behavior of similar switches can in many cases be used to detect or predict a behavior or malfunctional behavior of a relatively similar, first switch and/or to identify a probable fault cause. In particular, stored operating data of an older switch can in many cases be used to derive accurate forecasts of a future behavior of a relatively similar, more recent switch.

By means of the embodiment of the invention, fault finding and maintenance in a switch system can often be shortened considerably and fault-specific countermeasures can be initiated at an early time. The embodiment of the invention can be used in particular for the monitoring, testing, commissioning, maintenance, inspection, diagnosis, risk assessment and or control of a switch system, in particular even during operation.

Advantageous embodiments and developments of the invention are specified in the dependent claims.

According to an advantageous embodiment of the invention, the operating data acquired for a respective switch are transmitted to the central control by a transmitting device assigned to this switch. The transmission may take place in a wire-bound or wireless manner and/or by way of a data network, in particular the Internet. The transmitting device assigned to a respective switch may be arranged at the switch concerned or in a signal tower of the switch system. Because of the switch-specific transmitting devices, in many cases no service technician is required on site.

According to a further embodiment of the invention, an indication of a fault cause may be sought in the operating data of the selected second switch. If such an indication is found, the indication found may be output as a fault indication for the first switch. In cases where the selected second switch has similar operating data to the first switch, a fault cause indicated in the operating data of the selected second switch can often also be assumed as the probable fault cause for the second switch.

Furthermore, an indication of a fault that has occurred may be sought in the operating data of the selected second switch. If such an indication is found, the indication found may be output as a fault forecast for the first switch. In cases where the selected second switch has similar operating data to the first switch, a fault stored in the operating data of the selected second switch may serve as a fault forecast for the first switch.

According to an advantageous embodiment of the invention, the operating data of a respective switch may comprise a variation over time of a power consumption of a switch operating mechanism. An increased current or power consumption may in this case suggest a sluggishness of the switch concerned, which may be caused for example by icing, by an obstacle or by an obstruction, for example a stone.

According to a development of the embodiment of the invention, the operating behavior and/or the set behavior of the respective first switch may be simulated by means of a physical simulation model of the first switch on the basis of the operating data of the first switch. In particular, a switch operating mechanism, a power transmission and/or a switch blade of the first switch may be simulated by means of the physical simulation model.

According to a further embodiment of the invention, multiple second switches of which the operating data are similar to the operating data of the first switch may be selected. The operating data of the selected second switches may then be combined to derive the fault indication, in order for example to interpolate or extrapolate an operating parameter.

Furthermore, when comparing operating data of a respective first switch with operating data of a respective second switch, a distance measure for a distance between the respectively compared operating data may be determined. Alternatively or in addition, operating data patterns of the operating data to be compared may be determined by a pattern recognition method and compared, and the distance measure may be determined depending on the pattern comparison. Depending on the distance measure determined, that second switch of which the operating data have a smaller or smallest distance from the operating data of the first switch may be selected.

According to a particularly advantageous development of the embodiment of the invention, a knowledge graph of which the nodes are respectively assigned to a switch of the switch system may be managed by the central control. Operating data of a respective switch may then be stored in assignment to a node assigned to this switch or be stored in this node. Preferably, along with the operating data of the assigned switch acquired and transmitted at the time, historical operating data and other indications about this switch may preferably be collected as comprehensively as possible in assignment to a node. A respective node can consequently be understood as it were as a data twin of the respectively assigned switch. Such a knowledge graph allows operating data and other data originating from different sources, such as for example sensor data, maintenance data, ambient data, weather data or data concerning an installation location of a respective switch, to be brought together centrally. The data brought together may then be linked, correlated and evaluated in a particularly diverse way.

According to a further advantageous development of the embodiment of the invention, virtual operating data for a multiplicity of virtual switches may be generated by means of a physical simulation model. The virtual operating data may then be used as operating data of the second switches. A number of the virtual switches may be determined here depending on a number and/or an age of the switches of the switch system. In particular, the number of virtual switches may be increased in the case of recent and/or small switch systems. By generating virtual fleet data for the switch system, a volume of comparative data for comparing with the operating data of the first switch can be increased. This allows better fault detection, and generally improves an application of data-driven evaluation methods, in particular in the case of recent or small switch systems.

BRIEF DESCRIPTION

Some of the embodiments will be described in detail, with reference to the following figures, wherein like designations denote like members, wherein:

FIG. 1 shows a switch system controlled by a central control and

FIG. 2 shows an arrangement according to embodiments of the invention for fault detection in the switch system.

DETAILED DESCRIPTION

FIG. 1 illustrates a switch system WS of a track-bound means of transport, for example a railroad, that is controlled by a central control CTL. Alternatively or in addition, the switch system WS may also be provided for other rail-bound means of transport, such as for example streetcars and switchbacks, trolley coaches, cable cars or pneumatic mail systems.

The switch system WS comprises a multiplicity of switches W, which are arranged in spatial proximity to one another along track routes ST1 and ST2 or in some other way. The track route ST1 here comprises the tracks SP1 and SP2 and the track route ST2 comprises the tracks SP3 and SP4.

Switches W arranged on the same track or track route are often passed by the same vehicles one after the other, and are consequently subjected to loading and wear in a similar way. In addition, it can be assumed that switches W that are located in spatial proximity to one another are exposed to similar ambient conditions, in particular similar weathering influences. Consequently, behaviors, and consequently malfunctions, of the switches W concerned are often strongly correlated with one another. As a consequence, a behavior or a malfunction of a first switch can in many cases be detected, and in particular predicted, more accurately by a known behavior of neighboring or similar switches being evaluated. In particular, a fault cause can often be identified in this way.

For monitoring the switches W, operating data BD of the switches WR are continuously and switch-individually acquired by sensors. In the present exemplary embodiment, the operating data BD comprise in particular a variation over time of a current consumption or power consumption of a respective switch operating mechanism and an installation location of a respective switch W. Many types of fault leave a characteristic signature in this variation over time, on the basis of which the fault can be detected and/or identified. Thus, an increased current consumption of a switch operating mechanism often suggests a sluggishness of the switch, for example as a result of icing, an obstacle and/or by an obstruction, for example a trapped stone.

The variation over time of the current consumption or power consumption of a switch operating mechanism may be acquired at the switch W concerned or in a signal tower of the switch system WS.

Alternatively or in addition, the operating data BD may comprise in particular physical, technical-control, technical-effect and/or type-dependent operating parameters, property data, performance data, effect data, state data, configuration data, system data, default values, control data, sensor data, measured values, ambient data, weather data, temperature data, monitoring data, forecast data, analysis data, maintenance data, data concerning an operating time of the switch and/or other data occurring during the operation of the switch or relevant to the operation of the switch.

The operating data BD are transmitted to a central control CTL of the switch system WS by transmitting devices of the switch system WS in a wire-bound and/or wireless manner. The transmitting devices are respectively assigned to a switch W and may be arranged at the assigned switch W or in the signal tower of the switch system WS.

The central control CTL may preferably be implemented as a central computer in the signal tower of the switch system WS and/or at least partially in a cloud. To control the switches W, the central control CTL generates control data ST, preferably depending on the operating data BD received, and transmits these control data to the switch system WS.

FIG. 2 shows an arrangement according to the embodiment of the invention for fault detection in a switch system in a schematic representation. A first switch W1 of the switch system, for which a fault indication is to be derived, and a multiplicity of other, second switches W2 of the switch system, on the basis of which this fault indication is derived, are represented. Switches D1 and D2 respectively have sensors, which continuously acquire operating data BD1 for the switch W1 and operating data BD2 respectively for the switches W2.

The switch W1 is assigned a transmitting device SE1, which transmits the acquired operating data BD1 to a central control CTL in a wire-bound and/or wireless manner. Correspondingly, the second switches W2 are respectively assigned a transmitting device SE2, which transmits the operating data BD2 of the respective switch W2 to the central control CTL in a wire-bound and/or wireless manner. Because of the transmitting devices SE1 and SE2, no service technician is required on site for the acquisition and evaluation of the operating data BD1 and BD2.

The aforementioned switch system, the first switch W1, the second switches W2, the operating data BD1 and BD2 and the central control CTL may preferably be designed as described in connection with FIG. 1.

The central control CTL has one or more processors PROC for carrying out the method steps of the embodiment of the invention and has one or more memories MEM coupled to the processor PROC for storing the data to be processed by the central control CTL.

The central control CTL also has a simulation module SIM, which serves for simulating on the basis of the transmitted operating data of a switch a dynamic operating behavior of this switch by means of a physical switch simulation model SM. The simulation module may in particular simulate a switch operating mechanism, a drive motor, a force transmission and/or a switch blade of the switch. For this purpose, the operating data, here BD1, of the switch to be simulated, here W1, are fed to the simulation module SIM. The simulation is preferably carried out in parallel with the operation of the switch, and advantageously in real time.

In particular in the case of switch systems which comprise only a few or relatively recent switches, it may be provided that virtual operating data are generated for a multiplicity of virtual switches by means of the simulation model SM or some other physical switch simulation model. With these generated virtual operating data, the operating data BD2 of the second switches W2 can be supplemented, in order to increase a volume of comparative data with which the operating data BD1 of the first switch W1 can be compared. In this way, data-driven simulating and forecasting methods can often be improved significantly. In this case, the number of virtual switches may be determined depending on a number of switches in the switch system and/or depending on an operating age of these switches.

A knowledge graph KG, which is preferably implemented in a cloud C, is managed by the central control CTL. Along with the knowledge graph KG, components of the central control CTL may also be transferred into the cloud C or implemented there.

The knowledge graph KG comprises as data structures a multiplicity of nodes, which are interlinked by edges of the knowledge graph KG. The nodes are respectively uniquely assigned to a switch of the switch system or its switch operating mechanism, for example on the basis of a serial number of the switch or of the switch operating mechanism.

For all of the switches, here W1 and W2, of the switch system, their operating data, here BD1 and BD2, are respectively stored in the node assigned to the respective switch W1 or W2 and/or are assigned to it. A respective node of the knowledge graph KG is intended to act as it were as a data twin of the assigned switch.

In the present exemplary embodiment, in particular a variation over time of the current or power consumption of the switch operating mechanism of a respective switch and/or a force consumption or a time taken for changing over a switch are acquired as operating data and stored in the respectively assigned node. Along with the operating data of the respective switch acquired and transmitted at the time, in particular data concerning its installation location, maintenance data, state data, ambient data, data concerning the operating age or aging state, technical parameters and/or historical operating data of the respective switch are stored in the assigned node. The historical operating data in this case preferably include indications about a load history and or about malfunctions that have occurred of the switch concerned and preferably indications about fault causes. With otherwise similar operating data, the operating data of an older switch or an older switch operating mechanism stored in the knowledge graph KG can in many cases be used to derive accurate forecasts of a future behavior of a more recent switch.

The edges of the knowledge graph KG respectively connect nodes of the knowledge graph KG. An edge between two or more nodes may preferably be assigned operating relationships between the assigned switches. Such operating relationships between multiple switches may in particular comprise their sequence on a track route and/or some other neighborhood relationship or similarity relationship.

The central control CTL also has a monitoring module MON for determining a deviation of an operating behavior of a respective switch, here W1, from a set behavior of the switch. For this purpose, the operating data BD1 are transmitted to the monitoring module MON. The operating behavior of the first switch W1 is determined on the basis of the operating data BD1 and/or at least partially simulated by the simulation module SIM. The set behavior of the first switch W1 may be preset and/or at least partially simulated by the simulation module SIM on the basis of the operating data BD1. For the simulation of the operating behavior or the set behavior, the monitoring module MON is coupled to the simulation module SIM. The operating behavior and/or set behavior may be respectively represented by a variation over time or a signature of the current or power consumption of the switch concerned or its switch operating mechanism and/or by a force consumption or a time taken for changing over a switch or by other behavioral patterns of the switch concerned.

The monitoring module MON may establish the deviation of the operating behavior of the first switch W1 from the set behavior for example by comparing the operating behavior with the set behavior, determining a distance measure and checking whether the distance measure lies outside a preset tolerance range or above a preset threshold value.

If a deviation of the operating behavior of the first switch W1 from its set behavior is established by the monitoring module MON, the monitoring module MON instigates an inquiry of the operating data BD2 of the second switches W2 and a comparison of the operating data BD1 of the first switch W1 with the inquired operating data BD2 of the second switches W2. The initiation of the inquiry and the comparison is indicated in FIG. 2 by a dashed arrow.

In the course of the comparison, operating data patterns or signatures may be acquired by pattern recognition methods and compared. In a respective comparison, a distance measure D for a distance of the respectively compared operating data or operating data patterns is in each case determined as the comparison result. A respective distance measure D is in this case determined for a respective distance between the operating data BD1 of the first switch W1 and the operating data BD2 of a respective second switch W2. The distance measures D may also be understood as similarity measures.

In the present exemplary embodiment, distances or similarities between operating data and/or operating data patterns are determined in particular with regard to a current or power consumption of a switch operating mechanism, an installation location, an association with a track route, an operating age and/or an aging state of a respective switch. For the calculation of the distance measure D, an Euclidean or weighted distance between operating data vectors or operating data sub-vectors may be determined. In the case of a weighted distance, preferably operating-data-specific or operating-data—type-specific weights may be used. Alternatively or in addition, logical distances may also be used for the calculation of the distance measure D. Thus, in particular in the case of a highly branched track network, for example before a station, conditional probabilities or correlations for a train that travels over a first switch then also traveling over a second switch may be calculated. Switches that are strongly correlated in this respect can then be assigned a smaller distance than correspondingly more weakly correlated switches.

The above comparisons serve the purpose of finding that or those of the second switches W2 that is/are particularly similar to the first switch W1, behave(s) particularly similarly to it and/or has/have the same or similar signatures in the operating data. In particular, those of the second switches W2 that have behaved similarly to the first switch W1 in an earlier time period are searched. On the basis of a known further behavior of the second switches thus found, in many cases accurate forecasts of the further behavior of the first switch W1 can be derived. In addition, fault causes detected in the case of the second switches can often be identified as the probable fault cause for the first switch W1.

The comparisons or the similarity search are preferably carried out on the knowledge graph KG in the cloud C.

The distance measures D are fed to a selection device SEL of the central control CTL. In addition, the operating data BD2 of the second switches W2 are also transmitted to the selection device SEL. One or more second switches W2 of which the operating data BD2 have a smaller or a smallest distance D from the operating data BD1 of the first switch W1 are selected by the selection device SEL on the basis of the distance measures D and the operating data BD2. As a selection criterion, it may be checked here for example whether a respective distance D is smaller than a preset threshold value or smaller than all of the other distances D.

In the present exemplary embodiment, the distance measure D is specifically calculated such that, with otherwise similar operating data, preferably that or those switches W2 that lie on the same track or track route as the first switch W1 and/or is/are located in the spatial proximity of the first switch W1 is/are selected. As already mentioned above, because of the similar operating conditions, the behavior, and consequently the malfunctions, of those second switches W2 are often strongly correlated with the behavior and malfunctions of the first switch W1.

The selection device SEL transmits operating data SBD2 of the at least one second switch W2 that is selected, and is consequently similar to the first switch W1, to a fault detection module FDM. In cases where multiple similar second switches W2 are selected, their operating data may for example be combined by means of interpolation or extrapolation.

Furthermore, the operating data BD1 of the first switch W1 are transmitted to the fault detection module FDM. On the basis of the operating data SBD2 of the at least one selected second switch W2 and the operating data BD1 of the first switch W1, the fault detection module FDM derives a fault indication FA1 for the first switch W1. Here, the fault indication FA1 preferably comprises a fault cause for a fault that has occurred and/or a fault forecast for one or more faults to be expected. The fault indication FA1 is output by the central control and can be used for the anticipatory control of the switch system.

A fault cause for the first switch W1 may for example be derived by the operating data SBD2 similar to the operating data BD1 being searched through for a fault cause stored there and this being output as the probable fault cause for the first switch W1. In a similar way, the operating data SBD2 may be searched through for faults that have occurred and faults that are found output as a fault forecast for the first switch W1.

The embodiment of the invention described above easily allows an efficient and detailed detection of faults and their probable cause and a forecast of faults to be expected. In this way, fault finding and maintenance of a switch system can in many cases be shortened considerably and fault-specific countermeasures can be initiated at an early time. In particular, on the basis of the fault indication output, maintenance personnel can be informed as to which fault-specific tool is required for maintenance. Furthermore, the fault indications output may serve for estimating a risk of failure, a still remaining operating time and/or a degree of severity of a fault.

Although the present invention has been disclosed in the form of preferred embodiments and variations thereon, it will be understood that numerous additional modifications and variations could be made thereto without departing from the scope of the invention.

For the sake of clarity, it is to be understood that the use of ‘a’ or ‘an’ throughout this application does not exclude a plurality, and ‘comprising’ does not exclude other steps or elements.

Claims

1. A method for fault detection in a switch system that is distributed over at least one track route and comprises multiple switches, the method comprising:

acquiring operating data of the switch concerned for the switches of the switch system, and transmitting the operating data to a central control; and

b) for a respective first switch of the switch system, determining a deviation of an operating behavior of the first switch from a set behavior by the central control on a basis of the operating data of the first switch, if there is a deviation, comparing the operating data of the first switch with operating data of other, second switches of the switch system, depending on the results of the comparison, selecting a second switch with similar operating data, on the basis of the operating data of the selected second switch, deriving a fault indication for the first switch, and outputting the fault indication for the first switch.

2. The method as claimed in claim 1, wherein the operating data acquired for a respective switch are transmitted to the central control by a transmitting device assigned to the respective switch.

3. The method as claimed in claim 1, wherein an indication of a fault cause is sought in the operating data of the selected second switch and, if such an indication is found, the indication found is output as a fault indication for the first switch.

4. The method as claimed in claim 1, wherein an indication of a fault that has occurred is sought in the operating data of the selected second switch and, if such indication is found, the indication found is output as a fault forecast for the first switch.

5. The method as claimed in claim 1 wherein the operating data of a respective switch comprise a variation over time of a current consumption and/or a variation over time of a power consumption of a switch operating mechanism of the respective switch.

6. The method as claimed in claim 1, wherein the operating behavior and/or the set behavior of the respective first switch is simulated by means of a physical simulation model of the first switch on the basis of the operating data of the first switch.

7. The method as claimed in claim 1, wherein multiple second switches of which the operating data are similar to the operating data of the first switch are selected, and the operating data of the selected second switches are combined to derive the fault indication.

8. The method as claimed in claim 1, wherein, when comparing operating data of a respective first switch with operating data of a respective second switch, a distance measure for a distance between the respectively compared operating data is determined and/or operating data patterns are determined by a pattern recognition method and compared.

9. The method as claimed in claim 1, wherein a knowledge graph of which the nodes are respectively assigned to a switch of the switch system is managed by the central control, and the operating data of a respective switch are stored in assignment to a node assigned to this switch.

10. The method as claimed in claim 1, wherein virtual operating data for a multiplicity of virtual switches are generated by means of a physical simulation model, and the virtual operating data are used as operating data of the second switches.

11. The method as claimed in claim 10, wherein a number of virtual switches is determined depending on a number and/or an age of the switches of the switch system.

12. An arrangement for fault detection in a switch system that is distributed over at least one track route, designed for carrying out a method as claimed in claim 1.

13. A computer program product, comprising a computer readable hardware storage device having computer readable program code stored therein, said program code executable by a processor of a computer system to implement a method as claimed in claim 1.

14. A computer-readable storage medium with a computer product as claimed in claim 13.