A method, device, equipment and medium for detecting wheel tread faults

Abstract

The present application discloses a method, device, equipment and medium for detecting wheel tread faults, including: obtaining vehicle operation state data of each target type in the current cycle; the vehicle operation state data of each target type includes vertical vibration data of axle boxes and tread polygon impact data; determining any vehicle operation state data of the target type in the current cycle as target data, and determining the trend change state of the target data in the current cycle based on the historical cycle data corresponding to the target data; outputting a polygon fault detection result of the wheel tread based on the trend change state of the vehicle operation state data of each target type. In this way, the problem of inaccurate detection when the runout is small is avoided, the accuracy of detecting wheel tread polygon faults can be improved, and the detection cost can be reduced.

Description

Technical Field

[0001] The present application relates to the technical field of wheel tread fault detection, and particularly relates to a wheel tread fault detection method, device, equipment and medium. Background Art

[0002] The wheel tread is one of the key components of the running gear of high-speed EMUs, and its operating state has a very important impact on vehicle vibration, train operation safety, etc. Polygonal out-of-roundness fault is a common failure form of wheel treads. After such a fault occurs, the vehicle is often accompanied by strong vibration and noise, which not only directly harms the tread, but also deteriorates the operating conditions of other related components such as motors, gears, bogies, and shock absorbers.

[0003] Currently, the detection schemes for polygonal faults include two methods: offline detection by a lathe and online detection. Among them, offline detection by a lathe means that the treads of the entire train set are statically detected offline at a fixed maintenance cycle, and the tread polygonal faults are turned and repaired according to the detection results. This method is relatively general, lacks purposefulness, is time-consuming and laborious, and has a high cost. The conventional method of online detection is to monitor the axle box vibration, calculate the runout by integrating the vertical vibration acceleration, and judge the out-of-roundness. However, for a high-speed running train, due to the high operating speed, when there is a tread polygonal fault during operation, even when the runout is small, high-frequency vibration will be generated, which will harm the frame and axle box related components. This scheme has the problem of inaccurate detection when the runout is small. Summary of the Invention

[0004] In view of this, the purpose of the present application is to provide a wheel tread fault detection method, device, equipment and medium, which can improve the accuracy of detecting wheel tread polygonal faults and reduce the detection cost. The specific scheme is as follows:

[0005] In a first aspect, the present application discloses a wheel tread fault detection method, including:

[0006] Obtaining vehicle operation state data of each target type in the current cycle; the vehicle operation state data of each target type includes axle box vertical vibration data and tread polygonal impact data;

[0007] Determining any vehicle operation state data of the target type in the current cycle as target data, and determining the trend change state of the target data in the current cycle based on the historical cycle data corresponding to the target data;

[0008] Outputting a detection result of the polygonal fault of the wheel tread based on the trend change state of the vehicle operation state data of each target type.

[0009] Optionally, determining the trend change state of the target data in the current period based on the historical period data corresponding to the target data includes:

[0010] Performing a sliding window process on the target data to obtain a sliding window process result for the current period;

[0011] Determining the trend change state of the target data in the current period based on the sliding window process result for the current period and the sliding window process result for the historical period;

[0012] Among them, the sliding window process result includes the window mean, window variance, and fitting slope of the window.

[0013] Optionally, determining the trend change state of the target data in the current period based on the sliding window process result for the current period and the sliding window process result for the historical period includes:

[0014] Determining a state change judgment threshold for the target data in the current period based on the sliding window process result for the historical period;

[0015] Judging the trend change state of the target data in the current period based on the state change judgment threshold and the sliding window process result for the current period.

[0016] Optionally, determining the state change judgment threshold for the target data in the current period based on the sliding window process result for the historical period includes:

[0017] Determining windows with a window variance less than a first preset variance threshold from the sliding window process results of a preset number of most recent historical periods to obtain first target windows;

[0018] Calculating the mean of the window means of all the first target windows to obtain the state change judgment threshold for the target data in the current period.

[0019] Optionally, the trend change state includes any one of an ascending state, a non-ascending state, and a continuous ascending state.

[0020] Optionally, if the target data meets the first ascending state determination condition and / or the second ascending state determination condition, it is determined that the trend change state of the target data in the current period is an ascending state;

[0021] Among them, the first ascending state determination condition is: the target data is greater than the state change judgment threshold, the window slope in the sliding window process result for the current period is greater than a first preset slope threshold, and the window variance in the sliding window process result for the current period is greater than a second preset variance threshold;

[0022] The second rising state determination condition is that the number of second target windows meeting the preset conditions reaches a quantity threshold, and the fitting slope of the target data and the historical period data corresponding to the target data within the first preset number of consecutive windows is greater than the second preset slope threshold; the second target window is a window among the preset number of consecutive windows, and the preset condition is that the fitting slope of the window is within a preset range and the window variance is less than the third preset variance threshold.

[0023] Optionally, if the target data meets the first non-rising state determination condition and / or the second non-rising state determination condition, it is determined that the trend change state of the target data in the current period is a non-rising state;

[0024] Among them, the first non-rising state determination condition is that there is a window with a window slope lower than the third preset slope threshold among the second preset number of consecutive windows; among them, the second preset number of consecutive windows are all windows before the current period; and the second non-rising state determination condition is that the window variance corresponding to the current period is greater than the fourth preset variance threshold, and the window change rate corresponding to the current period is lower than the preset change rate threshold; the window change rate is the change rate of the start and end data of the window corresponding to the current period.

[0025] Optionally, if the target data is greater than the preset multiple of the state change judgment threshold, and the trend change states of the vehicle operation state data of this target type in consecutive M periods are all rising states or continuously rising states, it is determined that the trend change state of the target data in the current period is a continuously rising state;

[0026] Among them, the consecutive M periods include the current period.

[0027] Optionally, the determining the trend change state of the target data in the current period based on the state change judgment threshold and the sliding window processing result of the current period includes:

[0028] Determining whether the target data meets the first rising state determination condition based on the state change judgment threshold and the sliding window processing result of the current period. If the target data meets the first rising state determination condition, it is preliminarily determined that the trend change state of the target data in the current period is a rising state, and it is determined whether the target data meets the first non-rising state determination condition. If the target data does not meet the first rising state determination condition, it is determined whether the target data meets the second rising state determination condition. If the target data meets the second rising state determination condition, it is preliminarily determined that the trend change state of the target data in the current period is a rising state, and it is determined whether the target data meets the first non-rising state determination condition;

[0029] If the target data meets the first non-rising state determination condition, it is finally determined that the trend change state of the target data in the current period is a non-rising state. If the target data does not meet the first non-rising state determination condition, it is determined whether the target data meets the second non-rising determination condition. If the target data meets the second non-rising determination condition, it is finally determined that the trend change state of the target data in the current period is a non-rising state. If the target data does not meet the second non-rising determination condition, it is determined whether the target data meets the preset continuous rising determination condition;

[0030] If the target data meets the preset continuous rising determination condition, it is finally determined that the trend change state of the target data in the current period is a continuous rising state. If it does not meet the preset continuous rising determination condition, it is finally determined that the trend change state of the target data in the current period is a rising state.

[0031] Optionally,

[0032] The first rising state determination condition is: the target data is greater than the state change judgment threshold, the window slope in the sliding window processing result of the current period is greater than the first preset slope threshold, and the window variance in the sliding window processing result of the current period is greater than the second preset variance threshold;

[0033] The second rising state determination condition is: the number of second target windows that meet the preset conditions reaches the quantity threshold, and the fitting slope of the target data and the historical period data corresponding to the target data within the first preset number of consecutive windows is greater than the second preset slope threshold; the second target window is the window in the preset number of consecutive windows, and the preset condition is that the fitting slope of the window is within the preset range and the window variance is less than the third preset variance threshold;

[0034] The first non-rising state determination condition is: there is a window in the second preset number of consecutive windows whose window slope is lower than the third preset slope threshold; among them, the windows before the current period of the second preset number of consecutive windows;

[0035] The second non-rising state determination condition is: the window variance corresponding to the current period is greater than the fourth preset variance threshold, and the window change rate corresponding to the current period is lower than the preset change rate threshold; the window change rate is the change rate of the start and end data of the window corresponding to the current period;

[0036] The preset continuous rising determination condition is: the target data is greater than the state change judgment threshold multiplied by a preset multiple, and the change state of the vehicle operation state data of this target type in M consecutive periods is a rising state or a continuous rising state, then it is determined that the change state of the target data in the current period is a continuous rising state; among them, M consecutive periods include the current period.

[0037] Optionally,

[0038] The process of obtaining the vertical vibration data of the axle box in the current cycle is as follows: calculate the effective value of the vertical vibration acceleration signal of the axle box at each moment within the current cycle; determine the mean value of all the effective values within the current cycle to obtain the vertical vibration data of the axle box in the current cycle;

[0039] The process of obtaining the vertical vibration data of the axle box in the current cycle is as follows: calculate the range difference of the tread polygon impact signal at each moment within the current cycle; determine the mean value of all the range differences within the current cycle to obtain the tread polygon impact data of the current cycle.

[0040] Optionally, it further includes:

[0041] Determine the stable operating speed per hour in the current cycle;

[0042] Correspondingly, the output of the polygon fault detection result of the wheel tread based on the trend change state of the vehicle operating state data of each target type includes:

[0043] Output the polygon fault detection result of the wheel tread based on the stable operating speed per hour and the trend change state of the vehicle operating state data of each target type.

[0044] Optionally, the determination of the stable operating speed per hour in the current cycle includes:

[0045] Count the frequencies in each stable operating speed range in the current cycle to obtain statistical data;

[0046] Determine the speed level corresponding to the maximum frequency in the statistical data as the stable operating speed per hour.

[0047] Optionally, the cycle is one day, and the polygon fault detection result includes a fault reminder. The output of the polygon fault detection result of the wheel tread based on the stable operating speed per hour and the trend change state of the vehicle operating state data of each target type includes:

[0048] If the stable operating speed per hour is at the 350 km / h speed level, output the fault reminder of the wheel tread based on the first preset reminder strategy and the trend change state of the vehicle operating state data of each target type;

[0049] If the stable operating speed per hour is at the 300 km / h speed level, output the fault reminder of the wheel tread based on the second preset reminder strategy and the trend change state of the vehicle operating state data of each target type;

[0050] Among them, the first preset reminder strategy includes outputting first-level reminders, second-level reminders, and third-level reminders, and the reminder level increases with the increase in severity; if the trend change states of the vertical vibration data of the axle box and the polygon impact data of the tread in the current cycle are in the rising state, rising state, or rising state, non-rising state, or non-rising state, rising state, then a first-level reminder is output; if the trend change states of the vertical vibration data of the axle box and the polygon impact data of the tread in the current cycle are in the continuously rising state, rising state, or continuously rising state, non-rising state, or rising state, continuously rising state, or non-rising state, continuously rising state, then it is determined that the reminder determination result in the current cycle is a second-level reminder. If the reminder determination result in the current cycle is not the Nth consecutive second-level reminder, then a second-level reminder is output; if the trend change states of the vertical vibration data of the axle box and the polygon impact data of the tread in the current cycle are in the continuously rising state, continuously rising state, or the reminder determination result in the current cycle is the Nth consecutive second-level reminder, then a third-level reminder is output.

[0051] Moreover, the second preset reminder strategy includes outputting first-level reminders, second-level reminders, and third-level reminders, and the reminder level increases with the increase in severity; if the trend change states of the vertical vibration data of the axle box and the polygon impact data of the tread in the current cycle are in the rising state, rising state, or non-rising state, rising state, then a first-level reminder is output; if the trend change states of the vertical vibration data of the axle box and the polygon impact data of the tread in the current cycle are in the continuously rising state, rising state, or rising state, continuously rising state, or non-rising state, continuously rising state, then it is determined that the reminder determination result in the current cycle is a second-level reminder. If the reminder determination result in the current cycle is not the Nth consecutive second-level reminder, then a second-level reminder is output; if the trend change states of the vertical vibration data of the axle box and the polygon impact data of the tread in the current cycle are in the continuously rising state, continuously rising state, or the reminder determination result in the current cycle is the Nth consecutive second-level reminder, then a third-level reminder is output.

[0052] In a second aspect, the present application discloses a wheel tread fault detection device, including:

[0053] A state data acquisition module, configured to acquire the vehicle operation state data of each target type in the current cycle; the vehicle operation state data of each target type includes the vertical vibration data of the axle box and the polygon impact data of the tread;

[0054] A change state determination module, configured to determine any one of the vehicle operation state data of the target type in the current cycle as target data, and determine the trend change state of the target data in the current cycle based on the historical cycle data corresponding to the target data;

[0055] The detection result output module is configured to output the detection result of the polygonal fault of the wheel tread based on the trend change state of the vehicle operation state data of each target type.

[0056] In a third aspect, the present application discloses an electronic device, which includes a memory and a processor, wherein:

[0057] The memory is used to store a computer program;

[0058] The processor is configured to execute the computer program to implement the aforementioned method for detecting the fault of the wheel tread.

[0059] In a fourth aspect, the present application discloses a computer-readable storage medium for storing a computer program, wherein the computer program, when executed by a processor, implements the aforementioned method for detecting the fault of the wheel tread.

[0060] It can be seen that the present application obtains the vehicle operation state data of each target type in the current cycle. The vehicle operation state data of each target type includes the vertical vibration data of the axle box and the impact data of the tread polygon. Any vehicle operation state data of the target type in the current cycle is determined as the target data, and the trend change state of the target data in the current cycle is determined based on the historical cycle data corresponding to the target data. And the detection result of the polygonal fault of the wheel tread is output based on the trend change state of the vehicle operation state data of each target type. That is to say, in the process of detecting the polygonal fault of the wheel tread, the present application uses the vertical vibration data of the axle box in the historical cycle to judge the trend change state of the vertical vibration data of the axle box in the current cycle, and uses the impact data of the tread polygon in the historical cycle to judge the trend change state of the impact data of the tread polygon in the current cycle. In this way, the trend change state of the vertical vibration data of the axle box and the impact data of the tread polygon is judged through the historical cycle data, and then the detection result of the polygonal fault is determined, avoiding the problem of inaccurate detection when the runout is small, improving the accuracy of detecting the polygonal fault of the wheel tread, and reducing the detection cost.

[0061] Moreover, the present application determines the state change judgment threshold of the target data in the current cycle based on the sliding window processing result of the historical cycle, and judges the trend change state of the target data in the current cycle based on the state change judgment threshold and the sliding window processing result of the current cycle. In this way, the state change judgment threshold of the current cycle is dynamically determined, which can make an adaptive adjustment for different lines and different running speeds, so as to better adapt to application scenarios such as the line allocation of the multiple unit train. Description of the Drawings

[0062] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the accompanying drawings required in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only the embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can also be obtained based on the provided drawings.

[0063] Figure 1 Flowchart of a wheel tread fault detection method disclosed in the present application;

[0064] Figure 2 Schematic diagram of a sliding window processing disclosed in the present application;

[0065] Figure 3 Schematic diagram of calculating the state change judgment threshold corresponding to the vertical vibration data of the axle box disclosed in the present application;

[0066] Figure 4 Schematic diagram of calculating the state change judgment threshold corresponding to the tread polygon impact data disclosed in the present application;

[0067] Figure 5 Schematic diagram of the structure of a wheel tread fault detection device disclosed in the present application;

[0068] Figure 6 Schematic diagram of the structure of an electronic device disclosed in the present application. Detailed implementation manners

[0069] The following will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments of the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present application.

[0070] Currently, the detection schemes for polygonal faults include two methods: offline detection on a lathe and online detection. Among them, offline detection on a lathe means that the treads of the entire train set are statically detected offline every fixed maintenance cycle, and the polygonal faults of the treads are turned and repaired according to the detection results. This method is relatively general, lacks purpose, is time-consuming and laborious, and has a high cost. The conventional method of online detection is to monitor the vibration of the axle box. By integrating the vertical vibration acceleration to calculate the radial runout, the out-of-roundness is judged. However, for a high-speed running train, due to its high operating speed, when there are polygonal faults on the tread during operation, even when the radial runout is small, high-frequency vibrations will be generated, which will cause harm to the frame and related components of the axle box. This scheme has the problem of inaccurate detection when the radial runout is small. Therefore, this application provides a detection scheme for wheel tread faults, which can improve the accuracy of detecting polygonal faults on the wheel tread and reduce the detection cost.

[0071] See Figure 1 As shown, an embodiment of this application discloses a method for detecting wheel tread faults, including:

[0072] Step S11: Obtain the vehicle operation state data of each target type in the current cycle; the vehicle operation state data of each target type includes axle box vertical vibration data and tread polygon impact data.

[0073] In one implementation, the process of obtaining the axle box vertical vibration data in the current cycle is as follows: calculate the effective value of the axle box vertical vibration acceleration signal at each moment in the current cycle; determine the mean value of all the effective values in the current cycle to obtain the axle box vertical vibration data in the current cycle; the process of obtaining the tread polygon impact data in the current cycle is as follows: calculate the range difference of the tread polygon impact signal at each moment in the current cycle; determine the mean value of all the range differences in the current cycle to obtain the tread polygon impact data in the current cycle.

[0074] It should be noted that in the embodiments of the present application, the vertical vibration acceleration signal of the axle box can be collected through an information collection system. In one embodiment, the collected data can be analyzed offline, and in another embodiment, the collected data can be analyzed online. Further, the information collection system includes a vibration and shock detection composite sensor installed on the wheel axle box for collecting the vertical vibration acceleration signal of the axle box. In the embodiments of the present application, multiple vertical vibration acceleration signal values of the axle box can be collected at each moment, and the root mean square value, that is, the effective value, of the vertical vibration acceleration signal of the axle box at each moment can be calculated based on the multiple vertical vibration acceleration signal values. Then, the mean value of the effective values at all moments within a period is determined to obtain the vertical vibration data of the axle box for this period. Moreover, the tread polygon impact signal is calculated using the vertical vibration acceleration signal of the axle box. The tread polygon impact signal at each moment within any period is normalized, and the extreme difference value is calculated. The mean value of the extreme difference values at all moments within this period is determined to obtain the tread polygon impact data for this period, and the tread polygon impact data is in dB value.

[0075] Step S12: Determine the vehicle operation state data of any of the target types in the current period as the target data, and determine the trend change state of the target data in the current period based on the historical period data corresponding to the target data.

[0076] It can be understood that in the present application, the vertical vibration data of the axle box and the tread polygon impact data are processed with the same logic to obtain the corresponding trend change states respectively.

[0077] In one embodiment, the target data can be processed by a sliding window to obtain the sliding window processing result for the current period; the trend change state of the target data in the current period is determined based on the sliding window processing result for the current period and the sliding window processing result for the historical period; wherein, the sliding window processing result includes the window mean value, window variance, and fitting slope of the window. Moreover, in the embodiments of the present application, the target data and the data of multiple historical periods can be processed by a sliding window to obtain the sliding window processing result for the current period, where the multiple historical periods can be multiple consecutive historical periods closest to the current period, and the number of the multiple historical periods is determined based on the window length of the sliding window. For example, if the window length is 5, then the multiple historical periods are 4 historical periods. Also, in the embodiments of the present application, when the vertical vibration data of the axle box and the tread polygon impact data for the current period are obtained, the vertical vibration data of the axle box and the tread polygon impact data of multiple historical periods are obtained from the storage component. The period in the embodiments of the present application can be 1 day or 0.5 day, etc.

[0078] Next, taking one day as a period as an example, the window length of the sliding window is 5 and the sliding step is 1, and the mean value, variance, and fitting slope of the data within the window are calculated. See Figure 2As shown Figure 2 This is a schematic diagram of sliding window processing disclosed in an embodiment of the present application. Among them, the data within the window in Figure a is the daily average value of effective vibration days from July 21st to July 25th. The average value, variance, and fitting slope of the data within the window are calculated. After the window slides with a step size of 1, as shown in Figure b, the data within the window is the daily average value of effective vibration days from July 22nd to July 26th. Among them, the daily average value of effective vibration days is also the aforementioned vertical vibration data of the axle box.

[0079] Furthermore, determining the trend change state of the target data in the current period based on the sliding window processing result of the current period and the sliding window processing result of the historical period may specifically include the following steps:

[0080] Step 00: Determine the state change judgment threshold of the target data in the current period based on the sliding window processing result of the historical period.

[0081] In one implementation manner, windows with window variances less than the first preset variance threshold can be determined from the sliding window processing results of a preset number of recent historical periods to obtain the first target windows; the average values of the window means of all the first target windows are calculated to obtain the state change judgment threshold of the target data in the current period. It should be noted that the vertical vibration data of the axle box and the polygon impact data of the tread surface may correspond to different first preset variance thresholds.

[0082] For example, if the preset number of recent historical periods is 10 days, for the vertical vibration data of the axle box, the window means corresponding to the dates with window variances less than α before the current day can be screened out, and their average value is the state change judgment threshold corresponding to the vertical vibration data of the axle box. As Figure 3 shown Figure 3 This is a schematic diagram of calculating the state change judgment threshold corresponding to the vertical vibration data of the axle box disclosed in an embodiment of the present application. Screen out all window variances less than α before August 1st, and calculate the average value of the corresponding window means, that is, the state change judgment threshold corresponding to the vertical vibration data of the axle box on August 1st. Similarly, for the polygon impact data of the tread surface, the window means corresponding to the dates with window variances less than β before the current day are screened out, and their average value is the state change judgment threshold corresponding to the polygon impact data of the tread surface. As Figure 4 shown Figure 4 This is a schematic diagram of calculating the state change judgment threshold corresponding to the polygon impact data of the tread surface disclosed in an embodiment of the present application. Screen out all window variances less than β before August 1st, and calculate the average value of the corresponding window means, that is, the state change judgment threshold corresponding to the polygon impact data of the tread surface on August 1st. Among them, the daily average value of impact dB is the aforementioned polygon impact data of the tread surface. That is, in the embodiment of the present application, the threshold is calculated based on the latest data every day. After receiving the daily average value of effective vibration days and the daily average value of impact dB data of the latest day, the vibration threshold and the impact dB threshold of the current day are calculated.

[0083] It should be noted that it is difficult to form a judgment standard when the same vehicle is allocated on different lines. Since the vertical vibration of the vehicle is related to the wheel-rail relationship and is greatly affected by the track state, there are obvious differences in axle box vibration due to the track state differences on different lines. In each cycle of the embodiment of the present application, the state change judgment threshold is recalculated. In this way, the dynamic calculation threshold can be adaptively adjusted for different lines and different running speeds, and can better adapt to application scenarios such as the line allocation of EMUs.

[0084] Step 01: Based on the state change judgment threshold and the result of the sliding window processing in the current cycle, judge the trend change state of the target data in the current cycle.

[0085] Among them, the trend change state includes any one of an ascending state, a non-ascending state, and a continuous ascending state.

[0086] Furthermore, if the target data meets the first ascending state determination condition and / or the second ascending state determination condition, it is determined that the trend change state of the target data in the current cycle is an ascending state; among them, the first ascending state determination condition is: the target data is greater than the state change judgment threshold, and the window slope in the result of the sliding window processing in the current cycle is greater than the first preset slope threshold, and the window variance in the result of the sliding window processing in the current cycle is greater than the second preset variance threshold; the second ascending state determination condition is: the number of second target windows that meet the preset conditions reaches the quantity threshold, and the fitting slope of the target data and the historical cycle data corresponding to the target data within the first preset number of consecutive windows is greater than the second preset slope threshold; the second target window is the window in the preset number of consecutive windows, and the preset condition is that the fitting slope of the window is within the preset range and the window variance is less than the third preset variance threshold. And the quantity threshold can be determined based on the first preset number. For example, if the first preset number is 20, the quantity threshold can be 15.

[0087] Moreover, if the target data meets the first non-ascending state determination condition and / or the second non-ascending state determination condition, it is determined that the trend change state of the target data in the current cycle is a non-ascending state; among them, the first non-ascending state determination condition is: there is a window in the second preset number of consecutive windows whose window slope is lower than the third preset slope threshold; among them, the second preset number of consecutive windows are all the windows before the current cycle; the second non-ascending state determination condition is: the window variance corresponding to the current cycle is greater than the fourth preset variance threshold, and the window change rate corresponding to the current cycle is lower than the preset change rate threshold; the window change rate is the change rate of the head and tail data of the window corresponding to the current cycle.

[0088] Moreover, if the target data is greater than the state change judgment threshold of a preset multiple, and the trend change state of the vehicle operation state data of the target type is in an ascending state or a continuously ascending state in consecutive M cycles, it is determined that the trend change state of the target data in the current cycle is a continuously ascending state; where the consecutive M cycles include the current cycle.

[0089] In one implementation, the method for determining the trend change state of the target data in the current cycle based on the state change judgment threshold and the sliding window processing result of the current cycle may specifically include the following steps:

[0090] Based on the state change judgment threshold and the sliding window processing result of the current cycle, determine whether the target data meets the first ascending state determination condition or the second ascending state determination condition. If the target data meets the first ascending state determination condition or the second ascending state determination condition, preliminarily determine that the trend change state of the target data in the current cycle is an ascending state;

[0091] If the trend change state of the target data in the current cycle is preliminarily determined to be an ascending state, and the target data meets the first non-ascending state determination condition or the second non-ascending determination condition, then finally determine that the trend change state of the target data in the current cycle is a non-ascending state;

[0092] If the change state of the target data in the current cycle is preliminarily determined to be an ascending state, and the target data does not meet the first non-ascending state determination condition and the second non-ascending determination condition, and meets the preset continuous ascending determination condition, then finally determine that the trend change state of the target data in the current cycle is a continuously ascending state. If the preset continuous ascending determination condition is not met, then finally determine that the trend change state of the target data in the current cycle is an ascending state.

[0093] In another implementation, the method for determining the trend change state of the target data in the current cycle based on the state change judgment threshold and the sliding window processing result of the current cycle may specifically include the following steps:

[0094] Based on the state change judgment threshold and the sliding window processing result of the current cycle, determine whether the target data meets the first ascending state determination condition. If the target data meets the first ascending state determination condition, preliminarily determine that the trend change state of the target data in the current cycle is an ascending state, and determine whether the target data meets the first non-ascending state determination condition. If the target data does not meet the first ascending state determination condition, then determine whether the target data meets the second ascending state determination condition. If the target data meets the second ascending state determination condition, preliminarily determine that the trend change state of the target data in the current cycle is an ascending state, and determine whether the target data meets the first non-ascending state determination condition;

[0095] If the target data meets the first non-rising state determination condition, it is finally determined that the trend change state of the target data in the current period is a non-rising state. If the target data does not meet the first non-rising state determination condition, it is determined whether the target data meets the second non-rising determination condition; if the target data meets the second non-rising determination condition, it is finally determined that the trend change state of the target data in the current period is a non-rising state. If the target data does not meet the second non-rising determination condition, it is determined whether the target data meets the preset continuous rising determination condition;

[0096] If the target data meets the preset continuous rising determination condition, it is finally determined that the trend change state of the target data in the current period is a continuous rising state. If it does not meet the preset continuous rising determination condition, it is finally determined that the trend change state of the target data in the current period is a rising state.

[0097] Among them, the first rising state determination condition is: the target data is greater than the state change judgment threshold, the window slope in the sliding window processing result of the current period is greater than the first preset slope threshold, and the window variance in the sliding window processing result of the current period is greater than the second preset variance threshold; the second rising state determination condition is: the number of second target windows that meet the preset conditions reaches the number threshold, and the fitting slope of the target data and the historical period data corresponding to the target data within the first preset number of consecutive windows is greater than the second preset slope threshold; the second target window is the window in the preset number of consecutive windows, and the preset condition is that the fitting slope of the window is within the preset range and the window variance is less than the third preset variance threshold; the first non-rising state determination condition is: there is a window in the second preset number of consecutive windows whose window slope is lower than the third preset slope threshold; among them, the second preset number of consecutive windows are the windows before the current period; the second non-rising state determination condition is: the window variance corresponding to the current period is greater than the fourth preset variance threshold, and the window change rate corresponding to the current period is lower than the preset change rate threshold; the window change rate is the change rate of the start and end data of the window corresponding to the current period; the preset continuous rising determination condition is: the target data is greater than the state change judgment threshold multiplied by a preset multiple, and the change state of the vehicle operation state data of this target type in M consecutive periods is a rising state or a continuous rising state, then it is determined that the change state of the target data in the current period is a continuous rising state; among them, the M consecutive periods include the current period.

[0098] Taking the trend change state judgment process of the vertical vibration data of the axle box as an example, the trend change state judgment process of the embodiment of the present application is introduced as follows:

[0099] Step a: The data of the current day simultaneously meets three conditions: ① The window slope is greater than α1 (e.g., 0.18), ② the window variance is greater than β 1 (e.g., 0.8), ③ the daily average value of the effective vibration days is greater than the vibration threshold. It is preliminarily judged as the rising state and step c is carried out, otherwise step b is carried out;

[0100] Step b: Among N (e.g., 20) windows including the current day, X (e.g., 15) or more windows meet the conditions: the slope is in the range of [-δ, δ] (e.g., [-0.18, 0.18]), and the variance is lower than β 2 (e.g., 0.4). Then calculate the fitting slope of the daily average value of all effective vibration days within these N windows. If this slope is greater than α 2 (e.g., 0.05), it is preliminarily judged as the rising state and step c is carried out, otherwise the process ends;

[0101] Step c: If among the consecutive Y (e.g., 7) windows before the current day, there is a window with a slope lower than α 3 (e.g., 0.1), it is preliminarily judged as the non-rising state, otherwise step d is carried out;

[0102] Step d: If the variance of the current day's window is greater than β 3 (e.g., 0.4), and the change rate of the effective daily average value of the head and tail vibration days of this window is lower than μ (e.g., 0.04), it is judged as the non-rising state, otherwise step e is carried out. Here, the calculation formula for the change rate is:

[0103]

[0104] where, Vib 1 represents the effective daily average value of the vibration days at the tail of the window, Vib 0 represents the effective daily average value of the vibration days at the head of the window, T 1 represents the time at the tail of the window, T 0 represents the time at the head of the window.

[0105] Step e: If it is judged as the rising state for M consecutive times, and the daily average value of the vibration days of the current day is greater than λ times (e.g., 1.6 times) the threshold, it is judged as the continuous rising state. Record the rising state as 1, the continuous rising state as 2, and the non-rising state as 0, and end the process.

[0106] Furthermore, the trend change state judgment process of the tread polygon impact data can refer to the trend change state judgment process of the vertical vibration data of the axle box.

[0107] Step S13: Output the polygon fault detection result of the wheel tread based on the trend change state of the vehicle operation state data of each target type.

[0108] In one embodiment, the embodiments of the present application can also determine the stable operating speed of the current cycle; and output a polygon fault detection result of the wheel tread based on the stable operating speed and the trend change state of the vehicle operating state data of each target type.

[0109] Further, the embodiments of the present application can count the frequencies in each stable operating speed range in the current cycle to obtain statistical data; and determine the speed level corresponding to the maximum frequency in the statistical data as the stable operating speed. For example, count the frequencies in the operating speed ranges corresponding to the 300 km / h speed level and the 350 km / h speed level, and take the speed level corresponding to the maximum frequency in the statistical data as the stable operating speed.

[0110] Among them, the stable operating speed is the stable speed when the train set is running on the main line. For high-speed EMUs, it is generally 350±10 km / h (i.e., the 350 km / h speed level) or 300±10 km / h (i.e., the 300 km / h speed level). Moreover, the aforementioned information acquisition system of the embodiments of the present application further includes a rotational speed sensor installed on the wheel axle. After the embodiments of the present application obtain the operating rotational speed through the rotational speed sensor and obtain the speed data of the current cycle, that is, the operating rotational speed, they obtain the stable operating speed of the current cycle.

[0111] In one embodiment, the cycle is one day, and the polygon fault detection result includes a fault reminder. If the stable operating speed is the 350 km / h speed level, a fault reminder of the wheel tread is output based on the first preset reminder strategy and the trend change state of the vehicle operating state data of each target type; if the stable operating speed is the 300 km / h speed level, a fault reminder of the wheel tread is output based on the second preset reminder strategy and the trend change state of the vehicle operating state data of each target type.

[0112] Among them, the first preset reminder strategy includes outputting first-level reminders, second-level reminders, and third-level reminders, and the reminder level increases with the increase in severity; if the trend change states of the vertical vibration data of the axle box and the polygon impact data of the tread surface in the current cycle are in an increasing state, an increasing state, or an increasing state, a non-increasing state, or a non-increasing state, an increasing state, then a first-level reminder is output; if the trend change states of the vertical vibration data of the axle box and the polygon impact data of the tread surface in the current cycle are in a continuously increasing state, an increasing state, or a continuously increasing state, a non-increasing state, or an increasing state, a continuously increasing state, or a non-increasing state, a continuously increasing state, then it is determined that the reminder determination result in the current cycle is a second-level reminder. If the reminder determination result in the current cycle is not the Nth consecutive second-level reminder, then a second-level reminder is output; if the trend change states of the vertical vibration data of the axle box and the polygon impact data of the tread surface in the current cycle are in a continuously increasing state, a continuously increasing state, or the reminder determination result in the current cycle is the Nth consecutive second-level reminder, then a third-level reminder is output;

[0113] Moreover, the second preset reminder strategy includes outputting first-level reminders, second-level reminders, and third-level reminders, and the reminder level increases with the increase in severity; if the trend change states of the vertical vibration data of the axle box and the polygon impact data of the tread surface in the current cycle are in an increasing state, an increasing state, or a non-increasing state, an increasing state, then a first-level reminder is output; if the trend change states of the vertical vibration data of the axle box and the polygon impact data of the tread surface in the current cycle are in a continuously increasing state, an increasing state, or an increasing state, a continuously increasing state, or a non-increasing state, a continuously increasing state, then it is determined that the reminder determination result in the current cycle is a second-level reminder. If the reminder determination result in the current cycle is not the Nth consecutive second-level reminder, then a second-level reminder is output; if the trend change states of the vertical vibration data of the axle box and the polygon impact data of the tread surface in the current cycle are in a continuously increasing state, a continuously increasing state, or the reminder determination result in the current cycle is the Nth consecutive second-level reminder, then a third-level reminder is output.

[0114] Furthermore, the polygon fault detection result can include a fault reminder and a maintenance suggestion. Based on the trend change states of the vertical vibration data of the axle box and the polygon impact data of the tread surface in the current cycle and the speed during stable operation, a fault reminder and a maintenance suggestion are output. The output strategies for the 350 km / h speed level are shown in Table 1, and the output strategies for the 300 km / h speed level are shown in Table 2, where the increasing state is recorded as 1, the continuously increasing state is recorded as 2, and the non-increasing state is recorded as 0.

[0115] Table 1

[0116]

[0117] Table 2

[0118]

[0119] It should be noted that the existing analysis methods for wheel tread polygons usually analyze the short-term tread conditions, whether online or offline. When a wheel tread polygon fault occurs during the operation of high-speed EMUs, the vertical vibration acceleration and the wheel tread polygon impact signal value increase with the severity of the fault. In the embodiments of the present application, by analyzing the relatively long-term vibration data and impact data during the operation of high-speed EMUs, the development degree of the wheel tread polygon fault is judged, and a fault reminder and maintenance suggestions are output in a timely manner. Even if the runout is small, the severity of the wheel tread polygon fault can be accurately judged, making the maintenance work of the wheel tread polygon more targeted and accurate, thereby reducing the maintenance cost and reducing the impact of the abnormal high-frequency vibration generated by the wheel tread polygon on the relevant components of the frame and axle box.

[0120] That is, if the period is one day, after the embodiments of the present application obtain the speed data and the vertical vibration acceleration signal of the axle box for the latest day, that is, the current period, the stable operating speed is judged, and the daily average value data of the vibration effective value and the daily average value data of the impact dB for the day are calculated. Furthermore, the vertical vibration data of the axle box corresponding to the current period and the wheel tread polygon impact data are determined, and a sliding window process (window length is 5, sliding step is 1) is performed, and the average value, variance, and fitting slope within the window are calculated. Then, the state change judgment thresholds corresponding to the vertical vibration data of the axle box and the wheel tread polygon impact data for the day are calculated, and the trend change state of the vertical vibration data of the axle box for the day and the trend change state of the wheel tread polygon impact data for the day are judged. Finally, based on the trend change states of the above two types of data, a polygon fault reminder is output, and corresponding maintenance suggestions are provided according to the reminder level.

[0121] It can be seen that the embodiments of the present application obtain the vehicle operation state data of each target type in the current period, and the vehicle operation state data of each target type includes the vertical vibration data of the axle box and the wheel tread polygon impact data. Any vehicle operation state data of the target type in the current period is determined as the target data, and the trend change state of the target data in the current period is determined based on the historical period data corresponding to the target data. And based on the trend change states of the vehicle operation state data of each target type, a polygon fault detection result of the wheel tread is output. That is, in the process of detecting the polygon fault of the wheel tread, the embodiments of the present application use the vertical vibration data of the axle box in the historical period to judge the trend change state of the vertical vibration data of the axle box in the current period, and use the wheel tread polygon impact data in the historical period to judge the trend change state of the wheel tread polygon impact data in the current period. In this way, by judging the trend change states of the vertical vibration data of the axle box and the wheel tread polygon impact data through the historical period data, and then determining the polygon fault detection result, the problem of inaccurate detection when the runout is small is avoided, the accuracy of detecting the wheel tread polygon fault can be improved, and the detection cost can be reduced.

[0122] Moreover, the embodiment of the present application determines the state change judgment threshold of the target data in the current period based on the sliding window processing result of the historical period, and judges the trend change state of the target data in the current period based on the state change judgment threshold and the sliding window processing result of the current period. In this way, the state change judgment threshold of the current period is dynamically determined, which can make adaptive adjustments for different lines and different operating speeds, so as to better adapt to application scenarios such as the line allocation of bullet trains.

[0123] See Figure 5 As shown, the embodiment of the present application discloses a wheel tread fault detection device, including:

[0124] A state data acquisition module 11, configured to acquire vehicle operation state data of each target type in the current period; the vehicle operation state data of each target type includes axle box vertical vibration data and tread polygon impact data;

[0125] A change state determination module 12, configured to determine any of the vehicle operation state data of the target type in the current period as target data, and determine the trend change state of the target data in the current period based on the historical period data corresponding to the target data;

[0126] A detection result output module 13, configured to output a polygon fault detection result of the wheel tread based on the trend change state of the vehicle operation state data of each target type.

[0127] The change state determination module includes:

[0128] A sliding window processing sub-module, configured to perform sliding window processing on the target data to obtain a sliding window processing result of the current period;

[0129] A change state determination sub-module, configured to determine the trend change state of the target data in the current period based on the sliding window processing result of the current period and the sliding window processing result of the historical period; wherein, the sliding window processing result includes the window mean, window variance, and fitting slope of the window.

[0130] Further, the change state determination sub-module specifically includes:

[0131] A state change judgment threshold determination unit, configured to determine the state change judgment threshold of the target data in the current period based on the sliding window processing result of the historical period;

[0132] A change state determination unit, configured to judge the trend change state of the target data in the current period based on the state change judgment threshold and the sliding window processing result of the current period.

[0133] In one embodiment, the state change judgment threshold determination unit is specifically configured to determine, from the sliding window processing results of a preset number of most recent historical periods, a window whose window variance is less than a first preset variance threshold to obtain a first target window; calculate the mean value of the window means of all the first target windows to obtain the state change judgment threshold of the target data in the current period.

[0134] Wherein, the trend change state includes any one of an ascending state, a non-ascending state, and a continuously ascending state.

[0135] The change state determination unit is configured to determine that the trend change state of the target data in the current period is an ascending state if the target data meets the first ascending state determination condition and / or the second ascending state determination condition; wherein, the first ascending state determination condition is: the target data is greater than the state change judgment threshold, the window slope in the sliding window processing result of the current period is greater than a first preset slope threshold, and the window variance in the sliding window processing result of the current period is greater than a second preset variance threshold; the second ascending state determination condition is: the number of second target windows meeting a preset condition reaches a quantity threshold, and the fitting slope of the target data and the historical period data corresponding to the target data within a first preset number of consecutive windows is greater than a second preset slope threshold; the second target window is a window in the preset number of consecutive windows, and the preset condition is that the fitting slope of the window is within a preset range and the window variance is less than a third preset variance threshold.

[0136] The change state determination unit is further configured to determine that the trend change state of the target data in the current period is a non-ascending state if the target data meets the first non-ascending state determination condition and / or the second non-ascending state determination condition; wherein, the first non-ascending state determination condition is: there is a window in a second preset number of consecutive windows whose window slope is lower than a third preset slope threshold; wherein, all the second preset number of consecutive windows are windows before the current period; and, the second non-ascending state determination condition is: the window variance corresponding to the current period is greater than a fourth preset variance threshold, and the window change rate corresponding to the current period is lower than a preset change rate threshold; the window change rate is the change rate of the head and tail data of the window corresponding to the current period.

[0137] The change state determination unit is further configured to determine that the trend change state of the target data in the current period is a continuously ascending state if the target data is greater than a preset multiple of the state change judgment threshold, and the trend change states of the vehicle operation state data of the target type in consecutive M periods are all in an ascending state or a continuously ascending state; wherein, the consecutive M periods include the current period.

[0138] In one embodiment, the change state determination unit is specifically configured to:

[0139] Based on the state change judgment threshold and the sliding window processing result of the current period, determine whether the target data meets the first rising state determination condition. If the target data meets the first rising state determination condition, preliminarily determine that the trend change state of the target data in the current period is a rising state, and determine whether the target data meets the first non-rising state determination condition. If the target data does not meet the first rising state determination condition, determine whether the target data meets the second rising state determination condition. If the target data meets the second rising state determination condition, preliminarily determine that the trend change state of the target data in the current period is a rising state, and determine whether the target data meets the first non-rising state determination condition;

[0140] If the target data meets the first non-rising state determination condition, finally determine that the trend change state of the target data in the current period is a non-rising state. If the target data does not meet the first non-rising state determination condition, determine whether the target data meets the second non-rising determination condition; if the target data meets the second non-rising determination condition, finally determine that the trend change state of the target data in the current period is a non-rising state. If the target data does not meet the second non-rising determination condition, determine whether the target data meets the preset continuous rising determination condition;

[0141] If the target data meets the preset continuous rising determination condition, finally determine that the trend change state of the target data in the current period is a continuous rising state. If it does not meet the preset continuous rising determination condition, finally determine that the trend change state of the target data in the current period is a rising state.

[0142] The first rising state determination condition is: the target data is greater than the state change judgment threshold, the window slope in the sliding window processing result of the current period is greater than the first preset slope threshold, and the window variance in the sliding window processing result of the current period is greater than the second preset variance threshold;

[0143] The second rising state determination condition is: the number of second target windows that meet the preset conditions reaches the number threshold, and the fitting slope of the target data and the historical period data corresponding to the target data within the first preset number of consecutive windows is greater than the second preset slope threshold; the second target window is the window in the preset number of consecutive windows, and the preset condition is that the fitting slope of the window is within the preset range and the window variance is less than the third preset variance threshold;

[0144] The first non-rising state determination condition is: there is a window in the second preset number of consecutive windows whose window slope is lower than the third preset slope threshold; where, the windows before the current period of the second preset number of consecutive windows;

[0145] The second non-rising state determination condition is that the window variance corresponding to the current period is greater than the fourth preset variance threshold, and the window change rate corresponding to the current period is lower than the preset change rate threshold; the window change rate is the change rate of the head and tail data of the window corresponding to the current period.

[0146] The preset continuous rising determination condition is that the target data is greater than the preset multiple of the state change judgment threshold, and the change states of the vehicle operation state data of the target type in consecutive M periods are all in a rising state or a continuously rising state, then it is determined that the change state of the target data in the current period is a continuously rising state; where consecutive M periods include the current period.

[0147] The state data acquisition module is specifically used for:

[0148] Calculating the effective value of the vertical vibration acceleration signal of the axle box at each moment within the current period; determining the mean value of all the effective values within the current period to obtain the vertical vibration data of the axle box in the current period.

[0149] Calculating the difference value of the tread polygon impact signal at each moment within the current period; determining the mean value of all the difference values within the current period to obtain the tread polygon impact data in the current period.

[0150] Furthermore, the device further includes:

[0151] The stable operating speed determination module is used to determine the stable operating speed in the current period;

[0152] Correspondingly, the detection result output module is specifically used to output the polygon fault detection result of the wheel tread based on the stable operating speed and the trend change state of the vehicle operation state data of each target type.

[0153] In one implementation, the stable operating speed determination module is specifically used for:

[0154] Counting the frequencies in each stable operating speed range in the current period to obtain statistical data;

[0155] Determining the speed level corresponding to the maximum frequency in the statistical data as the stable operating speed.

[0156] In one implementation, the period is one day, and the detection result output module is specifically used for:

[0157] If the stable operating speed is the 350 km / h speed level, then output a fault reminder for the wheel tread based on the first preset reminder strategy and the trend change state of the vehicle operation state data of each target type.

[0158] If the stable operating speed is at the 300 km / h speed level, a fault reminder for the wheel tread is output based on the second preset reminder strategy and the trend change state of the vehicle operating state data of each target type.

[0159] Among them, the first preset reminder strategy includes outputting a first-level reminder, a second-level reminder, and a third-level reminder, and the reminder level increases with the increase in severity; if the trend change states of the vertical vibration data of the axle box and the polygon impact data of the tread in the current cycle are respectively an upward state, an upward state, or an upward state, a non-upward state, or a non-upward state, an upward state, then a first-level reminder is output; if the trend change states of the vertical vibration data of the axle box and the polygon impact data of the tread in the current cycle are respectively a continuously upward state, an upward state, or a continuously upward state, a non-upward state, or an upward state, a continuously upward state, or a non-upward state, a continuously upward state, then it is determined that the reminder determination result in the current cycle is a second-level reminder. If the reminder determination result in the current cycle is not the Nth consecutive second-level reminder, then a second-level reminder is output; if the trend change states of the vertical vibration data of the axle box and the polygon impact data of the tread in the current cycle are respectively a continuously upward state, a continuously upward state, or the reminder determination result in the current cycle is the Nth consecutive second-level reminder, then a third-level reminder is output.

[0160] Moreover, the second preset reminder strategy includes outputting a first-level reminder, a second-level reminder, and a third-level reminder, and the reminder level increases with the increase in severity; if the trend change states of the vertical vibration data of the axle box and the polygon impact data of the tread in the current cycle are respectively an upward state, an upward state, or a non-upward state, an upward state, then a first-level reminder is output; if the trend change states of the vertical vibration data of the axle box and the polygon impact data of the tread in the current cycle are respectively a continuously upward state, an upward state, or an upward state, a continuously upward state, or a non-upward state, a continuously upward state, then it is determined that the reminder determination result in the current cycle is a second-level reminder. If the reminder determination result in the current cycle is not the Nth consecutive second-level reminder, then a second-level reminder is output; if the trend change states of the vertical vibration data of the axle box and the polygon impact data of the tread in the current cycle are respectively a continuously upward state, a continuously upward state, or the reminder determination result in the current cycle is the Nth consecutive second-level reminder, then a third-level reminder is output.

[0161] It can be seen that in the embodiment of the present application, the vehicle operation state data of each target type in the current cycle is obtained. The vehicle operation state data of each target type includes vertical vibration data of the axle box and polygon impact data of the tread. Any vehicle operation state data of the target type in the current cycle is determined as target data, and the trend change state of the target data in the current cycle is determined based on the historical cycle data corresponding to the target data. And a polygon fault detection result of the wheel tread is output based on the trend change state of the vehicle operation state data of each target type. That is, in the process of detecting the polygon fault of the wheel tread in the embodiment of the present application, the trend change state of the vertical vibration data of the axle box in the current cycle is judged by using the vertical vibration data of the axle box in the historical cycle, and the trend change state of the polygon impact data of the tread in the current cycle is judged by using the polygon impact data of the tread in the historical cycle. In this way, the trend change state of the vertical vibration data of the axle box and the polygon impact data of the tread is judged through the historical cycle data, and then the polygon fault detection result is determined, avoiding the problem of inaccurate detection when the runout is small, improving the accuracy of detecting the polygon fault of the wheel tread, and reducing the detection cost.

[0162] Moreover, in the embodiment of the present application, the state change judgment threshold of the target data in the current cycle is determined based on the sliding window processing result in the historical cycle, and the trend change state of the target data in the current cycle is judged based on the state change judgment threshold and the sliding window processing result in the current cycle. In this way, the state change judgment threshold in the current cycle is dynamically determined, which can make adaptive adjustments for different lines and different running speeds, so as to better adapt to application scenarios such as the line allocation of EMUs.

[0163] See Figure 6 As shown, the embodiment of the present application discloses an electronic device 20, including a processor 21 and a memory 22. Among them, the memory 22 is used to store a computer program, and the processor 21 is used to execute the computer program, which is the wheel tread fault detection method disclosed in the foregoing embodiment.

[0164] For the specific process of the above wheel tread fault detection method, reference can be made to the corresponding content disclosed in the foregoing embodiment, and details will not be elaborated here.

[0165] Moreover, as a carrier for resource storage, the memory 22 can be a read-only memory, a random access memory, a magnetic disk or an optical disc, etc., and the storage method can be temporary storage or permanent storage.

[0166] In addition, the electronic device 20 further includes a power supply 23, a communication interface 24, an input / output interface 25, and a communication bus 26. Among them, the power supply 23 is used to provide operating voltage for each hardware device on the electronic device 20. The communication interface 24 can create a data transmission channel between the electronic device 20 and external devices, and the communication protocol it follows can be any communication protocol applicable to the technical solution of this application, and no specific limitation is imposed here. The input / output interface 25 is used to obtain external input data or output data to the outside, and its specific interface type can be selected according to specific application needs, and no specific limitation is made here.

[0167] Further, an embodiment of this application also discloses a computer-readable storage medium for storing a computer program, wherein when the computer program is executed by a processor, it implements the wheel tread fault detection method disclosed in the foregoing embodiment.

[0168] For the specific process of the above wheel tread fault detection method, reference can be made to the corresponding content disclosed in the foregoing embodiment, and details are not described herein again.

[0169] In this specification, the various embodiments are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. For the same or similar parts between the various embodiments, reference can be made to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and reference can be made to the description of the method part for related parts.

[0170] The steps of the method or algorithm described in combination with the embodiments disclosed in this article can be directly implemented by hardware, a software module executed by a processor, or a combination of the two. The software module can be placed in a random access memory (RAM), memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, register, hard disk, removable disk, CD-ROM, or any other form of storage medium well-known in the technical field.

[0171] The above has introduced in detail a wheel tread fault detection method, device, equipment, and medium provided by this application. Specific examples are used in this article to elaborate on the principle and implementation manner of this application. The description of the above embodiments is only used to help understand the method and its core idea of this application. At the same time, for those of ordinary skill in the art, based on the idea of this application, there will be changes in the specific implementation manner and application scope. In summary, the content of this specification should not be construed as a limitation to this application.

Claims

1. A method for detecting wheel tread faults, characterized in that, it includes: Obtaining the vehicle operation state data of each target type in the current cycle; The vehicle operation state data of each target type includes axle box vertical vibration data and tread polygon impact data; Determining the vehicle operation state data of any target type in the current cycle as target data, and determining the trend change state of the target data in the current cycle based on the historical cycle data corresponding to the target data; Outputting the polygon fault detection result of the wheel tread based on the trend change state of the vehicle operation state data of each target type; Wherein, determining the trend change state of the target data in the current cycle based on the historical cycle data corresponding to the target data includes: performing a sliding window process on the target data to obtain the sliding window process result in the current cycle; determining the windows with window variances less than the first preset variance threshold from the sliding window process results of a preset number of recent historical cycles to obtain the first target windows; calculating the mean value of the window means of all the first target windows to obtain the state change judgment threshold of the target data in the current cycle; judging the trend change state of the target data in the current cycle based on the state change judgment threshold and the sliding window process result in the current cycle; the sliding window process result includes the window mean value, window variance and fitting slope of the window.

2. The method for detecting wheel tread faults according to claim 1, characterized in that, The trend change state includes any one of an ascending state, a non-ascending state and a continuously ascending state.

3. The method for detecting wheel tread faults according to claim 2, characterized in that, If the target data meets the first ascending state determination condition and / or the second ascending state determination condition, it is determined that the trend change state of the target data in the current cycle is an ascending state; Wherein, the first ascending state determination condition is: the target data is greater than the state change judgment threshold, the window slope in the sliding window process result in the current cycle is greater than the first preset slope threshold, and the window variance in the sliding window process result in the current cycle is greater than the second preset variance threshold; The second ascending state determination condition is: the number of second target windows meeting the preset condition reaches the quantity threshold, and the fitting slopes of the target data and the historical cycle data corresponding to the target data within the first preset number of consecutive windows are greater than the second preset slope threshold; the second target window is the window in the preset number of consecutive windows, and the preset condition is that the fitting slope of the window is within the preset range and the window variance is less than the third preset variance threshold.

4. The method for detecting wheel tread faults according to claim 2, characterized in that, If the target data meets the first non-ascending state determination condition and / or the second non-ascending state determination condition, it is determined that the trend change state of the target data in the current cycle is a non-ascending state; Among them, the first non-rising state determination condition is that there is a window whose window slope is lower than the third preset slope threshold among the second preset number of consecutive windows; among them, the second preset number of consecutive windows are all the windows before the current period; and, the second non-rising state determination condition is that the window variance corresponding to the current period is greater than the fourth preset variance threshold, and the window change rate corresponding to the current period is lower than the preset change rate threshold; the window change rate is the change rate of the head and tail data of the window corresponding to the current period.

5. The wheel tread fault detection method according to claim 2, characterized in that if the target data is greater than the preset multiple of the state change judgment threshold, and the trend change state of the vehicle operation state data of this target type is in an ascending state or a continuous ascending state in the consecutive M periods, it is determined that the trend change state of the target data in the current period is a continuous ascending state; wherein, the consecutive M periods include the current period.

6. The wheel tread fault detection method according to claim 1, characterized in that judging the trend change state of the target data in the current period based on the state change judgment threshold and the sliding window processing result of the current period includes: judging whether the target data meets the first ascending state determination condition based on the state change judgment threshold and the sliding window processing result of the current period. If the target data meets the first ascending state determination condition, it is preliminarily determined that the trend change state of the target data in the current period is an ascending state, and it is judged whether the target data meets the first non-ascending state determination condition. If the target data does not meet the first ascending state determination condition, it is judged whether the target data meets the second ascending state determination condition. If the target data meets the second ascending state determination condition, it is preliminarily determined that the trend change state of the target data in the current period is an ascending state, and it is judged whether the target data meets the first non-ascending state determination condition; if the target data meets the first non-ascending state determination condition, it is finally determined that the trend change state of the target data in the current period is a non-ascending state. If the target data does not meet the first non-ascending state determination condition, it is judged whether the target data meets the second non-ascending state determination condition; if the target data meets the second non-ascending state determination condition, it is finally determined that the trend change state of the target data in the current period is a non-ascending state. If the target data does not meet the second non-ascending state determination condition, it is judged whether the target data meets the preset continuous ascending determination condition; if the target data meets the preset continuous ascending determination condition, it is finally determined that the trend change state of the target data in the current period is a continuous ascending state. If it does not meet the preset continuous ascending determination condition, it is finally determined that the trend change state of the target data in the current period is an ascending state.

7. The wheel tread fault detection method according to claim 6, characterized in that The first rising state determination condition is: the target data is greater than the state change judgment threshold, the window slope in the sliding window processing result of the current period is greater than the first preset slope threshold, and the window variance in the sliding window processing result of the current period is greater than the second preset variance threshold; The second rising state determination condition is: the number of second target windows that meet the preset conditions reaches the number threshold, and the fitting slope of the target data and the historical period data corresponding to the target data within the first preset number of consecutive windows is greater than the second preset slope threshold; The second target window is the window in the preset number of consecutive windows, and the preset condition is that the fitting slope of the window is within the preset range and the window variance is less than the third preset variance threshold; The first non-rising state determination condition is: there is a window in the second preset number of consecutive windows whose window slope is lower than the third preset slope threshold; where, the windows before the current period of the second preset number of consecutive windows; The second non-rising state determination condition is: the window variance corresponding to the current period is greater than the fourth preset variance threshold, and the window change rate corresponding to the current period is lower than the preset change rate threshold; the window change rate is the change rate of the head and tail data of the window corresponding to the current period; The preset continuous rising determination condition is: the target data is greater than the preset multiple of the state change judgment threshold, and the change state of the vehicle operation state data of the target type is in an ascending state or a continuously ascending state in consecutive M periods, then it is determined that the change state of the target data in the current period is a continuously ascending state; where, the consecutive M periods include the current period.

8. The wheel tread fault detection method according to claim 1, wherein, The process of obtaining the vertical vibration data of the axle box in the current period is: calculating the effective value of the vertical vibration acceleration signal of the axle box at each moment within the current period; determining the mean value of all the effective values within the current period to obtain the vertical vibration data of the axle box in the current period; The process of obtaining the vertical vibration data of the axle box in the current period is: calculating the difference value of the tread polygon impact signal at each moment within the current period; determining the mean value of all the difference values within the current period to obtain the tread polygon impact data in the current period.

9. The wheel tread fault detection method according to any one of claims 1 to 8, wherein, It further includes: Determining the stable operating speed in the current period; Correspondingly, the output of the polygon fault detection result of the wheel tread based on the trend change state of the vehicle operation state data of each target type includes: Outputting the polygon fault detection result of the wheel tread based on the stable operating speed and the trend change state of the vehicle operation state data of each target type.

10. The wheel tread fault detection method according to claim 9, wherein, The determination of the stable operating speed in the current period includes: Counting the frequencies in each stable operating speed interval in the current period to obtain statistical data; Determining the speed level corresponding to the maximum frequency in the statistical data as the stable operating speed.

11. The wheel tread fault detection method according to claim 9, characterized in that, with a cycle of one day, the polygon fault detection result includes a fault reminder, and the output of the polygon fault detection result of the wheel tread based on the stable operation speed and the trend change state of the vehicle operation state data of each target type includes: If the stable operation speed is at the 350 km / h speed level, a fault reminder of the wheel tread is output based on the first preset reminder strategy and the trend change state of the vehicle operation state data of each target type; If the stable operation speed is at the 300 km / h speed level, a fault reminder of the wheel tread is output based on the second preset reminder strategy and the trend change state of the vehicle operation state data of each target type; Among them, the first preset reminder strategy includes outputting a first-level reminder, a second-level reminder, and a third-level reminder, and the reminder level increases with the increase of the severity; if the trend change states of the vertical vibration data of the axle box and the polygon impact data of the tread in the current cycle are respectively an upward state, an upward state, or an upward state, a non-upward state, or a non-upward state, an upward state, then a first-level reminder is output; if the trend change states of the vertical vibration data of the axle box and the polygon impact data of the tread in the current cycle are respectively a continuously upward state, an upward state, or a continuously upward state, a non-upward state, or an upward state, a continuously upward state, or a non-upward state, a continuously upward state, then it is determined that the reminder determination result in the current cycle is a second-level reminder. If the reminder determination result in the current cycle is not the Nth consecutive second-level reminder, then a second-level reminder is output; if the trend change states of the vertical vibration data of the axle box and the polygon impact data of the tread in the current cycle are respectively a continuously upward state, a continuously upward state, or the reminder determination result in the current cycle is the Nth consecutive second-level reminder, then a third-level reminder is output; And, the second preset reminder strategy includes outputting a first-level reminder, a second-level reminder, and a third-level reminder, and the reminder level increases with the increase of the severity; if the trend change states of the vertical vibration data of the axle box and the polygon impact data of the tread in the current cycle are respectively an upward state, an upward state, or a non-upward state, an upward state, then a first-level reminder is output; if the trend change states of the vertical vibration data of the axle box and the polygon impact data of the tread in the current cycle are respectively a continuously upward state, an upward state, or an upward state, a continuously upward state, or a non-upward state, a continuously upward state, then it is determined that the reminder determination result in the current cycle is a second-level reminder. If the reminder determination result in the current cycle is not the Nth consecutive second-level reminder, then a second-level reminder is output; if the trend change states of the vertical vibration data of the axle box and the polygon impact data of the tread in the current cycle are respectively a continuously upward state, a continuously upward state, or the reminder determination result in the current cycle is the Nth consecutive second-level reminder, then a third-level reminder is output.

12. A wheel tread fault detection device, characterized in that, comprising: a state data acquisition module for acquiring the vehicle operation state data of each target type in the current cycle; The vehicle operation state data of each target type includes axle box vertical vibration data and tread polygon impact data; A change state determination module, configured to determine the vehicle operation state data of any of the target types in the current period as target data, and determine the trend change state of the target data in the current period based on the historical period data corresponding to the target data; A detection result output module, configured to output a polygon fault detection result of the wheel tread based on the trend change state of the vehicle operation state data of each target type; Among them, the change state determination module is specifically configured to: perform a sliding window process on the target data to obtain a sliding window process result in the current period; determine windows with window variances less than a first preset variance threshold from the sliding window process results of a preset number of most recent historical periods to obtain first target windows; calculate the mean value of the window means of all the first target windows to obtain a state change judgment threshold of the target data in the current period; Judge the trend change state of the target data in the current period based on the state change judgment threshold and the sliding window process result in the current period; the sliding window process result includes the window mean value, window variance, and fitting slope of the window.

13. An electronic device, Characterized in that, It includes a memory and a processor, wherein: The memory is used to store a computer program; The processor is configured to execute the computer program to implement the wheel tread fault detection method according to any one of claims 1 to 11.

14. A computer-readable storage medium, Characterized in that, It is used to store a computer program, wherein the computer program, when executed by a processor, implements the wheel tread fault detection method according to any one of claims 1 to 11.

Images