Method for detecting abnormalities in railway vehicles

Abstract

To provide a method for detecting abnormality of railway vehicles which enables reduction of the number of sensors and also enables detection of abnormality of railway vehicles to be performed without conducting complicated signal processing such as a fast fourier transformation (FFT) analysis.SOLUTION: In a method for detecting abnormality of railway vehicles, an abnormality detection device acquires a strain amount detected by a strain sensor attached to one link of a railway vehicle, performs filter processing to the strain amount, performs predetermined processing to the strain amount which has been subject to the filter processing, and compares a value obtained by the predetermined processing with a predetermined value to detect abnormality of the railway vehicle.SELECTED DRAWING: Figure 7

Description

Technical Field

[0001] The present invention relates to a method for detecting abnormalities in railway vehicles.

Background Art

[0002] Regarding the detection of abnormalities occurring in the bogies of running railway vehicles, in the prior art, for example, regarding the detection of abnormalities in the bogie frames, 2 to 4 acceleration sensors were attached per bogie. Also, regarding the detection of abnormalities in bearings, 4 temperature sensors were attached per bogie in axle boxes for axles, and 2 temperature sensors were attached per bogie in gearbox devices. Further, as a measurement and evaluation of braking force, it has been proposed to attach strain gauges to the main part of a single link to measure and evaluate the braking force (see, for example, Patent Document 1).

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, in the prior art, there were problems such as an increase in the number of sensors per bogie, resulting in higher costs, and a decrease in reliability due to an increase in the number of sensors. In particular, the processing of acceleration sensors involved complex and special processing devices such as FFT (Fast Fourier Transformation) analysis. Thus, in the prior art, the detection of abnormalities in railway vehicle bogies had the problem that the number of sensors per bogie increased and the cost of the sensors became high.

[0005] The present invention has been made in view of the above problems, and an object thereof is to provide a method for detecting abnormalities in railway vehicles that can reduce the number of sensors and enable the detection of abnormalities in railway vehicles without performing complex signal processing such as FFT analysis. [Means for solving the problem]

[0006] To achieve the above objective, in one aspect of the present invention, an abnormality detection device acquires the amount of strain detected by a strain sensor attached to one link of a railway vehicle, filters the amount of strain, performs a predetermined process on the filtered amount of strain, compares the value obtained by the predetermined process with a predetermined value, and detects an abnormality in the railway vehicle.

[0007] In one aspect of the present invention, the RMS value of the filtered strain is calculated by the predetermined process, an evaluation variable is calculated using the RMS value, and the value obtained by the predetermined process is the evaluation variable.

[0008] One aspect of the present invention is that the value obtained by the predetermined process is the value obtained by full-wave rectification of the filtered distortion amount, and in the predetermined process, full-wave rectification is performed on the filtered distortion amount.

[0009] One aspect of the present invention is that the filtering process is a low-pass filter process, and the abnormality detection device calculates a first evaluation variable by dividing the effective value of the distortion amount processed by the low-pass filter by the required braking force, which is a command value for the bogie of the railway vehicle, when the brakes of the railway vehicle are in the ON state, and detects an abnormality in the braking force of the railway vehicle by comparing the first evaluation variable with the upper and lower limits set for the first evaluation variable.

[0010] One aspect of the present invention involves comparing the first evaluation variable with an upper and lower limit set for the first evaluation variable, and detecting that the braking force of the railway vehicle is within the normal range when the first evaluation variable is less than or equal to the upper limit and greater than or equal to the lower limit.

[0011] One aspect of the present invention is that the filtering process is a high-pass filter process, and when the train speed used in controlling the railway vehicle is greater than or equal to a predetermined value, the abnormality detection device calculates a second evaluation variable by dividing the effective value of the distortion amount processed by the high-pass filter by the train speed used in controlling the railway vehicle, and detects an abnormality in the wheel rotation system of the railway vehicle by comparing the second evaluation variable with the upper and lower limits set for the second evaluation variable. It should be noted that the train speed is not the speed of a single axle, but rather a calculation that avoids the influence of wheel slippage on each axle. Furthermore, the wheel rotation system refers to rotating equipment such as wheels, axles, gears, couplings, and motors.

[0012] One aspect of the present invention involves comparing the second evaluation variable with an upper and lower limit set for the second evaluation variable, and detecting that the wheel rotation system of the railway vehicle is within the normal range when the second evaluation variable is less than or equal to the upper limit and greater than or equal to the lower limit.

[0013] One aspect of the present invention is that, before the anomaly detection device performs filtering on the strain amount, it performs anti-aliasing filtering and absolute value processing on the acquired strain amount. [Effects of the Invention]

[0014] According to the present invention, the number of sensors can be reduced, and abnormalities in railway vehicles can be detected without performing complex signal processing such as FFT analysis. [Brief explanation of the drawing]

[0015] [Figure 1] This figure shows an example configuration of an abnormality detection device for a railway vehicle according to the first embodiment. [Figure 2] This is a front view showing an example of a mounting location for a strain sensor in a railway vehicle according to the first embodiment. [Figure 3] This diagram conceptually shows the location where a single link is attached. [Figure 4] This is a conceptual diagram showing the location where the acceleration sensor is attached. [Figure 5] It is a diagram showing an example of the frequency analysis result of the acceleration sensor (vertical and horizontal). [Figure 6] It is a diagram showing an example of the frequency analysis result of the strain sensor attached to a single link. [Figure 7] It is a flowchart of the preprocessing and the normal-abnormal determination detection process of the braking force according to the first embodiment. [Figure 8] It is a flowchart of the normal-abnormal determination detection process of the wheel rotation system according to the first embodiment. [Figure 9] It is a diagram showing examples of the upper limit value and the lower limit value for the variable a and examples of the upper limit value and the lower limit value for the variable b according to the first embodiment. [Figure 10] It is a timing chart of an example of data processing according to the first embodiment. [Figure 11] It is a diagram showing a configuration example of the abnormality detection device for a railway vehicle according to the second embodiment. [Figure 12] It is a diagram showing a sensor for detecting bending strain and an example of the direction of generation of bending stress. [Figure 13] It is a diagram showing an example of a bridge circuit. [Figure 14] It is a diagram showing the relationship between bending strain and yaw angular velocity. [Figure 15] It is a timing chart of an example of data processing according to the second embodiment. [Figure 16] It is a flowchart of the normal-abnormal determination detection process of the truck behavior according to the second embodiment.

Mode for Carrying Out the Invention

[0016] Hereinafter, embodiments of the present invention will be described with reference to the drawings. In the drawings used in the following description, the scales of the respective members are appropriately changed in order to make the respective members recognizable in size.

[0017] <First Embodiment> [Configuration Example of Abnormality Detection Device for Railway Vehicle] First, a configuration example of the abnormality detection device 1 for a railway vehicle will be described. Figure 1 shows an example of the configuration of an abnormality detection device for a railway vehicle according to this embodiment. As shown in Figure 1, the abnormality detection device 1 for a railway vehicle includes a strain sensor G and an abnormality detection device 2. The abnormality detection device 2 comprises a pre-processing unit 201, an LPF (Low-pass filter) 202, an RMS value calculation unit 203, a variable a calculation unit 204, an HPF (How-pass filter) 205, an RMS value calculation unit 206, a variable b calculation unit 207, a determination unit 208, an output unit 209, and a vehicle information acquisition unit 210.

[0018] The strain sensor G detects the amount of strain as a detected value. The strain sensor G is a sensor that applies the 4-active gauge method (orthogonal arrangement method) to measure the tensile and compressive loads acting on the trunk of a single link L, for example. Compared to the 1-gauge method, this measurement method offers advantages such as the elimination of bending components, improved strain sensitivity, and temperature compensation function.

[0019] The abnormality detection device 2 starts processing when the railway vehicle 10 is traveling at, for example, a train speed of 5 km / h or more, based on the speed information acquired by the vehicle information acquisition unit 210. The abnormality detection device 2 uses the detected values ​​from the strain sensor G to detect whether the braking force is normal or abnormal, and whether the rotation system is normal or abnormal. Note that the train speed is not the speed of a single axle, but is calculated in a way that does not affect the slippage of each axle.

[0020] The preprocessing unit 201 performs anti-aliasing filtering and absolute value processing on the detected value detected by the strain sensor G, and outputs the preprocessed detected value to the LPF202 and HPF205.

[0021] The LPF202 extracts components from the pre-processed detected values, for example, DC (direct current) ~ 3 Hz, using low-pass filtering, and outputs the extracted components to the RMS value calculation unit 203. Note that the frequency components extracted by the LPF202 are just examples and are not limited to these. The extraction bandwidth can be any bandwidth set according to the target railway vehicle or location, for example.

[0022] The RMS value calculation unit 203 calculates the RMS value based on the detected values ​​of the bandwidth extracted by the LPF 202.

[0023] The variable a calculation unit 204 calculates variable a (first evaluation variable) by dividing the effective value calculated by the effective value calculation unit 203 by the required brake force for the bogie, which is obtained by the vehicle information acquisition unit 210, when the information acquired by the vehicle information acquisition unit 210 indicates that the brakes are on. The required brake force per bogie is, for example, the vehicle's control command value. If the information acquired by the vehicle information acquisition unit 210 is a control command value, the variable a calculation unit 204 may calculate the required brake force for the bogie using the control command value. In this case, the variable a calculation unit 204 stores the relationship between the control command value and the required brake force for the bogie in a mathematical formula or tabular format.

[0024] The HPF205 extracts components of, for example, 5 Hz or higher from the pre-processed detected values ​​using high-pass filtering, and outputs the extracted components to the RMS value calculation unit 206. Note that the frequency components extracted by the HPF205 are just examples and are not limited to these. The extraction bandwidth can be any bandwidth set according to, for example, the railway vehicle or location being detected.

[0025] The RMS value calculation unit 206 calculates the RMS value based on the detected values ​​of the bandwidth extracted by the HPF 205.

[0026] The variable b calculation unit 207 calculates variable b (second evaluation variable) by dividing the effective value calculated by the effective value calculation unit 206 by the train formation speed acquired by the vehicle information acquisition unit 210. The train formation speed is, for example, the speed of the train formation (as displayed in the driver's cab, or used for train formation control). If the information acquired by the vehicle information acquisition unit 210 is a control command value, the variable b calculation unit 207 may calculate the vehicle speed using the control command value. In this case, the variable b calculation unit 207 stores the relationship between the control command value and the vehicle speed in a mathematical formula or tabular format.

[0027] The determination unit 208 compares variable a calculated by variable a calculation unit 204 with an upper limit (predetermined value) or lower limit (predetermined value) for variable a to determine whether the braking force is normal or above. The upper limit and lower limit for variable a are values ​​set according to, for example, the railway vehicle being detected or its location. The determination unit 208 also compares variable b calculated by variable b calculation unit 207 with an upper limit or lower limit for variable b to determine whether the wheel rotation system is normal or above. The upper limit and lower limit for variable b are values ​​set according to, for example, the railway vehicle being detected or its location.

[0028] The output unit 209 outputs the result of the determination by the determination unit 208 to, for example, the control device 3.

[0029] The vehicle information acquisition unit 210 acquires vehicle information from, for example, the control device 3 that controls the railway vehicle. This vehicle information includes, for example, control command values, speed information, information indicating whether the brakes are on or off, and the required braking force for the bogie.

[0030] Note that the configuration shown in Figure 1 is just one example and is not limited to it. The anomaly detection device 2 may include, for example, an acquisition unit that acquires detected values ​​from the strain sensor G, and a storage unit that stores the date and time information of the determination in association with the determination result. Alternatively, the storage unit may store upper and lower limits for variable a, and upper and lower limits for variable b.

[0031] [Examples of mounting locations for strain sensors] Next, we will explain examples of mounting locations for strain sensors. Figure 2 is a front view showing an example of the mounting location of a strain sensor in a railway vehicle according to this embodiment. As shown in Figure 2, in the railway vehicle 10, the car body 12 is mounted on a bogie 11. A connecting member 13 is fixed to the car body 12 side. A bogie frame 14 is fixed to the bogie 11 side. A single link L is provided between the connecting member 13 and the bogie frame 14. The strain sensor G is attached, for example, by being glued around the single link L.

[0032] In recent years, railway vehicles 10 have undergone weight reduction in both the car body 12 and the bogie 11, and the use of air-spring type bolsterless bogies has become the mainstream for the bogie 11. This railway vehicle 10 directly connects the car body 12 and the bogie frame 11 with air springs 16 (see Figure 3) that have a large allowable displacement, and transmits the driving force and braking force through a traction device consisting of a single link L.

[0033] Figure 3 is a conceptual diagram showing the position where the single link is attached. As shown in Figure 3, the single link L is attached to each bogie frame 20, for example, between the wheels 18. Note that the example shown in Figure 3 is an example of the position and is approximate, and is not limited to this.

[0034] [Consider] Here, we will explain an example of the results of performing FFT processing on data acquired from a strain sensor. Figure 4 is a conceptual diagram showing the mounting position of the acceleration sensors. In the example in Figure 4, the acceleration sensors (g902) (top and bottom) are mounted horizontally on the upper surface of the axle box (g901). The acceleration sensors are, for example, MEMS (Micro Electro Mechanical Systems) acceleration sensors. The strain sensors are mounted on the trunk of the single link, as in Figure 1.

[0035] Figure 5 shows an example of frequency analysis results for an acceleration sensor (up and down). In Figure 5, the horizontal axis represents time (s (seconds)), the left vertical axis represents frequency (Hz), and the contour lines represent vibration level (dB). Note that 0 (dB) = 1 × 10⁻⁶. -5 (m / s 2 ) was used. Note that the results in Figure 5 are an example of the results of FFT analysis performed after anti-aliasing filtering of the accelerometer's detected values. As shown in Figure 5, the FFT analysis results of the accelerometer show no DC component (g911). Furthermore, the FFT analysis results of the accelerometer show that the frequency corresponding to the wheel rotation speed is intermittent, resulting in low detection sensitivity (g912). In addition, the FFT analysis results of the accelerometer show that a frequency approximately four times the wheel rotation speed is dominant (g913). Thus, it is difficult to evaluate the wheel rotation speed using detection with an accelerometer.

[0036] Figure 6 shows an example of frequency analysis results for a strain sensor attached to a single link. In Figure 6, the horizontal axis and the left vertical axis are the same as in Figure 5, and the contours represent strain levels (dB). Note that 0 (dB) = 1 × 10⁻¹⁰ -1 (με) was used. Note that the results in Figure 6 are an example of the results of FFT analysis performed after anti-aliasing filtering of the accelerometer's detected values. As shown in Figure 6, the FFT analysis results of the strain sensor attached to the single link show a DC (direct current) component, which includes the braking force (g921). Furthermore, the FFT analysis results of the strain sensor show a dominant frequency corresponding to the wheel rotation speed (g922), and the gear ratio multiplication component can also be detected (g923). Therefore, it is suggested that using detection with a strain sensor makes it easy to evaluate the wheel rotation speed (e.g., wheel, axle (not shown), large gear (not shown)), and the small gear (not shown), coupling (not shown), and motor (not shown) that multiply the gear ratio. For this reason, in this embodiment, the detected value of the strain sensor attached to the single link is processed to perform anomaly detection.

[0037] [Example of processing procedure] Next, an example of the processing procedure for the railway vehicle abnormality detection device 1 will be explained using Figures 7 and 8. Figure 7 is a flowchart of the pre-processing and brake force normal / abnormal determination detection process according to this embodiment. Figure 8 is a flowchart of the wheel rotation system normal / abnormal determination detection process according to this embodiment.

[0038] (Step S1) The anomaly detection device 2 acquires the detected value detected by the strain sensor G.

[0039] (Step S2) The vehicle information acquisition unit 210 acquires vehicle information from the control device.

[0040] (Step S3) The preprocessor 201 determines whether the train speed is greater than or equal to a predetermined value (for example, 5 km / h) based on the speed information included in the vehicle information. If the preprocessor 201 determines that the speed is greater than or equal to the predetermined value, it proceeds to step S4. If the preprocessor 201 determines that the speed is less than the predetermined value, it repeats the process in step S3.

[0041] (Step S4) The preprocessing unit 201 performs anti-aliasing filtering and absolute value processing on the detected value detected by the strain sensor G.

[0042] The abnormality detection device 2 performs detection related to the braking force in steps S5 to S13 and detection related to the wheel rotation system in steps S21 to S28. The abnormality detection device 2 may perform the braking force detection process and the wheel rotation system detection process simultaneously, or in a time-division multiplexer, or one of them may be performed first.

[0043] First, the detection process procedure for braking force will be explained with reference to Figure 7. (Step S5) The LPF202 extracts components from the pre-processed detected values, for example, DC (direct current) ~ 3 Hz.

[0044] (Step S6) The RMS value calculation unit 203 calculates the RMS value for the detected value of the bandwidth extracted by the LPF 202.

[0045] (Step S7) When the information acquired by the vehicle information acquisition unit 210 indicates that the brakes are on, the variable a calculation unit 204 calculates variable a by dividing the effective value calculated by the effective value calculation unit 203 by the required brake force of the bogie acquired by the vehicle information acquisition unit 210.

[0046] (Step S8) The determination unit 208 determines whether the brake is in the ON state or not. If the brake is not in the ON state (Step S8; NO), the determination unit 208 repeats the process in step S8. If the brake is in the ON state (Step S8; YES), the determination unit 208 proceeds to the processes in steps S9 and S10.

[0047] (Step S9) The determination unit 208 compares the variable a calculated by the variable a calculation unit 204 with the upper limit value for variable a to determine whether variable a is greater than or equal to the upper limit value for variable a. If the determination unit 208 determines that variable a is greater than the upper limit value for variable a (Step S9; YES), it proceeds to the process in step S11. If the determination unit 208 determines that variable a is less than or equal to the upper limit value for variable a (Step S9; NO), it proceeds to the process in step S13.

[0048] (Step S10) The determination unit 208 compares the variable a calculated by the variable a calculation unit 204 with the lower limit value for variable a to determine whether the lower limit value for variable a is greater than or equal to variable a. If the determination unit 208 determines that the lower limit value for variable a is greater than or equal to variable a (Step S10; YES), it proceeds to the process in step S12. If the determination unit 208 determines that the lower limit value for variable a is not greater than or equal to variable a (Step S10; NO), it proceeds to the process in step S13.

[0049] The determination unit 208 may perform the processing in step S8, for example, between steps S3 and S4. Furthermore, the determination unit 208 may perform the processing in steps S9 and S10 in parallel, perform time-sharing processing, perform the processing in step S10 after step S9, or perform the processing in step S9 after step S10.

[0050] (Step S11) The determination unit 208 determines that the braking force is abnormal (excessive) and terminates the determination detection process.

[0051] (Step S12) The determination unit 208 determines that the braking force is abnormal (insufficient) and terminates the determination detection process.

[0052] (Step S13) The determination unit 208 determines that the braking force is within the normal range and terminates the determination detection process.

[0053] Next, the detection processing procedure for the wheel rotation system will be explained with reference to Figure 8. (Step S21) The HPF205 extracts components with frequencies of 5 Hz or higher from the pre-processed detection values.

[0054] (Step S22) The RMS value calculation unit 206 calculates the RMS value for the detected value of the bandwidth extracted by the HPF 205.

[0055] (Step S23) The variable b calculation unit 207 calculates variable b by dividing the effective value calculated by the effective value calculation unit 206 by the vehicle speed acquired by the vehicle information acquisition unit 210.

[0056] (Step S24) The determination unit 208 compares the variable b calculated by the variable b calculation unit 207 with the upper limit value for variable b to determine whether variable b is greater than or equal to the upper limit value for variable b. If the determination unit 208 determines that variable b is greater than the upper limit value for variable b (Step S24; YES), it proceeds to the process in step S26. If the determination unit 208 determines that variable b is less than or equal to the upper limit value for variable b (Step S24; NO), it proceeds to the process in step S28.

[0057] (Step S25) The determination unit 208 compares the variable b calculated by the variable b calculation unit 207 with the lower limit value for variable b to determine whether the lower limit value for variable b is greater than or equal to variable b. If the determination unit 208 determines that the lower limit value for variable b is greater than or equal to variable b (Step S25; YES), it proceeds to the process in step S27. If the determination unit 208 determines that the lower limit value for variable b is not greater than or equal to variable b (Step S25; NO), it proceeds to the process in step S28.

[0058] The determination unit 208 may perform the processes of step S24 and step S25 in parallel, perform time-division processing, perform the process of step S25 after step S24, or perform the process of step S24 after step S25.

[0059] (Step S26) The determination unit 208 determines that the wheel rotation system is abnormal and terminates the determination detection process.

[0060] (Step S27) The determination unit 208 determines that the wheel rotation system is abnormal and terminates the determination detection process.

[0061] (Step S28) The determination unit 208 determines that the wheel rotation system is within the normal range and terminates the determination detection process.

[0062] Note that the processing procedures shown in Figures 7 and 8 are just examples and are not limited to them.

[0063] [Examples of upper and lower limits] Next, we will explain examples of upper and lower limits. Figure 9 shows examples of upper and lower limits for variable a and variable b according to this embodiment. Graph g100 shows an example of upper and lower limits for variable a. Graph g110 shows an example of upper and lower limits for variable b. In graphs g100 and g110, the horizontal axis represents time (e.g., seconds), and the vertical axis represents the variable and its upper or lower limit.

[0064] As shown in graph g100, in this embodiment, the ratio of the effective value (με) to the required braking force (kN) of the trolley is used as the reference g101, and threshold values ​​(for example, ±10%) for the upper limit g102 and the lower limit g103 are set. Furthermore, as shown in graph g110, in this embodiment, the ratio of the effective value (με) to the speed (km / h) is used as the reference g111, and threshold values ​​(for example, ±10%) for the upper limit g112 and the lower limit g113 are set.

[0065] Note that the method and range for setting the upper and lower limits explained using Figure 9 are just examples, and these may be set according to, for example, the vehicle to be detected, the season, the temperature, the environment in which the vehicle is driving, etc. Furthermore, the anomaly detection device 2 may be configured to switch between upper and lower limits depending on the conditions such as the environment and location mentioned above.

[0066] [Data Processing Timing Chart] Next, the above-mentioned process will be explained using a timing chart to describe the process content and timing. Figure 10 is a timing chart of a data processing example according to this embodiment. In Figure 10, the horizontal axis represents time (seconds), the vertical axis of graph g200 is the train speed (km / h), the vertical axis of graph g210 is the power braking notch value, the vertical axis of graph g220 is the required brake force for the bogie (kN), the vertical axis of graph g230 is the amount of strain (με), the vertical axis of graph g240 is the amount of strain after LPF (με), the vertical axis of graph g250 is the effective value after LPF (με), the vertical axis of graph g260 is variable a (evaluation variable a), the vertical axis of graph g270 is the amount of strain after HPF (με), the vertical axis of graph g280 is the effective value after HPF (με), and the vertical axis of graph g290 is variable b (evaluation variable b). Note that the unit of the effective value is μεrms, but it is abbreviated as με in Figure 10.

[0067] In the example shown in Figure 10, as shown in graph g290, the anomaly detection device 2 performs evaluations over the entire range where the train formation speed of the railway vehicles is above a predetermined value (for example, 5 km / h). As shown in graph g230, multiplying the strain by a coefficient converts it to the actual braking force (kN). Furthermore, as shown in graphs g220 and g250, the result obtained by applying an LPF (Low-Pass Filter) to the strain and then calculating the RMS value is similar to the actual braking force. The abnormality detection device 2 then detects that the brake is ON, as shown in graph g210, and performs an evaluation of variable a within the range where the brake is ON. Note that in the example of graph g260, the upper and lower limits are not shown.

[0068] Furthermore, as shown in graphs g200 and g280, the results obtained by applying HPF processing to the distortion amount and then calculating the RMS value are similar to the train formation speed. The anomaly detection device 2 then performs an evaluation of variable b over the entire range where the speed is above a predetermined value, as shown in graph g290. Note that the upper and lower limits are not shown in the example of graph g290.

[0069] Note that the example shown in Figure 10 is just one example and is not limited to this.

[0070] As described above, in this embodiment, in a railway vehicle, the brake force and wheel rotation system are determined and detected as normal or abnormal by LPF processing, HPF processing, RMS calculation processing, variable calculation processing, and comparison of upper or lower limits with respect to the detected value detected by a strain sensor attached to a single link.

[0071] In other words, the railway vehicle abnormality detection method of this embodiment is The anomaly detection device, The amount of strain detected by a strain sensor attached to one link of a railway vehicle is acquired. The aforementioned amount of distortion is subjected to filtering, The effective value of the filtered strain is calculated, The evaluation variable is calculated using the above effective value. The system detects abnormalities in the railway vehicle by comparing the evaluation variable with a predetermined value. This is a method for detecting abnormalities in railway vehicles.

[0072] As a result, according to this embodiment, only one strain sensor is needed per bogie frame, thus reducing the number of sensors. Furthermore, according to this embodiment, by filtering the amount of strain, it is possible to extract and evaluate the excess or deficiency of the braking force (driving force) applied to the bogie during deceleration (or acceleration) without performing complex signal processing such as FFT analysis. Moreover, according to this embodiment, since the rotational speed of the wheelsets and other components corresponding to the speed is extracted, it is possible to detect abnormalities in the rotational system from the wheels and axles to the small gears, large gears, couplings, and motors (for example, wheel flattening, hunting oscillation, derailment, bogie frame cracks, gear system failure, coupling separation, bearing peeling, and damage).

[0073] In the examples explained using Figures 7 and 8, we described an example where detection and output are performed even when the system is determined to be normal. However, it is also possible to detect and output only when an abnormality is detected, and not detect or output when the system is normal.

[0074] <Second Embodiment> Figure 11 shows an example of the configuration of an abnormality detection device for a railway vehicle according to this embodiment. As shown in Figure 11, the abnormality detection device 1A for a railway vehicle includes a strain sensor GA and an abnormality detection device 2A. The abnormality detection device 2A comprises a pre-processing unit 201, an LPF 202A, a determination unit 208A, an output unit 209, a vehicle information acquisition unit 210, a full-wave rectifier unit 211, and a storage unit 212. Note that the number of gauges used in the first embodiment is four, while the number of gauges used in this embodiment is two.

[0075] The strain sensor GA detects the amount of strain as a detected value. The strain sensor GA consists of two sensors that detect bending strain in a direction perpendicular to a single link, i.e., in the direction of the sleeper.

[0076] The LPF202A performs low-pass filtering (LPF) processing against bending strain. The bandwidth of the LPF202A is, for example, DC to 3 Hz.

[0077] The full-wave rectifier unit 211 is, for example, a bridge circuit, and performs full-wave rectification processing on bending strain. This full-wave rectification processing by the full-wave rectifier unit 211 is performed instead of the RMS value calculation in the first embodiment.

[0078] The determination unit 208A compares the bending strain value after full-wave rectification with a threshold value and detects an anomaly if the bending strain value after full-wave rectification exceeds the threshold value. Alternatively, the determination unit 208A may compare the bending strain value after LPF processing with an upper threshold and a lower threshold value and detect an anomaly if the bending strain value after LPF processing exceeds either the lower threshold or the upper threshold value.

[0079] The memory unit 212 stores the upper and lower threshold values ​​for the value after the LPF, or the threshold value for the value after the bridge circuit.

[0080] Although Figure 11 omits the detection of abnormalities in the wheel rotation system, the abnormality detection device 2A may also include an HPF 205, an RMS value calculation unit 206, and a variable b calculation unit 207, similar to the abnormality detection device 2.

[0081] Here, an example of a method for measuring bending strain will be described. In this embodiment, the two-active gauge method is used as the method for measuring bending strain. Figure 12 shows an example of a sensor for detecting bending strain and the direction of bending stress generation. Rg1 is the first strain sensor, and Rg2 is the second strain sensor. The strain sensors detect stress in the direction of the white arrows.

[0082] Figure 13 shows an example of a bridge circuit. The outputs e0 and E of the bridge circuit in Figure 13 are given by e0 = (E / 2)K s The relationship ε0 exists. The detected value of the first bending strain Rg1 is ε0, and the detected value of the second bending strain Rg2 is -ε0. Also, R is a fixed resistance.

[0083] Figure 14 shows the relationship between bending strain and yaw angular velocity. The horizontal axis represents yaw angular velocity (rad / s), and the vertical axis represents bending strain (με). As shown in Figure 14, experiments have confirmed a correlation between the bending strain of a single link and the yaw angular velocity of the bogie around the vertical z-axis. Therefore, in this embodiment, this relationship is used to detect signs of the bogie derailing from the rails or changes in yaw angular velocity (corresponding to the attack angle) that occur during derailment. Specifically, the determination unit 208A determines that an increase in yaw angular velocity is abnormal, that is, it detects and determines an abnormality based on the amount of strain corresponding to the threshold of yaw angular velocity. Note that the relationship in Figure 14 is just one example and may differ depending on the vehicle, bogie, etc.

[0084] While it is possible to determine the curvature of the rail from the yaw angular velocity and compare it with a database of rail curvatures, the calculated value may diverge. Therefore, in this embodiment, anomaly detection is performed using a strain threshold. If the calculated value does not diverge, the curvature of the rail may be determined from the yaw angular velocity.

[0085] As a result, according to this embodiment, the attack angle detection device, which was conventionally installed, for example, in the axle box, becomes unnecessary.

[0086] Figure 15 is a timing chart of the data processing according to this embodiment. In Figure 15, the horizontal axis represents time (seconds), the vertical axis of graph g300 is the train speed (km / h), the vertical axis of graph g310 is the power braking notch value, the vertical axis of graph g320 is the yaw angular velocity (dag / s), the vertical axis of graph g330 is the bending strain (με), the vertical axis of graph g340 is the curvature (l / m), the vertical axis of graph g350 is the full-wave rectified output of the yaw angular velocity (dag / s), and the vertical axis of graph g360 is the full-wave rectified output of the bending strain (με). Note that in Figure 15, in order to show the relationship between the values ​​for reference, a sensor for detecting yaw angular velocity, which is not provided by the abnormality detection device 2A, was installed on the trolley for experimental purposes to acquire values, and then full-wave rectification processing was performed on the acquired values ​​(graphs g320, g350). Then, the curvature was calculated for reference from the yaw angular velocity set up for the experiment (graph g340). In other words, graphs g320, g340, and g350 are reference values.

[0087] Furthermore, line g321 is the value of the yaw angular velocity after LPF, line g322 is the upper threshold for the yaw angular velocity, and line g323 is the lower threshold for the yaw angular velocity. Line g331 is the value of the bending strain after LPF, line g332 is the upper threshold for the bending strain, and line g333 is the lower threshold for the bending strain. Line g341 is the curvature obtained from the yaw angular velocity. Line g351 is the result of full-wave rectification of the yaw angular velocity, and line g352 is the threshold for the result of full-wave rectification of the yaw angular velocity. Line g361 is the result of full-wave rectification of the bending strain, and line g362 is the threshold for the result of full-wave rectification of the bending strain.

[0088] Figure 15 shows an example of a vehicle in motion as its speed increases. As shown in graph g320, when passing over a curved rail, for example, the yaw angular velocity fluctuates between positive and negative values ​​around 0. Furthermore, as shown in graphs g320 and g330, there is a correlation between the change in yaw angular velocity after the LPF and the change in bending strain after the LPF. For this reason, upper and lower thresholds can be set for the yaw angular velocity after the LPF, and based on a relationship such as that shown in Figure 15, upper and lower thresholds can be set for the corresponding bending strain after the LPF to detect abnormalities.

[0089] Furthermore, as shown in graph g340, the curvature value obtained from the yaw angular velocity is only three decimal places, and therefore, as mentioned above, it may diverge, for example, when the velocity is 0. For this reason, in this embodiment, anomaly detection is performed using a threshold value instead of curvature.

[0090] Furthermore, as shown in graphs g350 and g360, there is a correlation between the value obtained by full-wave rectification with respect to yaw angular velocity and the value obtained by full-wave rectification with respect to bending strain, such as the appearance of peaks. For this reason, the determination unit 208A compares the value obtained by full-wave rectification with respect to bending strain with a threshold value, and detects an anomaly if the value obtained by full-wave rectification with respect to bending strain exceeds the threshold value.

[0091] [Example of processing procedure] Next, an example of the processing procedure for the railway vehicle anomaly detection device 1A will be explained using Figure 16. Figure 16 is a flowchart of the process for determining whether the trolley behavior is normal or abnormal according to this embodiment.

[0092] (Step S101) The anomaly detection device 2A acquires the detected value detected by the strain sensor GA.

[0093] (Step S102) The vehicle information acquisition unit 210 acquires vehicle information from the control device.

[0094] (Step S103) The preprocessor 201 determines whether the train speed is greater than or equal to a predetermined value (for example, 5 km / h) based on the speed information included in the vehicle information. If the preprocessor 201 determines that the speed is greater than or equal to the predetermined value, it proceeds to the process in step S104. If the preprocessor 201 determines that the speed is less than the predetermined value, it repeats the process in step S103.

[0095] (Step S104) The preprocessing unit 201 performs, for example, anti-aliasing filtering on the detected value detected by the strain sensor GA.

[0096] (Step S105) The LPF202A extracts components from the pre-processed detected values, for example, DC (direct current) ~ 3 Hz.

[0097] (Step S106) The full-wave rectifier unit 211 performs full-wave rectification on the detected values ​​of the bandwidth extracted by the LPF 202A.

[0098] (Step S107) The determination unit 208A compares the full-wave rectified bending strain with a threshold value to determine whether the full-wave rectified bending strain is greater than the threshold value. If the determination unit 208A determines that the full-wave rectified bending strain is greater than the threshold value (Step S107; YES), it proceeds to step S108. If the determination unit 208A determines that the full-wave rectified bending strain is less than or equal to the threshold value (Step S107; NO), it proceeds to step S109.

[0099] (Step S108) The determination unit 208A determines that the trolley behavior is abnormal and terminates the determination detection process.

[0100] (Step S109) The determination unit 208A determines that the trolley behavior is within the normal range and terminates the determination detection process.

[0101] In this embodiment as well, if detection related to the brake force system is performed, the processing in steps S5 to S13 (Figure 7) may be carried out in the same manner as in the first embodiment.

[0102] In the example described above, the full-wave rectifier unit 211 performs full-wave rectification, but this is not the only example. For example, if positive and negative values ​​are not considered, the RMS value may be used. In this case, the anomaly detection device 2A may be equipped with an RMS value calculation unit instead of the full-wave rectifier unit 211.

[0103] As described above, in this embodiment, anomaly detection is performed by comparing the result of full-wave rectification processing of the value acquired by the sensor with a threshold value.

[0104] As a result, according to this embodiment, the configuration of the anomaly detection device 2A can be simplified.

[0105] Furthermore, a program to implement all or part of the functions of the anomaly detection device 2 (or 2A) in this invention may be recorded on a computer-readable recording medium, and all or part of the processing performed by the anomaly detection device 2 (or 2A) may be performed by loading the program recorded on this recording medium into a computer system and executing it. Herein, "computer system" includes hardware such as an OS and peripheral devices. Furthermore, "computer system" also includes a WWW system equipped with a homepage provisioning environment (or display environment). Furthermore, "computer-readable recording medium" refers to portable media such as flexible disks, magneto-optical disks, ROMs, CD-ROMs, and storage devices such as hard disks built into a computer system. Moreover, "computer-readable recording medium" also includes volatile memory (RAM) inside a computer system that acts as a server or client when a program is transmitted via a network such as the Internet or a communication line such as a telephone line, which holds the program for a certain period of time.

[0106] Furthermore, the above program may be transmitted from a computer system that stores the program in a memory device or the like to another computer system via a transmission medium or by transmission waves within the transmission medium. Here, the "transmission medium" for transmitting the program refers to a medium that has the function of transmitting information, such as a network (communication network) such as the Internet or a communication line (communication line) such as a telephone line. In addition, the above program may be for the purpose of realizing a part of the functions described above. Furthermore, it may be a so-called differential file (differential program) that can realize the functions described above in combination with a program already recorded in the computer system.

[0107] Although embodiments for carrying out the present invention have been described above using examples, the present invention is not limited in any way to these embodiments, and various modifications and substitutions can be made without departing from the spirit of the present invention. [Explanation of Symbols]

[0108] 1,1A…Railway vehicle anomaly detection device, G,GA…Strain sensor, 2,2A…Anomaly detection device, 201…Pre-processing unit, 202,202A…LPF, 203…RMS value calculation unit, 204…Variable a calculation unit, 205…HPF, 206…RMS value calculation unit, 207…Variable b calculation unit, 208,208A…Determination unit, 209…Output unit, 210…Vehicle information acquisition unit, 211…Full-wave rectification unit, 212…Storage unit

Claims

1. The anomaly detection device, A strain sensor attached to a single link of a railway vehicle detects the tensile and compressive loads acting on the main body of the single link, or the amount of strain representing a direction perpendicular to the single link, and acquires these values. The aforementioned distortion amount is subjected to filtering, including low-pass filtering that extracts components from at least DC to 3 Hz. With respect to the filtered distortion amount, one or both of the following predetermined processes are performed: a process to calculate the effective value and a full-wave rectification process. The value obtained by the aforementioned predetermined process is compared with a predetermined value to detect an abnormality in the railway vehicle. Methods for detecting abnormalities in railway vehicles.

2. The aforementioned strain amount represents the strain amount that represents the tensile and compressive loads acting on the main body of the single link. The filtering process includes a high-pass filter that extracts components of 5 Hz or higher. In the predetermined process described above, the effective value of the filtered strain is calculated, and the evaluation variable is calculated using the effective value. The value obtained by the predetermined process is the evaluation variable. The method for detecting abnormalities in a railway vehicle according to claim 1.

3. The aforementioned strain amount represents the strain amount in a direction perpendicular to the single link, The value obtained by the predetermined process is the value obtained by full-wave rectifying the filtered distortion amount. In the predetermined process, full-wave rectification is performed on the filtered distortion amount. The method for detecting abnormalities in a railway vehicle according to claim 1.

4. The filtering process is the low-pass filtering process, When the brakes of the railway vehicle are in the ON state, the abnormality detection device calculates a first evaluation variable by dividing the effective value of the low-pass filter-processed distortion by the required braking force, which is the command value for the bogie of the railway vehicle. By comparing the upper and lower limits set for the first evaluation variable with the first evaluation variable, an abnormality in the braking force of the railway vehicle is detected. The method for detecting abnormalities in a railway vehicle according to claim 2.

5. The system detects that the braking force of the railway vehicle is within the normal range when the result of comparing the first evaluation variable with the upper and lower limits set for the first evaluation variable is such that the first evaluation variable is less than or equal to the upper limit and greater than or equal to the lower limit. The method for detecting abnormalities in a railway vehicle according to claim 4.

6. The filtering process is the high-pass filtering process, When the abnormality detection device detects that the train formation speed used in controlling the railway vehicle is greater than or equal to a predetermined value, it calculates a second evaluation variable by dividing the effective value of the distortion amount processed by the high-pass filter by the train formation speed used in controlling the railway vehicle. By comparing the upper and lower limits set for the second evaluation variable with the second evaluation variable, an abnormality in the wheel rotation system of the railway vehicle is detected. A method for detecting abnormalities in a railway vehicle according to any one of claims 2, 4, or 5.

7. The system detects that the wheel rotation system of the railway vehicle is within the normal range when the result of comparing the upper and lower limits set for the second evaluation variable with the second evaluation variable is such that the second evaluation variable is less than or equal to the upper limit and greater than or equal to the lower limit. The method for detecting abnormalities in a railway vehicle according to claim 6.

8. Before the anomaly detection device performs the filtering process on the strain amount, it performs anti-aliasing filtering and absolute value filtering on the acquired strain amount. A method for detecting abnormalities in a railway vehicle according to any one of claims 2, 4, or 5.

Images