Wheel abrasion detection method and system
By using NURBS curve fitting and genetic algorithm optimization, the problem of inaccurate wheel wear in existing technologies has been solved, achieving precise quantification of wheel wear and supporting wheel life assessment and driving safety assurance.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- EAST CHINA JIAOTONG UNIVERSITY
- Filing Date
- 2026-01-28
- Publication Date
- 2026-05-12
AI Technical Summary
Existing train wheel inspection methods cannot accurately reflect the specific condition of wheel wear, nor can they provide detailed information on wheel wear, resulting in an inability to effectively assess wheel lifespan and driving safety.
The wheel tread data was fitted using non-uniform rational B-spline curves (NURBS). By calculating the coordinates of control points and the curve weight factors, the tread curve after wheel wear was constructed. The weight factors were then optimized using a genetic algorithm to establish a quantitative mapping relationship between wear amount and wear depth.
It achieves precise quantitative characterization of wheel wear, and can automatically calculate wear value after obtaining operating mileage, providing high-precision wear amount and wear depth information, supporting wheel life assessment and driving safety assurance.
Smart Images
Figure CN122015735A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of wheel inspection technology, and in particular to a method and system for detecting wheel wear. Background Technology
[0002] As the core load-bearing component in contact with the track, the wear condition of the wheel directly determines the safety, stability, and economy of train operation. With the increase in operating speed, the accumulation of operating mileage, and the frequent passage of complex track conditions (such as small-radius curves and turnout areas), the dynamic interaction between the wheel and rail becomes increasingly intense, and wheel wear exhibits complex characteristics of nonlinearity and multi-factor coupling. Even slight wear (on the order of a few tenths of a millimeter) can alter the wheel-rail contact relationship, leading to a decrease in critical speed, increased vibration, and in severe cases, faults such as excessive flange thickness and tread peeling. This not only shortens the wheel's service life (the normal turning cycle is usually 200,000-300,000 kilometers) but also requires huge maintenance costs and may even threaten driving safety.
[0003] In the existing train wheel inspection process, a wheelset fault dynamic detection system is generally used to inspect the vehicle. However, it only measures values such as wheel flange thickness, height, wheel diameter, inner distance, and coaxial wheel diameter difference, and fails to reflect the specific wear depth and wear amount of the wheel, thus failing to reflect the specific wear situation of the wheel. Summary of the Invention
[0004] The purpose of this invention is to provide a wheel wear detection method and system to solve the problem that the prior art cannot reflect the specific situation of wheel wear.
[0005] The present invention is implemented as follows: a wheel wear detection method, the method comprising: Acquire wheel tread measurement data of the sample train and calculate the measurement parameter values of the wheel tread measurement data; The control point coordinates and curve weight factors are determined based on the measured parameter values, and the wheel wear tread curve is constructed based on the control point coordinates and the curve weight factors. Calculate the area of difference between the wear curve of the wheel and the measured data of the wheel tread, and iteratively optimize the weight factor of the curve based on the area of difference to obtain the weight factor combination; The operating mileage of the train wheels to be tested is obtained, and the wheel wear value is determined based on the combination of the operating mileage and the weighting factor. The wheel wear value includes wear amount and / or wear depth.
[0006] Preferably, the formula used to calculate the measurement parameter values of the wheel tread measurement data includes: in, This indicates the value of the measured parameter. This represents the ordered sequence of points in the wheel tread measurement data. This represents the forward difference vector.
[0007] Preferably, the formula used to determine the control point coordinates and curve weight factors based on the measured parameter values includes: in, Indicates the coordinates of the control point. Indicates the target point for fitting. This represents the curve weight factor corresponding to the control point. Describes the k-th degree B-spline basis function. This represents the j-th data point in the wheel tread measurement data, and n represents the total number of data points in the wheel tread measurement data.
[0008] Preferably, constructing the wheel wear tread curve based on the control point coordinates and the curve weighting factor includes: A non-uniform rational B-spline curve is constructed based on the control point coordinates and the curve weight factor. The control point coordinates are constrained in the non-uniform rational B-spline curve according to the control point constraint conditions to obtain the tread curve after wheel wear. The control point constraint condition is that at the curve connection point of the tread curve after wheel wear, the coordinates of the left and right adjacent control points are collinear.
[0009] Preferably, the formula used to calculate the area of difference between the worn tread curve and the measured data of the wheel tread includes: in, This represents the area of difference in the region of difference. This represents the number of measured scatter points in the wheel tread wear section of the wheel tread measurement data. This indicates the number of scatter points in the region of difference. This represents the x-coordinate of the i-th vertex in the region of difference. This represents the x-coordinate of the (i+1)th vertex in the region of difference. This represents the ordinate of the i-th vertex in the region of difference. This represents the ordinate of the (i+1)th vertex in the region of difference. Indicates the first difference region The x-coordinates of the vertices, Indicates the first difference region The y-coordinates of the vertices, This represents the x-coordinate of the first vertex in the region of difference. This represents the ordinate of the first vertex in the region of difference.
[0010] Preferably, the curve weight factors are iteratively optimized based on the area of difference to obtain a weight factor combination, including: Using the curve weight factors as variables and the minimum difference area as the target value, a genetic algorithm is used to iteratively optimize the curve weight factors to obtain the weight factor combination, wherein the weight factor combination includes at least one internal weight factor.
[0011] Preferably, the wheel wear value is determined based on the combination of the operating mileage and the weighting factors, and the wheel wear value includes the wear amount and / or wear depth, using the following formula: in, Indicates the amount or depth of wear. , , This represents the internal weight factor in the weight factor combination. Indicates the error term. , , , , , , , , , This represents the fitting coefficient corresponding to the operating mileage.
[0012] Another objective of this invention is to provide a wheel wear detection system, the system comprising: The parameterization module is used to acquire wheel tread measurement data of the sample train and calculate the measurement parameter values of the wheel tread measurement data. The tread curve construction module is used to determine the control point coordinates and curve weight factors based on the measured parameter values, and to construct the tread curve after wheel wear based on the control point coordinates and the curve weight factors. The weight factor optimization module is used to calculate the area of difference between the wear-out tread curve and the measured data of the wheel tread, and to iteratively optimize the weight factor of the curve based on the area of difference to obtain a weight factor combination. The wear output module is used to acquire the operating mileage of the train wheels under test, and determine the wheel wear value based on the combination of the operating mileage and the weighting factor. The wheel wear value includes wear amount and / or wear depth.
[0013] Preferably, the tread curve construction module is further configured to: construct a non-uniform rational B-spline curve based on the control point coordinates and the curve weight factor, and constrain the control point coordinates in the non-uniform rational B-spline curve according to the control point constraint conditions to obtain the tread curve after wheel wear. The control point constraint condition is that at the curve connection point of the tread curve after wheel wear, the coordinates of the left and right adjacent control points are collinear.
[0014] Preferably, the weight factor optimization module is further configured to: use the curve weight factor as a variable, take the minimum difference area as the target value, and use a genetic algorithm to iteratively optimize the curve weight factor to obtain the weight factor combination, wherein the weight factor combination includes at least one internal weight factor.
[0015] In this embodiment of the invention, by calculating the measurement parameter values of wheel tread measurement data, the coordinates of control points and curve weight factors can be effectively determined based on the measurement parameter values. Based on the control point coordinates and curve weight factors, the tread curve after wheel wear can be effectively constructed. By calculating the difference area between the tread curve after wheel wear and the wheel tread measurement data, the curve weight factors can be effectively iteratively optimized based on the difference area to obtain a weight factor combination. By using the weight factor combination as wear characteristic parameters, a quantitative mapping relationship between wear characteristic parameters and wheel wear values is constructed, so that after obtaining the operating mileage of the train wheel to be tested, the corresponding wheel wear value can be automatically calculated, which can intuitively reflect the specific situation of wheel wear. Attached Figure Description
[0016] Figure 1 This is a flowchart of the wheel wear detection method provided in the first embodiment of the present invention; Figure 2 This is a schematic diagram of the NURBS curve fitting region division of the wheel tread provided in the first embodiment of the present invention; Figure 3 This is a schematic diagram of the selection of control points for wheel wear sections and the constraint conditions of the control points provided in the first embodiment of the present invention; Figure 4 This is a schematic diagram of iterative optimization of weight factors based on a genetic algorithm provided in the first embodiment of the present invention; Figure 5 This is a schematic diagram of the wheel wear detection system provided in the second embodiment of the present invention; Figure 6 This is a schematic diagram of the structure of the terminal device provided in the third embodiment of the present invention. Detailed Implementation
[0017] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.
[0018] To illustrate the technical solution described in this invention, specific embodiments are described below.
[0019] Example 1 Please see Figure 1 This is a flowchart of a wheel wear detection method provided in the first embodiment of the present invention. This wheel wear detection method can be applied to any device or system, and includes the following steps: Step S10: Obtain the wheel tread measurement data of the sample train and calculate the measurement parameter values of the wheel tread measurement data; The wheel tread measurement data consists of a series of discrete points. To facilitate further analysis, the discrete point data will be fitted. The wear of the wheel tread is mainly concentrated in three areas: the first detection section (the range of 20 mm to the left and right of the nominal rolling circle of the wheel tread), the second detection section (the connection between the tread and the rim), and the third detection section (the rim area). These three sections will be the main research objects. That is, the measurement parameter values of the corresponding data in the first, second, and third detection sections of the wheel tread measurement data will be calculated.
[0020] Optionally, the formula used to calculate the measurement parameter values of the wheel tread measurement data includes: in, This indicates the value of the measured parameter. This represents the ordered sequence of points in the wheel tread measurement data. Represents the forward difference vector. This indicates the value of the first measured parameter.
[0021] Step S20: Determine the control point coordinates and curve weight factor based on the measured parameter values, and construct the wheel wear tread curve based on the control point coordinates and the curve weight factor; Non-uniform rational B-spline (NURBS) curves, manipulated by control vertices, offer high flexibility. The introduction of weight factors provides an effective means of adjusting the influence of control points on the curve shape. Specifically, a larger weight factor results in a curve that more closely approximates the corresponding control point. By appropriately changing the weight factors, the curve morphology can be precisely controlled, enabling it to describe various complex geometric shapes. Therefore, using NURBS curves to fit worn tread surfaces not only accurately expresses the discrete point information of the surface profile by adjusting control points and weight factors, but also facilitates the subsequent extraction of characteristic parameters. Based on these advantages, NURBS curves are used for fitting analysis of worn tread surfaces.
[0022] The goal of fitting is to make the NURBS curve To get as close as possible to these data points, construct the objective function using the least squares method: The least squares method is a commonly used approach for solving optimization problems. The measured wheel tread represents a series of data points. The core idea of the least squares method is to minimize the sum of squares of the errors between the NURBS fitted curve and the data points. For each data point, the difference between the fitted curve and the corresponding point on the NURBS fitted curve is calculated, and these differences are squared and summed. By adjusting the control points and weights of the NURBS fitted curve, the sum of squares is minimized, making the fitted curve as close as possible to the data points. This yields the wheel wear tread curve (the optimal fitted curve) in the sense of minimizing the sum of squares of the errors.
[0023] Optionally, the formulas used to determine the control point coordinates and curve weight factors based on the measured parameter values include: Among them, among them, Indicates the coordinates of the control point. Indicates the target point for fitting. This represents the curve weight factor corresponding to the control point. Describes the k-th degree B-spline basis function. Let j represent the j-th data point in the wheel tread measurement data, and n represent the total number of data points in the wheel tread measurement data. By solving the above system of equations, the optimal solution for the control point coordinates and curve weight factors is obtained, thereby determining the NURBS curve of the fitted wheel tread and obtaining the tread curve after wheel wear.
[0024] Further, the wheel wear tread curve is constructed based on the control point coordinates and the curve weighting factor, including: A non-uniform rational B-spline curve is constructed based on the control point coordinates and the curve weight factor. The control point coordinates are constrained in the non-uniform rational B-spline curve according to the control point constraint conditions to obtain the tread curve after wheel wear. The control point constraint condition is that at the curve connection point of the tread curve after wheel wear, the coordinates of the left and right adjacent control points are collinear. The wheel tread is described using a piecewise approximation fitting method; please refer to [link / reference]. Figure 2 The wheel flange end section AB and the tread end section EF show almost no wear throughout the train's entire service life. Tread wear mainly occurs at the flange, the junction of the tread and flange, and within 20mm to the left and right of the nominal rolling circle of the wheel tread. These three areas correspond to... Figure 2 The three sections BC, CD, and DE are the main research objects. The NURBS approximation fitting method is used to fit the wear sections. In the BC and CD sections, the wear depth gradually increases throughout the entire service life of the train, but the wear width does not increase significantly. Therefore, the same control vertex and the same node vector are used for fitting, and the weight factor of the NURBS curve is used as the characteristic parameter of the shape. In the DE section, the wear depth and wear width increase simultaneously with the increase of operating mileage. The same control endpoint is used, and the intermediate control points and weight factors of different NURBS curves are used as characteristic parameters to describe the wear characteristics.
[0025] In this embodiment, based on the characteristics of the wheel wear section, a third-order NURBS curve and five control points are selected for fitting. To ensure the smoothness and continuity of the fitted tread surface at the connection point, control point constraints are added to the second and fourth control points in the NURBS curve, namely, at the curve connection point, the left and right adjacent control points are collinear. Figure 3 The selection of control points and their constraints in the wheel wear section is shown in the figure. As can be seen from the figure, in the BC section of the wheel flange wear section, the control points... It needs to pass through the previous arc at the point tangent point Above, control point It needs to go through the control endpoint tangent Similarly, in the wear section CD, the intermediate control point and They must pass through the control endpoints respectively and tangent and In the wear section DE, intermediate control points P6-1 and P6-3 must pass through control endpoints. and tangent and .
[0026] To use NURBS curves to describe the characteristics and changes of all wheel tread profiles within a whole cycle, the selection of control points is crucial. The control points need to be outside the maximum wear limit of the tread. Therefore, the data of 64 wheel treads from 200,000 km of high-speed train were processed, and the coordinates corresponding to the maximum wear range of the tread section were taken and recombined to form the wear limit range of the wheel tread for 200,000 km.
[0027] By fitting NURBS curves to segments BC, CD, and DE of this limit range, and assuming all five control vertices have a weight factor of 1, the control vertices are calculated using the least squares method as follows: Figure 3 As shown. The coordinates of the control vertices are shown in Table 1; The NURBS curve node vector of segment BC is The NURBS curve node vector of segment CD is The NURBS curve node vector of segment DE is .
[0028] Table 1. Coordinates of NURBS Curve Control Points for Wheel Tread Wear Sections Using a triple NURBS curve, the wear-resistant tread surface showed good results after fitting five control points. The absolute value error of the normal in section BC was within 0.05 mm, the absolute value error of the normal in section CD was within 0.01 mm, and the absolute value error of the normal in section DE was within 0.0035 mm.
[0029] Step S30: Calculate the area of difference between the wear curve of the wheel and the measured data of the wheel tread, and iteratively optimize the weight factor of the curve based on the area of difference to obtain the weight factor combination; To accurately describe the evolution of high-speed train wheel treads within a single refinishing cycle and quantitatively analyze the macroscopic geometric changes of the wheel treads, feature extraction of the worn treads is proposed. The essence of feature extraction is to extract features that represent the information of this series of discrete points, using as few parameters as possible to accurately and comprehensively represent the worn tread information. The worn tread curve can be accurately fitted by a cubic NURBS curve with five control vertices. Since the first and last control vertices of the NURBS curve are determined, three internal weighting factors are proposed as tread feature parameters for sections BC and CD.
[0030] NURBS curves are defined by control vertices, node vectors, and weight factors. To use the weight factors as characteristic parameters of the tread under different wear conditions, the control vertices and node vectors must be identical. Within a refinishing cycle, the wheel tread at various wear stages needs to undergo NURBS curve fitting at the same set of control points (the x-coordinates of the four control endpoints B, C, D, and E are determined at different operating mileages; the y-coordinates have slight deviations due to wear and measurement errors, which are therefore negligible) and the same node vector, only changing the weight factors. At this point, the weight factors of the curves at each wear stage can be used as characteristic parameters of the wear profile. As the weight factors are known, the smaller the weight factor, the smaller the weight of the control point in determining the curve position, and the curve will gradually move away from the corresponding control point. However, changing any one of the three weight factors will alter the curve shape of that segment; therefore, a genetic algorithm is used to optimize the weight factors.
[0031] Optionally, the formula used to calculate the area of difference between the worn tread curve and the measured data of the wheel tread includes: in, This represents the area of difference in the region of difference. This represents the number of measured scatter points in the wheel tread wear section of the wheel tread measurement data. This indicates the number of scatter points in the region of difference. This represents the x-coordinate of the i-th vertex in the region of difference. This represents the x-coordinate of the (i+1)th vertex in the region of difference. This represents the ordinate of the i-th vertex in the region of difference. This represents the ordinate of the (i+1)th vertex in the region of difference. Indicates the first difference region The x-coordinates of the vertices, Indicates the first difference region The y-coordinates of the vertices, This represents the x-coordinate of the first vertex in the region of difference. This represents the ordinate of the first vertex in the region of difference.
[0032] Further, the curve weight factors are iteratively optimized based on the difference area to obtain a weight factor combination, including: using the curve weight factors as variables and the minimum difference area as the target value, a genetic algorithm is used to iteratively optimize the curve weight factors to obtain the weight factor combination, wherein the weight factor combination includes at least one internal weight factor; In this process, a genetic algorithm was used to optimize the weight factors, with the population size set to 80 and the number of iterations set to 100. A polygon was constructed using the `polyshape` function, and the minimum difference area between the measured curve segment and the NURBS fitted curve segment was used as the objective value of the objective function, calculated according to the shoelace formula.
[0033] Please see Figure 4 The method acquires measured tread data points for sections BC and CD under different wear mileages, performs chord length parameterization based on the accumulated measured data points, sets the NURBS curve order to 3, and the number of control points to 5. It inputs the coordinates of the 5 control vertices obtained by the least squares method, calculates the node vectors of the curve using the Debussy-Cockell recursive formula, calculates the NURBS curve fitting value, calculates the target fitness of the difference area, and determines whether the iteration count meets the iteration conditions. If it does, a genetic algorithm is used to optimize the weight factors. The optimization process of the genetic algorithm includes roulette wheel selection function calculation, multiple crossover and recombination operator processing, and outputting the actual population mutation operator. Based on the iterative optimization results of the genetic algorithm, the weight factor combination is output.
[0034] The weight factors of the wheel tread over the entire wear cycle are solved using the polyshape function. The weight factors of the BC segment NURBS curve of the seven-car two-position wheel are shown in Table 2, and the weight factors of the CD segment NURBS curve are shown in Table 3. Table 2 Table 3 For the BC section, the wear area at the wheel flange increases with increasing operating mileage. Wear is relatively gradual within 100,000 kilometers, but increases significantly beyond 100,000 kilometers. (Internal weight factor) w 1. The wear width at the wheel flange is 0.87955 at an operating mileage of 200,000 kilometers, and 0.9999 at other mileages. This indicates that within an operating mileage of 185,000 kilometers, the wear width at the wheel flange of the high-speed train is within the control point P. 4-1 Within the horizontal coordinate range of -42.79036mm, when the operating mileage reaches 200,000 kilometers, the wear width at the wheel flange increases accordingly, exceeding -42.79036mm. (Internal weight factor) w 2 and w 3. As the operating mileage increases, the wear depth gradually increases, and the wear amount gradually increases, with the maximum wear amount at the rim being 6.734356 mm2.
[0035] The three internal weight factors of segment CD w 1. w 2 and w3. All of these increase with the increase of wear mileage. Observing the changes in weight factor and wear amount, it can be seen that the wear amount at the connection between the tread and the flange is relatively large within the first 100,000 kilometers of operation, and the wear amount is relatively average above 100,000 kilometers.
[0036] With increasing operating mileage, the wear depth and width of the DE section of the wheel tread increase significantly. Wear in this section is divided into three stages: light wear within 80,000 km, moderate wear within 80,000-163,000 km, and heavy wear within 163,000-200,000 km, with wear intervals in the ranges of [-45, -35], [-35, -20], and [-20, 20], respectively. The NURBS curve control points for the DE section of one wheel on eight vehicles are selected as shown in Table 5, and the corresponding weighting factor values are shown in Table 4. Table 4 As shown in Table 4, in the light wear stage, the wheel wear depth is within 0.08 mm, and the wear amount is within 2.1 mm². Among the four internal weight factors, w The change in 4 is quite significant, increasing with the increase in operating mileage, and the internal weight factor... w 1. w 2. w 5. No significant changes were observed; wear mainly occurred near the nominal rolling circle of the wheel, and the wear width did not change significantly. In the moderate wear stage, when the operating mileage exceeded 116,000 kilometers, the wheel wear rate accelerated significantly, the wear depth exceeded 0.1 mm, and the wear amount also increased significantly. w 2. w 4. w The wear factor 6 increases with increasing operating mileage; in the severe wear stage, the wear depth exceeds 0.2mm, and the wear rate increases significantly, with the wear amount exceeding 6mm², reaching a maximum of 13.01754mm². w 4. w 4. w All 6 increased significantly with the increase of operating mileage.
[0037] Taking an eight-car, one-wheel configuration as an example, the range of the horizontal coordinate [-20, 20] is controlled by two control points at the beginning and end. The selection of the five control vertices for severe wear is shown in Table 5. The range of moderate wear is [-15, 20]. Point P0 is fixed, and control points P1 and P2 are added. The horizontal coordinates of points P1 and P7 are -15 and 20, respectively, indicating the size of the wear width. The range of light wear is [-15, 15]. Control point P5 is added, with its horizontal coordinate being 15mm. P1 and P5 together indicate the size of the wear width.
[0038] Table 5 Step S40: Obtain the operating mileage of the train wheels to be tested, and determine the wheel wear value based on the operating mileage and the weight factor combination; Among them, wheel wear value includes wear amount and / or wear depth. During the service of a vehicle, the geometric shape of the wheel profile changes continuously. There is obvious wear at the wheel flange, at the junction of the flange and the tread, and within 20mm to the left and right of the nominal rolling circle of the wheel. Among these, the wear near the nominal rolling circle of the wheel is the most significant. The weight factor best represents the information of the wheel profile under different wear conditions, so the weight factor is used as the characteristic parameter of the wheel profile after wear.
[0039] As shown in Table 3, taking the wear section CD of the first wheel of an eight-car vehicle as an example, since the actual measured data sample size is small, the interpolation method is used to expand the data of the three internal weight factors of the CD section, and then the corresponding 50 sets of wear amounts are calculated, and a total of 50 sets of sample data are collected.
[0040] To establish the relationship between wear and weight factors, different fitting methods can be selected for analysis. These include multiple linear regression, ridge regression, support vector regression, Lasso regression, binomial regression, and polynomial chaos expansion (PCE). Ridge regression, Lasso regression, and support vector regression models show significant errors between the data points and the ideal fitting line, resulting in poor fitting performance. Multiple linear regression shows better fitting performance but still has some errors. Binomial regression and PCE models offer the best fitting performance; therefore, the binomial regression model is used for analysis.
[0041] For different types of train wheels, a parametric descriptive model (binomial regression model) is established to describe the relationship between the wear amount of wheel tread sections BC and CD at different operating mileages, and the wear amount and wear depth of wheel tread sections DE in the heavy wear, moderate wear, and light wear stages, and three internal weight factors. This results in the fitting coefficients between the wear amount of wheel tread sections BC and CD at different operating mileages, and the wear amount and wear depth of wheel tread sections DE in the heavy wear, moderate wear, and light wear stages, and the three internal weight factors, thus obtaining a relational query list.
[0042] Optionally, the wheel wear value is determined based on the combination of the operating mileage and the weighting factors. The wheel wear value includes the wear amount and / or wear depth, and the formula used includes: in, Indicates the amount or depth of wear. , , This represents the internal weight factor in the weight factor combination. Indicates the error term. , , , , , , , , , This represents the fitting coefficient corresponding to the operating mileage.
[0043] The first objective of this embodiment is to address the problem that existing wheel wear detection methods can only obtain discrete geometric parameters or wear depth at a single feature point, and cannot accurately represent the evolution law of the entire surface profile. This embodiment provides a parameterized description method for the wheel profile. By fitting the actual wheel profile with a NURBS curve and extracting the weight factor corresponding to the curve, a precise quantitative representation of the overall profile shape of the wheel after wear is achieved, providing high-precision surface data support for subsequent wear calculation.
[0044] The second objective of this embodiment is to overcome the shortcomings of existing technologies, such as one-sided measurement of wear depth and lack of scientific calculation methods for wear amount. It proposes a wheel wear amount calculation method based on the NURBS parameterized model. By comparing the NURBS profile curves and weight factor changes of the wheel before and after wear, the total material loss of the entire wheel area can be accurately solved, realizing a comprehensive and objective quantification of wear amount. This provides a reliable basis for wheel service life assessment, rewinding cycle optimization, and driving safety assurance.
[0045] In this embodiment, a high-precision, continuous, and smooth parametric model of the tread wear of a high-speed train wheel throughout its entire life cycle within one refinishing cycle is achieved, taking into account both the wear depth and wear width. Using a cubic NURBS curve, only 5-6 control points are needed to accurately fit the measured tread profile. While strictly ensuring geometric continuity and smoothness, the model simultaneously characterizes the dual features of wear depth and wear width in three key wear regions: the flange area (BC segment), the flange-tread transition area (CD segment), and the area near the rolling circle (DE segment). This overcomes the shortcomings of traditional discrete measurements or simple geometric fitting, which struggle to fully describe complex concave wear morphologies, providing a high-fidelity geometric foundation for the study of full-cycle wear evolution.
[0046] This innovative approach uses the weighting factors of NURBS curves as wear characteristic parameters to construct an automatic inversion calculation model for "tread profile → wear amount". The weighting factors under the shared control vertices of segments BC / CD and the weighting factors of the collinear control points of the rolling circles in segment DE are defined as characteristic parameters of the wear profile. A quantitative mapping relationship between these parameters and the measured wear amount (including wear depth and wear width) is established through quadratic polynomial fitting. This allows for the automatic, rapid, and quantitative calculation of the current wear state simply by extracting the corresponding NURBS weighting factors after obtaining the wheel tread profile at any given time, solving the industry pain point of existing technologies that "can measure morphology but cannot calculate wear".
[0047] The established parametric model exhibits excellent generalization ability and engineering applicability, providing a reliable basis for wear prediction and rewinding decisions. After being trained on the wheel data of an eight-car, one-position wheel, the model was successfully applied to the profile reconstruction of a seven-car, two-position wheel, with the error controlled within 0.15 mm. Simultaneously, through wheel-rail contact geometry verification, the fitted profile and the measured statistical profile are highly consistent at the contact point position, with a normal fitting error ≤0.05 mm. This fully demonstrates that the method not only has high accuracy but also possesses the potential for cross-vehicle and cross-position application, laying a technical foundation for wheel health management, rewinding cycle optimization, and the development of intelligent operation and maintenance systems.
[0048] In this embodiment, by calculating the measurement parameter values of the wheel tread measurement data, the control point coordinates and curve weight factors can be effectively determined based on the measurement parameter values. Based on the control point coordinates and curve weight factors, the wheel wear tread curve can be effectively constructed. By calculating the difference area between the wheel wear tread curve and the wheel tread measurement data, the curve weight factors can be effectively iteratively optimized based on the difference area to obtain the weight factor combination.
[0049] Example 2 Please see Figure 5 This is a schematic diagram of the structure of the wheel wear detection system 100 provided in the second embodiment of the present invention, including: The parameterization module 10 is used to acquire the wheel tread measurement data of the sample train and calculate the measurement parameter values of the wheel tread measurement data.
[0050] The tread curve construction module 11 is used to determine the control point coordinates and curve weight factors based on the measured parameter values, and to construct the tread curve after wheel wear based on the control point coordinates and the curve weight factors.
[0051] The weight factor optimization module 12 is used to calculate the difference area between the wear-out tread curve of the wheel and the wheel tread measurement data, and to iteratively optimize the weight factor of the curve based on the difference area to obtain the weight factor combination.
[0052] The wear output module 13 is used to acquire the operating mileage of the train wheel under test, and determine the wheel wear value based on the combination of the operating mileage and the weighting factor. The wheel wear value includes wear amount and / or wear depth.
[0053] Optionally, the tread curve construction module 11 is further configured to: construct a non-uniform rational B-spline curve based on the control point coordinates and the curve weight factor, and constrain the control point coordinates in the non-uniform rational B-spline curve according to the control point constraint conditions to obtain the tread curve after wheel wear. The control point constraint condition is that at the curve connection point of the tread curve after wheel wear, the coordinates of the left and right adjacent control points are collinear.
[0054] Furthermore, the weight factor optimization module 12 is also used to: use the curve weight factor as a variable, take the minimum difference area as the target value, and use a genetic algorithm to iteratively optimize the curve weight factor to obtain the weight factor combination, wherein the weight factor combination includes at least one internal weight factor.
[0055] In this embodiment, by calculating the measurement parameter values of the wheel tread measurement data, the control point coordinates and curve weight factors can be effectively determined based on the measurement parameter values. Based on the control point coordinates and curve weight factors, the wheel wear tread curve can be effectively constructed. By calculating the difference area between the wheel wear tread curve and the wheel tread measurement data, the curve weight factors can be effectively iteratively optimized based on the difference area to obtain a weight factor combination. By using the weight factor combination as wear characteristic parameters, a quantitative mapping relationship between wear characteristic parameters and wheel wear values is constructed, so that after obtaining the operating mileage of the train wheel to be tested, the corresponding wheel wear value can be automatically calculated, which can intuitively reflect the specific situation of wheel wear.
[0056] Example 3 Figure 6 This is a structural block diagram of a terminal device 2 provided in the third embodiment of this application. For example... Figure 6 As shown, the terminal device 2 in this embodiment includes a processor 20, a memory 21, and a computer program 22 stored in the memory 21 and executable on the processor 20, such as a program for a wheel wear detection method. When the processor 20 executes the computer program 22, it implements the steps in the various embodiments of the wheel wear detection methods described above.
[0057] For example, the computer program 22 may be divided into one or more modules, which are stored in the memory 21 and executed by the processor 20 to complete this application. The one or more modules may be a series of computer program instruction segments capable of performing specific functions, which describe the execution process of the computer program 22 in the terminal device 2. The terminal device may include, but is not limited to, the processor 20 and the memory 21.
[0058] The processor 20 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.
[0059] The memory 21 can be an internal storage unit of the terminal device 2, such as a hard drive or memory of the terminal device 2. The memory 21 can also be an external storage device of the terminal device 2, such as a plug-in hard drive, Smart Media Card (SMC), Secure Digital (SD) card, or Flash Card equipped on the terminal device 2. Furthermore, the memory 21 can include both internal and external storage units of the terminal device 2. The memory 21 is used to store the computer program and other programs and data required by the terminal device. The memory 21 can also be used to temporarily store data that has been output or will be output.
[0060] Furthermore, the functional modules in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.
[0061] If an integrated module is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. This computer-readable storage medium can be non-volatile or volatile. Based on this understanding, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable storage medium can include: any entity or device capable of carrying computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the contents of a computer-readable storage medium may be appropriately added to or subtracted from the contents as required by the legislation and patent practice in a jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, a computer-readable storage medium may not include electrical carrier signals and telecommunication signals.
[0062] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.
Claims
1. A method for detecting wheel wear, characterized in that, The method includes: Acquire wheel tread measurement data of the sample train and calculate the measurement parameter values of the wheel tread measurement data; The control point coordinates and curve weight factors are determined based on the measured parameter values, and the wheel wear tread curve is constructed based on the control point coordinates and the curve weight factors. Calculate the area of difference between the wear curve of the wheel and the measured data of the wheel tread, and iteratively optimize the weight factor of the curve based on the area of difference to obtain the weight factor combination; The operating mileage of the train wheels to be tested is obtained, and the wheel wear value is determined based on the combination of the operating mileage and the weighting factor. The wheel wear value includes wear amount and / or wear depth.
2. The wheel wear detection method as described in claim 1, characterized in that, The formulas used to calculate the measurement parameter values of the wheel tread measurement data include: ； in, This indicates the value of the measured parameter. This represents the ordered sequence of points in the wheel tread measurement data. This represents the forward difference vector.
3. The wheel wear detection method as described in claim 2, characterized in that, The formulas used to determine the control point coordinates and curve weight factors based on the measured parameter values include: ； in, Indicates the coordinates of the control point. Indicates the target point for fitting. This represents the curve weight factor corresponding to the control point. Describes the k-th degree B-spline basis function. This represents the j-th data point in the wheel tread measurement data, and n represents the total number of data points in the wheel tread measurement data.
4. The wheel wear detection method as described in claim 1, characterized in that, Constructing the wheel wear tread curve based on the control point coordinates and the curve weighting factor includes: A non-uniform rational B-spline curve is constructed based on the control point coordinates and the curve weight factor. The control point coordinates are constrained in the non-uniform rational B-spline curve according to the control point constraint conditions to obtain the tread curve after wheel wear. The control point constraint condition is that at the curve connection point of the tread curve after wheel wear, the coordinates of the left and right adjacent control points are collinear.
5. The wheel wear detection method as described in claim 1, characterized in that, The formula used to calculate the area of difference between the worn tread curve and the measured tread data of the wheel includes: ； in, This represents the area of difference in the region of difference. This represents the number of measured scatter points in the wheel tread wear section of the wheel tread measurement data. This indicates the number of scatter points in the region of difference. This represents the x-coordinate of the i-th vertex in the region of difference. This represents the x-coordinate of the (i+1)th vertex in the region of difference. This represents the ordinate of the i-th vertex in the region of difference. This represents the ordinate of the (i+1)th vertex in the region of difference. Indicates the first difference region The x-coordinates of the vertices, Indicates the first difference region The y-coordinates of the vertices, This represents the x-coordinate of the first vertex in the region of difference. This represents the ordinate of the first vertex in the region of difference.
6. The wheel wear detection method as described in claim 1, characterized in that, The curve weight factors are iteratively optimized based on the area of difference to obtain a combination of weight factors, including: Using the curve weight factors as variables and the minimum difference area as the target value, a genetic algorithm is used to iteratively optimize the curve weight factors to obtain the weight factor combination, wherein the weight factor combination includes at least one internal weight factor.
7. The wheel wear detection method as described in claim 6, characterized in that, The wheel wear value is determined based on the operating mileage and the weighting factor combination. The wheel wear value includes the wear amount and / or wear depth, and the formula used includes: ； in, Indicates the amount or depth of wear. , , This represents the internal weight factor in the weight factor combination. Indicates the error term. , , , , , , , , , This represents the fitting coefficient corresponding to the operating mileage.
8. A wheel wear detection system, characterized in that, The system includes: The parameterization module is used to acquire wheel tread measurement data of the sample train and calculate the measurement parameter values of the wheel tread measurement data. The tread curve construction module is used to determine the control point coordinates and curve weight factors based on the measured parameter values, and to construct the tread curve after wheel wear based on the control point coordinates and the curve weight factors. The weight factor optimization module is used to calculate the area of difference between the wear-out tread curve and the measured data of the wheel tread, and to iteratively optimize the weight factor of the curve based on the area of difference to obtain a weight factor combination. The wear output module is used to acquire the operating mileage of the train wheels under test, and determine the wheel wear value based on the combination of the operating mileage and the weighting factor. The wheel wear value includes wear amount and / or wear depth.
9. The wheel wear detection system as described in claim 8, characterized in that, The tread curve construction module is also used for: A non-uniform rational B-spline curve is constructed based on the control point coordinates and the curve weight factor. The control point coordinates are constrained in the non-uniform rational B-spline curve according to the control point constraint conditions to obtain the tread curve after wheel wear. The control point constraint condition is that at the curve connection point of the tread curve after wheel wear, the coordinates of the left and right adjacent control points are collinear.
10. The wheel wear detection system as described in claim 8, characterized in that, The weight factor optimization module is also used for: Using the curve weight factors as variables and the minimum difference area as the target value, a genetic algorithm is used to iteratively optimize the curve weight factors to obtain the weight factor combination, wherein the weight factor combination includes at least one internal weight factor.