Method and monitoring device for monitoring the integrity of a rail transport system

Abstract

The invention relates to a method for monitoring the integrity of a rail transport system (2), having the steps of: detecting measurement signals from a plurality of strain sensors (21) on a track rail (6) of the rail transport system (2); and determining the integrity of at least one component (5, 6, 7, 9) of the rail transport system (2) on the basis of the measurement signals, wherein the plurality of strain sensors (21) are components of a linear detection means (19) attached to the track rail (6) and extend along the track rail (6) over a measurement section (31) with a length (L) of at least 100 m, and wherein the integrity is determined by means of an electronic evaluation unit (20) which is signal-connected to the plurality of strain sensors (21), in particular at least 100 of the strain sensors (21), in order to be able to receive the measurement signals. The invention also relates to a monitoring device (1) for monitoring the integrity of the rail transport system (2).

Description

[0001] Procedure and monitoring device for monitoring the integrity of a railway transport system

[0002] The content of the German patent application DE 102025 101 315.2 is incorporated herein by reference.

[0003] The invention relates to a method for monitoring the integrity of a railway system. Furthermore, the invention relates to a monitoring device for monitoring the integrity of a railway system.

[0004] US Patent 2008 / 0019701 A1 discloses a device for recording rail traffic at a specific location on a track, for example, to determine the number of axles, speed, time of passage, or wheel load of a rail vehicle. The information obtained at specific points can help to observe or document rail traffic. Such a device allows, at best, only limited conclusions to be drawn about the condition of the rail transport system, particularly regarding condition information that affects the operational safety of the rail transport system.

[0005] It is an object of the invention to create an improved method for monitoring the integrity of a railway transport system, which in particular ensures increased operational safety, provides relevant information on the railway transport system in a comprehensive manner and is economically feasible.

[0006] This problem is solved by a method for monitoring the integrity of a railway system with the features of claim 1. It has been recognized that the integrity of the railway system can be monitored particularly comprehensively and efficiently if the strain sensors are components of a linear detection device attached to the track rail and extend along the track rail over a measuring distance of at least 100 m, wherein the integrity is determined by means of an electronic evaluation unit that is in signal communication with the multiple strain sensors.Because the measuring section extends over a length of at least 100 m, in particular at least 200 m, in particular at least 500 m, in particular at least 1 km, and / or a maximum of 10 km, it becomes possible for the first time to monitor the integrity of the rail transport system, especially a track section, not just at specific points, but holistically and / or continuously, in particular continuously over time and / or space. Such a method leads to a broad information base on the rail transport system and, based on this, to a reliable determination of the integrity of the rail transport system. With correspondingly reliable knowledge about the integrity of the rail transport system, the operational safety of the rail transport system and its maintenance efficiency can be significantly increased.

[0007] The integrity of the rail transport system refers to its unaltered state. Monitoring this integrity can also be described as detecting damage.

[0008] The rail transport system comprises infrastructure components, in particular the track superstructure, especially a track grid, especially rails and sleepers, and / or a ballast bed and / or an overhead line, and / or at least one, in particular several, rail vehicles, especially transport vehicles and / or work vehicles. The method is preferably designed to monitor the integrity of at least one of the components of the rail transport system, in particular to detect measurement signals that correlate with the integrity of at least one of the components of the rail transport system.

[0009] The linear detection element can be a single piece, in particular not divisible without damage, and / or continuous. The linear detection element can have at least one and / or at most two connections for connecting to the electronic evaluation unit. The at least one connection can be configured for transmitting optical and / or electrical signals. The at least one connection can be located at a free end and / or between the free ends of the linear detection element. A connection located between the ends can be configured as a bending coupler. At least one connection can be configured as a monitoring connection.

[0010] The length of the measuring section is preferably measured along the linear detection element and / or along the track rails and / or is determined by the distance between the first and last strain sensor of the linear detection element. The length of the measuring section is preferably at least 50 m, in particular at least 100 m, in particular at least 250 m, in particular at least 500 m, in particular at least 1 km, in particular at least 3 km, and / or a maximum of 20 km, in particular a maximum of 10 km, in particular a maximum of 5 km.

[0011] The electronic evaluation unit preferably comprises an electronic processing unit, in particular a processor, for acquiring and / or processing the measurement signals from the multiple strain sensors. The evaluation unit can also be configured to determine the integrity of at least one component of the railway system. The evaluation unit preferably has at least one, in particular at least two, and / or a maximum of ten, in particular a maximum of five, in particular a maximum of three, connections for connecting to the acquiring device and / or for receiving the measurement signals.

[0012] The linear detection means preferably comprises at least 50, in particular at least 100, in particular at least 250, in particular at least 500, in particular at least 1,000, in particular at least 2,000, and / or a maximum of 10,000, in particular a maximum of 5,000, in particular a maximum of 3,000, strain sensors.

[0013] The evaluation unit can be designed to receive and / or evaluate the measurement signals from at least one, in particular at least two, and / or a maximum of 10, linear detection devices, in particular their strain sensors, in particular all strain sensors of the respective detection device.

[0014] The distance between two adjacent strain sensors can be in a range of 0.1 m to 10 m, particularly from 0.2 m to 5 m, particularly from 0.3 m to 2 m, and particularly from 0.4 m to 1.4 m. Preferably, the distance between two adjacent strain sensors is dimensioned such that they can be arranged on the track at the distance of two adjacent sleepers and / or at half the distance between two adjacent sleepers. The distance between the strain sensors can be increased by a compensation factor, particularly in a range of 1% to 50%, and particularly from 5% to 20%, of the target distance, especially the sleeper spacing, in order to ensure precise positioning of the strain sensors even with tolerances in the sleeper spacing.

[0015] The evaluation unit may have a data interface, in particular a network interface, in particular a wired interface, in particular a USB interface, and / or a wireless interface, in particular a GPRS interface and / or a WiFi interface, and / or its own power supply, in particular an electrochemical energy storage device and / or a photovoltaic system and / or an energy harvesting device, and / or an electrical connection, in particular for supplying the electrical energy required for operation, and / or a protective housing and / or a data storage device, in particular for temporarily storing the measurement signals and / or the information on the integrity of the railway system and / or reference values ​​of the measurement signals, and / or for storing a computer program product.in particular for automatically executing the procedure for determining the integrity of the rail transport system and / or for executing statistical procedure steps, in particular for calculating the average and / or for determining a standard deviation, and / or having a processor for processing the measurement signals, in particular with a clock frequency of at least 8 MHz, in particular at least 1 GHz and / or a maximum of 100 GHz.

[0016] The evaluation unit and / or a central processing unit can be configured to determine integrity based on a structural-mechanical and / or numerical model, in particular a finite element method (FEM) model and / or discrete element method (DEM) model, and / or analytical equations. These are preferably implemented in a computer program and / or stored on a memory unit of the evaluation unit and / or the central processing unit.

[0017] The linear detection means preferably comprises at least two, in particular at least five, in particular at least 10 and / or a maximum of 50, in particular a maximum of 20 signal lines, in particular optical and / or electrical signal lines. The linear detection means can be designed as a measuring tape, in particular with signal lines arranged adjacent to each other, in particular side by side.

[0018] The multiple signal lines can be materially bonded together to form the detection means and / or arranged within the same covering material, in particular insulating sheath, and in particular be bonded together by it.

[0019] Preferably, the linear detection means is further developed with at least one of the features described in PCT / EP2024 / 070372, the contents of which are incorporated herein by reference.

[0020] Preferably, the linear detection means, in particular the multiple strain sensors, is designed to detect strains of the track grid, in particular at least one of the track rails.

[0021] The linear detection device can be attached to at least one track rail, in particular by clamping and / or gluing it, especially in the area of ​​the rail fastening or the track sleepers and / or below and / or on and / or within the track sleepers, in particular by fastening it, especially in such a way that the condition of the track sleepers or their integrity can also be monitored.

[0022] To determine the integrity of at least one component of the rail transport system, the measurement signal from at least one strain sensor can be evaluated individually, particularly at a specific point in time and / or over time. A single value can provide information about the strain, for example, of the track rail. The time course can reveal a continuous change in integrity, such as the growth of a crack in a track rail and / or a displacement of the track grid and / or a sudden change, for example, due to the loosening of a rail fastening and / or a derailment and / or a rail break and / or buckling of at least one rail, particularly in the horizontal direction, especially inwards and / or outwards.

[0023] Preferably, when determining integrity, a distinction is made between a permanent strain change, particularly one with more serious effects on the integrity of the railway system, and / or a temporary or sudden strain change, particularly one with lesser effects on integrity, for example, due to a rail vehicle passing over the track, which leads to normal, sudden strain of the rails. Sudden strain can also be referred to as impulsive strain. Sudden strain is understood to mean that the strain increases and then decreases again, particularly to the original level before the increase, or vice versa.

[0024] The strain sensors can be configured, in particular exclusively, as fiber optic strain sensors, especially as fiber Bragg grating sensors. A multitude of such sensors can be integrated onto a single optical fiber. This allows for reliable signal transmission over particularly long distances, especially several hundred meters. Alternatively or additionally, at least one of the strain sensors can be configured as a piezoelectric and / or electrical strain sensor, especially as a strain gauge.

[0025] According to one aspect, integrity is determined by evaluating time-dependent changes, particularly the rate of change, of at least one of the measurement signals. Preferably, it is further considered whether the change is a steady or abrupt change and / or the magnitude of the change in the measurement signal. Large, steady changes are usually particularly critical, especially at high rates of change. A high strain, particularly a high proportion of steady strain, is present, for example, at at least 0.01%, particularly at least 0.1%, particularly at least 0.2%, particularly at least 0.5%, and particularly at least 1%. A high rate of change is present, for example, at at least 1% / day, particularly at least 1% / hour, particularly at least 1% / second, and particularly at least 100% / second.High, and especially permanent, strains at high rates of change can be caused, for example, by a crack in the track rail, a derailment, an object on the track, especially a tree trunk, and / or a fault in the subsoil and / or debris flow and / or a mudslide. Based on the measurement signals, a corresponding impairment of integrity, especially of the infrastructure component, can be determined.

[0026] Integrity can also be determined by evaluating the time-dependent change in the amplitude of an oscillating measurement signal, in particular the decay rate of the oscillation of the measurement signal and / or the duration of the after-oscillation after a load on the track rail.

[0027] The extent of the impairment of integrity can be determined more precisely by evaluating the local extent of the strain change, particularly its permanent nature. A fault in the subsurface would lead to a change in the measurement signals of strain sensors over a range of at least several meters and / or a maximum of 1 km, particularly a maximum of 500 m, whereas an obstacle, such as a tree trunk, would lead to a strain change with only a few strain sensors within a range of a few meters around the obstacle, particularly with a single strain sensor.

[0028] According to another aspect, the integrity of the track rails can be determined under the load of a passing rail vehicle, in particular a passenger vehicle and / or a work vehicle. The reaction of the track rail to the load from the rail vehicle allows conclusions to be drawn about the condition, in particular the stiffness, of the track superstructure, in particular the track grid, in particular the rails and / or the sleepers, and / or the track bed, and / or the track substructure, and / or the condition of the rail vehicle, in particular the bogie and / or the wheels. Preferably, the weight of the rail vehicle is known, so that the properties, in particular the integrity, of the track rails can be determined with particular precision. The rail vehicle can be a measuring vehicle with a known, in particular calculated, weight, and in particular a known, in particular calculated, wheel load.

[0029] A particular aspect, especially one considered independent or eligible for independent patent protection, concerns a method for determining the weight of a rail vehicle, comprising the steps of: acquiring measurement signals from at least one strain sensor on a track rail as the rail vehicle passes over the track rail, and determining the weight of the rail vehicle based on these measurement signals. The measurement signals preferably correlate with the strain, in particular the longitudinal strain and / or the bending strain, of the track rail and / or the vertical force exerted on the track rail by the rail vehicle. Information about the geometry of the track rail, in particular the rail profile, and about the material of the track rail, in particular its stiffness or modulus of elasticity, can be used to determine the weight of the rail vehicle.The weight of a rail vehicle can be determined by calculating the individual wheel loads and summing them. By running a rail vehicle of known weight along the track and recording the resulting measurement signals, the weight determination can be calibrated. This method of weight determination is particularly precise and requires minimal effort. The procedure for determining the weight of the rail vehicle can be further developed with at least one of the features described in connection with the procedure for determining the integrity of the rail transport system, and / or vice versa.According to another aspect, integrity can be determined using a combination of several measurement signals, in particular multiple measurement signals from the same strain sensor, especially at different times, and / or multiple strain sensors, especially at different times and / or at the same time. Evaluating the multiple measurement signals to determine integrity can be used, especially with statistical methods, to increase reliability and / or to more precisely determine the nature of the integrity impairment or damage. Alternatively, a single measurement signal can be evaluated to determine integrity. This reduces the required computing power. For example, the single measurement signal or the multiple measurement signals can be compared with one or more limit values ​​and / or reference values.An impairment of integrity can be determined if corresponding limit values ​​are exceeded and / or if there is a certain degree of change compared to the respective reference value.

[0030] According to one aspect, integrity can be determined, particularly using the electronic evaluation unit, based on the measurement signals and a machine learning method, especially artificial intelligence. For this purpose, a corresponding algorithm can be stored on or executed by the evaluation unit and / or the central processing unit. Such a method can be trained using training data, which enables the detection of anomalies in the measurement signals, particularly integrity anomalies. Specifically, the training data for various types of damage to the rail transport system can include the measurement signals associated with each type of damage, particularly those acquired using strain sensors. This allows for the particularly efficient and reliable evaluation of large data volumes, especially when dealing with a large number of strain sensors.Integrity can be determined in particular using a method of unsupervised learning.

[0031] For unsupervised learning, raw data is sufficient as training data; that is, data in which peculiarities, especially integrity anomalies and / or integrity errors, are not identified or marked. Such raw data can also include pre-processed data, in particular data consisting of pre-processed data, where pre-processing can include, for example, normalization of the data format and / or pre-sorting and / or pre-filtering, especially pre-sorting or pre-filtering according to measurement signal values ​​and / or their acquisition time and / or location. This eliminates the need for manual effort in providing appropriately marked training datasets. The training data can then be grouped (clustered), particularly automatically. Based on this grouping, an assessment can be made as to whether a peculiarity, especially an integrity anomaly, is present.Data points that cannot be assigned to any group can be flagged as peculiarities, particularly as anomalies. For example, a k-means algorithm can be applied, preferably with a predefined number of classes or groups, and / or a DB SC AN algorithm can be used, preferably with a density-based classification of the data, particularly the training data, especially without requiring a known or predefined number of groups. Those data points, particularly the training data, that cannot be assigned to any group can, for example, be identified as "noise," which can indicate a peculiarity, particularly an integrity anomaly. An autoencoder method can be applied. A neural network can be fed with or trained on data, particularly for identifying unknown patterns of peculiarities.In this process, the data, especially the training data, can be compressed and reconstructed. A peculiarity, particularly an integrity anomaly, can be identified if an error, especially a "reconstruction error," is determined, particularly if it exceeds an error threshold. This allows the identification of anomalies that arise from the measurement signals only based on complex relationships, especially when considering multidimensional dependencies.

[0032] The data, particularly the training data, preferably consists of measurement signals acquired from the track rail of the railway system, especially by multiple strain sensors. The data can include corresponding measurement signals for a specific track section and / or for a specific period.

[0033] Particularly using the linear detection device attached to the track rail, especially fiber optic strain and / or temperature sensors, very extensive data can be collected, especially over long distances along the track rail and / or over extended periods, particularly continuously. Such extensive data can be efficiently analyzed using machine learning, especially unsupervised learning, and / or used as training data for machine learning. The analysis can be performed in conjunction with other infrastructure measurement data, particularly based on a variety of different measurement data that correlate with infrastructure load or integrity, especially the infrastructure's response. For example, track alignment data, images, 3D survey data, ground-penetrating radar data, traffic load data, subsidence measurement data, and / or acceleration data can be analyzed.This allows for the identification of outliers or anomalies, preferably using at least one of the methods mentioned above. These areas can be flagged, particularly automatically, and / or reviewed by an expert based on automated seed mapping and / or pre-selection. This enables the evaluation and analysis of very large datasets, and in particular the reliable identification of previously unknown relationships. Specifically, it allows for the identification of anomalies that are not already known and / or that are based on complex relationships.

[0034] Alternatively or additionally, a supervised learning method can be considered, although this involves more effort and requires more extensive manual review. Preferably, manually processed data, particularly training data, is provided in which specific features, especially integrity anomalies and / or integrity errors, have been manually identified, particularly marked. Semi-automated processing of training data, especially with automated preselection – as described above – can also be advantageous.

[0035] These methods are equally suitable for determining the integrity of track and rolling stock. Preferably, the measurement signals from the strain sensors are initially acquired and stored, particularly in a state where the rail transport system, especially the at least one infrastructure component, is undamaged. Corresponding measurement signals or information derived from them can be stored as reference values, particularly in a storage unit of the evaluation unit or in a central processing unit, especially a control center.

[0036] Integrity can be determined, in particular, using multiple measurement signals from adjacent strain sensors. This allows the impact of a passing rail vehicle on the strain of the track rail to be tracked. This makes it possible to distinguish whether a defect is present in an infrastructure component and / or the rail vehicle. In contrast to an infrastructure defect, strains due to a defect in the rail vehicle, especially a wheel and / or wheel bearing, occur cyclically, particularly at intervals that are multiples of the wheel circumference. Local damage to an infrastructure component, especially the track rails or the track superstructure, on the other hand, does not generally lead to a cyclical exaggeration of the impact-like strain changes along the track rail as a vehicle passes by.

[0037] According to another aspect, the integrity can be determined by evaluating the frequency of a track rail vibration and / or by static or quasi-static loading of the track rail. The vibration can be excited by a rail vehicle traveling over the track rail and / or a test force acting on the track rail, for example, by impacts on the track rail and / or cyclic loading of the track rail, in particular by means of a dynamic track stabilizer. The excitation of the vibration and / or the static loading can be carried out by means of at least one test force device that provides the test force. The test force is preferably a defined force, in particular a known, in particular a set and / or adjustable, in particular controllable and / or adjustable, force. The test force preferably has a horizontal and / or upward and / or downward force component. In particular, the test force orThe specific load can be oriented exclusively horizontally and / or vertically. The test force or the specific load can be directed exclusively upwards and / or downwards, in particular with or without a horizontal force component. The test force can be static or quasi-static and / or dynamic, in particular as an oscillating and / or fluctuating load, or in particular only static or only dynamic. The at least one test force device can be designed accordingly. The test force can be applied directly to a track rail, a track sleeper, and / or the track bed, in particular via a test force contact device, for example, a gripper and / or a plunger and / or a roller. The recorded frequency provides information about the stiffness of the track superstructure, in particular the track grid.Stiffness, in turn, is an indicator of integrity, particularly of the track rail, especially whether there are cracks, depressions, breakouts and / or wave-like defects, and / or of the rail fastening and / or the track bed, especially whether the desired compaction state of the ballast bed is present.

[0038] Based on the wave propagation time in the ballast bed, which can be determined from the measurement signals, conclusions can be drawn about ballast degradation and / or the compaction state of the ballast bed. Furthermore, determining integrity can encompass determining the integrity of a running gear, in particular a rail wheel and / or a rail vehicle. For this purpose, the measurement signals from several adjacent strain sensors are preferably evaluated. A cyclical increase in strain change, especially compared to strain changes of neighboring strain sensors, along the track or measurement section can indicate that a cyclic impact of a rail wheel on the track rail is the cause. Bearing damage or a flat spot on the rail wheel can thus be detected based on the measurement signals.

[0039] The position of the defective bogie, especially the faulty wheel and / or wheel bearing, is particularly advantageous when determining the location of the damaged bogie on the rail vehicle. For this purpose, the axles and / or wheels of the passing rail vehicle can be counted using strain sensors, particularly based on the strain pulses. This allows the identification of which wheel is causing the vibration leading to the excessive strain. In a single maintenance step, the defective bogie, especially the wheel and / or wheel bearing, can thus be repaired in a targeted and therefore particularly economical manner.

[0040] According to another aspect, determining integrity can include determining the integrity of the track rail, in particular detecting surface irregularities, especially depressions and / or breaks, particularly periodic depressions or surface waviness, especially on the running surface of the track rail. Depressions on the running surface of the track rail are also referred to as squats. Damage to the track rail can be determined, in particular, by observing that, when a rail vehicle passes over it, a higher strain occurs at the location of the damage than at adjacent strain gauges, and / or by observing that, when several rail vehicles pass over it, the increased strain occurs in a region of the same strain gauges, especially the same strain gauge.A permanent stretching of individual strain sensors also suggests local damage to the track rail in the area of ​​these strain sensors.

[0041] Another aspect that can be identified as an impairment of integrity or damage in the form of an insufficiently compacted ballast bed, in particular a void beneath a track sleeper. An insufficiently compacted ballast bed is the cause of increased elongation of the track rail when a rail vehicle passes over it.

[0042] According to another aspect, a rail crack can be determined by detecting an interruption in the signal line via the linear detection device. An interruption in the signal line can be detected by regularly checking the continuity of a signal through a signal line of the detection device, particularly along the entire measuring section. This signal could be, for example, an electrical signal and / or a light signal, especially a signal pulse. For this purpose, a signal, particularly one reflected by the strain gauge, can be detected by the evaluation unit and / or at an end of the detection device opposite the evaluation unit, particularly by a test unit.If continuity is not given, there is an interruption which may be caused by a rail crack, especially if the detection device is attached to the track rail, in particular glued and / or clamped to it.

[0043] In the case of a detected rail crack, the crack growth or the future crack size can preferably be predicted, in particular based on the change in strain over time, and / or the end of the operational capability of the track and / or a suitable, in particular the latest possible, time for a rail replacement and / or a rail maintenance measure can be determined.

[0044] According to another aspect, a rail crack, in particular a partial and / or complete break, can be determined by comparing the measurement signal of at least one strain sensor with a strain threshold value and / or with at least one previous measurement signal of the at least one strain sensor and / or with the measurement signal of at least one other strain sensor. Preferably, the determination is carried out during a change in load, in particular due to the passage of a rail vehicle and / or due to a temperature change and the resulting change in thermal expansion.

[0045] According to another aspect, determining integrity can include determining the integrity of the rail fastening. In particular, a break in the rail fastening, a missing rail fastening, and / or a lack of clamping force in the rail fastening can be detected.

[0046] According to another aspect, the integrity of the rail fastening can be determined by loading the track rail, in particular with a test force oriented obliquely, especially perpendicularly, to the longitudinal direction of the rail, and / or by evaluating the resulting measurement signal from at least one of the strain gauges. Preferably, the integrity is determined by means of cross-correlation and / or normalized cross-correlation and / or by comparing the maximum values ​​of at least two measurement signals. The at least two measurement signals can be measurement signals from adjacent strain gauges, in particular along the same track rail and / or adjacent track rails. The evaluation of the measurement signals from adjacent strain gauges preferably takes place at essentially the same time. This makes it possible to determine how the load propagates along the track grid, in particular between two measuring points, and especially across at least one track sleeper.The two strain sensors, whose measurement signals are evaluated, can be located on the same track sleeper. If the rail fastening is damaged, particularly if it is loose, the load is transferred with reduced stiffness between the two strain sensors compared to an intact rail fastening. This load can then be applied to the track grid by a passing rail vehicle and / or a dynamic track stabilizer.

[0047] According to one aspect of the invention, a damage location can be determined using a data set that assigns positional information, particularly along the track and / or in a global coordinate system, to the respective strain sensor. For example, a specific kilometer marker of a track and / or a position in a geocentric coordinate system can be assigned to the respective strain sensor. The data set can be stored in the evaluation unit, in particular in a storage unit thereof. Alternatively, the data set can be stored in a central processing unit, in particular in a control center.

[0048] According to another aspect, maintenance measures can be initiated or carried out based on the specific integrity of the rail transport system, and / or the rail transport system can be controlled depending on this specific integrity. For example, a rail vehicle can be stopped and / or operated at reduced speed if damage is detected to an infrastructure component, particularly a track rail, and / or to a rail vehicle, particularly the undercarriage.In the event of a defect in the integrity of an infrastructure component, a speed limit may be imposed on the affected track section, the track may be closed, and / or maintenance measures, such as compacting the ballast bed and / or repairing the rail and / or the rail fastening, may be initiated, particularly manually by monitoring personnel and / or automatically based on the determined integrity level. Early initiation of maintenance measures can prevent damage from accelerating. Reduced speed or the closure of track sections can contribute to accident prevention.

[0049] In principle, the linear detection device, especially fiber optic, can include strain and / or temperature sensors.

[0050] To determine the integrity of the railway system, changes in the measurement signals, particularly strain, in the unloaded state of the track rail, especially over time, particularly over a period of several weeks and / or months, and / or changes in the measurement signals, particularly strain, in the loaded state, especially under a specific load, of the track rail compared to another loading state, especially under a different specific load, or compared to an unloaded state, can be evaluated. The relationship between the change in the measurement signal and a specific change in load is also referred to as sensitivity. Integrity can be determined, in particular, based on the sensitivity.

[0051] Preferably, other parameters that could potentially influence the measurement signals, such as temperature, are taken into account when determining integrity. Temperature measurement can therefore be an essential part of the method. Certain evaluations can be performed without temperature measurement. In particular, the temperature of the track sections can be inferred from measured strains.

[0052] According to an advantageous embodiment, the detection of strains, in particular the strain sensors, can be combined with one or more temperature sensors, in particular point and / or line temperature sensors, in particular by means of Distributed Fiber Optic Sensing, preferably in a single line-shaped detection means or in separate, in particular line-shaped, detection means.

[0053] Temperature measurement can be performed continuously along a line, particularly using distributed fiber optic sensing, and / or by evaluating, particularly validating, the signals from the fiber Bragg grating sensors, especially point-like ones. The measured temperatures can improve the evaluation and / or be used as additional information for a machine learning method.

[0054] A further object of the invention is to create an improved monitoring device for monitoring the integrity of a rail transport system, which in particular ensures increased operational safety, provides relevant information on the rail transport system in a comprehensive manner and is economically feasible.

[0055] This problem is solved by a monitoring device having the features of claim 16. The advantages of the monitoring device preferably correspond to the advantages of the method described above. The monitoring device can be further developed with at least one of the features described above in connection with the method.

[0056] The invention also relates to a computer program product comprising instructions that cause a computer to execute the method described above, and a storage unit containing such a computer program product.

[0057] Further features, details, and advantages will become apparent from the following description of an exemplary embodiment with reference to the figures. These show:

[0058] Fig. 1 shows a schematic representation of a railway system with a monitoring device for monitoring the integrity of the railway system, comprising a linear detection means on a track rail with several strain sensors and an electronic evaluation unit for determining the integrity based on measurement signals from the strain sensors.

[0059] Fig. 2 shows a schematic representation of the rail transport system with the detection device attached to the track rail, further details shown.

[0060] Figures 1 and 2 describe an embodiment of a monitoring device 1 for monitoring the integrity of a rail transport system 2.

[0061] The rail transport system 2 comprises a rail network with a multitude of track sections 3. Each track section 3 has a track bed 4 with track sleepers 5 and track rails 6 arranged on it. The track bed 4 rests on a ballast, in particular a ballast bed 7, and / or a solid, in particular ballastless, subsoil.

[0062] Preferably, the track section 3 includes an overhead line 8 for supplying rail vehicles 9 with electrical energy. The respective rail vehicle 9 can be a transport vehicle 9.1 or a work vehicle 9.2, in particular a track construction vehicle, a maintenance vehicle and / or a measuring vehicle. The respective rail vehicle 9 includes a bogie 10 with track wheels 11 on several axles 12.The rail transport system 2 may have damage, in particular to one of the rail vehicles 9, in particular a roundness defect, in particular a flat spot 13 and / or a bearing defect, to one of the wheels 11, and / or to an infrastructure component, in particular to one of the track rails 6, in particular a rail crack 14 and / or a surface irregularity, in particular a depression 15 and / or a breakout, in particular periodic depressions 16, and / or a faulty rail fastening 17, for example a loose or missing rail fastening.

[0063] The integrity of the rail transport system 2 can also be impaired by an object 18 blocking the track, in particular if it is located or lies on the track section 3, in particular on the track grid 4, in particular on the track sleepers 5. The object can be, for example, rubble 18, soil, ice, snow, a plant component, in particular a tree trunk, a tool, a machine, in particular a rail vehicle 9 and / or a maintenance machine 9.2 and / or a living being, in particular an animal and / or a person.

[0064] The ballast bed 7 may also exhibit a lack of integrity in the form of insufficient ballast bed compaction, in particular a void 28.

[0065] The monitoring device 1 has a linear detection means 19 and an electronic evaluation unit 20.

[0066] The linear sensing means 19 comprises a plurality of strain sensors 21. The strain sensors 21 are preferably designed as fiber-optic sensors, in particular as fiber Bragg grid sensors. Preferably, the linear sensing means 19 comprises at least 20, in particular at least 50, in particular at least 100, in particular at least 200, in particular at least 500, in particular at least 1,000, and / or a maximum of 10,000, in particular a maximum of 5,000, in particular a maximum of 3,000, of the strain sensors 21.

[0067] The detection device 19 can have one or more optical fibers. Preferably, at least one, and in particular exactly one and / or more, of the optical fibers is equipped with strain sensors 21 and / or, in particular for transmitting data, without a strain sensor 21.

[0068] The strain sensors 21, in particular the number mentioned above, are preferably formed on a one-piece or uninterrupted or continuous sensing means 19, in particular by a single or more optical fibers.

[0069] The detection device 19 can also include at least one, in particular several, electrical signal lines.

[0070] The multiple optical fibers and / or the at least one electrical signal line can be bonded together to form the sensing device 19, in particular by means of a carrier material, especially an insulating material, in particular by being embedded in it and / or encased by it and / or bonded together by it. The evaluation unit 20 can be in signal communication with the multiple strain sensors 21, in particular to evaluate the measurement signals acquired by them. For this purpose, the evaluation unit 20 is preferably in signal communication with all strain sensors 21 of the sensing device 19 and / or with at least 50, in particular at least 100, in particular at least 200, in particular at least 500, and / or a maximum of 10,000, of the strain sensors 21.

[0071] The evaluation unit 20 can be in signal communication with a central computing unit 22, in particular a control center 23, of the rail transport system 2. The signal communication 24 is preferably wireless. For this purpose, the evaluation unit 20 can have a radio module 25.

[0072] Preferably, the monitoring device 1 comprises two linear detection means 19, each attached to one of the track rails 6. The evaluation unit 20 can be in signal communication with the strain sensors 21 of one or more, in particular both, detection means 19.

[0073] The work vehicle 9.2 can be a measuring vehicle, which is specifically designed to apply a defined load to the track rails 6, particularly by means of the track wheels 11, particularly in a vertical direction, and particularly with a known weight force. The work vehicle 9.2 has a loading unit 26, which is separate from the track wheels 11, for exerting a test force F on the track section 3, particularly the track grid 4, and in particular on the at least one track rail 6, which has a force component oriented horizontally and perpendicular to the longitudinal direction 30 of the rail. The loading unit 26 can, for example, be a dynamic track stabilizer. The work vehicle 9.2 can have at least one track tamping unit 27 for compacting the ballast bed 7. Preferably, the work vehicle 9.2 is a track tamping machine.

[0074] The work vehicle 9.2 can be in signal communication with the control center 23, in particular via wireless signal communication.

[0075] The detection device 19, in particular the strain sensors 21, extends over a measuring distance 31 with a length L in a range of 100 m to 10 km, in particular from 500 m to 5 km, in particular from 1 km to 3 km.

[0076] The operation of the rail transport system 2, in particular the monitoring device 1, is as follows:

[0077] The linear detection devices 19 are attached to the track rails 6, particularly in the area of ​​the rail fastenings, and are in particular clamped and / or bonded in place. The strain sensors 21 provide measurement signals, in particular in the form of light signals. The measurement signals from the strain sensors 21 are read out and processed by the evaluation unit 20.

[0078] Preferably, the measurement signals from the strain sensors 21 are evaluated and stored in a state in which the rail transport system 2, in particular the infrastructure components 5, 6, 7, is undamaged. Corresponding measurement signals can be stored as reference values ​​in a storage unit 29 of the evaluation unit 20 or the control center 23. The measurement signals from the strain sensors 21 are preferably read out continuously and / or at specific and / or manually and / or automatically determined times and / or at fixed intervals, for example, at a frequency of at least 1 / day, in particular at least 1 / hour, in particular at least 1 / minute, in particular at least 1 Hz, in particular at least 10 Hz, in particular at least 100 Hz, in particular at least 1 kHz, and / or at a maximum of 1 MHz, in particular at most 1 kHz.

[0079] The integrity of the rail transport system 2 can be determined based on the measurement signals, in particular according to the frequency of the measurement signal acquisition.

[0080] Whether the integrity, i.e., the completeness, or damage of the rail transport system 2 exists can be determined by comparing the respective, current measurement signal of a strain sensor 21 with other measurement signals, in particular at least one adjacent strain sensor 21 and / or with a previous measurement signal, in particular at least one of the reference values, in particular from the same strain sensor 21 and / or another strain sensor 21 and / or with a time course of one of the measurement signals, in particular with the rate of change of the measurement signal. For example, the rail crack 14 results in an increase in strains in the longitudinal direction 30 of the rail, in particular across the remaining cross-section of the respective track rail 6. This increase in strain can be detected by the measurement signal of a strain sensor 21 located adjacent to the rail crack 14, in particular by comparison with the reference value of the same strain sensor 21.Damage to the rail transport system 2 can therefore be detected using the measurement signals.

[0081] Another way to detect damage, particularly a rail crack 14, using the measurement signals is to monitor the signal line through the linear detection element 19. Depending on the extent of the rail crack 14, the signal line may be severed by the linear detection element 19. Monitoring whether the linear detection element 19 is conductive or continuous thus enables the detection of corresponding rail cracks 14. This monitoring can be carried out by passing a test signal, particularly a light pulse, through the detection element 19, especially at regular intervals, and checking its continuity by detecting the test signal after it has passed through the detection element 19. For this purpose, a test unit can be provided at one end of the detection element 19 opposite the evaluation unit 20.Alternatively, a test and / or measurement signal transmitted to the evaluation unit 20, in particular transmitted back, in particular reflected, by at least one of the strain sensors 21, in particular the strain sensor 21 located furthest from the evaluation unit 20, can be tested.

[0082] In particular, the evaluation unit 20 can be used to monitor the rate of change, especially in addition to the absolute change in the measurement signals, particularly the strains. For example, a rapid change in the measurement signal, especially in the case of a large, sustained absolute change in the measurement signal, can be interpreted as indicating that a significant change has occurred in the structure of the rail transport system 2, especially in an infrastructure component 5, 6, 7, for example due to a rockfall or mudslide 18, a landslide, a derailment, a fallen tree, a loosened rail fastening 17 and / or a rail crack 14, especially a break or tear. Corresponding integrity defects can thus be detected by evaluating the measurement signals.

[0083] Slow changes in the measurement signals may be caused by thermal expansion and are not critical, especially if they rise and fall with temperature. Slow changes in the measurement signals, particularly expansion, with a one-sided tendency may indicate aging, especially crack growth and / or wear, particularly insufficient compaction and / or debris flow and / or undermining of the track superstructure, especially the ballast bed 7. Undermining of the track superstructure may also be present if a change in expansion, especially regardless of its rate, is detected without any discernible effect on the track section 3.

[0084] In the case of a detected rail crack 14, the crack growth or the future crack size can preferably be predicted, in particular based on the change in strain over time, and / or the end of the operational capability of the track section 3 and / or a suitable, in particular the latest possible, time for a rail replacement and / or a rail maintenance measure can be determined.

[0085] The location of each integrity defect can preferably be determined using a data set that assigns corresponding position information to the respective strain sensor 21. The data set can be stored in the memory unit 29. The integrity defect, in particular damage, determined on the basis of the at least one measurement signal can thus be uniquely assigned to the location of the respective detecting strain sensor 21. This makes maintenance measures particularly targeted and economical.

[0086] Another possibility for determining the integrity of the rail transport system 2 is to evaluate the measurement signals when the track section 3 is loaded by a rail vehicle 9, in particular by the weight of the rail vehicle 9, which is transferred in particular via the track wheels 11 to the track rails 6, and / or by the test force F, which is exerted on the track grid 4, in particular the track rail 6, by means of the loading unit 26.

[0087] A depression 15 can be detected if, during the passage of a rail vehicle 9, the measurement signal of a strain sensor 21 adjacent to the depression 15 exhibits a sudden, decreasing change that is, in particular, greater than the measurement signal changes of adjacent strain sensors 21 and / or than the reference value or the measurement signal changes of the same strain sensor 21 during passages of a rail vehicle 9 with similar force transmission from the rail wheel 11 at an earlier time when the depression 15 was not present. If the force transmission via the respective rail wheel 11 is known, a scaling, in particular weight-dependent, can be performed to make the current measurement signal comparable with previous measurement signals.

[0088] The presence of periodic depressions 16 can be determined by the fact that the excessive strains occur at several strain sensors 21, especially at the same distance from each other and / or with increasing or decreasing height.

[0089] A void 28 in the ballast bed 7, particularly under the track sleepers 5, and / or insufficient compaction of the ballast bed 7, can also be inferred from increased strain when a rail vehicle 9 passes over it. With insufficient compaction, the change in strain or the measurement signal is higher and / or slower than with a highly compacted, stiff ballast bed 7.

[0090] The wave travel time in the ballast bed 17, which can be determined from the measurement signals, can also be used to conclude that the ballast is degraded and / or that the ballast bed 7 is compacted.

[0091] A faulty rail fastening 17 can be detected, in particular, by measuring signals from the strain sensors 21 when a horizontal shear force and / or a vertical force, especially directed upwards, is applied to the respective track rail 6, particularly by means of the dynamic track stabilizer 26. The evaluation of the measuring signals can include cross-correlation of the measuring signals from several, especially adjacent, strain sensors 21.

[0092] To determine the integrity of the rail fastenings 17, the vibration frequency of the correspondingly excited track rail 6 can be evaluated. Particularly preferably, the measurement signal of a strain gauge 21 is evaluated on the track rail 6 opposite the track rail 6 subjected to a cyclic test force F. The test force F is transmitted to the opposite track rail 6 via the rail fastenings 17 and the track sleeper 5, thereby allowing fastening defects due to reduced transmission stiffness to be detected.

[0093] By evaluating the measurement signals, an integrity defect of the rail vehicle 9 can also be detected. The number of axles 12 displaced via the strain sensors 21 can be recorded, so that it is possible to determine whether a train separation has occurred.

[0094] It is also possible to determine whether the chassis 10 is damaged, in particular whether a wheel bearing and / or a rail wheel 11 is damaged. A flat spot 13 on a rail wheel 11, for example, leads to shocks that can be detected as sudden measurement signal deflections, which are greater in amplitude than the deflections due to the pure weight force. In contrast to the abrupt measurement signal deflections due to depressions 15, the measurement signal deflections due to flat spots 13 do not occur locally, but along the entire measuring section 31 and / or periodically according to the circumferential length of the respective rail wheel 11. This makes it clearly distinguishable whether there is damage to the rail vehicle 9 or to an infrastructure component 5, 6, 7.

[0095] Preferably, the rail traffic system 2, in particular the at least one rail vehicle 9, is controlled depending on the specified integrity, especially by means of the interlocking system 23. For this purpose, relevant information, in particular the measurement signals and / or integrity information, can be transmitted from the evaluation unit 20 to the control center 23. The integrity information can be checked to determine whether the safe operation of the rail traffic system 2 is ensured or not. If the integrity deficiency prevents safe operation, the affected rail vehicle 9 and / or the track 3 can be blocked from operation.

[0096] Depending on the result of the integrity assessment, track construction work, in particular maintenance work, can be initiated. For example, the work vehicle 9.2, in particular the track tamping machine, can be sent to the identified position of the insufficiently compacted ballast bed 7 in order to carry out ballast bed compaction.

[0097] In the case of flat spots 13, the corresponding rail wheel 11 can be repaired or replaced. In the case of depressions 15 or rail gaps 14, a repair can be carried out, which might no longer be possible if the damage worsens or could even lead to accidents. This reliably avoids more costly measures due to damage spreading and increasing.

[0098] Integrity monitoring can also be performed by detecting a derailment of a rail vehicle 9. If suddenly high, permanent strains are detected based on the measurement signals, and / or if the signal line of the detection device 19 is interrupted or disconnected, and / or if the expected strains do not occur due to the rail vehicle 9 moving along the track 3, a derailment can be assumed, particularly if several of the aforementioned conditions are present. Preferably, the derailment location is determined automatically based on the position of the strain sensors 21. Emergency measures, such as alerting rescue services, can be initiated automatically upon detection of a derailment. The method for monitoring the integrity of the rail transport system 2, or the monitoring device 1, ensures particularly safe operation.Because integrity defects can be detected early, maintenance measures can be implemented promptly and therefore particularly economically. Safety risks due to damage can be reliably avoided.

Claims

Patent claims 1. Procedure for monitoring the integrity of a railway transport system (2), comprising the steps: 1.1 Acquisition of measurement signals from several strain sensors (21) on a track rail (6) of the railway system (2) and 1.2 Determining the integrity of at least one component (5, 6, 7, 9) of the rail transport system (2) using the measurement signals, characterized in that 1.3 that the multiple deformation sensors (21) are components of a linear detection device (19) attached to the track rail (6) and extend over a measuring section (31) along the track rail (6) with a length L of at least 100 m and 1.4 that the integrity is determined by means of an electronic evaluation unit (20) which is in signal communication with the multiple damping sensors (21), in particular at least 100 of the damping sensors (21), to receive the measurement signals.

2. Method according to claim 1, characterized in that the integrity is determined by means of a machine learning method, in particular an unsupervised learning method, wherein this method is based on training data which makes it possible to determine an integrity anomaly.

3. A method according to claim 1 or 2, characterized in that the distance between two adjacent strain sensors (21) is in a range of 0.2 m to 5 m and that the linear detection means (19) comprises at least 500 of the strain sensors (21).

4. A method according to any one of the preceding claims, characterized in that the strain sensors (21) are fiber optic strain sensors (21), in particular fiber Bragg grating sensors.

5. Method according to one of the preceding claims, characterized in that the integrity is determined by evaluating the rate of change of at least one of the measurement signals.

6. Method according to one of the preceding claims, characterized in that the integrity is determined when the track rails (6) are loaded by a passing rail vehicle (9), in particular a transport vehicle (9.1) and / or a work vehicle (9.2).

7. Method according to one of the preceding claims, characterized in that the integrity is determined based on a combination of several of the measurement signals.

8. Method according to one of the preceding claims, characterized in that the integrity is determined by evaluating the frequency of a vibration of the track rail (6).

9. A method according to any one of the preceding claims, characterized in that determining the integrity comprises determining the integrity of a chassis (10) of a rail vehicle (9).

10. A method according to claim 9, characterized in that determining the integrity comprises detecting flat spots (13) of a rail wheel (11) of the rail vehicle (9).

11. Method according to one of the preceding claims, characterized in that determining the integrity includes determining the integrity of a track grid (4), in particular the track rail (6) and / or a track sleeper (5).

12. Method according to claim 11, characterized by determining a rail crack (14) by detecting an interruption of the signal line by the linear detection means (19).

13. Method according to claim 11 or 12, characterized by determining a rail crack (14) by comparing the measurement signal of at least one strain sensor (21) with a strain threshold value and / or with at least one previous measurement signal of the at least one strain sensor (21) and / or with the measurement signal of at least one other strain sensor (21).

14. Method according to one of the preceding claims, characterized in that determining the integrity includes determining the integrity of the rail fastening (17).

15. Method according to claim 14, characterized in that the integrity of the rail fastening (17) is determined by loading the track rail (6) with a test force (F) oriented obliquely to the longitudinal direction (30) of the rail and evaluating the resulting measurement signal of at least one of the strain sensors (21).

16. Method according to any of the preceding claims, characterized by determining a damage position based on a data set which assigns position information to each of the strain sensors (21).

17. Method according to one of the preceding claims, characterized by initiating maintenance measures based on the determined integrity of the rail transport system (2).

18. Method according to one of the preceding claims, characterized by acquiring temperature measurement signals from temperature sensors on the track rail (6) of the rail transport system (2) and determining the integrity taking into account the temperature measurement signals.

19. Method according to claim 18, characterized in that the temperature measurement signals are acquired by means of point and / or line-shaped fiber optic temperature sensors.

20. Monitoring device (1) for monitoring the integrity of a railway transport system (2), comprising 20.1 several strain sensors (21) on a track rail (6) of the rail transport system (2) and 20.2 an electronic evaluation unit (20) configured to determine the integrity of at least one component (5, 6, 7, 9) of the railway transport system (2) using measurement signals from the strain sensors (21), characterized in that, 20.3 the multiple strain sensors (21) are components of a linear detection means (19) attached to the track rail (6) and extend over a measuring section (31) along the track rail (6) with a length of at least 100 m, and 20.4 that the evaluation unit (20) is in signal communication with the multiple strain sensors (21), in particular at least 100 of the strain sensors (21), for receiving the measurement signals.

21. Monitoring device (1) according to claim 20, characterized in that the electronic evaluation unit (20) is configured to determine the integrity by means of a machine learning method, in particular an unsupervised learning method, wherein this method is based on training data that makes it possible to determine an integrity anomaly.

22. Monitoring device (1) according to claim 20 or 21, characterized in that the distance between two adjacent strain sensors (21) is in a range of 0.2 m to 5 m and that the linear detection means (19) comprises at least 500 strain sensors (21).

Images