Method and device for monitoring a railway network and rail network

Abstract

The method and the device serve to monitor a railway network (EN) which comprises stationary network units (NEs), mobile network units (NEm) and a monitoring device (CS) with stationary or mobile measuring devices (MFs; MFm) by means of which measured values ​​of at least one measured quantity of the assigned stationary network units (NEs) are recorded and transmitted to a stationary or mobile process computer (PRs; PRm) for evaluation.According to the invention, measured values ​​determined by the measuring devices (MFs; MFm) are supplied to a stationary and / or mobile extractor (EXs; EXm), which extracts from the measured values ​​of the associated stationary and / or mobile measuring device (MFs; MFm) those values ​​that lie outside a reference range or correspond to predefined rules or patterns and forwards them automatically or upon query; each stationary network unit (NEs) is assigned an identification unit (ID) containing identification data that is automatically or upon query; and the extracted measured values ​​and the associated identification data are transmitted to the stationary or mobile process computer (PRs; PRm), which subjects the extracted measured values ​​to an evaluation in order to determine the current state and / or a future expected state of the associated stationary network units (NEs).

Description

[0001] The invention relates to a method and a device for monitoring a railway network as well as a railway network equipped with such a monitoring device.

[0002] Railway networks consist of an extremely high number of different components and modules or network units, of which a relatively small number are monitored to ensure the safe operation of the railway network.

[0003] Control systems and safety systems, in particular interlocking systems, are described, for example, in [1], R. Hämmerli, "The Principles of Safety Systems for Railway Operations", Swiss Federal Railways SBB, Volume 1, February 1990, and Volume 2, November 1982. Furthermore, railway networks include communication systems, as described in [2], EP2631152A1.

[0004] For monitoring roadways, measuring vehicles equipped with measuring devices are normally used. From [3], DE19926164A1, a measuring vehicle is known in which the vibration behavior of a first vehicle component is measured in the frequency range below 500 Hz and of a second vehicle component in a frequency range above 500 Hz. The measured values ​​are evaluated to determine the condition of the vehicle or the roadway.

[0005] Measuring vehicles of this type travel at a considerable speed on the object being measured and are therefore subject to an interaction model with the track and the vehicle. All types of track and vehicle stress are cumulatively incorporated into a parameter representation. Evaluating the complex signals is therefore very time-consuming. The quality of the measurement results is further compromised by the fact that position data can deviate from the actual value by several meters. Moreover, the object being measured and the measuring instrument are subject to different influences during each measurement process. Measuring vehicles are also often not representative of the mobile units in the vehicle fleet and are technically complex, error-prone, and maintenance-intensive unique units.

[0006] Patent [3], WO2014044485A2, discloses a further method for diagnosing components of a railway network. A first measuring device acquires measured values ​​of at least one parameter to describe the operating state of the track component and transmits them wirelessly to a control center, which evaluates the acquired data. A second measuring device acquires measured values ​​and transmits them to the control center that are independent of the operating state of the track components.

[0007] The described methods enable the determination of condition data for network components. During test drives, the behavior of the network units being traversed is measured and recorded. The measured values ​​are combined with the respective position data of the test vehicle to subsequently determine the position of the network units. As mentioned, this position data is usually subject to errors, so that identifying a faulty network unit is often not possible. The data obtained during test drives typically do not provide a sufficient basis for statistical analysis.

[0008] However, stationary measuring units used sporadically generate larger amounts of data, the central processing of which requires a significant amount of time.

[0009] EP2894074A1 discloses a method for monitoring network units of a railway network, wherein self-contained area boundaries are provided within a reference system and status data is queried for network units located therein.

[0010] US2018048400A1 and US2017149603A1 disclose methods for fault detection in communication systems which include base stations for wireless communication.

[0011] The JP2013049395A reveals a train monitoring system.

[0012] EP1900597A1 discloses a diagnostic system for monitoring a rail system in which rail infrastructure-related sensor data and rail vehicle-related sensor data are recorded, combined and centrally evaluated over time.

[0013] EP2943001A1 discloses a method for identifying a source of error in a communication network which includes at least one participant which is configured to be connected to a base station.

[0014] WO2016115443A1 discloses a method for recording vibration data, which is centrally evaluated by a computer system.

[0015] The present invention is therefore based on the objective of creating an improved method and an improved device for monitoring a railway network, as well as an improved railway network that can be advantageously monitored with such a monitoring device.

[0016] The inventive method and device are intended to allow the acquisition of condition data and condition forecasts for the railway network with reduced effort and time expenditure.

[0017] The railway network should be precisely mapped and monitored in detail. The exact location of errors and defects should be possible.

[0018] Even with a high level of monitoring, the burden of large amounts of data should be avoided without losing relevant information.

[0019] The method according to the invention is intended to enable the determination of information on both the current state and the future expected state of the railway network and its network units.

[0020] Furthermore, the monitoring device should be self-learning, so that the procedure is automatically optimized during the operation of the railway network and the current state as well as the expected future state of network units can be determined in a shorter time and with higher precision.

[0021] The collected condition data should allow for the optimization of maintenance and upkeep, and if necessary, also the control of the railway network and operational safety.

[0022] This problem is solved by a method, a device, and a railway network, which have the features specified in claims 1, 11, and 14. Advantageous embodiments of the invention are specified in further claims.

[0023] The method and the device serve to monitor a railway network, which comprises stationary network units, mobile network units and a monitoring device with stationary measuring equipment, by means of which measured values ​​of at least one measured quantity of the assigned stationary network units are recorded and transmitted to a stationary process computer for evaluation.

[0024] According to the invention, stationary network units (NEs) are provided, which are equipped with identification units (IDs), measuring devices (MFs), and extractors (EXs); measured values ​​determined by the stationary measuring devices (MFs), which correspond to vibrations of the stationary network units (NEs) occurring during the passage of a train, are supplied to the associated stationary extractor (EXs), which extracts from the measured values ​​of the associated stationary measuring devices (MFs) and forwards automatically or upon querying those values ​​that lie outside a reference range or correspond to predefined rules or patterns; the extracted measured values ​​and the associated identification data are transmitted to the stationary process computer (PRs), which subjects the extracted measured values ​​to an evaluation.to determine the current state and / or a future expected state of the associated stationary network units (NEs).

[0025] Stationary network units, which, equipped with an identification unit, form an extended stationary network unit, are network components such as sleepers, rails, switches, signaling modules, safety modules, or other elements of the railway infrastructure. Preferably, all network units that are subject to wear and / or may fail are monitored by means of measuring devices. An existing railway network can be gradually transformed into a railway network according to the invention. The method for monitoring the railway network can therefore also be implemented gradually.Following the partial or complete transformation of an existing railway network or the construction of a new railway network according to the invention, identification and measurement data for the relevant extended stationary network units are recorded and transmitted to a centralized process computer. This computer determines the state of the stationary network units or of network sections comprising a group of network units based on the acquired data. Instead of a single centralized process computer, several process computers can also be provided, each performing different analysis tasks.

[0026] Depending on the assessed condition of network units or sections, necessary measures may be initiated. These measures may involve direct intervention in the railway network, particularly in the event of failure or serious damage to network units. For example, if significant defects are identified in a section of track, this section may be closed, with corresponding adjustments to the signaling systems and interlocking.

[0027] However, if defects occur that do not require immediate intervention, the relevant information is preferably transmitted to a maintenance computer. This computer can then register the defects and schedule maintenance work. Routine maintenance can therefore be brought forward or postponed as needed, for example, if track sections are reported to be in perfect working order.

[0028] Changes to the railway network can also be transmitted to a computer for administrative tasks. For example, based on condition reports of track sections, timetable changes can be implemented, which in turn alter the load on the railway network. For instance, heavy vehicles may be prohibited from using the damaged section of track. Freight trains, for example, may be rerouted.

[0029] The construction of a railway network or monitoring device according to the invention is relatively straightforward. Network units, passive and active modules of the railway network, can be easily equipped with identification units, or with identification units and measuring devices, or with identification units, measuring devices, and extractors. During the manufacturing of the network units, the identification units, measuring devices, and, if applicable, extractors, as well as any processor units or microcontrollers, can be inserted into designated recesses. The manufacturing effort remains practically the same, while the cost of the network unit increases only minimally by the cost of the integrated electronic units.The electrical or electronic units can be managed and / or controlled individually or jointly by a processor and can be equipped individually or jointly with a communication unit that allows determined data and signals, in particular identification data and measured values, to be sent wired or wirelessly to a higher-level data processing unit, e.g. directly or preferably via a stationary concentrator to a process computer.

[0030] The network unit, together with the mechanically coupled electronic units, at least one identification unit, preferably supplemented by at least one measuring device, and more preferably by an extractor, forms extended stationary units that can exist in various configurations. The electronic units can be interconnected by cables or wireless communication interfaces. For the collection of information and data, a wireless unidirectional or bidirectional data connection is preferably established between the stationary network units or the extended stationary network units and concentrators, which collect the transmitted data from their assigned network units and transfer it directly or indirectly to a process computer.

[0031] Concentrators can be assigned to a specific group of similar or dissimilar network units. Preferably, the concentrators are suitable for the dynamic setup of networks to which a large number of extended network units can be connected. This connection preferably occurs automatically, e.g., according to salutation procedures such as those developed for the Bluetooth system. That is, the extended network units and the associated concentrator form an ad hoc network that transmits the acquired data in a concentrated manner to the centralized processing computer.

[0032] Preferably, the condition of the concentrators is also recorded and monitored in the process computer or a maintenance computer to ensure that the proper functioning and sufficient capacity of the concentrators are always guaranteed.

[0033] The concentrators are preferably controllable, allowing data to be transmitted selectively. For example, a group of extended network units can be completely suppressed to free up resources for the analysis of another group of network units. Furthermore, it can be provided that the extended network units can transmit their identification to the process computer independently of their state and the corresponding measured values. In this way, it is possible to map the railway network or its network units independently of their current state. Once such a complete map has been created, it can display, for example, the current states of the extended network units on a first level and, for example, the future expected states of the network units on a second level.

[0034] The railway network according to the invention therefore practically possesses a nervous system extending to the periphery of the network, which provides feedback from all connected or extended network units virtually without delay. This "nervous system" delivers an enormous amount of information that can only be evaluated with considerable effort using data processing equipment. The invention, however, is based on the idea that the vast majority of this information, which inevitably occurs and reflects the normal state of the railway network, is not of interest. According to the invention, measured values ​​that deviate from a normal state are therefore extracted using extractors and fed to a downstream stage for processing. In this way, it is possible to detect "pathological" system behavior and to intervene in the system with maximum efficiency, if necessary.The inventive method thus records the "disease progression" of an identified network unit or a group of network units. Data describing the normal state of the railway network, on the other hand, are disregarded.

[0035] The extended network units can include their own power supply devices, e.g., with solar cells or piezoelectric elements, or they can be externally powered. The use of piezoelectric elements is particularly advantageous, as these are subjected to mechanical stresses during the operation of the railway network and generate corresponding voltages that can be used, for example, to charge a storage capacitor or a battery.

[0036] Preferably, energy-saving methods are employed. For example, the extended network units are activated when an event occurs, such as a passing train, so that energy is consumed only within short periods. For example, a microcontroller is used that has a minimal power consumption of preferably < 100 µA in operating mode, a practically negligible power consumption of preferably < 500 nA in standby mode, short delay times when transitioning from standby to operating mode (preferably < 1 µs), and all essential signal processing functions. For example, microcontrollers such as those described in the 2015 documentation "MSP Low-Power Microcontrollers" by Texas Instruments Incorporated are used.

[0037] During operation, and preferably upon the occurrence of a relevant event, possibly after a query, the extended network units transmit identification data. This identification data can be a network-wide unique identifier, similar to the IMEI number for mobile phones. In addition to the identification data, the coordinates and / or functionalities of the associated network unit are preferably stored in volatile or non-volatile memory (RAM / ROM). This data allows for a precise mapping of the railway network, including the network units and their positions and functionalities. The extended stationary network units, preferably equipped with a processor or microcontroller and communication means, are preferably programmed during manufacturing or on-site installation.

[0038] The measuring devices can be used to determine any static state data regarding properties and / or states of the network units that do not change over time or change only slowly, as well as dynamic state data regarding properties and / or states that change rapidly. For example, the vibration behavior of a network unit can be determined when a relevant event occurs. Considering additional measurement data, such as the temperature or humidity of supporting elements, allows for a more precise description of the state, since vibration amplitudes and frequencies can change depending on the prevailing temperature.

[0039] The measuring devices can be permanently or temporarily assigned to the network units. Any measuring devices, such as acceleration sensors, bending sensors, strain sensors, temperature sensors, and the like, can be used.

[0040] Regardless of the power supply, measured values ​​can only be acquired and transmitted during periods when a relevant event occurs. In this case, too, a processor or microcontroller can be switched from a standby state to an operating state, for example, when a train approaches.

[0041] In further preferred embodiments, the type of event is identified and taken into account in the process computer during signal evaluation. Identification is easily possible, for example, using timetable data. This allows not only the event itself, but also each of the vehicles that affected the network units to be identified. For example, the process computer uses timetable data to check whether a light passenger train or a heavily loaded freight train triggered the event. The corresponding parameters are then considered during the analysis of the measured values.

[0042] It was explained that the inventive method records the "disease progression" of an identified network unit or group of network units, which offers several advantages. Instead of evaluating all data potentially occurring in the railway network, the evaluation is limited to data that show conspicuous behavior of network units. That is, it checks whether symptoms of a disease or anomaly are present that could lead to limitations in the functionality or operational capability of the network units. These conspicuousnesses and / or symptoms can be detected using previously stored data.

[0043] Stationary extractors are used to check for measured values ​​that exhibit such symptoms. This involves checking whether the measured values ​​lie outside a reference range or correspond to rules or patterns that define anomalous behavior. Such measured values ​​are subsequently extracted, i.e., forwarded for further processing, while non-critical signals are preferably disregarded to conserve resources. The extractors are preferably controllable, allowing for the application of different extraction criteria. For example, the temperature of the track infrastructure is measured, and temperature-adapted rules or patterns are then activated in the extractors. The extractors are also preferably controllable in such a way that measurement data can be selectively queried, even if it lies within the reference range.

[0044] Particularly interesting are measurements that correspond to vibrations of the network units and occur, for example, during the passage of a train. As mentioned, these vibrations can be detected by stationary or mobile measuring devices. In an example not covered by the claimed invention, mobile measuring devices can be used; these are also linked to the identification data queried in parallel. Even when measurement data is obtained using a measuring vehicle, the network units can thus be precisely located.

[0045] Measurement values ​​representing vibrations of the network units can be subjected to various testing procedures. The reference range can, for example, define threshold values ​​or envelopes for vibration amplitudes. Maximum vibration amplitudes can be selectively defined for specific frequencies. Pattern recognition can be performed using methods described in [5], Heinrich Niemann, "Classification of Patterns", Springer Verlag 2003. Measurement values ​​can, for example, be subjected to the Fourier transform (FFT) to determine the individual frequencies and their intensities of the received signal mixture. In preferred embodiments, the power spectral density (PSD) is determined, which indicates the frequency-related power of a signal in an infinitesimal frequency band.By comparing the determined power density spectrum with known power density spectra, the state of the monitored stationary network unit NE can be determined.

[0046] Methods are available for determining the current state and for predicting the state of network units and network sections, which are described, for example, in [5], [6], J. Scott Armstrong, Principles of Forecasting, London 2002, and [7], P. Mertens, S. Rässler, Prognoserechnung, Kapitel 2, Michael Schröder, Einführung in die kurzreihenprognose und Vergleich der individuelle Verfahren, Springer-Verlag Berlin, Heidelberg 2012.

[0047] For example, extracted measurement values ​​available at discrete time intervals for individual stationary network units or for groups of identical or different stationary network units are checked against rules and / or patterns and / or currently determined comparative values, in particular currently determined comparative values ​​of similar neighboring stationary network units, in order to determine status data for the network units. The monitoring device itself continuously generates new information that can be used for future testing of network units.

[0048] The method for evaluating the extracted measured values ​​preferably uses a rule-based prediction as described in [6], Chapter 9. The rule-based prediction is preferably based on the extrapolation of time series determined for the monitored network units. In addition, expert knowledge is continuously and automatically determined and taken into account during the signal evaluation. For example, if the maintenance computer reports the failure of a network unit, the process computer can determine the most recently recorded time series for that network unit and use it as a pattern in the future, which can at least serve as an indicator of the imminent failure of another corresponding network unit. If the process computer itself detects the failure of a network unit, it can also use a time series created before the failure as a pattern for future fault detection.The monitoring device therefore continuously accumulates knowledge about the network units of the railway system, enabling it to monitor them ever faster and with greater precision, and to provide data for maintenance and servicing. Furthermore, information obtained by the manufacturer or user of the network units during laboratory tests can also be used as test criteria. Maintenance instructions thus do not need to be transmitted to personnel who would have to inspect the network units on-site at considerable expense, but can be implemented centrally for testing the relevant network units.

[0049] The invention is explained in more detail below with reference to the drawings. These show: Fig. 1 a part of a railway network EN with stationary network units NEs and a mobile network unit NEm operating on the railway network EN, as well as a monitoring device CS by means of which the railway network EN is monitored; Fig. 2 a railway network EN with stationary network units NEs in which a monitoring device CS according to the invention is implemented; and Fig. 3 the railway network EN of Fig. 2 in a further illustration.

[0050] Fig. 1 shows a part or route of a railway network EN with stationary network units NEs and a mobile network unit NEm operating on the railway network EN, as well as a monitoring device CS by means of which the railway network EN is monitored.

[0051] The monitoring device comprises identification units (IDs), each assigned to one of the stationary network units (NEs) and containing at least the identification data of the assigned stationary network unit (NE). The network units (NEs) shown in example are assigned the identities "0", "1", "2", and "3". The combination of a network unit (NE) and its associated identification unit (ID), as well as any additional modules, such as a stationary measuring device (MFs) and a stationary extractor (EXs), is referred to below as an extended stationary network unit (NEX). Extended stationary network units (NEXs) can be active or passive, intelligent or non-intelligent. Active network units (NEXs) transmit information permanently or sporadically and thus have the functionality of a beacon that wirelessly transmits signals. Passive extended stationary network units (NEXs) are typically transponders that receive query signals and transmit response signals.In RFID technology, query signals often also serve to power the transponders. Intelligent network units (NEs) are also equipped with a processor or microcontroller, which can be used to control, for example, the acquisition of measured values, the processing of these values, and the communication devices.

[0052] Fig. 1 further shows that the identification units ID preferably have a memory in which location data LOC or the coordinates K0, ..., K3 of the stationary network units NEs are stored.

[0053] Furthermore, characteristic data PROP of the stationary network units NEs are stored, which describe the specifications and functionalities of the stationary network units NEs. The variables X, Y, Z describe, for example, the type of stationary network unit NEs, e.g., a switch, a rail, or a sleeper. Preferably, extended stationary network units NEX are provided without gaps.

[0054] Furthermore, stationary measuring devices (MFs) are provided that record the static or dynamic physical or chemical states of the permanently assigned stationary network units (NEs) and automatically or upon request provide corresponding measured values. Static states are states that are practically constant for the duration of the signal sampling, such as the temperature of the stationary network units (NEs). Dynamic states are states in which the stationary network units (NEs) exhibit a time-varying response to an external influence. A network unit (NE), e.g., a rail, often shows vibrations after a train passes over it, which then slowly decay. These types of states are physical states. The chemical state of a network unit (NE) can be determined using chemical measuring devices, which are suitable, for example, for measuring humidity and pH values.If moisture and pH levels exceed certain values, damage to stationary network units (NEs) is to be expected. Stationary network units (NEs) can affect not only directly trafficked roads, but also the supporting subsoil, which can be significantly influenced by external factors.

[0055] In an example not covered by the claimed invention, dynamic state behavior, such as the vibration behavior of network units NEs, can also be measured by a mobile measuring device MFm mounted on a mobile network unit NEm or a rail vehicle. Due to the large number of stationary network units NEs, from which simple static and complex dynamic states are recorded, an enormous amount of data and signals are generated in the railway network EN according to the invention, which can hardly be processed with reasonable effort. Therefore, according to the invention, stationary extractors EXs are provided that extract from the measured values ​​of the associated stationary measuring devices MFs those values ​​that lie outside a reference range or correspond to predefined rules or patterns and output them automatically or upon query via a third interface.

[0056] Simple thresholds and limits can be defined for extracting the measured values. If measured values ​​do not exceed certain limits, e.g., humidity and pH values, they are not passed on but suppressed.

[0057] Simple barriers can also be provided for dynamic states. For example, a signal mixture is rectified and compared with a reference voltage or a reference value.

[0058] Alternatively, a signal mixture can be decomposed into its components, for example using a Fourier transform. These components can then be individually compared with reference values ​​or collectively with, for example, an envelope, to detect signals that exceed predefined limits.

[0059] It is also possible to use passive or active filter stages that filter out critical frequencies and suppress frequencies caused, for example, by a rail vehicle.

[0060] The relevant reference ranges, patterns, or rules can be specified by the manufacturer and stored in the ID identification unit along with the proprietary PROP data. Preferably, however, the relevant reference ranges, patterns, or rules are downloaded by a process computer (PRs, PRm), which ideally optimizes these reference ranges, patterns, or rules continuously.

[0061] Pattern recognition can be carried out using methods described in [5], Heinrich Niemann, "Classification of Patterns", Springer Verlag 2003. Measured values ​​can, for example, be subjected to the Fourier transform (FFT) to determine the individual frequencies and their intensities of the received signal mixture.

[0062] The data output by the stationary extractors (EXs) is collected in concentrators (XX) and fed to a stationary process computer (PRs). Data transmission between all modules preferably occurs wirelessly.

[0063] In an example that does not fall under the claimed invention, the data determined by the mobile extractor EXm are fed to a mobile process computer PRm.

[0064] The process computer PRs; PRm subjects the extracted measured values, linked with the associated identification data, to an evaluation in order to determine the current state and / or a future expected state of the associated stationary network units NEs.

[0065] For this purpose, extracted measured values ​​available at discrete time intervals for individual stationary network units NEs or for groups of identical or different stationary network units NEs are checked using rules and / or patterns and / or currently determined comparative values, in particular currently determined comparative values ​​of similar neighboring stationary network units NEs.

[0066] Preferably, time series are created for the discrete measured values. Measured values ​​relating to oscillations of a stationary network unit are preferably subjected to a Fourier transform, after which time series are created for the determined frequency components of the signal. If a Fourier transform has already been performed in the extractors, the transformed data are preferably used further. The elements of the time series are preferably weighted so that current data receives a higher weight and older data is reduced in importance and, if necessary, deleted. Expert knowledge, i.e., stored data, reveals, for example, which frequency components are critical. Subsequently, the relevant time series can be evaluated and forecasts for future developments can be generated. Based on these forecasts, interventions in the railway network can then be planned and implemented.

[0067] The creation of rule-based forecasts using the extrapolation of time series models and the application of expert knowledge is described in [6], Chapter 9. In [7], page 25, forecasts using moving average and raised moving average methods are illustrated. Furthermore, the principle of exponential smoothing is described.

[0068] Fig. 1 The diagram also shows extended stationary network units NEX0, NEX1, NEX2, and NEX3, which have different structures. The associated stationary network unit NEs only has an identification unit ID and its state is not monitored; that is, monitoring the identification data only verifies the existence and functionality of the identification unit ID.

[0069] The extended stationary network unit NEX1 comprises, in addition to the stationary network unit NEs and the identification unit ID, a stationary measuring device MFs, which monitors the stationary network unit NEs, and an extractor EXs. It transmits extracted measured values ​​to a concentrator CX via a preferably wireless bus SS. The identification data is transmitted to the stationary measuring device MFs via a first interface and, together with the measured values, to the extractor EXs via a second interface. The extractor EXs then transmits the extracted measured values, along with the identification data, to the concentrator CX via a third wireless interface. In contrast, with the extended stationary network unit NEX2, the identification data is transmitted directly to the associated stationary extractor EXs via the first interface.In contrast, the extended stationary network unit NEX3 comprises only an identification unit ID and a stationary measuring device MFs, from which the measured values ​​are wirelessly transmitted, linked with the identification data, to a centralized concentrator CX, which is equipped with an extractor EXs for this case.

[0070] The extended stationary network unit NEX16 comprises an identification unit ID, which, as a mobile network unit NEm passes by, transmits the identification data ID via an air interface to a mobile measuring device MFm, preferably installed on the underside of the mobile network unit NEm. The mobile measuring device MFm preferably comprises several sensor modules, a first of which serves to receive, optionally query, and receive the identification data. A second sensor module detects vibrations, including those of the stationary network unit NEm. The measured values ​​are fed to a mobile extractor EXm, which extracts those measured values ​​that lie outside a reference range or correspond to predefined rules or patterns and transmits them to the mobile process computer PRm, which evaluates the extracted measured values ​​taking into account the associated identification data.

[0071] Fig. 2 Figure 1 shows a railway network EN with stationary network units NEs in which a monitoring device CS according to the invention is implemented. The diagram shows, by way of example, various track sections with signal boxes ST and main tracks HG as well as secondary tracks NG, and identifies some of the installed extended stationary network units NEX00111010001, NEX00111010010, NEX00111010011, NEX00111010100, NEX00111010101, which are of type NEX1. Fig. 1 The system corresponds to the following: Identification data from some of the extended stationary network units (NEX) is wirelessly transmitted along with extracted measurement values ​​to associated concentrators (CX), which consolidate and concentrate the data, also via a wireless interface, to the stationary process computer (PRs). The data is processed in the process computer (PRs) and, if necessary, stored in a database (PRD). Information required for analyzing the measured data is also extracted from the database (PRD).

[0072] Fig. 3 shows the EN railway network of Fig. 2 In another illustration, several track sections are shown that are already equipped with extended stationary network units (NEX). Along the relevant track sections, the installed extended stationary network units (NEX) are symbolized by series of short vertical lines. These units are equipped with extractors (EXs) and wirelessly connected to concentrators (XX), which in turn are wirelessly, directly or indirectly, connected to the stationary process computer (PRs).

[0073] In Fig. 3 This illustrates that the behavior of stationary network units (NEs) can be recorded not only by stationary measuring devices (MFs) but also by mobile measuring devices (MFm). The extraction of the relevant measured values ​​takes place at the level of the extractors (EXs, EXm), for example, using application programs (Ax).

[0074] Subsequently, the measured values ​​obtained in the stationary process computer PRs and the mobile process computer PRm are evaluated, for example, using application programs Av, to obtain current status information or predictions for future developments of the states. Preferably, time series are created and evaluated using time series analyses and expert knowledge. If, through expert knowledge, a fault in the relevant extended stationary network unit NEX is identified, for example, by detecting an exceedance of a tolerance value or a match with a fault pattern, a message SX ID is sent to the network computer NR. This message contains the identity of the relevant stationary network unit NEX and preferably fault information. The network computer NR transmits this information to a maintenance module or a maintenance computer NM, which then initiates measures to correct the fault in the relevant network unit NEX.

[0075] Predictions can also be used to estimate when tolerance values ​​are expected to be exceeded in individual network units NEX. These predictions SV ID are also transmitted via the network computer NR to the maintenance computer NM, which subsequently schedules the corresponding maintenance operations. Preferably, the maintenance computer NM checks in which sections of the railway network EN maintenance work should be carried out within a specific period. Through appropriate planning and coordination of the maintenance work, costs and effort in carrying out the maintenance work can be reduced.

[0076] Time series data collected for network units (NEX) are stored in a database (PRD) and preferably continuously updated. Of particular interest are time series for which a current error or failure of a network unit (NEX) has been registered. This error or failure can be detected by the process computer (PRs). Alternatively, the failure of a network unit (NEX) can be reported to the process computer. Fig. 3This shows that an error message F IDX was transmitted to the maintenance computer NM for network unit NEs with the identification number IDX. For example, the failure of network unit NEX was detected during a route inspection and reported to the maintenance computer NM via a communication interface. The error message F IDX is subsequently transmitted to the process computer(s) PRs, PRm, which retrieves the last saved data ID X for the faulty network unit NEX from the database PRD and re-evaluates it based on the error message F IDX. If anomalies are found that indicate the fault, the error pattern FM IDX is used. The error pattern FM IDX can then be used as a fingerprint or error pattern in the extractors EXs, EXm and the database PRD to extract or evaluate measured values. Similarly, error messages F IDX can be obtained by the process computers PRs, PRm and used as the error pattern FM IDX.The arrow RG symbolizes that the monitoring device CS is thus capable of generating rules and patterns itself. Rules R2 are loaded into the extractors EXs and used to extract measured values, while rules R1 are loaded into the process computers PRs and PRm and used to evaluate measured values. The system is therefore self-learning and can continuously optimize itself over time. Bibliography

[0077] [1] R. Hämmerli, The Principles of Safety Systems for Railway Operations, Swiss Federal Railways SBB, Volume 1, February 1990, and Volume 2, November 1982 [2] EP2631152A1 [3] DE19926164A1 [4] WO2014044485A2 [5] Heinrich Niemann, "Classification of Patterns", Springer Verlag 2003 [6] J. Scott Armstrong, Principles of Forecasting, London 2002 [7] P. Mertens, S. Rässler, Forecasting, Chapter 2, Michael Schröder, Introduction to Short-Term Time Series Forecasting and Comparison of the Individual Methods, Springer-Verlag Berlin Heidelberg 2012

Claims

1. Method for monitoring a railway network (EN), comprising stationary network units (NEs), which represent elements of a railway infrastructure, mobile network units (NEm) and a monitoring device (CS) with stationary measuring devices (MFs), by means of which measurement values of at least one measurement parameter of the associated stationary network units (NEs) are captured and transmitted to a stationary process computer (PRs) for analysis, wherein, for each stationary network unit (NE), an identification unit (ID) is assigned, which contains identification data that are transmitted automatically or upon request, wherein stationary network units (NEs) are provided, which are equipped with identification units (ID), measuring devices (MFs) and extractors (EXs), wherein measurement values determined by the stationary measuring devices (MFs), which correspond to vibrations of the stationary network units (NEs) occurring during the passage of a train, are fed to the associated stationary extractor (EXs), which extracts from the measurement values of the assigned stationary measuring devices (MFs) those that lie outside a reference range or correspond to predefined rules or patterns, and forwards them automatically or upon request, and wherein the extracted measurement values and the associated identification data are transmitted to the stationary process computer (PRs), which subjects the extracted measurement values to an analysis in order to determine the current state and / or a future expected state of the associated stationary network units (NEs).

2. Method according to claim 1, characterized in, that the measurement values extracted by several stationary extractors (EXs), as well as, preferably, the associated identification data, are transmitted wirelessly or via a wired connection to a stationary concentrator (CXs), which concentrates the transmitted data and forwards it to the stationary process computer (PRs).

3. Method according to claim 1 or 2, characterized in, that stationary network elements (NEs) fitted with identification units (IDs) are network components such as sleepers, rails, points or other elements of the railway infrastructure, and that the identification units (IDs) are permanently attached to the stationary network elements (NEs) or form an integral part thereof.

4. Method according to claim 1, 2 or 3, characterized in, that the identification units (IDs), which are optionally equipped with a power supply unit, comprise a memory in which the identification data and / or data relating to the installation location and / or to the properties and / or functionalities of the stationary network units (NEs) are stored permanently or temporarily, and which are automatically transmitted or retrieved.

5. Method according to one of the claims 1 - 4, characterized in, that the extractors (EXs; EXm) determine whether frequency components of the recorded vibrations lie outside the reference range in terms of frequency and / or intensity, or whether the measurement values correspond to a specified pattern.

6. Method according to one of the claims 1 - 5, characterized in, that the extractors (EXs; EXm) are controllable so that the reference range, the patterns and / or rules can be selectively adjusted, preferably as a function of a parameter such as time, weather conditions or temperature.

7. Method according to one of the claims 1 - 6, characterized in, that identification units (ID) and stationary measuring devices (MFs) and stationary extractors (EXs), which are each assigned to one another, together form a first stationary network unit (NEX1); or that identification units (ID) and stationary measuring devices (MFs), which are each assigned to each other, together form a second stationary network unit (NEX2), wherein the identification units (ID) and / or the stationary measuring devices (MFs) and / or the stationary extractors (EXs) and / or the stationary concentrators (CXs) communicate with one another wirelessly or via a wired connection.

8. Method according to one of the claims 1 - 7, characterized in, that the stationary process computer (PRs) examines, at discrete time intervals, the extracted measurement values available for individual stationary network units (NEs) or for groups of identical or different stationary network units (NEs) on the basis of rules and / or patterns and / or currently determined reference values, in particular currently determined reference values of similar neighbouring stationary network units (NEs), in order to determine status data for the network units (NEs).

9. Method according to one of the claims 1 - 8, characterized in, that extracted measurement values available at discrete time intervals are assigned, on the basis of the associated identification data, to a respective time series for which, preferably taking expert data into account, status data regarding the current state and / or predictions regarding the future development of the state of the concerned stationary network units (NEs) are determined.

10. Method according to one of the claims 1 - 9, characterized in, that, on the basis of status data, in particular status data indicating an anomaly in stationary network units (NEs), which have been determined by process computers (PRs) for one or more stationary network units (NEs) or supplied by an external computer, expert data, such as patterns and / or rules, are generated and stored in a database, which expert data are used for processing the measurement values in the extractors (EXs) and / or in the stationary process computers (PRs).

11. A device for monitoring a railway network (EN), comprising stationary network units (NEs) representing elements of a railway infrastructure, mobile network units (NEm) and a monitoring device (CS) operating according to a method as per one of claims 1-10, wherein the monitoring device comprises stationary measuring devices (MFs) which are configured such that they capture measurement values of at least one measurement parameter of the associated stationary network units (NEs) and transmit them to a stationary process computer (PRs) for analysis, wherein each stationary network unit (NEs) is assigned an identification unit (ID) containing identification data which is transmitted automatically or upon request, in which stationary network units (NEs) are provided, which are equipped with identification units (ID), measuring devices (MFs) and extractors (EXs), wherein the device is designed such that measurement values determined by the stationary measuring devices (MFs), which correspond to vibrations of the stationary network units (NEs) occurring during the passage of a train, are fed to the associated stationary extractor (EXs), which extracts from the measurement values of the assigned stationary measuring devices (MFs) those that lie outside a reference range or correspond to predefined rules or patterns, and forwards them automatically or upon request, and that the extracted measurement values and the associated identification data are transmitted to the stationary process computer (PRs), which subjects the extracted measurement values to an analysis in order to determine the current state and / or a future expected state of the associated stationary network units (NEs).

12. Device according to claim 11, characterized in, that several stationary extractors (EXs) are each connected wireless or wired to a stationary concentrator (CX), through which extracted measurement values and associated identification data are transferable in concentrated form to the stationary process computer (PRs).

13. Device according to claim 11 or 12, characterized in, that programme modules (Ax) are provided for the stationary extractors (EXs) to detect measurement values that lie outside a reference range or correspond to specified rules or patterns, and that programme modules (Ax) are provided for the stationary process computers (PRs), which serve to detect faults or defects that occur currently or may occur in the future in the monitored stationary network units (NEs).

14. Railway network (EN) with a device according to claim 11.

Images