Title of the invention: Anomaly detection device and method for railway vehicles
Technical field
[0001]
　The present invention relates to an abnormality detection device and method for railway vehicles.
Background technology
[0002]
　Demand for improved maintainability is increasing among railway operators, and technology for realizing CBM (Condition Based Maintenance) is required. As one of them, in recent years, in railway vehicles, in order to reduce vehicle operation / maintenance costs and risk management, for example, downtime reduction and maintenance, which is the time when systems and services are stopped during operation, are reduced. For the purpose of cost reduction, there is a need for an abnormality detection device or monitoring system that can detect signs of abnormalities in vehicle vibration comfort, for example, vehicle (vehicle body) vibrations at an early stage, and diagnose and determine the causes of the abnormalities. There is.
　As an abnormality detection method and apparatus, there is conventionally, for example, a technique described in Japanese Patent Application Laid-Open No. 2004-170080 (Patent Document 1).
Prior art literature
Patent documents
[0003]
Patent Document 1: Japanese Unexamined Patent Publication No. 2004-170080
Outline of the invention
Problems to be solved by the invention
[0004]
　Patent Document 1 describes, "A vehicle abnormality detection method for detecting vehicle vibration and detecting an abnormality in the vehicle and the track of the vehicle based on the detected vibration, and detecting the vibration of the bogie and the wheel set of the vehicle. , A vehicle abnormality detection method characterized by identifying whether the cause of vibration is in the bogie or in the track based on the detected vibration value of the bogie and the detected vibration value of the wheel set. " is there. That is, an abnormality detection method for a railroad vehicle and a vehicle for detecting an abnormality in a track is shown based on the vibration of the vehicle.
　That is, the method described in Patent Document 1 detects the acceleration of both the bogie and the wheel set of the railroad vehicle, filters the detected acceleration of the bogie and the detected wheel set, and the acceleration after noise processing is a predetermined threshold value. It is to identify whether the cause of the vibration is in the bogie or in the track depending on whether or not it exceeds.
[0005]
　However, the method described in Patent Document 1 is intended for the case where it is possible to clearly distinguish whether the abnormal factor is on the bogie or on the track, and the abnormal factor on the vehicle side and the abnormal factor on the track side are combined. The case where abnormal vibration occurs is not particularly considered, and there is a possibility that the abnormal cause may be erroneously detected.
　In order to improve maintainability, it is necessary to estimate the cause of the abnormality in a short time after detecting the sign of the abnormality.
　In order to detect and estimate the anomaly sign, conventionally, a dedicated sensor is required for each device in which an abnormality is expected, which causes a problem of high cost.
[0006]
　Therefore, in the present invention, when an abnormal vibration of a vehicle occurs, the vibration may be caused not only by an abnormal factor on the vehicle side or an abnormal factor on the track side, but also by a combined abnormal factor of both. It is an object of the present invention to provide an abnormality detection technology for railway vehicles that can detect and estimate, and can estimate an abnormality factor in a short time with a small number of sensors.
Means to solve problems
[0007]
　In order to solve the above object, one of the typical railcar abnormality detection devices and methods of the present invention acquires vehicle vibration data and operation data, and uses the acquired vibration data and operation data. , A device having a function of deriving an abnormal factor element by data analysis (for example, correlation analysis) and estimating or identifying an abnormal vibration factor of a railroad vehicle based on the abnormal factor element.
[0008]
　For example, in an abnormality detection device for detecting an abnormality of a railroad vehicle traveling on a track,
　a vehicle vibration factor estimation device for estimating a vehicle vibration factor of the 　railroad vehicle
　is provided, and the vehicle vibration factor estimation device is
the vibration of the railroad vehicle. The vibration data and the operation data detected by the vibration data detection device including the data detection unit for detecting the data and the operation data detection device including the data detection unit for detecting the operation data of the railway vehicle are acquired and analyzed. The data analysis unit that derives the abnormal factor element of the vibration of the railroad vehicle,
　the abnormal factor element estimation unit that estimates the abnormal factor element of the vibration derived by the data analysis unit based on the vibration data and the operation data, and the
　above.
　It is characterized by including an anomalous factor estimation unit that estimates anomalous factors of vibration in the railway vehicle based on the anomalous factor elements estimated by the anomalous factor element estimation unit .
Effect of the invention
[0009]
　According to the present invention, not only when the abnormal factor of the railway vehicle is on the vehicle side or the track side, but also when the abnormal factor on the vehicle side and the abnormal factor on the track side are combined, the number of sensors is small and the time is short. The cause of the abnormality can be estimated.
　For example, by analyzing vibration data and operation data to derive anomalous factor elements, it is possible to narrow down the anomalous factor elements, reduce the number of parameter survey cases, and shorten the time. In addition, the current value of the characteristics of the derived abnormal factor element is identified, and the abnormal factor is determined based on the comparison result between the current value and the normal value. When determining this anomalous factor, a mechanical quantity that has not been measured can be estimated by using a mechanical model, for example, a vehicle dynamic model (vehicle model) or a trajectory dynamic model (trajectory model).
A brief description of the drawing
[0010]
FIG. 1 is a system configuration diagram of an abnormality detection device for a railway vehicle showing the first embodiment of the present invention.
FIG. 2 is a system configuration diagram of an abnormality detection device for a railway vehicle showing another embodiment of the first embodiment.
FIG. 3 is a system configuration diagram of an abnormality detection device for a railroad vehicle showing still another embodiment of the first embodiment.
[Fig. 4] Fig. 4 shows the time history waveform of the left-right vibration acceleration of the vehicle body when abnormal vibration occurs in car b of the train set X due to an abnormality factor on the vehicle side in the abnormality detection device of the railway vehicle of the first embodiment. It is a figure shown.
FIG. 5 is a flowchart illustrating a processing procedure in the data analysis unit in the abnormality detection device for a railway vehicle according to the first embodiment.
FIG. 6 is a flowchart showing a processing procedure in the parameter estimation unit in the abnormality detection device for the railway vehicle of the first embodiment.
FIG. 7 is a diagram showing the results of comparing the measured value of the vehicle body left-right acceleration PSD with the analysis value of the vehicle body left-right acceleration PSD before and after the parameter estimation in the parameter estimation unit of the first embodiment. ..
FIG. 8 is a flowchart showing a processing procedure in an abnormality factor estimation unit in the abnormality detection device for a railway vehicle according to the first embodiment.
FIG. 9 is a diagram showing an abnormality factor estimation result in the abnormality factor estimation unit of the first embodiment.
[Fig. 10] Fig. 10 shows the time history of left-right vibration acceleration of the vehicle body when abnormal vibration occurs in each car of the train set X due to an abnormality factor on the track side in the abnormality detection device of the railway vehicle of the first embodiment. It is a figure which showed the waveform.
FIG. 11 is a diagram showing an abnormality factor estimation result in the abnormality factor estimation unit of the first embodiment.
FIG. 12 is a system configuration diagram of an abnormality detection device for a railway vehicle showing the second embodiment of the present invention.
[Fig. 13] Fig. 13 shows a vehicle body when an abnormal vibration occurs in a train set B due to a combination of an abnormal factor on the vehicle side and an abnormal factor on the track side in the abnormality detecting device for a railway vehicle of the second embodiment. It is a figure which showed the time history waveform of the left-right vibration acceleration.
FIG. 14 is a diagram showing an abnormality factor estimation result in the abnormality factor estimation unit of the second embodiment.
FIG. 15 is a system configuration diagram of an abnormality detection device for a railroad vehicle showing a third embodiment of the present invention.
FIG. 16 is a processing flow diagram in a parameter estimation unit for an abnormality detection device for a railway vehicle according to a third embodiment.
Mode for carrying out the invention
[0011]
　Hereinafter, the abnormality detection device for a railway vehicle of the present invention will be described with reference to the drawings.
Example 1
[0012]
　First, the configuration of the abnormality detection device for railway vehicles will be described with reference to FIGS. 1 to 3.
[0013]
　FIG. 1 is a functional block diagram showing a system configuration of an abnormality detection device for a railway vehicle, and FIGS. 2 and 3 are block diagrams of an abnormality detection device showing an improved example thereof.
　A railroad vehicle traveling on a track (rail) 20 is composed of a vehicle body 1 and a bogie 16.
[0014]
　The vehicle body 1 is mounted on the bogie 16 via an air spring 8. On the floor surface of the vehicle body 1, for example, a vibration data detection unit 60 including a vehicle body accelerometer (not shown) for measuring the left-right vibration acceleration of the vehicle body 1 is installed.
　The vehicle body accelerometer detects the voltage of the vehicle body accelerometer, and the voltage of the vehicle body accelerometer is detected by the vibration data detection unit 60, for example, the vehicle body left and right vibration acceleration signal as vibration data.
[0015]
　The vibration data detection unit 60 includes a vibration data detection unit 61 including a vibration data detection unit that detects vibration data indicating vibration of a railway vehicle.
　The vibration data detecting means 61 will be described later, but any means such as an acceleration sensor or a gyro sensor that can acquire vibration data may be used.
[0016]
　Further, although the vibration data detecting means 61 is installed on the vehicle body 1, it can be installed not only on the vehicle body 1 but also on the bogie frame 11 or the axle box body 12. In this embodiment, it will be described as being installed on the vehicle body 1.
[0017]
　The vibration acquisition direction of the vibration data detecting means 61 is a translation (forward / backward translation / left / right translation / vertical translation) direction and a rotation (roll / pitch / yaw) direction with respect to the front-back, left-right, and vertical directions with respect to the traveling direction of the railway vehicle. It is applicable to any of the above. In this embodiment, it will be described as a horizontal translation direction.
　Further, although a plurality of vibration data detecting means 61 may be installed in one vehicle, it will be described as one in this embodiment. Further, the installation position of the vibration data detecting means 61 may be the central position of the vehicle body 1 or the end portion of the vehicle body 1 such as the bogie position or the vehicle equipment room position.
[0018]
　Further, the vehicle body 1 includes, for example, an operation data detection unit 50 having a function of detecting operation data from an operation management system that manages operation data.
　The operation data detection unit 50 includes, for example, an operation data detection unit 51 including an operation data detection unit that detects operation data such as a travel speed, a travel position, and a occupancy rate.
[0019]
　In addition, although FIG. 1 shows an example in which the vibration data detecting means 61 and the operation data detecting means 51 are provided in only one vehicle, for example, as shown in FIG. 2, in preparation for a plurality of cars in one rolling stock. Alternatively, as shown in FIG. 3, it may be prepared for a plurality of cars of a plurality of trains.
　Further, in FIGS. 1, 2 and 3, the data analysis unit 100, the parameter estimation unit 120 and the abnormality factor estimation unit 150 are each provided as one, but each car or each trained vehicle may be provided with one.
[0020]
　The bogie 16 is composed of a bogie frame 11, an air spring 8, a yaw damper 4, an axle box body 12, an axle 13, an axle spring device 14, an axle box support rubber 15 for an axle box serving as a bearing housing for the axle 13, and the like. ..
　The wheel set 13 is rotatably held with respect to the axle box body 12, and the axle box body 12 and the bogie frame 11 are elastically supported in the vertical direction by the axle spring device 14, and are horizontally supported by the axle box support rubber 15. It is elastically supported in the direction. An air spring 8 is arranged between the vehicle body 1 and the bogie frame 11, and the vehicle body 1 is elastically supported by the bogie frame 11 by the air spring 8.
[0021]
　The abnormality detection device includes a data analysis unit 100, a parameter estimation unit 120, an abnormality factor estimation unit 150, and the like, and constitutes a vehicle vibration factor estimation device that estimates the cause of abnormal vibration of a railroad vehicle.
[0022]
　The data analysis unit 100, the parameter estimation unit 120, and the abnormality factor estimation unit 150 are composed of, for example, an arithmetic unit that executes each process described later according to a program stored inside.
[0023]
　The data analysis unit 100 acquires vibration data from the vibration data detecting means 61 and operation data from the operation data detecting means 51, and causes anomalous factor elements (for example, wheel shape, axle box support device, yaw damper, etc.) that vibrate the railway vehicle. It has a function of analyzing, identifying, and outputting (passage deviation, gap deviation, etc.), and is mainly composed of data analysis means 101.
[0024]
　The data analysis means 101 inputs vibration data and operation data detected by the vibration data detecting means 61 and the operation data detecting means 51, and derives an abnormality factor element of abnormal vibration based on the vibration data and the operation data. For example, a filtering function that filters the vibration data of the left-right acceleration of the vehicle body 1 detected by the vibration data detecting means 61 and extracts data effective for analysis, filtered vibration data (left-right acceleration), and operation data. It is an apparatus having an analysis function of performing analysis processing and extracting the feature amount of data, and a function of deriving anomalous factor elements based on the feature amount of the extracted data.
[0025]
　This abnormal factor element is an element that may be a factor of the abnormality when an abnormal vibration occurs in the railway vehicle. Broadly speaking, there are two types of abnormal factor elements, one on the vehicle side and the other on the track side.
[0026]
　Examples of the abnormal factor elements on the vehicle side include elements such as the axle box support rubber 15 and the yaw damper 4 of the axle box support device installed in the railway vehicle.
　The element of the abnormality factor on the track side is the wheel shape such as the wheel tread of the wheel set 13, and for example, there are elements on the track side such as the deviation of the track 20 and the deviation of the gauge.
[0027]
　The alignment irregularity is the unevenness of the side surface of the track in the length direction, and the gauge irregularity is the basic dimension (Shinkansen: 1435 mm, present) in the normal state without distortion of the gauge (left and right rail spacing). This is the amount of deviation when strain is generated for the incoming line: 1067 mm (in the curved portion, the amount of deviation with respect to the amount obtained by adding slack to the basic dimensions).
　The specific processing flow of the data analysis unit 100 will be described later.
[0028]
　The parameter estimation unit 120 selects the dynamic model (vehicle model or track model) and the parameter of the dynamic model that are previously associated with the abnormal factor element from the abnormal factor elements output by the data analysis means 101 of the data analysis unit 100. It is a device having a function of receiving the selected mechanical model, the parameters of the dynamic model, and the vibration data and the operation data, and estimating the parameters of the dynamic model in the abnormal state of the railway vehicle.
　The parameter estimation unit 120 includes a mechanical model selection means 121, a dynamic model parameter selection means 122, and a parameter estimation means 123.
[0029]
　The mechanical model selection means 121 of the parameter estimation unit 120 and the parameter selection means 122 of the dynamic model receive the anomalous factor elements derived by the data analysis means 101 of the data analysis unit 100, respectively, and store them in a database (not shown), for example. The dynamic model associated with the anomalous factor element and the parameters of the dynamic model are selected.
[0030]
　This mechanical model is a model (vehicle model or track model) for predicting the vibration characteristics of a railway vehicle. As an example, on a virtual track that simulates track irregularities such as track 20 and track deviation, the vehicle body 1, bogie 16 and wheel set 13 are rigid bodies, and the axle box support rubber 15 and yaw damper 4 are springs and dampers. Vibrations that you want to predict, such as a vehicle left-right system dynamic model that predicts vehicle body left-right vibration and a vehicle vertical system dynamic model that predicts vehicle body vertical vibration when a virtual vehicle that connects each rigid body with a spring or damper runs. There are various dynamic models depending on the type of.
[0031]
　Further, the parameters of the mechanical model are characteristic values ​​of factor elements included in this mechanical model, and are the spring constant of the axle box support rubber 15 which is the vehicle model on the vehicle side, the damping coefficient of the yaw damper 4, and the track on the track side. It corresponds to the wheel tread gradient of the wheelset 13 which is a model, the amount of deviation of the track 20 and the amount of deviation of the gauge.
[0032]
　The parameter estimation means 123 inputs the dynamic model from the mechanical model selection means 121 and the parameters of the mechanical model from the parameter selection means 122 of the mechanical model.
　Further, in addition to the dynamic model and the parameters of the dynamic model, the vibration data and the operation data detected by the vibration data detecting means 61 and the operation data detecting means 51 are input.
　Then, based on the vibration data and the operation data, the estimated values ​​of the parameters of the dynamic model in the abnormal state are derived.
　The specific processing flow of the parameter estimation unit 120 will be described later.
[0033]
　The anomalous factor estimation unit 150 compares the parameters of the dynamic model in the abnormal state estimated by the parameter estimation unit 120 with the parameters of the dynamic model in the normal state, and based on the comparison result, determines the abnormal factors of the vibration of the railway vehicle. It is a device having a function of estimating.
　The abnormal factor estimation unit 150 is composed of a normal value storage means 151, a normal value comparison means 152, and an abnormal factor estimation means 153.
[0034]
　The normal value storage means 151 stores, for example, the spring constant of the axle box support rubber 15, the damping coefficient of the yaw damper 4, the wheel tread gradient of the wheelset 13, and the normal parameter values ​​such as the deviation amount and the gauge deviation amount of the track 20. It is a device including a storage unit.
[0035]
　In the normal value comparing means 152, the estimated value of the parameter of the mechanical model derived by the parameter estimating means 123 of the parameter estimating unit 120 and the parameter normal value of the normal value storage means 151 are input.
[0036]
　The normal value comparison means 152 of the abnormal factor estimation unit 150 is based on the estimated value of the parameter in the abnormal state and the normal value of the parameter in the normal state stored in the normal value storage means 151, and is normal with the estimated value of the parameter. Calculate the ratio of values ​​(comparison of normal values). Based on this calculated ratio, the abnormal factor estimating means 153 estimates the abnormal factor, that is, the normality and abnormality of the track and the vehicle when vibration data is generated in the railway vehicle.
　The specific processing flow of the abnormal factor estimation unit 150 will be described later.
[0037]
　The determination result output means 200 has a function of notifying the driver of the vehicle body 1 and the operation manager or maintenance staff on the ground of the estimation result of the abnormality factor estimated and derived by the abnormality factor estimation means 153 of the abnormality factor estimation unit 150. It is a device to have.
[0038]
　2 and 3 are improved examples of the first embodiment, and FIG. 2 shows that the operation data detecting means 51 is installed on one vehicle body (car No. a) and 1 on each of the other vehicle bodies (cars b and c). This is an example in which one vibration data detecting means 61 is installed. Further, in FIG. 3, one operation data detecting means 51 and one vibration data detecting means 61 are installed on one vehicle body (car No. a) of each of the trained vehicles X, Y, and Z, and the other vehicle bodies (car No. b, No. 3) are installed. This is an example in which one vibration data detecting means 61 is installed in each c).
[0039]
　In FIG. 2, the data of the vibration data detecting means 61 installed in each of the cars a, b, and c of one train set and the operation data detecting means 51 installed in the car a a are the data analysis means of the data analysis unit 100. It is supplied to 101. Further, in FIG. 3, the data of the vibration data detecting means 61 installed in each of the plurality of trains X, Y, and Z and the operation data detecting means 51 installed in one of the trains of each train are data. It is supplied to the data analysis means 101 of the analysis unit 100.
　In FIGS. 2 and 3, the same parts as those in FIG. 1 are designated by the same reference numerals, and the description thereof will be omitted.
[0040]
　Next, each specific processing flow of data analysis, parameter estimation, and abnormality factor estimation in the abnormality detection device for the railway vehicle of this embodiment will be described with reference to FIGS. 4 to 9.
[0041]
　FIG. 4 shows a case where abnormal vibration occurs in a railroad vehicle in car No. b of the formation vehicle X due to a decrease in the damping coefficient of the yaw damper 4 due to an oil leak of the yaw damper 4 (see FIG. 1) (vehicle dynamics model) (vehicle dynamics model). FIG. 4B is a diagram schematically showing)).
　In this embodiment, the case where the acceleration sensor 62 in the left-right translation direction (hereinafter referred to as the left-right acceleration sensor) installed at the trolley position is used as the vibration data detecting means 61 will be described.
[0042]
　In FIG. 4, the horizontal axis shows the distance [m] and the vertical axis shows the vehicle body lateral acceleration [m / s 2 ]. FIG. 4 (A) shows the vehicle body lateral acceleration [m / s 2 ] measured by the lateral acceleration sensor 62 installed in the car No. a. A characteristic diagram showing the waveform characteristic 621 of [ m / s 2 ], FIG. 4 (B) is a characteristic diagram showing the waveform characteristic 622 of the vehicle body lateral acceleration [m / s 2 ] measured by the left-right acceleration sensor 62 installed in the car No. b . 4 (C) is a characteristic diagram showing a waveform characteristic 623 of the vehicle body lateral acceleration [m / s 2 ] measured by the lateral acceleration sensor 62 installed in the car No. c .
　From the figure, it can be seen that the amplitude of the waveform characteristic 622 of the vehicle body left-right acceleration of the vehicle body b vibrates significantly to the left and right as compared with the vehicle bodies a and c, and the vehicle body 1 (car No. b) is abnormally vibrated.
[0043]
　FIG. 5 is a flowchart illustrating a processing procedure of the data analysis unit in the abnormality detection device for the railway vehicle of the first embodiment.
[0044]
　The operation based on the flowchart of FIG. 5 is as follows. In this flowchart, the steps are displayed as S.
　First, in S1, the vehicle body left-right acceleration is acquired from the left-right acceleration sensor 62, and the operation data such as the traveling position, the traveling speed, and the occupancy rate are acquired from the operation data detecting means 51.
[0045]
　S2 to S5 are processes in the data analysis means 101.
　Next, in S2, the waveform characteristic 622 (see FIG. 4B) showing abnormal vibration is filtered from the vehicle body lateral acceleration [m / s 2 ] acquired in S1 using a well-known filtering technique, and the cause of the abnormality is Extract valid data for estimation.
　Here, preprocessing for the analysis process performed in S3 is included, such as a process of extracting only the frequency band that is easily perceived by humans from the left-right acceleration of the vehicle body and a process of extracting only the data of the traveling section where the vibration is particularly large.
[0046]
　In S3, the vibration data and the operation data extracted in S1 and S2 are analyzed, and the feature amount (characteristic value of the factor element) of the data is extracted.
　Here, for example, a process of extracting a frequency band in which vibration is particularly large by frequency analysis of vibration data, a process of statistically analyzing vibration data and operation data, and a process of extracting a dependency between a vibration waveform and a traveling position are performed. Do.
[0047]
　Further, when the plurality of cars or the plurality of trains are provided with the plurality of vibration data detecting means 61 and the operation data detecting means 51 as in the embodiment of FIGS. 2 and 3, the space between the cars or the trains is the same. Performs processing such as extracting the difference in the vibration waveform in.
[0048]
　The vibration data detecting means 61 detects a large amount of data from a plurality of cars or a plurality of trains, and the operation data detecting means 51 detects not only the traveling speed, the traveling position and the occupancy rate, but also various data such as the combined state of the vehicles. By detecting, it is possible to obtain a wider variety of highly accurate data features. Further, as a method for extracting the feature amount of the data, a well-known statistical analysis technique or the like may be used, or a big data analysis using an artificial intelligence technique or the like may be used.
[0049]
　In the case of this embodiment, it is assumed that "there is a difference in car number" is extracted as a feature amount for the abnormal vibration generated only in car number b of the trained vehicle X.
[0050]
　In S4, anomalous factor elements are derived based on the feature amount of the data extracted in S3.
　In the case of this embodiment, since the feature amount of the data is "there is a difference in car number", it is determined that the abnormal factor is not the track side but the vehicle side.
　As a result, it is assumed that the axle box support rubber 15, the yaw damper 4, the wheel treads of the wheel axle 13 and the like are derived as abnormal factor elements from the elements on the vehicle side. Then, these abnormal factor elements are output to the parameter estimation unit 120 in S5.
[0051]
　Next, the processing flow of the parameter estimation unit 120 will be described with reference to FIG. The operation based on the flowchart of FIG. 6 is as follows.
　First, in S11, the data analysis means 101 of the data analysis unit 100 acquires information such as the axle box support rubber 15, the yaw damper 4, and the wheel treads of the wheel sets 13 which are listed as abnormal factor elements.
[0052]
　In S12, the mechanical model selection means 121 selects a mechanical model previously associated with the abnormal factor element based on the abnormal factor element acquired in S11.
　In this embodiment, since the vehicle body left-right acceleration is targeted, it is assumed that the vehicle left-right system dynamic model is selected.
[0053]
　In S13, the parameter selection means 122 of the dynamic model selects the parameter related to the anomalous factor element from the parameters of the vehicle left-right system dynamic model.
　In this embodiment, since the abnormality factor element is the element on the vehicle side, it is assumed that the spring constant of the axle box support rubber 15, the damping coefficient of the yaw damper 4, the wheel tread gradient of the wheel axle 13 and the like are selected.
　Normally, a vehicle left-right system dynamic model contains a large number of parameters, but as described above, by narrowing down the parameters to be estimated in advance by the abnormality factor elements, the number of parameter survey cases is reduced and the abnormality factor estimation is shortened. It can be timed.
[0054]
　S14 to S19 are processes in the parameter estimation means 123.
　In S14, the initial value (initial estimated value) of the parameter selected in S13 is set.
　In this embodiment, initial values ​​such as the spring constant of the axle box support rubber 15, the damping coefficient of the yaw damper 4, and the wheel tread gradient of the wheel axle 13 are set. The initial value of the parameter may be a normal value of the parameter or a randomly generated value.
[0055]
　In S15, vibration analysis (vehicle motion analysis) is performed based on the mechanical model of S14 and the initial values ​​of the set parameters, and the vehicle body lateral acceleration under the set parameters is calculated.
[0056]
　In carrying out the vibration analysis, data on the virtual track such as the amount of deviation of the track 20 and the amount of deviation of the gauge in the vehicle left-right system dynamic model are required.
　In this embodiment, since it is determined that the track 20 is in a normal state, the data of the virtual track includes, for example, a management reference value prepared in advance in the normal value storage means 151, an actually measured value measured by an inspection vehicle, and the like. Use normal values.
[0057]
　In S16, the analysis value of the vehicle body left-right acceleration obtained in the analysis of S15 is compared with the actually measured value (normal value), and the error is calculated. For example, when the analysis value 612 and the measured value 611 match, it is possible to estimate (specify) the anomalous factor element with the parameter.
　There are no restrictions on the error evaluation index as long as the error can be evaluated (error comparison), but in this embodiment, as shown in FIG. 7, the measured value of the vehicle body lateral acceleration PSD (Power Spectrum Density) for each frequency. The value obtained by integrating the difference between the 711 and the analysis value 712 (hereinafter, the integrated difference value) is used.
[0058]
　In S17, if the integrated difference value of the vehicle body left-right acceleration PSD calculated in S16 exceeds the preset threshold value, the estimation accuracy is regarded as insufficient and the estimation is continued, and if it becomes less than the threshold value or the preset update is performed. If the number of times is exceeded, the estimation is terminated.
　If the estimation is continued, the parameters are updated in S18. There are no restrictions on the parameter update method, but for example, there is a method of repeatedly updating the parameters so that the integrated difference value of the vehicle body left-right acceleration PSD is minimized by using an optimization method such as a genetic algorithm.
　In this embodiment, parameters such as the spring constant of the axle box support rubber 15, the damping coefficient of the yaw damper 4, and the wheel tread gradient of the wheel axle 13 are updated, and as shown in FIG. 7, the calculated vehicle body lateral acceleration PSD is actually measured. Obtain an estimated value of the parameter such that the value 711 and the analysis value 712 match (see FIG. 7B).
[0059]
　FIG. 7 shows the frequency [Hz] on the horizontal axis, the vehicle body lateral acceleration PSD [(m / s 2 ) 2 / Hz] on the vertical axis, and FIG. 7 (A) shows the measured value 711 shown by the solid line before estimation. The characteristic diagram showing the analysis value 712 shown by the dotted line, FIG. 7B is a characteristic diagram showing the actually measured value 711 and the analysis value 712 after estimation.
　In the figure, before the estimation, there is a large difference between the measured value and the analyzed value in some frequency bands where the abnormal vibration occurs, but after the estimation, there is almost no difference between them.
[0060]
　When the estimation is completed, the estimated value of this parameter is output to the abnormal factor estimation unit 150 in S19.
[0061]
　As described above, by using the mechanical model, physical quantities such as the spring constant of the axle box support rubber 15, the damping coefficient of the yaw damper 4, and the wheel tread gradient of the wheel set 13 that are not directly measured by the sensor are estimated, that is, the physical quantities. Since there is no need for a sensor to measure the number of sensors, the number of sensors can be reduced.
[0062]
　In this embodiment, one left / right acceleration sensor 62 is installed in the vehicle body 1, but when a plurality of left and right acceleration sensors 62 are installed in the vehicle body 1, the carriage 16, the axle box body 12, and the like, a plurality of left and right acceleration sensors 62 are detected from them. By updating the parameters so that the measured values ​​of the left and right accelerations of the above and the analysis values ​​match, the estimation of the parameters can be made highly accurate.
[0063]
　Next, the processing flow of the abnormal factor estimation unit 150 will be described with reference to FIG. The operation based on the flowchart of FIG. 8 is as follows.
　First, in S21, the estimated value of the parameter is acquired from the parameter estimation unit 120.
　In this embodiment, the abnormal factor elements, that is, the spring constant of the axle box support rubber 15 in the abnormal state, the damping coefficient of the yaw damper 4, and the estimated value of the wheel tread gradient of the wheel axle 13 are acquired.
[0064]
　Next, in S22, the normal values ​​of these parameters are acquired from the normal value storage means 151 in the abnormal factor estimation unit 150.
　For normal values, the spring constant of the axle box support rubber 15, the design value and element test value for the damping coefficient of the yaw damper 4, the value of the wheel tread slope of the wheelset 13 in the new state, the control value at the time of maintenance, etc. Is used.
[0065]
　In S23, the ratio of the estimated value and the normal value of each parameter is calculated by the normal value comparing means 152.
　Then, in S24, the abnormality factor estimating means 153 compares and determines this ratio with the preset threshold value. This comparison determination is repeated a predetermined number of times.
　A parameter having this ratio below the threshold value is determined to be in the normal state because it can be said to be close to the normal state, and a parameter exceeding the threshold value is determined to be in the abnormal state. Then, the parameter determined to be in the abnormal state is set as the abnormal factor and output to the determination result output means 200.
[0066]
　FIG. 9 schematically shows the determination of normality and abnormality of the parameters in S24, and shows the abnormal factor elements such as the spring rigidity of the axle box support rubber 15 on the horizontal axis, the damping coefficient of the yaw damper 4, and the wheel tread gradient and the vertical axis. It is a schematic diagram which showed the relationship between the ratio of the estimated value and the normal value of a parameter on an axis, and a threshold.
　In this embodiment, as shown in the figure, the attenuation coefficient of the yaw damper 4 in which the ratio of the estimated value to the normal value exceeds the threshold value 91 is determined to be an abnormal state. Then, the decrease in the attenuation coefficient of the yaw damper 4 is output to the determination result output means 200 as an abnormal factor. The determination result output means 200 notifies the driver on the vehicle, the operation manager on the ground, the maintenance staff, and the like of the estimation result of the abnormal factor by using a well-known communication technique.
[0067]
　As described above, in the railroad vehicle abnormality detection device of this embodiment, it is possible to estimate an abnormality factor on the vehicle side such as a decrease in the damping coefficient of the yaw damper 4.
[0068]
　Next, as another example of estimating the abnormality factor in the abnormality detection device for the railway vehicle of this embodiment, a case where abnormal vibration is generated due to the abnormality factor on the track side will be described with reference to FIGS. 10 and 11.
[0069]
　FIG. 10 is a diagram schematically showing a case where abnormal vibration occurs in a part of the track 20 due to an increase in the amount of gauge deviation from the control reference value. Since the basic processing flow is the same as the description described with reference to FIGS. 1 to 9, only the differences will be described.
[0070]
　First, the data analysis unit 100 acquires the vehicle body lateral acceleration from the lateral acceleration sensor 62, and acquires operation data such as a traveling position, traveling speed, and occupancy rate from the operation data detecting means 51.
　The data analysis unit 100 extracts the feature amount of the data from the vehicle body lateral acceleration and the operation data acquired as described above.
　In this embodiment, as shown in FIGS. 10A, 10B, and 10C, only in a part of the section 620, with respect to the abnormal vehicle body left-right vibration (abnormal vibration) generated in all the cars. It is assumed that "with section difference" is extracted as a feature quantity.
　Since the feature amount of the data is "there is a section difference", it is judged that the abnormal factor is not the vehicle side but the track side. As a result, it is assumed that the deviation of the track 20 and the deviation of the gauge are derived as anomalous factor elements from the elements on the orbit side.
[0071]
　Next, the parameter estimation unit 120 selects the mechanical model and the parameters of the mechanical model from the derived anomalous factor elements, and estimates the parameters in the abnormal state.
　In this embodiment, as in the example of FIG. 4 described above, since the vehicle body left-right acceleration is targeted, it is assumed that the vehicle left-right system dynamic model is selected. Further, since the anomalous factor element is the element on the track side from the parameters of the vehicle left-right system dynamic model, the deviation amount and the gauge deviation amount of the track 20 are selected.
[0072]
　Next, the initial values ​​of these parameters are set, vibration analysis is performed, and the vehicle body lateral acceleration under the set parameters is calculated.
[0073]
　In carrying out the vibration analysis, virtual vehicle data such as the spring constant of the axle box support rubber 15 and the damping coefficient of the yaw damper 4 in the vehicle left-right system dynamic model are required.
　In this embodiment, since it is determined that the vehicle is in a normal state, a normal value such as a design value is used as the data of this virtual vehicle.
　Then, an estimated value of the deviation amount of the track 20 and the deviation amount of the gauge is obtained so that the error of the analysis value of the vehicle body left-right acceleration and the measured value match.
[0074]
　Next, the abnormal factor estimation unit 150 calculates the ratio between the estimated value of each parameter and the normal value, and compares it with the preset threshold value 111 as shown in FIG.
[0075]
　FIG. 11 is a schematic diagram showing the relationship between the amount of deviation and gauge deviation on the horizontal axis and the ratio and threshold value of the parameter estimated value and the normal value on the vertical axis.
　The amount of gauge deviation in which the ratio of the estimated value of the parameter to the normal value exceeds the threshold value 111 is determined to be an abnormal state, and the increase in the amount of gauge deviation in the track 20 is output to the determination result output means 200 as an abnormal factor.
　The determination result output means 200 notifies the driver on the vehicle, the operation manager on the ground, the maintenance staff, and the like of the estimation result of the abnormal factor.
[0076]
　As described above, the railroad vehicle abnormality detection device of the present embodiment can estimate an abnormality factor on the track side such as an increase in the amount of gauge deviation of the track 20.
　That is, in the railroad vehicle abnormality detection device of this embodiment, when an abnormal vibration occurs, it is possible to determine which element on the vehicle side or the track side is the cause of the abnormality.
　Further, in the railroad vehicle abnormality detection device of the present embodiment, the data analysis unit 100 narrows down the abnormality factor elements in advance, and the parameter estimation unit 120 associates the narrowed down abnormality factor elements with the abnormality factor elements in advance. After selecting the parameters of the dynamic model and the dynamic model, the parameters are estimated.
[0077]
　As a result, the number of parameter survey cases can be reduced and the estimation of abnormal factors can be shortened.
　In addition, by using a mechanical model, the spring constant of the axle box support rubber 15 that is not directly measured by the sensor, the damping coefficient of the yaw damper 4, the parameters on the vehicle side such as the wheel tread gradient of the wheel set 13, and the track 20 It is possible to estimate the parameters on the track side such as the amount of deviation and the amount of deviation, and it has the feature that the number of sensors can be reduced.
[0078]
　In this embodiment, since the vibration data detecting means 61 is provided in the vehicle body 1, the vehicle left-right system dynamic model is used, but when the vibration data detecting means 61 is provided in the bogie, the bogie A mechanical model of a vehicle element such as a left-right system dynamic model may be used.
Example 2
[0079]
　FIG. 12 is a diagram showing a system configuration of an abnormality detection device for another railway vehicle of the present invention.
　In this embodiment, as the processing of the abnormality detection device, in addition to the abnormality factor on the vehicle side (vehicle dynamics model) of the first embodiment, the abnormality factor on the track side (track dynamics model) is combined to generate abnormal vibration. That is, it is an example in the case where the anomalous factor element is on both the track side and the vehicle side.
[0080]
　First, the configuration of the abnormality detection device will be described.
　In this embodiment, the parameter estimation unit 120 is provided with a plurality of parameter estimation units such as the parameter estimation unit 120a and the parameter estimation unit 120b, and the parameter estimation unit 120a is dynamic based on the anomalous factor element, vibration data, and operation data. The parameters of the model A are estimated, the parameters of the dynamic model B are estimated by the parameter estimation unit 120b based on the parameter estimation values, vibration data, and operation data of the mechanical model A, and the mechanical model is estimated by the abnormal factor estimation unit 150. Based on the parameter estimated values ​​of A and B and the parameter normal values, the difference value between the estimated values ​​and the normal values ​​is obtained, and the cause of the abnormality is specified.
[0081]
　The parameter estimation unit 120a includes the mechanical model A selection means 121a, the parameter selection means 122a of the mechanical model A, and the parameter estimation means 123a, and the parameter estimation unit 120b includes the mechanical model B selection means 121b and the parameter selection of the mechanical model B. The means 122b and the parameter estimation means 123b are provided.
[0082]
　As shown in FIG. 13B, in the formation vehicle B, in addition to the decrease in the damping coefficient of the yaw damper 4 due to the oil leakage of the yaw damper 4, a part of the track 20 (the vehicle body of the formation vehicle B) is shown. It is a figure which shows typically the case where the abnormal vibration occurs in the section 620') of the lateral acceleration due to the fact that the amount of the gauge deviation increased more than a control reference value.
[0083]
　Next, the specific processing for estimating the cause of abnormality in the abnormality detection device for the railway vehicle of this embodiment will be described. The same members as those described in Example 1 are designated by the same reference numerals. A detailed description will be omitted, and only the operation of the portion changed from the first embodiment will be described below.
[0084]
　First, the data analysis unit 100 acquires operation data such as vehicle body left-right acceleration, traveling position, traveling speed, and occupancy rate from the vibration data detecting means 61 (for example, the left-right acceleration sensor 62) and the operation data detecting means 51.
[0085]
　The data analysis unit 100 extracts the feature amount of the data from the acquired vehicle body lateral acceleration and operation data.
　In this embodiment, as shown in FIG. 13B, it is assumed that "there is a formation difference" is extracted as a feature amount for the abnormal left-right vibration of the vehicle body that occurs only in the formation vehicle B. Further, it is assumed that "there is a section difference" is extracted as a feature amount for the abnormal left-right vibration of the vehicle body that occurs only in a part of the section of the track 20.
　Since the feature amount of the data is "there is a difference in formation", it is judged that the abnormal factor is the vehicle side.
　As a result, it is assumed that the axle box support rubber 15, the yaw damper 4, the wheel treads of the wheel axle 13 and the like are derived as abnormal factor elements from the elements on the vehicle side. Further, since "there is a section difference" is a feature quantity, it is assumed that the deviation of the track 20 and the deviation of the gauge are derived as anomalous factor elements from the elements on the orbit side.
[0086]
　From the above, it is determined that the abnormal vibration is caused by a combination of both the abnormal factor on the vehicle side and the abnormal factor on the track side.
[0087]
　Next, the parameter estimation unit 120a selects the parameters of the mechanical model A by the mechanical model A selection means 121a and the parameters of the mechanical model A by the parameter selection means 122a from the anomalous factor elements derived by the data analysis unit 100. Further, based on the vibration data and the operation data, the parameter estimation means 123a estimates the parameters in the abnormal state.
[0088]
　Further, in the parameter estimation unit 120b, the parameter selection of the mechanical model B by the mechanical model B selection means 121b and the parameters of the mechanical model B by the parameter selection means 122b are performed from the parameter estimated values ​​of the mechanical model A estimated by the parameter estimation unit 120a. Is selected, and further, the parameter is estimated by the parameter estimation means 123b in an abnormal state based on the vibration data and the operation data.
[0089]
　The anomalous factor estimation unit 150 obtains the difference value between the estimated value and the normal value by the normal value comparing means 152 based on the parameter estimated value and the parameter normal value of the mechanical model A estimated by the parameter estimation unit 120b, and estimates the abnormal factor. The anomalous factor is estimated by means 153.
[0090]
　In this embodiment, there are abnormal factor elements on both the vehicle side and the track side, but when these are estimated at the same time, both the vehicle side and the track side are estimated, and a large amount of parameters are estimated. This can reduce the accuracy of parameter estimation.
　Therefore, the railroad vehicle abnormality detection device according to the present embodiment is characterized in that the abnormality factor elements on the vehicle side and the track side are separated and estimated.
[0091]
　First, the parameter estimation unit 120a that estimates the abnormal factor elements on the vehicle side will be described.
　In this embodiment, first, since the vehicle body lateral acceleration is targeted as in the first embodiment, it is assumed that the vehicle left-right system dynamic model is selected by the dynamic model A selection means 121a. Further, in the parameter selection means 122a of the dynamic model A, the spring constant of the axle box support rubber 15 and the damping coefficient of the yaw damper 4 and the wheel axle 13 are selected from the parameters of the vehicle left-right system dynamic model based on the abnormal factor element on the vehicle side. Wheel tread slope, etc. are selected.
[0092]
　Next, the parameter estimation unit 120a executes the same processing as in S14 to S19 of FIG.
　That is, the initial values ​​of these parameters are set, vibration analysis is performed, the vehicle body left-right acceleration in the set parameters is calculated, and the parameters are updated so that the error between the analysis value of the vehicle body left-right acceleration and the measured value matches. I do.
[0093]
　Here, the measured value of the vehicle body left-right acceleration is characterized in that the vehicle body left-right acceleration detected in another section where the abnormal vibration is not generated is used instead of the section where the abnormal vibration is generated.
　Regarding another section in which abnormal vibration does not occur, it can be determined that only the vehicle side is in an abnormal state and the track side is in a normal state.
　Therefore, when performing vibration analysis, the virtual track data such as the amount of deviation of the track 20 and the amount of deviation of the gauge in the vehicle left-right system dynamic model, which is required, includes, for example, the control reference value and the track measured by the inspection vehicle. Use normal values ​​such as actual measurement data of.
　In this way, the parameters on the vehicle side can be estimated accurately by performing the estimation using the vibration data of the section in which only the vehicle side is in the abnormal state and the track side is in the normal state.
　In this embodiment, the spring constant of the axle box support rubber 15, the damping coefficient of the yaw damper 4, and the estimated value of the wheel tread gradient of the wheel axle 13 are finally obtained.
[0094]
　Next, the parameter estimation unit 120b that estimates the abnormal factor elements on the orbit side will be described.
　Here, since the vehicle body lateral acceleration is targeted as described above, the vehicle left-right system dynamic model is selected by the dynamic model B selection means 121b. Further, in the parameter B selection means 121b of the dynamic model B, the deviation amount and the gauge deviation amount of the track 20 are selected from the parameters of the vehicle left-right system dynamic model based on the abnormal factor element on the track side.
[0095]
　Next, the parameter estimation unit 120a executes the same processing as in S14 to S19 of FIG. That is, the initial values ​​of these parameters are set, vibration analysis is performed, the vehicle body left-right acceleration in the set parameters is calculated, and the parameters are updated so that the error between the analysis value of the vehicle body left-right acceleration and the measured value matches. I do.
[0096]
　Here, as the measured value of the vehicle body left-right acceleration, the vehicle body left-right acceleration detected in the section where abnormal vibration is generated is used. In carrying out the vibration analysis, virtual vehicle data such as the spring constant of the axle box support rubber 15 and the damping coefficient of the yaw damper 4 in the vehicle left-right system dynamic model are required, but these are in the abnormal state estimated earlier. Use the estimated values ​​of the parameters on the vehicle side. Since the estimation of the vehicle-side parameters of the vehicle left-right system dynamic model has already been completed, the track-side parameters can be estimated accurately.
　In this embodiment, finally, an estimated value of the deviation amount and the gauge deviation amount of the track 20 is obtained.
[0097]
　Then, the abnormal factor estimation unit 150 calculates the ratio between the estimated value of each parameter and the normal value, and compares it with a preset threshold value as shown in FIG.
[0098]
　FIG. 14 (A) is a diagram showing the relationship between the axle box support rubber spring constant on the horizontal axis, the yaw damper damping coefficient, the ratio of the estimated value and the normal value of the parameter in the wheel tread gradient, and the threshold value, and FIG. 14 (B) is a diagram. It is a figure which shows the relationship between the amount of deviation, the amount of deviation, the ratio of the estimated value of the parameter on the vertical axis, and the normal value, and the threshold.
　Estimated value of parameter and normal value Among the factors on the vehicle side, the damping coefficient of the yaw damper 4 that exceeds the threshold value is determined to be in an abnormal state, and among the factors on the track side, the amount of deviation of the track 20 that exceeds the threshold value is in an abnormal state. Is determined.
　As a result, it is determined that the abnormal factor is a combination of a decrease in the damping coefficient of the yaw damper 4 and an increase in the amount of deviation of the track 20.
[0099]
　As described above, in the railroad vehicle abnormality detection device of this embodiment, even when abnormal vibration occurs due to a combination of abnormal factors on the vehicle side and abnormal factors on the track side, the abnormal factor is the vehicle. It has the feature that it can determine which element on the side or the orbit side.
Example 3
[0100]
　FIG. 15 is a diagram showing a system configuration of an abnormality detection device for another railway vehicle of the present invention.
　The abnormality detection device of the present embodiment includes an abnormality factor element determining means 161 as a process when the parameter estimation accuracy is insufficient in the parameter estimation unit 120 with respect to the abnormality detection device of the first embodiment. The factor element determination unit 160 is added. The same members as those described in the first embodiment are designated by the same reference numerals, and detailed description thereof will be omitted.
[0101]
　FIG. 16 is a flowchart showing a processing procedure of the parameter estimation unit 120 in this embodiment. Only the differences from the first embodiment will be described below.
　In this embodiment, as in the first embodiment, in S15 to S18, the error between the measured value and the analysis value of the vehicle body lateral acceleration is calculated, and the parameter update is repeated until this error becomes sufficiently small. However, in S15 to S18, even if the number of updates exceeds the preset number of updates, if the error between the measured value and the analysis value of the vehicle body left-right acceleration does not become less than or equal to the preset threshold value in S17', it is estimated. Judge that the accuracy is insufficient. Then, the result (insufficient estimation) is output to the abnormal factor element determining means 161.
[0102]
　In S30, the abnormality factor element determining means 161 has an abnormality factor on the vehicle side, on the track side, or on the vehicle side and the track side from the abnormality factor elements already obtained by the data analysis unit 100. Whether it is in both or not, the abnormality factor element is determined, and only the result (track abnormality, vehicle abnormality) is output.
[0103]
　As described above, in the railroad vehicle abnormality detection device of this embodiment, even when the parameter estimation unit 120 has insufficient parameter estimation accuracy, the abnormality factor determination is not interrupted and the abnormality factor is determined. The judgment result can be output.
[0104]
　According to the above-described embodiment, the following effects can be expected.
　(1) By using a mechanical model, parameters that are not directly measured can be estimated, so that the number of sensors can be reduced.
　(2) By narrowing down the anomalous factor elements in advance from the correlation analysis of vibration data and operation data, the number of parameter survey cases can be reduced and factor estimation can be performed in a short time.
　(3) By utilizing not only the vehicle model but also the track model, not only the abnormal factor on the vehicle side but also the abnormal factor on the track side can be identified.
　(4) Even when a vibration abnormality occurs due to both the vehicle side and the track side as factors, the abnormality factor can be identified.
[0105]
　The above-described embodiment has been described in detail in order to explain the present invention in an easy-to-understand manner, and is not necessarily limited to the one including all the described configurations.
　Further, it is possible to replace a part of the configuration of one embodiment with the configuration of another embodiment, and it is also possible to add the configuration of another embodiment to the configuration of one embodiment.
　Further, it is possible to add / delete / replace other configurations with respect to a part of the configurations of each embodiment.
Code description
[0106]
1 Body
4 Yaw damper
8 Air spring
11 Bogie frame
12 Axle box
13 Wheel shaft
14 Axle spring device
15 Axle box support rubber
16 Bogie
20 Track
50 Operation data detection unit
51 Operation data detection means
60 Vibration data detection unit
61 Vibration data detection means
62 Left / right acceleration sensor
100 Data analysis unit
101 Data analysis means
120, 120a, 120b Parameter estimation unit
121 Mechanical model selection means
122 Mechanical model parameter selection means
123 Parameter estimation means
150 Abnormal factor estimation unit
151 Normal value storage means
152 Normal value comparison means
153 Abnormal factor estimation means
160 Abnormal factor element discriminating unit
161 Abnormal factor element discriminating means
200 Judgment result output means
The scope of the claims
[Claim 1]
　An abnormality detection device for detecting an abnormality of a railroad vehicle traveling on a track
　includes a vehicle vibration factor estimation device for estimating a vehicle vibration factor that vibrates the railroad vehicle, and the
　vehicle vibration factor estimation device is
　the vibration of the railroad vehicle. The vibration data and the operation data detected by the vibration data detection device including the data detection unit for detecting the vibration data indicating the above and the operation data detection device including the data detection unit for detecting the operation data of the railway vehicle are acquired. A data analysis unit that analyzes and derives anomalous factor elements of the vibration of the railcar,
　and an anomaly factor element estimation unit that estimates anomalous factor elements of vibration derived by the data analysis unit based on the vibration data and the operation data. 　An abnormality detection device for a railroad vehicle ,
　comprising an anomaly factor estimation unit that estimates an abnormal factor of vibration in the railcar based on the anomaly factor element estimated by the anomaly factor element estimation unit
.
[Claim 2]
　A failure detection device for railway vehicle according to claim 1,
　wherein the abnormal factor element estimation unit is made from the parameter estimation unit for estimating the parameters of the dynamic model of the vehicle side and the raceway side,
　the parameter estimation unit,
　wherein A mechanical model selection means for selecting a mechanical model associated with the abnormal factor element in advance from anomalous factor elements,
　a parameter selection means for selecting parameters of the
mechanical model, the mechanical model, the parameters, the vibration data, receiving the operational data includes a parameter estimation means for estimating parameters in an abnormal state of the railway vehicle,
　the abnormality factor estimating unit,
　and estimates of the parameters in an abnormal state estimated by the parameter estimation section, normal state
　Anomalous factors that estimate the anomalous factor that causes the railcar to vibrate abnormally based on the ratio that is the comparison result of the normal value comparison means that compares the normal values ​​of the parameters in the above and calculates the ratio between them.
　An abnormality detection device for rolling stock , including an estimation means .
[Claim 3]
　The abnormality detection device for a railroad vehicle according to claim 2,
　wherein the parameter estimation unit includes first and second parameter estimation units, and the
　abnormality factor element is on both the vehicle side and the track side.
　The first parameter estimation unit
　estimates the parameters in the abnormal state based on the abnormality factor on the vehicle side based on the abnormality factor element from the data analysis unit, the vibration data, and the operation data, and estimates
　the second parameter. parts are
　the parameters and the vibration data was estimated by the first parameter estimation unit, based on the operational data, to estimate the parameters in an abnormal state based on the abnormal factor of the track-side,
　the abnormality factor estimating unit,
　wherein
　An abnormality of a railroad vehicle, characterized in that an abnormality factor is estimated based on an estimated value of a parameter in an abnormal state estimated by the first parameter estimation unit and the second parameter estimation unit and a normal value of a parameter in a normal state. Detection device.
[Claim 4]
　The abnormality detection device for a railway vehicle according to claim 2 or 3,
　further
　comprising an abnormality factor element determination unit , wherein the abnormality factor element determination unit
　estimates accuracy of the estimated value of the parameter in the parameter estimation unit. If the estimation accuracy is less than or equal to a preset threshold value, it is determined that the estimation accuracy is insufficient, and the abnormal factor is on the vehicle side, on the track side, or with the vehicle side.
　An abnormality detection device for a railroad vehicle, characterized in that it outputs only or is on both sides of the track .
[Claim 5]
　The
　vibration data according to any one of claims 2 to 4, wherein the vibration data detecting device detects vibration of the vehicle body, the carriage frame, or the axle box of the railway vehicle. An abnormality detection device for railroad vehicles, which comprises a detection unit.
[Claim 6]
　The abnormality detection device for a railcar according to any one of claims 2 to 5,
　wherein the vibration data detection device and the operation data detection device are provided in a plurality of cars of one train set or a plurality of train cars. An abnormality detection device for railcars, which is characterized by this.
[Claim 7]
　The abnormality of a railway vehicle according to any one of claims 2 to 6,
　wherein the dynamic model is a vehicle left-right system dynamic model or a vehicle vertical system dynamic model. Detection device.
[Claim 8]
　The abnormality detection device for a railroad vehicle according to any one of claims 2 to 7,
　wherein the parameter estimation means of the parameter estimation unit estimates a parameter in the abnormality factor element by using a genetic algorithm. Anomaly detection device for railroad vehicles.
[Claim 9]
　In the railroad vehicle abnormality detection device according to claim 1, the
　data analysis unit
　acquires vibration data of the railroad vehicle and operation data of the railroad vehicle, analyzes the acquired vibration data and operation data, and describes the data. Anomalous factor elements in the vibration state of a railroad vehicle are derived, and the
　anomaly factor element estimation unit
　calculates the current values ​​of the characteristics of the anomalous factor elements in the vibration data and operation data derived by the data analysis unit based on a mechanical model. The
　estimation unit that identifies,
　estimates the abnormal factor element, and estimates the abnormal factor compares the estimated value of the abnormal factor element estimated by the abnormal factor element estimation unit with the normal value in the normal state, and makes the comparison.
　An abnormality detection device for a railway vehicle, which estimates an abnormality factor of the railway vehicle based on the result of the data.
[Claim 10]
　In the abnormality detection device for a railroad vehicle according to claim 9, the
　anomaly factor element estimation unit includes a parameter estimation unit, and the
　parameter estimation unit includes
　a mechanical model selection means for selecting the mechanical model and the
　mechanical model.
　Includes a parameter selection means for selecting the parameters of the above, and a parameter estimation means for estimating the parameters of the anomalous factor element based on the dynamic model, the parameters, the vibration data, and the operation data, and outputting the estimated values ​​of the parameters. The
　abnormality factor estimation unit includes
　a storage means for storing normal values ​​of parameters in the normal state,
　an estimated value of parameters estimated by the parameter estimation means, and normality of parameters stored in advance in the storage means. The 　dynamic model
　includes a normal value comparing means for comparing values ​​and an abnormal factor estimating means for estimating an abnormal factor of the railroad vehicle based on the ratio of the estimated value and the normal value of the parameter, and the
dynamic model includes a vehicle dynamics model and track dynamics.
　An abnormality detection device for railway vehicles , which is characterized by being a model .
[Claim 11]
　In the abnormality detection method in the abnormality detection device for a railway vehicle according to claim 1,
　a step of acquiring vibration data of the railway vehicle and operation data of the railway vehicle,
　and analysis of the acquired vibration data and operation data are performed.
　The current values ​​of the characteristics of the anomalous factor elements in the vibration data and the operation data derived in the data analysis step and the data analysis step for deriving the anomalous factor elements in the vibration state of the railroad vehicle are identified based on the mechanical model. , The anomalous factor of the
　railway vehicle is estimated based on the comparison result between the anomaly factor element estimation step for estimating the anomaly factor element and the estimated value of the anomaly factor element estimated by the anomaly factor element estimation step with the normal value in the normal state. An abnormality
　detection method for a railroad vehicle, which comprises an abnormality factor estimation step .
[Claim 12]
　In the method for detecting anomalies in a railroad vehicle according to claim 11, the
　anomaly factor element estimation step includes a parameter estimation step, and the
　parameter estimation step includes
　a mechanical model selection step for selecting the mechanical model and the
　dynamic model. a parameter selection step of selecting a parameter,
　wherein the dynamic model, the parameters and the vibration data, the parameters of the error factors element estimated based on the operational data, the estimating step of outputting an estimated value of the parameter,
　the The abnormal factor estimation step includes
　a normal value comparison step for comparing
　the estimated value of the parameter estimated in the parameter estimation step with the normal value of the parameter stored in the storage unit in advance, and the estimated value and the normal value of the parameter. A 　method for detecting anomalies in a railroad vehicle ,
　comprising an estimation step of estimating anomalous factors of the railroad vehicle based on a ratio, wherein the dynamics model is a vehicle dynamics model or a track dynamics model