Wheel tread off-track detection method and system for a rail vehicle

Abstract

The application provides a wheel tread off-track detection method and system for a rail vehicle. The method comprises symmetrically arranging a jacking driving device with a jacking wheel on both longitudinal sides of the rail, arranging a pressure sensing unit between the two jacking wheels, installing an optical detection unit with an orthogonal light path on the outside, integrating a displacement measuring unit on the jacking wheel, and forming a three-dimensional detection network. During detection, the jacking wheel gradually lifts the wheel, and the blocking signal, tread pressure and vertical displacement parameters are synchronously collected to establish a space mapping relationship between the pressure decreasing path and the displacement abnormal point. The blocking signal is converted into a blocking removal rate, and the displacement cumulative value is calculated in combination with the space mapping relationship. When the displacement cumulative value continuously increases and the blocking removal rate exceeds the threshold value, it is determined that the wheel and rail are separated, and the jacking is stopped. Through multi-sensor cooperative detection and analysis, the application realizes accurate determination of the wheel and rail separation state and adaptive jacking control, and effectively solves the wheel diameter adaptability problem of the traditional fixed jacking mode.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, and in particular to a wheel tread off-rail detection method and system for a rail vehicle. BACKGROUND

[0002] The wheel of a rail vehicle is subjected to alternating stress in long-term operation and is prone to fatigue defects, which directly affects the driving safety. In order to ensure safety, the wheel needs to be periodically detected by ultrasonic flaw detection. Since the cost of disassembling and assembling the wheel is high and the efficiency is low, currently, an online flaw detection device is generally used, that is, the detection is completed while the wheel is kept in the installed state. However, there is a difference in the wheel diameter (such as the minimum wheel diameter deviation from the standard wheel diameter due to wear), which puts forward a precise control requirement for the jacking mechanism of the flaw detection device: it is necessary to ensure that the wheel with different diameters can reliably be separated from the rail, while avoiding the jacking being too high to affect the detection process.

[0003] The existing online ultrasonic flaw detection device generally uses a jacking wheel device, and its working process is as follows: the jacking mechanism jacks up the wheel to a fixed height to separate it from the rail, and then drives the wheel to rotate to make the ultrasonic probe adhere to the wheel surface to complete the detection. However, the existing scheme relies on a preset fixed jacking distance and does not consider the influence of the wheel diameter difference. SUMMARY

[0004] The present application provides a wheel tread off-rail detection method and system for a rail vehicle to solve the problem of poor off-rail detection effect caused by the fixed jacking mode in the prior art.

[0005] In a first aspect, the present application provides a wheel tread off-rail detection method for a rail vehicle, jacking drive devices are symmetrically arranged on both longitudinal sides of the rail, each jacking drive device includes two jacking wheels, the jacking wheels are used to lift the wheel, a pressure sensing unit is arranged between the two jacking wheels, an optical detection unit is installed on the outside of the jacking drive device, the optical path direction of the optical detection unit is orthogonal to the movement plane of the wheel, and a displacement measuring unit is integrated on the jacking wheel, the displacement measuring unit moves synchronously to generate the vertical displacement parameter of the wheel when the jacking wheel lifts the wheel; the pressure sensing unit, the optical detection unit and the displacement measuring unit form a three-dimensional detection coverage area in the wheel-rail contact area; when the wheel contacts the rail, the optical path of the optical detection unit is completely blocked by the wheel.

[0006] The method comprises:

[0007] In the process of gradually lifting the wheel by the jacking wheels of the jacking drive device, the blocked signal state of the optical detection unit, the tread pressure change parameter output by the pressure sensing unit and the vertical displacement parameter generated by the displacement measuring unit are acquired to establish a spatial mapping relationship between the pressure decreasing area expansion path and the distribution characteristics of the displacement abnormal points.

[0008] mapping the blocking signal state to a blocking release rate of the optical detection channel, calculating a displacement accumulation value when the wheel tread separates from the rail according to the spatial mapping relationship combined with the blocking release rate of the optical detection channel;

[0009] when the displacement accumulation value presents a one-way increasing characteristic in the continuous lifting stage, and the blocking release rate of the optical detection channel is greater than a preset threshold, determining that the wheel rail separates, and generating a control instruction to control the jacking driving device to terminate the operation.

[0010] Optionally, further comprising:

[0011] synchronously collecting the contact surface images of the wheel and the rail during the lifting process, extracting the illumination reflection features and the contour deformation parameters of the contact area from the contact surface images;

[0012] differentially comparing the illumination reflection features with a preset standard reflection template under the wheel rail fitting state to generate a deformation compensation factor;

[0013] dynamically correcting the displacement accumulation value based on the deformation compensation factor, and optimizing the boundary accuracy of the spatial mapping relationship by fusing the contour deformation parameters.

[0014] Optionally, the acquisition of the blocking signal state of the optical detection unit, the tread pressure change parameter output by the pressure sensing unit, and the vertical displacement parameter generated by the displacement measuring unit to establish a spatial mapping relationship between the pressure decreasing area expansion path and the displacement abnormal point distribution characteristic, comprises:

[0015] According to the blocking signal state of the optical detection unit, the change rate of the corresponding blocking release rate of the optical detection channel is calculated, and the peak coordinate information corresponding to the change rate is extracted;

[0016] extracting the expansion direction and the decreasing gradient of the pressure decreasing area in the tread pressure change parameter output by the pressure sensing unit to determine the pressure decreasing starting point and the pressure decreasing propagation path;

[0017] identify the spatial distribution of the displacement abnormal points in the vertical displacement parameter generated by the displacement measuring unit;

[0018] Based on the peak coordinate information, the pressure decreasing starting point and the pressure decreasing propagation path, and the spatial distribution of the displacement abnormal points, a spatial mapping relationship between the pressure decreasing area expansion path and the displacement abnormal point distribution characteristic is established.

[0019] Optionally, the spatial mapping relationship between the pressure-decrement region expansion path and the displacement anomaly point distribution characteristics is established based on the peak coordinate information, the pressure-decrement starting point, the pressure-decrement propagation path, and the spatial distribution of the displacement anomaly point, and includes:

[0020] The time point in the peak coordinate information is matched with the occurrence time of the pressure-decrement starting point. If the time difference between the two is less than a set tolerance, the pressure-decrement starting point is marked as an effective trigger point.

[0021] Starting from the effective trigger point, a plurality of directionally associated regions are demarcated in the wheel-rail contact area according to the extension direction of the pressure-decrement propagation path, and the spatial distribution of the displacement anomaly points in each directionally associated region is counted.

[0022] The spatial distribution of the displacement anomaly points is proportionally distributed with the pressure-decrement gradient of the corresponding directionally associated region. If the ratio of the density to the gradient in any directionally associated region exceeds a dynamic threshold, the directionally associated region is determined as a spatial mapping area of the pressure-decrement expansion path and the displacement anomaly point.

[0023] The coverage range and connection relationship of all spatial mapping areas are taken as the spatial mapping relationship between the pressure-decrement region expansion path and the displacement anomaly point distribution characteristics.

[0024] Optionally, the blocking signal state is mapped to a blocking release rate of the optical detection channel. According to the spatial mapping relationship, the displacement cumulative value when the wheel tread separates is calculated by combining the blocking release rate of the optical detection channel, and includes:

[0025] The duration of the blocking signal of each optical detection channel during lifting is counted, and the blocking release weight factor of each optical detection channel is calculated according to the duration of the blocking signal.

[0026] The proportion of the number of optical detection channels in which the blocking signal disappears to the total number of optical detection channels is converted into a blocking release rate sequence according to the lifting sequence.

[0027] According to the pressure-decrement expansion path marked in the spatial mapping relationship, the wheel tread area that coincides with the displacement anomaly point distribution area is screened out.

[0028] The vertical displacement increments detected by all displacement measurement units in the wheel tread area during the lifting stage are accumulated, and each vertical displacement increment is weighted and corrected according to the blocking release rate sequence and the corresponding blocking release weight factor, to generate a displacement cumulative value.

[0029] Optionally, the change rate of the blocking release rate of the optical detection channel corresponding to the blocking signal state of the optical detection unit is calculated, and the peak coordinate information corresponding to the change rate is extracted, including:

[0030] segmenting the blockage removal rate sequence according to preset time windows, calculating a difference value of the blockage removal rate in each time window, and dividing the difference value by a time window length to obtain a change rate sequence;

[0031] traversing the change rate sequence, identifying a local maximum value point with a rate value greater than rate values of adjacent front and rear time windows, recording a time stamp and an associated optical detection channel position number corresponding to the local maximum value point, and forming peak value coordinate information.

[0032] Optionally, the extension direction and the decreasing gradient of the pressure decreasing region in the tread pressure change parameter output by the pressure sensing unit are extracted to determine a pressure decreasing starting point and a pressure decreasing propagation path, including:

[0033] dividing the wheel tread into grid units, and counting a drop amplitude of the pressure change parameter in each grid unit;

[0034] taking the grid unit with the largest drop amplitude as the pressure decreasing starting point, and starting from the pressure decreasing starting point, tracking a path in a direction in which the pressure drop amplitude of the adjacent grid unit increases until a pressure rising or stable region is detected, to form a pressure decreasing propagation path;

[0035] calculating a decreasing gradient of the pressure decreasing propagation path according to a ratio of the pressure drop amplitude to the path length of each grid unit on the pressure decreasing propagation path.

[0036] Optionally, the spatial distribution of the displacement abnormal points in the vertical displacement parameter generated by the displacement measuring unit is identified, including:

[0037] continuously sampling the vertical displacement parameter generated by the displacement measuring unit, detecting a displacement increment difference value between adjacent sampling points, and if the displacement increment difference value exceeds a set mutation threshold, marking a time period between the adjacent sampling points as a nonlinear change period;

[0038] recording a starting time stamp of all nonlinear change periods, calculating a maximum time difference between the starting time stamps detected by different displacement measuring units as a time sequence deviation;

[0039] grouping the nonlinear change periods according to the time sequence deviation, screening out a nonlinear change period group with a time difference less than a synchronization threshold, and in each nonlinear change period group, screening out a unidirectional sudden increase displacement point according to the direction consistency of the displacement increment, marking the unidirectional sudden increase displacement point as a displacement abnormal point and recording a position coordinate corresponding to the displacement abnormal point;

[0040] determining the spatial distribution of the displacement abnormal points according to the position coordinates corresponding to the plurality of displacement abnormal points.

[0041] Optionally, when the displacement cumulative value presents a one-way increasing characteristic in a continuous lifting phase, and the unblocking rate of the optical detection channel is greater than a preset threshold, it is determined that the wheel rail separation occurs, and a control instruction is generated to control the jacking driving device to terminate the operation, including:

[0042] The displacement cumulative values are recorded in sequence of lifting phases, and if the displacement cumulative value of the current phase is greater than that of the previous phase and the difference between the displacement cumulative values continues to increase, it is marked as one-way increasing;

[0043] The proportion of the number of channels in the optical detection channel whose unblocking rate is greater than a preset threshold is counted, and if the channel number proportion meets the preset threshold in a continuous number of lifting phases, it is marked as effective proportion meeting the standard;

[0044] When the one-way increasing mark and the effective proportion meeting the standard mark exist at the same time, a control instruction is generated to control the jacking driving device to terminate the operation;

[0045] The method further includes:

[0046] In response to the control instruction, the driving rotating unit is started to drive the wheel to rotate, and the acoustic wave detection unit is adjusted to perform defect detection work based on the channel activation sequence of the optical detection unit adhering to the wheel surface.

[0047] In a second aspect, the application provides a wheel tread off-track detection system for a rail vehicle, including a jacking driving device, an optical measurement unit and a processing unit; wherein each jacking driving device includes two top rotating wheels, a pressure sensing unit and a displacement detection unit, the jacking driving device is arranged on the longitudinal sides of the track, and each jacking driving device includes two top rotating wheels, which are used to lift the wheel;

[0048] A pressure sensing unit is arranged between the two top rotating wheels, and the pressure sensing unit is used to output the tread pressure change parameter;

[0049] The optical detection unit is installed on the outside of the jacking driving device, and the optical path direction thereof is orthogonal to the wheel movement plane, when the wheel contacts the track, the optical path of the optical detection unit is completely blocked by the wheel;

[0050] The displacement measurement unit is integrated on the top rotating wheel, and when the top rotating wheel lifts the wheel, the displacement measurement unit moves synchronously to generate the vertical displacement parameter of the wheel; the pressure sensing unit, the optical detection unit and the displacement measurement unit form a three-dimensional detection coverage area in the wheel rail contact area;

[0051] The processing unit is configured to acquire the blocking signal state of the optical detection unit, the tread pressure change parameter output by the pressure sensing unit, and the vertical displacement parameter generated by the displacement measuring unit during the step-by-step lifting of the top wheel of the jacking driving device, so as to establish a spatial mapping relationship between the pressure-decreasing area expansion path and the distribution characteristics of the displacement abnormal points; map the blocking signal state to a blocking removal rate of the optical detection channel, calculate the displacement cumulative value when the wheel tread separates according to the spatial mapping relationship and the blocking removal rate of the optical detection channel; when the displacement cumulative value presents a one-way increasing characteristic in the continuous lifting stage, and the blocking removal rate of the optical detection channel is greater than a preset threshold, it is determined that the wheel rail separates, and a control instruction is generated to control the jacking driving device to terminate the operation.

[0052] In the technical scheme of the present application, a three-dimensional detection network is formed in the wheel rail contact area by the pressure sensing unit, the optical detection unit and the displacement measuring unit, and real-time acquisition of the blocking signal, the tread pressure change parameter and the vertical displacement parameter is realized, so as to realize multi-dimensional data cooperative perception and improve detection comprehensiveness and reliability. Based on the correlation analysis of the pressure-decreasing area expansion path and the distribution characteristics of the displacement abnormal points, the physical propagation characteristics of the wheel rail separation process are accurately modeled to provide high-confidence data support for determination. The blocking signal state is mapped to a blocking removal rate, the displacement cumulative value is corrected in combination with the spatial mapping relationship, noise interference is eliminated, and the effective separation amount of the wheel lifting is truly reflected. Through the dual verification of the one-way increasing characteristic of the displacement cumulative value and the threshold compliance of the blocking removal rate, the absolute reliability of the wheel rail separation determination is ensured, and false determination and equipment damage are avoided.

[0053] Further, the technical scheme of the present application also synchronously acquires the contact surface images of the wheel and the rail during the lifting process, extracts the illumination reflection characteristics and the contour deformation parameters, generates a deformation compensation factor through difference comparison, dynamically corrects the displacement cumulative value and optimizes the boundary accuracy of the spatial mapping relationship. Through image feature analysis and deformation compensation mechanism, the displacement detection error caused by wheel surface deformation or foreign matter attachment is significantly reduced, the abnormal area is accurately positioned through illumination reflection difference comparison, the contour parameter optimizes the spatial mapping boundary, and the adaptability and detection accuracy of the system under complex working conditions are improved.

[0054] These and other aspects of the present application will become more apparent in the following detailed description of embodiments. BRIEF DESCRIPTION OF DRAWINGS

[0055] In order to more clearly illustrate the technical solutions of the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiment or prior art description. Obviously, the drawings in the following description are some embodiments of the present application, and those skilled in the art can also obtain other drawings according to these drawings without creating any inventive labor.

[0056] Figure 1 An architectural diagram of a wheel tread off-track detection system of a rail vehicle is provided for an embodiment of the present application.

[0057] Figure 2 A structural schematic diagram of a wheel tread off-track detection system of a rail vehicle is provided for an embodiment of the present application.

[0058] Figure 3 A structural schematic diagram of another wheel tread off-track detection system of a rail vehicle is provided for an embodiment of the present application.

[0059] Figure 4 A structural schematic diagram of another wheel tread off-track detection system of a rail vehicle is provided for an embodiment of the present application.

[0060] Figure 5 A flowchart of a wheel tread off-track detection method of a rail vehicle is provided for an embodiment of the present application. DETAILED DESCRIPTION

[0061] In order to enable persons skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application.

[0062] In some of the processes described in the specification and claims of the present application and the above-described drawings, a plurality of operations appear in a specific order. However, it should be clearly understood that these operations can be performed in the order in which they appear in this text or in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes can include more or fewer operations, and the operations can be performed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this text are used to distinguish different messages, devices, modules, etc., and do not represent the order of precedence. Also, "first" and "second" are not of different types.

[0063] Research has found that in the online ultrasonic flaw detection technology of the wheels of a rail vehicle, the existing equipment relies on a fixed jacking height of a jacking wheel device, which leads to the following key defects in the case of wheel diameter difference: the minimum wheel diameter wheel is not fully off the track due to insufficient jacking, and the tread is damaged when rotating due to friction with the rail; and the maximum wheel diameter wheel is jacked too high, which affects the precision of the probe sticking, resulting in distortion of the detection signal. This contradiction is due to the indiscriminate treatment of wheel diameter changes by the fixed jacking strategy, and a dynamic adaptive separation determination method is urgently needed.

[0064] To address the aforementioned problems, this invention proposes a wheel tread derailment detection method based on multi-source sensor fusion. Its core lies in the real-time capture of changes in wheel-rail contact state through three-dimensional collaborative monitoring of a pressure sensing unit, an optical detection unit, and a displacement measurement unit. Specifically, during the lifting process, the critical state of wheel-rail separation is dynamically determined by analyzing the spatial expansion characteristics of the tread pressure reduction path, the release ratio of the optical blocking signal, and the cumulative trend of vertical displacement, and the lifting height is adaptively controlled. This method abandons the traditional fixed lifting distance mode. Through spatiotemporal correlation and fusion calculation of multi-sensor data, it ensures that wheels of different diameters can reliably derail with the minimum necessary height. This avoids insufficient lifting in small-diameter scenarios and eliminates excessive lifting in large-diameter scenarios, fundamentally solving the detection failure and equipment damage problems caused by wheel diameter differences in existing technologies, and significantly improving the safety and process consistency of flaw detection operations.

[0065] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0066] Figure 1 This application provides a schematic diagram of the structure of a wheel tread off-rail detection system for a rail vehicle, as shown in the embodiment of the present application. Figure 1 As shown, the system includes a lifting drive device 11, an optical measurement unit 12, and a processing unit 13; wherein each lifting drive device 11 includes two top rollers 11a and 11b, a pressure sensing unit 11c, and a displacement detection unit 11d.

[0067] The connection relationships between the various units of the system are as follows: Figure 2 As shown, in Figure 2 middle( Figure 2 (Only one side of the track is shown as an example) Lifting drive devices 11 are symmetrically arranged on both sides of the track in the longitudinal direction. Each lifting drive device includes a lifting wheel 11a and a lifting wheel 11b, which are used to lift the wheel on the same side.

[0068] like Figure 3 As shown, a pressure sensing unit 11c is disposed between the two top rollers 11a and 11b. The pressure sensing unit 11c can be a piezoelectric, strain gauge, or capacitive sensor. When the wheel is not lifted, the pressure sensing unit 11c can collect the pressure on the wheel-rail contact surface. During the wheel lifting process, it can detect the pressure gradient change on the wheel-rail contact surface and output the tread pressure change parameter, which is then used by the subsequent processing unit 14 to determine whether the wheel and rail have separated.

[0069] Optical detection units 12a and 12b are mounted on the outer side of the lifting drive device 11, wherein the optical path direction between optical detection units 12a and 12b is orthogonal to the wheel motion plane. Figure 4 As shown, when the wheel contacts the track, the optical path of the optical detection unit is completely blocked by the wheel. The top rotating wheel is configured with a hollow structure. When the top rotating wheels 11a and 11b lift the wheel, the optical signal emitted by the optical detection unit 12a can be received by the optical detection unit 12b through the hollow structure of the top rotating wheel. When the optical detection unit 12b receives the optical signal, the processing unit 13 determines that the optical path is not blocked by the wheel. The optical detection unit 12 employs one or a combination of through-beam, specular reflection, or diffuse reflection optical sensors. The processing unit 13 determines whether the vehicle and track are separated based on the state of optical path obstruction by emitting light through the optical detection unit 12.

[0070] Simultaneously, a displacement measuring unit 11d is integrated on the top rotating wheel. When the top rotating wheel lifts the wheel, the displacement measuring unit 11d moves synchronously to generate the vertical displacement parameters of the wheel. The displacement measuring unit 11d calculates the lifting height by measuring the rotation angle of the top rotating wheel screw through an encoder, or directly detects the stroke of the hydraulic cylinder through a linear potentiometer, ensuring that the vertical displacement parameters are strictly synchronized with the actual lifting amount of the wheel. It should be noted that... Figure 2 The position of the displacement measuring unit 11d shown is only an example and can be set according to specific needs. It is only necessary to integrate the displacement measuring unit 11d onto the top rotating wheel.

[0071] With the above settings, the pressure sensing unit 11c, the optical detection unit 12, and the displacement measurement unit 11d in the system form a three-dimensional detection coverage area in the wheel-rail contact region.

[0072] In the three-dimensional detection coverage area, the processing unit 13 is used to acquire the blocking signal status of the optical detection unit 12, the tread pressure change parameters output by the pressure sensing unit 11c, and the vertical displacement parameters generated by the displacement measurement unit 11d during the process of the lifting drive device 11 raising the wheel step by step. By fusing the expansion direction of the pressure reduction area, the spatial distribution of displacement anomalies, and the correlation characteristics of the optical path blocking release rate, a spatial mapping relationship between the expansion path of the pressure reduction area and the distribution characteristics of displacement anomalies is established. The blocking signal status is mapped to the blocking release rate of the optical detection channel. Based on the spatial mapping relationship and the blocking release rate of the optical detection channel, the cumulative displacement value when the wheel tread separates is calculated. When the cumulative displacement value shows a unidirectional increasing characteristic during the continuous lifting stage, and the blocking release rate of the optical detection channel is greater than a preset threshold, wheel-rail separation is determined, and a control command is generated to control the lifting drive device 11 to terminate the operation.

[0073] Optionally, the system further comprises a driving rotation unit (not shown in the figure) for driving the wheel to rotate after the wheel-rail separation, and an acoustic wave detection unit (not shown in the figure) for dynamically adjusting the probe fitting track based on the channel activation sequence of the optical detection unit.

[0074] The processing unit 13 is further configured to, in response to the control instruction, start the driving rotation unit to drive the wheel to rotate, and control the acoustic wave detection unit to perform the defect detection operation.

[0075] For the above system, an embodiment of the present application provides a wheel tread off-rail detection method for a railway vehicle, Figure 5 For the above system, an embodiment of the present application provides a wheel tread off-rail detection method for a railway vehicle, Figure 5 As shown in the figure, the method comprises:

[0076] 501. In the process of gradually lifting the wheel by the lifting driving device, the blocking signal state of the optical detection unit, the tread pressure change parameter output by the pressure sensing unit, and the vertical displacement parameter generated by the displacement measuring unit are acquired to establish a spatial mapping relationship between the pressure decreasing area expansion path and the displacement abnormal point distribution characteristics.

[0077] Optionally, step 501 can specifically include the following steps:

[0078] 5011. According to the blocking signal state of the optical detection unit, the change rate of the blocking removal rate corresponding to the optical detection channel is calculated, and the peak coordinate information corresponding to the change rate is extracted.

[0079] In the process, the blocking removal rate sequence is divided according to the preset time window, the difference value of the blocking removal rate in each time window is calculated, and the difference value is divided by the time window length to obtain the change rate sequence; the change rate sequence is traversed, the local maximum value point whose rate value is greater than the rate value of the adjacent front and rear time windows is identified, the corresponding time stamp and associated optical detection channel position number are recorded, and the peak coordinate information is formed.

[0080] 5012. The expansion direction and decreasing gradient of the pressure decreasing area in the tread pressure change parameter output by the pressure sensing unit are extracted to determine the pressure decreasing starting point and the pressure decreasing propagation path.

[0081] The step 5012 can specifically include the following processes: dividing the wheel tread into grid units, and counting the drop range of the pressure change parameter in each grid unit; taking the grid unit with the largest drop range as the pressure decrease starting point, and starting from the pressure decrease starting point, tracking the path in the direction of increasing pressure drop range of adjacent grid units until the pressure rises or a stable region is detected, forming a pressure decrease propagation path; and calculating the decrease gradient of the pressure decrease propagation path according to the ratio of the pressure drop range to the path length of each grid unit on the pressure decrease propagation path.

[0082] 5013, identifying the spatial distribution of displacement abnormal points in the vertical displacement parameters generated by the displacement measurement unit.

[0083] The step 5013 can specifically include the following processes: continuously sampling the vertical displacement parameters generated by the displacement measurement unit, detecting the displacement increment difference between adjacent sampling points, and marking the time period between adjacent sampling points as a nonlinear change period if the displacement increment difference exceeds a set mutation threshold; recording the start time stamp of all nonlinear change periods, calculating the maximum time difference between the start time stamps detected by different displacement measurement units as a time sequence deviation; grouping the nonlinear change periods according to the time sequence deviation, screening out the nonlinear change period groups with a time difference less than a synchronization threshold, and in each nonlinear change period group, screening out the displacement points with unidirectional sudden increase according to the consistency of the displacement increment direction, marking them as displacement abnormal points and recording the position coordinates corresponding to the displacement abnormal points; and determining the spatial distribution of displacement abnormal points according to the position coordinates corresponding to the displacement abnormal points.

[0084] 5014, establishing a spatial mapping relationship between the pressure decrease region expansion path and the distribution characteristics of displacement abnormal points based on the peak coordinate information, the pressure decrease starting point, the pressure decrease propagation path, and the spatial distribution of displacement abnormal points.

[0085] The step 5014 can specifically include the following processes: matching the time point in the peak coordinate information with the occurrence time of the pressure decrease starting point, and marking the pressure decrease starting point as an effective trigger point if the time difference between them is less than a set tolerance; defining a plurality of directionally associated regions in the wheel-rail contact area according to the extension direction of the pressure decrease propagation path starting from the effective trigger point, and counting the spatial distribution of displacement abnormal points in each directionally associated region; proportionally distributing the spatial distribution of displacement abnormal points and the pressure decrease gradient of the corresponding directionally associated region, and determining that the directionally associated region is a spatial mapping area of the pressure decrease expansion path and displacement abnormal points if the ratio of the density to the gradient in any directionally associated region exceeds a dynamic threshold; and taking the coverage range and connection relationship of all spatial mapping areas as the spatial mapping relationship between the pressure decrease region expansion path and the distribution characteristics of displacement abnormal points.

[0086] In the above scheme, the blocking signal state refers to a binary signal set indicating whether the light path is blocked or unblocked by the wheel, the tread pressure change parameter is a dynamic numerical sequence representing the change of the wheel-rail contact surface pressure with the lifting height, and the vertical displacement parameter is a physical quantity quantifying the lifting height of the wheel. The pressure reduction area expansion path describes the propagation direction and range of the wheel-rail contact surface pressure dissipation, the displacement abnormal point distribution characteristic refers to a set of spatial positions where the displacement suddenly increases or jitters during the lifting of the wheel, and the spatial mapping relationship is a correlation matching model of multi-dimensional data on the wheel-rail contact surface.

[0087] In the embodiment of the present application, first, the blocking signal state of the optical detection unit is collected in real time through step 5011, the number of unblocked channels is counted in a preset time window (for example, every 0.1 second), and the proportion of the number of unblocked channels in the total number of channels in each window is calculated; then the difference value of the unblocking rate between adjacent time windows is calculated, and the difference value is divided by the time window length to generate a sequence of unblocking rate change rates, which is used to represent the dynamic intensity of the light path unblocking; subsequently, the change rate sequence is traversed, and the local maximum points are identified, that is, the rate value of a certain time window needs to be greater than the rate values of the previous window and the next window, so as to exclude noise interference; finally, the time stamp corresponding to the local maximum value and the associated optical detection channel position number are recorded to form peak coordinate information, for example, the peak coordinate information includes the time stamp 5.2 seconds and the optical detection channel position number 3 to 8, which is used to mark the key event node in the separation process, such as the rapid unblocking stage of the light path.

[0088] Subsequently, the tread pressure parameter output by the pressure sensing unit is received through step 5012, the wheel-rail contact surface is divided into grid units, such as 10 mm by 10 mm square grids, the drop amplitude of the pressure value in each grid unit, that is, the difference between the initial pressure and the current pressure, is counted; then the grid unit with the largest drop amplitude is selected as the pressure reduction starting point, which represents the separation starting position; starting from the starting point, the path is tracked in the direction of increasing pressure drop amplitude of adjacent grid units, such as the right side unit, until the pressure is detected to recover to a stable state, such as the pressure rising to 80% of the initial value or fluctuating less than 5%; the expansion direction of the path, such as extending outward from the center of the rim, and the gradient are recorded, the gradient is calculated by dividing the total pressure drop amplitude of all grid units on the path by the path length, for example, a total drop of 50 MPa divided by a path length of 25 mm gives a gradient of 2 MPa per mm; finally, the complete path is determined as the pressure reduction propagation path, which is used to represent the dynamic characteristics of the wheel-rail contact surface pressure dissipation.

[0089] Meanwhile, the vertical displacement parameters of the displacement measuring units are acquired through step 5013, the displacement increment difference between adjacent sampling points is calculated, and if the difference exceeds a preset mutation threshold (such as 0.5 mm / s), the point is marked as a displacement abnormal point, and the lateral position coordinates of all abnormal points are counted to form a spatial distribution map of displacement abnormal points. Specifically, the vertical displacement parameters generated by the displacement measuring units are continuously sampled, for example, 1000 times per second, the displacement increment difference between adjacent sampling points is calculated, for example, the displacement value of the next sampling point minus the displacement value of the previous sampling point. If the difference exceeds the set mutation threshold, for example, 0.5 mm / s, the time period between the adjacent sampling points is marked as a non-linear change section, representing displacement sudden increase or jitter phenomenon. The starting time stamp of all non-linear change sections is recorded, and the maximum time difference between the starting time stamps detected by different displacement measuring units, for example, left and right sensors, is calculated as the time sequence deviation, for example, the left sensor detects the mutation 0.05 seconds later than the right sensor. According to the time sequence deviation, the non-linear change sections are grouped, for example, sections with a time difference less than 0.1 second are grouped together, and sections with high synchronization are screened out. Within each section group, the direction consistency of the displacement increment is analyzed, for example, all increments are positive values, and displacement points with one-way sudden increase are screened out, marked as displacement abnormal points, and their corresponding track lateral position coordinates are recorded, for example, X=120 mm, Y=40 mm. Finally, a spatial distribution map is generated according to the position coordinates of all abnormal points, reflecting the uneven separation area.

[0090] Finally, the time point (for example, 5.2 seconds) in the peak coordinate information generated by step 5011 is aligned with the pressure decrease starting point occurrence time (for example, 5.1 seconds) determined by step 5012 through step 5014. If the time difference between the two is less than a set tolerance, for example, 50 milliseconds, the pressure decrease starting point is marked as an effective trigger point, ensuring the spatiotemporal consistency of the light path release and pressure drop; taking the effective trigger point as the starting point, a plurality of directionally associated regions are demarcated on the wheel-rail contact surface according to the extension direction of the pressure decrease propagation path, for example, extending outward from the center of the rim; the density of displacement abnormal points in each directionally associated region is counted, for example, the number of abnormal points per square centimeter, and the density is proportionally distributed with the pressure decrease gradient of the corresponding region (for example, 2 MPa per mm) to calculate the ratio of density to pressure decrease gradient; if the ratio in a certain region exceeds a dynamic threshold, the region is determined as a spatial mapping area of the pressure decrease expansion path and displacement abnormal point, indicating that the pressure dissipation and displacement mutation are strongly related here, wherein the dynamic threshold can be set according to requirements, for example, set to 0.8; finally, the coverage range and connection relationship of all spatial mapping areas are integrated to generate an overlapping track of the pressure decrease area expansion path and the displacement abnormal point distribution characteristics on the wheel-rail contact surface, forming a complete spatial mapping relationship.

[0091] That is, in the integration of the spatial mapping area, first, based on the ratio of the density and pressure gradient in the direction of the correlation area, the effective mapping area is screened out, for example, the left wheel rim fan-shaped area density / gradient = 1.2> dynamic threshold (0.8) is marked as an effective mapping area; then according to the extension direction of the pressure propagation path, the spatial connection relationship between each mapping area is analyzed, for example, the left wheel rim fan-shaped area and the center area path are continuous, and the adjacent or path continuous mapping areas are combined into an extended sub-path; finally, according to the coverage range and topological connection characteristics of all sub-paths, the overlapping trajectory of the pressure propagation path and the displacement anomaly point distribution on the wheel-rail contact surface is generated. For example, the fan-shaped trajectory radiating from the wheel rim center is the complete spatial mapping relationship, which is used to represent the strong correlation area of pressure dissipation and displacement mutation.

[0092] In practical application, in the wheel detection scene with wheel diameter of 860 mm, the lifting driving device starts to lift the wheel step by step after the top rotating wheel is started, and the specific process is as follows: the optical detection unit monitors the light path blocking state in real time, when it is lifted to the 5th second, the blocking removal rate change rate of the 3rd to 8th channels reaches the peak value, the time stamp at this time is recorded as 5.2 seconds and the corresponding channel position; the pressure sensing unit synchronously collects the wheel-rail contact surface pressure data, it is found that the pressure in the wheel rim center area continuously decreases from 12 MPa to 4 MPa, the pressure drop direction is the outside direction of the wheel, and the gradient is 2 MPa per mm, so the pressure decrease starting point is determined as the wheel rim center coordinate (X = 120 mm, Y = 0); at the same time, the displacement measuring unit detects that the left wheel rim area has a displacement increase of 0.8 mm when the wheel is lifted to 5 mm, which is marked as a displacement anomaly point and its position coordinate (X = 80 mm, Y = -40 mm) is recorded. The processing unit aligns the peak time 5.2 seconds of the optical detection with the pressure decrease starting point time 5.1 seconds in time sequence, the time difference between the two is 0.1 seconds, which is less than the set tolerance threshold 0.2 seconds, and it is determined as an effective correlation event; then a fan-shaped detection area is drawn along the pressure propagation direction (extending to the outside), the density of the displacement anomaly points in the left wheel rim area is 5 points per square centimeter, and the pressure gradient in this area is 2 MPa per mm, so the ratio of the density and the gradient is 2.5 (5 / 2), which is greater than the dynamic threshold 0.8, and it is determined that this area is a strong correlation mapping area of pressure dissipation and displacement mutation; finally, the spatial position and path continuity of all mapping areas are integrated to generate an overlapping trajectory extending from the wheel rim center to the left outside, forming a complete spatial mapping relationship, which directly represents the dynamic correlation of the pressure dissipation path and the displacement anomaly distribution in the wheel-rail separation process.

[0093] The overall scheme of step 501 realizes fine modeling of the wheel-rail separation process through spatio-temporal correlation and dynamic matching of multi-source data, accurate capture of separation trigger nodes by optical signals, revelation of contact force dissipation rules by pressure data, positioning of uneven separation areas by displacement abnormal points, and finally construction of a spatial mapping relationship that intuitively reflects the physical propagation characteristics of the separation state, providing a high-confidence data basis for subsequent determination and avoiding the risk of false judgment by a single sensor, thereby significantly improving detection accuracy and reliability.

[0094] 502. Map the blocking signal state to the blocking removal rate of the optical detection channel, calculate the displacement cumulative value when the wheel tread separates according to the spatial mapping relationship combined with the blocking removal rate of the optical detection channel.

[0095] Optionally, step 502 includes:

[0096] 5021. Statistically analyze the duration of the blocking signal of each optical detection channel during the lifting process, and calculate the blocking removal weight factor of each optical detection channel according to the duration of the blocking signal.

[0097] 5022. Convert the proportion of the number of optical detection channels with the blocking signal disappearing to the total number of optical detection channels into a blocking removal rate sequence according to the lifting time sequence.

[0098] 5023. According to the pressure reduction expansion path marked in the spatial mapping relationship, filter out the wheel tread area coinciding with the displacement abnormal point distribution area.

[0099] 5024. Accumulate the vertical displacement increments detected by all displacement measurement units in the wheel tread area during the lifting stage, and weight and correct each vertical displacement increment according to the blocking removal rate sequence and the corresponding blocking removal weight factor to generate a displacement cumulative value.

[0100] In the above scheme, the blocking signal state refers to a real-time signal set in which the light path is blocked or removed by the wheel, and the blocking removal rate is a proportion sequence in which the number of channels with the blocking signal disappearing accounts for the total number of channels. The blocking signal duration is the length of time during which the light path of a single channel is continuously blocked during the lifting process, and the blocking removal weight factor is calculated according to the blocking signal duration and is used to represent the reliability weight of channel removal blocking. The displacement cumulative value is the total displacement after accumulating the vertical displacement increments in the wheel tread area, which is corrected by the blocking removal rate sequence and the weight factor, and reflects the actual lifting amount of the wheel-rail separation process. The pressure reduction expansion path marked in the spatial mapping relationship is the propagation direction and range of the wheel-rail contact surface pressure dissipation, and the displacement abnormal point distribution area is a set of spatial positions with sudden displacement increase or jitter.

[0101] In the embodiment of the present application, step 5021 is used to traverse the blocking signal state of each optical detection channel, and the duration of the optical path being blocked during the lifting process is counted, for example, channel A is blocked for 5 seconds. The weight factor is calculated according to the duration, and the formula is weight factor = 1 / (1+duration), which represents the credibility of the channel unblocking.

[0102] Secondly, step 5022 is used to divide the lifting process into fixed time windows, for example, each 0.1 second is a node, and the number of optical detection channels in which the blocking signal disappears in each window is counted in real time. The proportion of the number to the total number of channels is calculated, for example, the total number of channels is 20, and if 15 channels are unblocked in a window, the proportion is 15 / 20=75%; the proportion values of the windows are recorded in sequence according to the lifting time sequence (such as 0 seconds, 0.1 seconds, 0.2 seconds, …), and a sequence of unblocking rates is generated. For example, if the lifting process lasts for 1 second, a sequence is generated at an interval of 0.1 second: the unblocking rate at 0.1 second is 50% (10 / 20), the unblocking rate at 0.2 second is 75% (15 / 20), the unblocking rate at 0.3 second is 90% (18 / 20), and so on to form an unblocking rate sequence that increases with time, which is used to represent the dynamic trend of the separation process.

[0103] Then, step 5023 is used to obtain the generated spatial mapping relationship, extract a pressure-decreasing expansion path, for example, a fan-shaped path extending outward from the center coordinates of the rim X=120, mm Y=0, and obtain a displacement anomaly point distribution area, for example, a displacement sudden increase point in the left rim area X=80 to 160 mm, Y=-40 to 0; the overlapping area of the two is screened through spatial superposition analysis, for example, the left rim outside X=100 to 140 mm, Y=-20 to 0, which is marked as an effective separation area, to ensure that the subsequent displacement cumulative value only accumulates the vertical displacement data of this area. For example, if the pressure path covers the center to the outside of the rim and the displacement anomaly points are concentrated on the left rim outside, the overlapping area is the overlapping fan-shaped area of the two, and the subsequent calculation is only for this area.

[0104] Finally, the vertical displacement increments of all displacement measurement units in the overlapping area detected during the lifting stage are accumulated, for example, 0.2 mm each time, and the displacement increments are weighted and corrected according to the unblocking rate sequence, for example, the unblocking rate at a certain moment is 80%, and the corresponding weight factor, for example, the channel weight factor is 0.8, for example, the increment x the unblocking rate x the weight factor, to finally generate the displacement cumulative value.

[0105] In practical applications, for example, taking the detection of a wheel with a diameter of 840 mm as an example, after the lifting driving device is started, the optical detection unit detects that the blockage signals of 15 channels disappear during the lifting process, the duration of the blockage signals of each channel is counted, and the weight factor is calculated. Taking channel 3 as an example, the duration is 2 seconds, and the weight factor is 0.33. The blockage removal rate sequence is generated according to the lifting sequence, for example, the removal rate at the third second is 70%. The processing unit screens out the area where the left wheel flange outside is the pressure decreasing expansion path and the displacement abnormal point according to the spatial mapping relationship, and accumulates the displacement increment detected by the displacement measurement unit in the area, for example, 0.3 mm each time, combined with the current blockage removal rate of 70% and the weight factor of 0.33, the weighted displacement increment is calculated as 0.3*0.7*0.33=0.07 mm, which is accumulated in the total displacement cumulative value. Finally, when the cumulative value continuously increases and the blockage removal rate meets the standard, it is determined that the wheel rail is separated.

[0106] In the complete scheme of the above step 502, the reliability weight of the optical blockage signal, the dynamic removal rate sequence and the spatial mapping relationship are fused to realize the accurate correction of the displacement cumulative value. The blockage removal weight factor reduces the weight of the short-term interference signal, the blockage removal rate sequence reflects the dynamic trend of separation, and the spatial mapping relationship ensures that the displacement calculation focuses on the effective area. The finally generated displacement cumulative value not only eliminates noise interference, but also reflects the actual lifting amount in the separation process, providing a high credibility quantitative basis for wheel rail separation judgment, and significantly improving the stability and accuracy of the detection result.

[0107] 503、When the displacement cumulative value presents a one-way increasing characteristic in the continuous lifting stage, and the blockage removal rate of the optical detection channel is greater than a preset threshold, it is determined that the wheel rail is separated, and a control instruction is generated to control the lifting driving device to terminate the work.

[0108] Optionally, step 503 comprises:

[0109] 5031、Record the displacement cumulative value in the order of the lifting stage. If the displacement cumulative value of the current stage is greater than that of the previous stage and the difference between the displacement cumulative values continuously increases, it is marked as one-way increasing.

[0110] 5032、Statistical the proportion of the number of channels whose blockage removal rate is greater than a preset threshold in the optical detection channel. If the proportion of the number of channels meets the preset threshold in a continuous number of lifting stages, it is marked as effective proportion meeting the standard.

[0111] 5033、When the one-way increasing mark and the effective proportion meeting the standard mark exist at the same time, a control instruction is generated to control the lifting driving device to terminate the work.

[0112] In the above scheme, the displacement cumulative value is the total sum of the corrected vertical displacement during the wheel tread separation process, and the one-way increasing characteristic means that the displacement cumulative value continuously increases in the continuous lifting stage and the increment difference gradually expands. The blockage removal rate is the proportion sequence of the number of optical detection channels in which the blockage signal disappears to the total number of channels, and the preset threshold is the minimum standard value of the blockage removal rate. The effective proportion standard means that the proportion of the number of channels in which the blockage removal rate is greater than the preset threshold remains stable in continuous lifting stages. The control instruction is a control signal for triggering the stop of the lifting driving device.

[0113] In the embodiments of the present application, first, the displacement cumulative value is recorded in the order of lifting stages, for example, each stage is to lift the lifting mechanism by 1 millimeter. If the displacement cumulative value of the current stage is greater than that of the previous stage and the difference between the current stage and the previous stage is greater than that between the previous stage and the earlier stage, it is marked as one-way increasing, indicating that the lifting process has no rollback and the separation trend is stable.

[0114] Secondly, the number of channels in which the blockage removal rate is greater than the preset threshold in the optical detection channel is counted, and the proportion of the total number of channels is calculated, for example, the preset threshold is 80%. If the proportion of the number of channels in three continuous stages (such as stages 5, 6, and 7) is all ≥80%, it is marked as effective proportion standard, which ensures the continuity of the separation state.

[0115] Finally, when the one-way increasing mark and the effective proportion standard mark exist at the same time, that is, the displacement cumulative value continuously increases and the blockage removal rate stably meets the standard, a control instruction is generated and sent to the lifting driving device to terminate the lifting operation, avoiding excessive lifting or incomplete separation.

[0116] For example, taking the detection of a wheel with a diameter of 860 millimeters as an example, the lifting driving device lifts the wheel step by step, and the processing unit records the displacement cumulative value of each stage. The displacement cumulative value of stage 3 is 3.2 millimeters, that of stage 4 is 4.0 millimeters, and that of stage 5 is 5.1 millimeters, and the difference gradually expands, which is marked as one-way increasing. At the same time, the blockage removal rates of the optical detection channels in stages 4 to 6 are 85%, 88%, and 83% respectively, and the proportion of the number of channels exceeding the preset threshold of 80% in the three continuous stages is marked as effective proportion standard. After the processing unit detects that the two kinds of marks coexist, it immediately generates a control instruction to stop lifting, and the wheel is precisely lifted to the complete off-track position, and the subsequent flaw detection operation is normally executed.

[0117] In the complete scheme of the above step 503, through the one-way increasing characteristic of the displacement cumulative value and the double verification of the continuous standard of the blockage removal rate, the high reliability of the wheel-rail separation judgment is ensured. The one-way increasing characteristic excludes displacement fluctuation interference, and the blockage removal rate threshold verifies the sufficiency of the separation range, and the synergistic effect of the two avoids premature or late termination of the lifting operation, significantly improves the detection accuracy and equipment safety, and at the same time guarantees the stable execution of the subsequent flaw detection process.

[0118] Optionally, the method further comprises:

[0119] 504、in response to the control instruction, starting the driving rotating unit to drive the wheel to rotate, and controlling the sound wave detection unit to perform defect detection on the wheel surface based on the channel activation sequence of the optical detection unit.

[0120] In this step, the control instruction is a signal triggering the device to perform subsequent actions, the driving rotating unit drives the wheel to rotate through the mechanical transmission device, the sound wave detection unit is a device for detecting internal defects of the wheel using an ultrasonic probe, the channel activation sequence is a set of optical channel position numbers with a standard block removal rate in the optical detection unit, which is used to guide the probe moving track. The wheel surface refers to the constant contact pressure between the ultrasonic probe and the wheel surface to ensure the quality of signal transmission.

[0121] In the embodiment of the present application, after determining that the wheel and rail are separated, the processing unit sends a control instruction to the driving rotating unit and the sound wave detection unit; the servo motor of the driving rotating unit engages the wheel shaft end through the gear to set the rotating speed, for example, 5 revolutions per minute, to drive the wheel to rotate at a constant speed, ensuring full coverage of the wheel surface; at the same time, the optical channel position numbers with a standard block removal rate, for example, channels 3 to 15, in the optical detection unit are extracted, and a probe moving priority sequence is generated from high to low according to the removal rate; the mechanical arm of the sound wave detection unit dynamically adjusts the probe position according to the sequence, for example, the outer side area of the wheel rim corresponding to channel 3, the probe is attached to the wheel surface with constant pressure and emits ultrasonic pulses; after the probe receives the reflected sound wave signal, the processing unit analyzes the echo amplitude and time delay characteristics to identify internal defects such as cracks or cavities, and generates a defect distribution map in real time. For example, after a certain wheel with a diameter of 860 mm is lifted to completely separate from the rail, the driving rotating unit drives the wheel to rotate at 5 revolutions per minute, and the sound wave detection unit preferentially scans the wheel surface area corresponding to channels 5 to 12 with a block removal rate of 90%, detects a crack with a depth of 2 mm on the inner side of the wheel rim and marks the coordinates, and completes the defect detection of the full wheel surface.

[0122] In actual application, taking the detection of a wheel with a diameter of 860 mm as an example, the processing unit sends a control instruction after determining that the wheel and rail are separated, and the driving rotating unit is started and drives the wheel to rotate at a constant speed of 5 revolutions per minute; the channel activation sequence of the optical detection unit shows that the block removal rates of channels 5 to 12 are all more than 90%, corresponding to the middle to outer area of the wheel surface; the mechanical arm of the sound wave detection unit first moves to the wheel surface position corresponding to channel 5 according to the priority, the probe is attached to the wheel surface and emits ultrasonic pulses, detects an abnormal echo on the inner side of the wheel rim, and marks it as a potential crack with a depth of 2 mm; as the wheel continues to rotate, the probe successively covers the areas corresponding to channels 6 to 12, and generates a defect report containing the crack position and size after completing the full wheel surface detection.

[0123] In the scheme of step 504, through the cooperative control of driving the rotating unit and the sound wave detection unit, the automatic detection operation after the wheel is off the track is realized. The channel activation sequence guides the probe to preferentially cover the high-probability defect area, improving the detection efficiency; the constant pressing pressure guarantees the stability of ultrasonic signal transmission, avoiding missed detection; the rotating motion ensures the full circumferential wheel surface without dead angle detection. This step seamlessly connects the separation judgment and defect detection, forms a complete maintenance operation closed loop, and significantly improves the detection accuracy and operation automation level.

[0124] The following is a complete embodiment for steps 501-504:

[0125] In the online ultrasonic flaw detection of a certain type of rail vehicle wheel with a wheel diameter of 860 mm, after the lifting driving device is started, the pressure sensing unit detects the pressure change of the wheel-rail contact surface in real time. The initial pressure is 12 MPa, and the pressure decreases to 3 MPa along the wheel rim center to the outside during lifting.

[0126] The optical detection unit synchronously monitors the light path blocking state. When the wheel contacts the track, 15 light paths in 20 channels are blocked, and when the lifting is 4 mm, the blocked channels increase to 18.

[0127] The displacement measurement unit records the vertical displacement parameters, and detects a sudden increase of 0.5 mm in displacement of the left wheel rim when the lifting is 5 mm. The processing unit fuses the pressure decreasing path, the displacement abnormal point distribution, and the peak value of the unblocking rate, among which the unblocking rate reaches 90% at the 5th second, constructs the spatial mapping relationship and locates the separation core area. Then, the duration of each optical channel blocking is counted, and the weight factor is calculated, for example, channel 3 lasts for 3 seconds, and the weight factor is 1 / (1+3)=0.25. Combined with the unblocking rate sequence of 70% at the 3rd second, 85% at the 4th second, and 90% at the 5th second, the displacement increment is weighted and corrected, and the displacement increment of the left wheel rim area is corrected to 0.5×0.9×0.25≈0.11 mm, which is added to the total displacement cumulative value.

[0128] When the displacement cumulative value presents a one-way increase from 4.0 mm to 4.5 mm and then to 5.1 mm in stages 3 to 5, and the unblocking rate meets the standard for three consecutive stages, a control instruction is generated to terminate the lifting, and the wheel is lifted to 5.1 mm and completely off the track. Then the driving rotating unit drives the wheel to rotate at a uniform speed of 5 revolutions per minute, and the sound wave detection unit activates channels 5 to 18 based on the unblocking rate meeting the standard, and preferentially detects the outer side area of the wheel rim. The probe emits ultrasonic waves and identifies an internal crack with a depth of 2 mm, and completes the nondestructive detection of the full circumferential wheel surface.

[0129] The application realizes accurate determination and self-adaptive lifting of wheel-rail separation state through dynamic fusion of multi-source sensing data and real-time feedback control, and the displacement cumulative value reflects the effective lifting amount after deformation compensation and weight correction, and the double verification mechanism ensures the high reliability of separation determination. In the flaw detection stage, the optical channel priority coverage and rotation cooperative control are used to improve the defect detection efficiency and comprehensiveness, and a full-closed-loop operation process of separation detection, dynamic control and defect identification is formed, which significantly solves the wheel diameter adaptability defect of the traditional fixed lifting strategy, and balances the detection accuracy, equipment safety and maintenance automation level.

[0130] Optionally, the method further comprises: synchronously collecting the contact surface images of the wheel and the rail during the lifting process, extracting the illumination reflection features and the contour deformation parameters of the contact area from the contact surface images; comparing the illumination reflection features with a preset standard reflection template under the wheel-rail fitting state to generate a deformation compensation factor; dynamically correcting the displacement cumulative value based on the deformation compensation factor, and optimizing the boundary accuracy of the space mapping relationship by fusing the contour deformation parameters.

[0131] In this step, the contact surface image is real-time image data of the contact area of the wheel and the rail, the illumination reflection feature refers to the light intensity distribution pattern in the image caused by surface material or deformation, and the contour deformation parameter is the geometric distortion quantity of the edge of the wheel contact surface. The standard reflection template is a pre-stored reflection feature reference image under the condition of complete fitting of the wheel and the rail, and the deformation compensation factor is a correction coefficient generated by comparing the real-time reflection feature with the template difference, which is used to eliminate the interference of deformation on displacement detection.

[0132] In the embodiment of the application, during the lifting process, the contact surface images of the wheel and the rail are synchronously photographed by a high-resolution industrial camera, the images are preprocessed by greying, denoising and edge enhancement, the illumination reflection features such as high light area distribution and reflection intensity gradient are extracted, and the contour deformation parameters such as local bending degree of the rim edge are calculated; the real-time illumination reflection features are compared with the pre-stored standard reflection template to generate a difference matrix, the deformation compensation factor is generated according to the amplitude distribution of the difference matrix, the weight of the compensation factor is higher in the area with larger difference, for example, the difference value 0.8 corresponds to the compensation factor 0.6.

[0133] The displacement cumulative value is weighted and corrected based on the compensation factor, for example, the deformation area displacement increment 0.3 millimeters multiplied by the compensation factor 0.6 is corrected to 0.18 millimeters.

[0134] Finally, the fusion contour deformation parameters, such as the left rim curvature of 2 degrees, are adjusted to improve the boundary accuracy of the pressure-decreasing path and the displacement abnormal point in the space mapping relationship, for example, the left rim mapping area boundary is inlaid by 5 mm. Taking a certain wheel detection as an example, the contact surface image shows that the left rim has a high light reflection abnormality and the edge is curved. After generating a compensation factor of 0.6, the virtual displacement of this area is corrected, and the mapping boundary is optimized to avoid misjudgment of the non-contact area, thereby improving the detection accuracy.

[0135] In actual application, taking a certain wheel detection as an example, the industrial camera captures the contact surface image during lifting, and the image shows that there is a high light reflection area on the left side of the rim and the edge contour is slightly curved. The processing unit extracts the light reflection features of this area, compares them with the pre-stored standard reflection template at the pixel level to generate a difference matrix, and the difference value of the left area is 0.8. The corresponding compensation factor is set to 0.6. The increment of 0.3 mm from this area in the displacement cumulative value is corrected to 0.3 x 0.6 = 0.18 mm. At the same time, the contour deformation parameter detects that the left rim curvature is 2 degrees, and the space mapping relationship accordingly inlays the left mapping area boundary by 5 mm to avoid misjudgment of the non-contact area. Through deformation compensation and boundary optimization, the system accurately corrects the virtual displacement error and improves the detection boundary accuracy, ensuring the reliability of the wheel-rail separation judgment.

[0136] In the above steps, through image feature analysis and deformation compensation mechanism, the displacement detection error caused by wheel surface deformation or foreign matter attachment is significantly reduced. The light reflection difference comparison accurately locates the abnormal area, the deformation compensation factor dynamically suppresses the virtual displacement, and the contour parameter optimizes the space mapping boundary. The three work together to improve the environmental adaptability and boundary judgment accuracy of the detection system, ensuring the reliability of the wheel-rail separation detection result under complex working conditions.

[0137] Finally, it should be noted that: the above examples are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing examples, those skilled in the art should understand that they can still modify the technical solutions described in the foregoing examples, or make equivalent substitutions for some of the technical features; and these modifications or substitutions do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A wheel tread derailing detection method for a railway vehicle, characterized in that, The lifting driving device is symmetrically arranged on both longitudinal sides of the track, each lifting driving device includes two lifting rollers for lifting the wheel, a pressure sensing unit is arranged between the two lifting rollers, an optical detection unit is mounted on the outside of the lifting driving device, the optical path direction of the optical detection unit is orthogonal to the wheel movement plane, and a displacement measuring unit is integrated on the lifting roller, when the lifting roller lifts the wheel, the displacement measuring unit moves synchronously to generate the vertical displacement parameter of the wheel; the pressure sensing unit, the optical detection unit and the displacement measuring unit form a three-dimensional detection coverage area in the wheel-rail contact area; When the wheel contacts the track, the optical path of the optical detection unit is completely blocked by the wheel; The method comprises: During the process of gradually lifting the wheel by the lifting roller of the lifting driving device, the blocking signal state of the optical detection unit, the tread pressure change parameter output by the pressure sensing unit and the vertical displacement parameter generated by the displacement measuring unit are acquired to establish the spatial mapping relationship between the pressure decreasing area expansion path and the displacement abnormal point distribution characteristics; The blocking signal state is mapped to the unblocking rate of the optical detection channel, and the displacement cumulative value when the wheel tread separates is calculated according to the spatial mapping relationship and the unblocking rate of the optical detection channel; When the displacement cumulative value presents a one-way increasing characteristic in the continuous lifting stage, and the unblocking rate of the optical detection channel is greater than a preset threshold, it is determined that the wheel-rail separates, and a control instruction is generated to control the lifting driving device to terminate the work.

2. The method of claim 1, wherein, Further comprising: Synchronously collecting the contact surface images of the wheel and the track during the lifting process, extracting the illumination reflection characteristics and the contour deformation parameters of the contact area from the contact surface images; Differentially comparing the illumination reflection characteristics with a preset standard reflection template in the wheel-rail fitting state to generate a deformation compensation factor; Based on the deformation compensation factor, the displacement cumulative value is dynamically modified, and the contour deformation parameters are fused to optimize the boundary accuracy of the spatial mapping relationship.

3. The method of claim 1, wherein, The acquisition of the blocking signal state of the optical detection unit, the tread pressure change parameter output by the pressure sensing unit and the vertical displacement parameter generated by the displacement measuring unit to establish the spatial mapping relationship between the pressure decreasing area expansion path and the displacement abnormal point distribution characteristics comprises: According to the blocking signal state of the optical detection unit, the change rate of the unblocking rate of the optical detection channel is calculated, and the peak coordinate information corresponding to the change rate is extracted; The expansion direction and the decreasing gradient of the pressure decreasing area in the tread pressure change parameter output by the pressure sensing unit are extracted to determine the pressure decreasing starting point and the pressure decreasing propagation path; The spatial distribution of the displacement abnormal points in the vertical displacement parameter generated by the displacement measuring unit is recognized; Based on the peak coordinate information, the pressure decreasing starting point and the pressure decreasing propagation path and the spatial distribution of the displacement abnormal points, the spatial mapping relationship between the pressure decreasing area expansion path and the displacement abnormal point distribution characteristics is established.

4. The method of claim 3, wherein, The space mapping relationship between the pressure reduction region expansion path and the displacement anomaly point distribution characteristics is established based on the peak coordinate information, the pressure reduction starting point, the pressure reduction propagation path, and the spatial distribution of the displacement anomaly point, and includes: The time point in the peak coordinate information is matched with the occurrence time of the pressure reduction starting point. If the time difference between the two is less than the set tolerance, the pressure reduction starting point is marked as an effective trigger point; Taking the effective trigger point as the starting point, a plurality of directionally associated regions are demarcated in the wheel-rail contact area according to the extension direction of the pressure reduction propagation path, and the spatial distribution of the displacement anomaly points in each directionally associated region is counted; The spatial distribution of the displacement anomaly points is proportionally distributed with the pressure reduction gradient of the corresponding directionally associated region. If the ratio of the density to the gradient in any directionally associated region exceeds the dynamic threshold, the directionally associated region is determined as the space mapping area of the pressure reduction expansion path and the displacement anomaly point. The coverage range and connection relationship of all space mapping areas are taken as the space mapping relationship between the pressure reduction region expansion path and the displacement anomaly point distribution characteristics.

5. The method of claim 1, wherein, The blocking signal state is mapped to the blocking release rate of the optical detection channel. According to the space mapping relationship, the displacement cumulative value when the wheel tread separates is calculated by combining the blocking release rate of the optical detection channel, including: The blocking signal duration of each optical detection channel during lifting is counted, and the blocking release weight factor of each optical detection channel is calculated according to the blocking signal duration; The proportion of the number of optical detection channels in which the blocking signal disappears to the total number of optical detection channels is converted into a blocking release rate sequence according to the lifting sequence; According to the pressure reduction expansion path marked in the space mapping relationship, the wheel tread area that coincides with the displacement anomaly point distribution area is screened out; The vertical displacement increments detected by all displacement measurement units in the wheel tread area during the lifting stage are accumulated, and each vertical displacement increment is weighted and corrected according to the blocking release rate sequence and the corresponding blocking release weight factor to generate a displacement cumulative value.

6. The method of claim 3, wherein, According to the blocking signal state of the optical detection unit, the change rate of the blocking release rate corresponding to the optical detection channel is calculated, and the peak coordinate information corresponding to the change rate is extracted, including: The blocking release rate sequence is divided into preset time windows, the difference value of the blocking release rate in each time window is calculated, and the difference value is divided by the time window length to obtain a change rate sequence; The change rate sequence is traversed, the local maximum points with a rate value greater than the rate values of adjacent previous and subsequent time windows are identified, the corresponding time stamps and associated optical detection channel position numbers are recorded, and peak coordinate information is formed.

7. The method of claim 3, wherein, The extension direction and the reduction gradient of the pressure reduction region in the tread pressure change parameter output by the pressure sensing unit are extracted to determine the pressure reduction starting point and the pressure reduction propagation path, including: The wheel tread is divided into grid units, and the drop amplitude of the pressure change parameter in each grid unit is counted; The grid cell with the largest pressure decrease is taken as the pressure reduction starting point, and the path is traced from the pressure reduction starting point along the direction of increasing pressure decrease of adjacent grid cells until a pressure increase or stable region is detected, thus forming a pressure reduction propagation path. The decreasing gradient of the pressure decreasing propagation path is calculated based on the ratio of the pressure drop magnitude of each grid cell to the path length.

8. The method of claim 3, wherein, The process of identifying the spatial distribution of displacement anomalies in the vertical displacement parameters generated by the displacement measurement unit includes: The vertical displacement parameters generated by the displacement measurement unit are continuously sampled, and the displacement increment difference between adjacent sampling points is detected. If the displacement increment difference exceeds a set abrupt change threshold, the time period between adjacent sampling points is marked as a nonlinear change segment. Record the start timestamp of all nonlinear change segments, and calculate the maximum time difference between the start timestamps detected by different displacement measurement units as the timing deviation; The nonlinear change segments are grouped according to the time deviation, and groups of nonlinear change segments with a time difference less than the synchronization threshold are selected. In each group of nonlinear change segments, displacement points with unidirectional sudden increases are selected according to the consistency of the direction of displacement increment, marked as displacement anomaly points, and the position coordinates corresponding to the displacement anomaly points are recorded. The spatial distribution of displacement anomalies is determined based on the location coordinates of multiple displacement anomalies.

9. The method of claim 1, wherein, When the cumulative displacement value exhibits a unidirectional increasing characteristic during the continuous lifting phase, and the blocking release rate of the optical detection channel is greater than a preset threshold, wheel-rail separation is determined, and a control command is generated to control the lifting drive device to terminate operation, including: Record the cumulative displacement value in the order of the lifting stages. If the cumulative displacement value of the current stage is greater than that of the previous stage and the difference between the cumulative displacement values ​​continues to increase, it is marked as unidirectional increase. The percentage of channels with a blocking release rate greater than a preset threshold in the optical detection channels is counted. If the percentage of channels reaches the preset threshold in a consecutive number of lifting stages, it is marked as an effective proportion meeting the standard. When a unidirectional increment marker and an effective proportion target marker coexist, a control command is generated to control the lifting drive device to terminate operation. The method further includes: In response to the control command, the drive rotation unit is activated to drive the wheel to rotate, and the acoustic detection unit is adjusted to perform defect detection work on the wheel surface based on the channel activation sequence of the optical detection unit.

10. A wheel tread derailment detection system for a railway vehicle, characterized in that It includes a lifting drive device, an optical measurement unit, and a processing unit; wherein, each lifting drive device includes two top rollers, a pressure sensing unit, and a displacement detection unit, the lifting drive devices are arranged on both sides of the longitudinal direction of the track, and each lifting drive device includes two top rollers, the top rollers being used to lift the wheels; A pressure sensing unit is provided between the two top rollers, and the pressure sensing unit is used to output the tread pressure change parameters; An optical detection unit is installed on the outside of the lifting drive device. Its optical path direction is orthogonal to the wheel movement plane. When the wheel contacts the track, the optical path of the optical detection unit is completely blocked by the wheel. The top rotating wheel is integrated with a displacement measuring unit, which moves synchronously to generate a vertical displacement parameter of the wheel when the top rotating wheel lifts the wheel; the pressure sensing unit, the optical detection unit and the displacement measuring unit form a three-dimensional detection coverage area in a wheel-rail contact area; The processing unit is configured to acquire a blocking signal state of the optical detection unit, a tread pressure change parameter output by the pressure sensing unit and a vertical displacement parameter generated by the displacement measuring unit during a process in which the top rotating wheel of the jacking driving device lifts the wheel step by step, so as to establish a spatial mapping relationship between a pressure decreasing area expansion path and a displacement abnormal point distribution characteristic; map the blocking signal state to a blocking removal rate of an optical detection channel, calculate a displacement cumulative value when the wheel tread separates according to the spatial mapping relationship and the blocking removal rate of the optical detection channel; when the displacement cumulative value presents a one-way increasing characteristic in a continuous lifting stage and the blocking removal rate of the optical detection channel is greater than a preset threshold, determine that the wheel-rail separates, and generate a control instruction to control the jacking driving device to terminate the operation.

Images