Railway wheel life prediction method and system based on sample expansion

Abstract

The invention discloses a railway wheel life prediction method and system based on sample expansion, and relates to the technical field of railway wheel life prediction, and the method comprises the steps: collecting a railway wheel wear data set of a railway line within a preset age limit; solving according to the railway wheel abrasion data set transverse and longitudinal expansion data set model; constructing a railway wheel life prediction model; and making a maintenance plan in combination with the prediction result and other railway wheel fault diagnosis information. According to the railway wheel service life prediction method based on sample expansion provided by the invention, by collecting railway wheel wear data within a certain age limit on a railway line, a transverse expansion method and a longitudinal expansion method are adopted, and technologies of data synthesis, data enhancement, sample copying and the like are utilized, so that the data sample size and diversity are increased. The data sample size and diversity are increased through data expansion, so that the prediction model has better generalization ability and can better adapt to data under various operation conditions, and the prediction result is more reliable and stable.

Description

Technical Field

[0001] The present invention relates to the technical field of railway wheel life prediction, and in particular to a railway wheel life prediction method and system based on sample expansion. Background Art

[0002] During railway operation, due to the limitations of detection conditions and the interval between maintenance cycles, the amount of data on railway wheel wear actually measured is relatively small. This makes it difficult to predict the life of railway wheels based on actual measurement data, and it is impossible to cover various operating conditions and time periods, which affects the reliability of the prediction model. In order to overcome the problem of insufficient actual measurement data, it is necessary to expand the existing data to increase the data volume and sample diversity. Through data expansion, the wear of railway wheels under different operating conditions can be better covered, and the generalization ability and accuracy of the prediction model can be improved.

[0003] Data-based prediction has many advantages over model-based prediction. First, data prediction relies on actual observation data. Through the analysis and learning of a large amount of historical data, it can capture the actual performance under various operating conditions. Model prediction is usually based on preset physical or mathematical models, which may not fully reflect the actual complex operating environment. Secondly, data prediction can continuously optimize and update the prediction model through machine learning and other methods. As the amount of data increases, the accuracy of the prediction will continue to improve. Traditional model prediction is often difficult to dynamically adjust and optimize after it is established. Finally, data prediction can better handle the diversity and complexity of data. Through big data analysis, complex patterns and trends hidden in the data can be identified, thereby improving the accuracy and reliability of the prediction. In summary, the application of data-based prediction methods in railway wheel life prediction can more comprehensively and accurately reflect the actual operating conditions and provide more reliable decision support for railway operations. Summary of the invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by the present invention is: the existing prediction methods based on actual measurement data have the problem of small amount of actual data, low accuracy, and how to enhance the generalization ability to cover various operating conditions and time periods.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: a railway wheel life prediction method based on sample expansion, comprising collecting a railway wheel wear data set within a preset period of time of a railway line. A data set model is expanded horizontally and vertically according to the railway wheel wear data set to solve. A railway wheel life prediction model is constructed. A maintenance plan is formulated by combining the prediction results with other railway wheel fault diagnosis information.

[0007] As a preferred solution of the railway wheel life prediction method based on sample expansion described in the present invention, the railway wheel wear data set collected within a preset number of years of the railway line includes collecting the running mileage of railway vehicles, railway wheel diameter values, circumferential wear, rim thickness and rim height.

[0008] As a preferred solution of the railway wheel life prediction method based on sample expansion described in the present invention, wherein: the solution of the model of the horizontal and vertical expansion data set of the railway wheel wear data set includes solving the model of the horizontal expansion data set of the railway wheel wear data set, inputting the detection data collected by the vehicle type to be modeled, extracting the detection data of the same car type, for each feature, using a polynomial function to fit the existing data points, assuming a feature x and a corresponding observation value y, using an n-order polynomial to fit these data points, expressed as:

[0009] f(x)=a0+a1x+a2x 2 +…+a n x n

[0010] Among them, a0, a1, a2…a n are the coefficients of the polynomial and n is the order of the polynomial.

[0011] New data points are inserted between existing data points. The data points are evenly selected within the feature range. For each interpolation point, the corresponding eigenvalue is calculated using the fitted polynomial function. The x value of the interpolation point is substituted into the polynomial function to obtain the y value corresponding to the interpolation point.

[0012] The polynomial interpolation steps are repeated for each feature to generate corresponding new data points.

[0013] Use statistical methods to validate the newly generated data points,

[0014] The mean and variance between the newly generated data points and the original data points are calculated, and the change in the mean is controlled within ±5%, and the change in the variance is controlled within ±10%.

[0015] Conduct hypothesis testing using a t-test or z-test to ensure that the new data points are not statistically significantly different from the original data points. The test result requires a p-value greater than 0.05, indicating that at a 95% confidence level, the new data points are not significantly different from the original data points.

[0016] Use confidence intervals to assess whether the new data points fall within ±1.96 times the standard error of the mean of the original data points.

[0017] By drawing a data distribution graph to compare the distribution characteristics of the new data points and the original data points, we ensure that the new data points conform to the distribution characteristics of the original data points. The distribution shape of the new data points is required to be highly consistent with the distribution shape of the original data points, and the changes in skewness and kurtosis should be within ±0.1.

[0018] If the verification is successful, the new data points that have passed the verification are merged with the original data set to form an expanded data set. If the verification fails, the parameters of the polynomial are adjusted, the data points are regenerated, and the polynomial function is refitted and interpolated for each feature to generate new data points until the verification passes.

[0019] As a preferred solution of the railway wheel life prediction method based on sample expansion described in the present invention, the method of solving the model of the horizontally and vertically expanded data sets of the railway wheel wear data set also includes solving the model of the vertically expanded data set of the railway wheel wear data set, using the horizontally expanded data set as input, fitting the multivariate Gaussian distribution model, and estimating the parameters of the model by maximum likelihood estimation, including the mean vector and the covariance matrix, and the probability density function of the multivariate Gaussian distribution is:

[0020]

[0021] where x is a d-dimensional feature vector, u is the mean vector, and Σ is the covariance matrix.

[0022] Based on the fitted multivariate Gaussian distribution model, new data points are generated using the model. For each eigenvector, new data points are randomly sampled from the corresponding multivariate Gaussian distribution, and then new data points are generated according to the required amount of data.

[0023] After generating new data, perform the same data validation as horizontal expansion to ensure its rationality and accuracy. Calculate the mean and variance between the newly generated data points and the original data points. The change in the mean is controlled within ±5% of the original data mean, and the change in the variance is controlled within ±10%.

[0024] Conduct hypothesis testing and use t-test or z-test to ensure that there is no statistically significant difference between the new data points and the original data points. Set the significance level to 0.05. When the p-value between the new data points and the original data points is greater than 0.05, it means that there is no statistically significant difference between the two.

[0025] Use confidence intervals to assess whether a new data point falls within the 95% confidence interval of the original data mean, ensuring that the statistical properties of the new data point are as expected.

[0026] By drawing a data distribution graph to compare the distribution characteristics of the new data points and the original data points, we can ensure that the new data points conform to the distribution characteristics of the original data points, and the distribution shape of the new data points is highly consistent with the distribution shape of the original data points, so that the changes in skewness and kurtosis should be within ±0.1.

[0027] If the generated data deviates from the actual situation, the parameters of the multivariate Gaussian distribution model are adjusted. After verification, the generated new data are merged with the existing data set to obtain the expanded railway wheel wear data set.

[0028] As a preferred solution of the railway wheel life prediction method based on sample expansion described in the present invention, wherein: the construction of the railway wheel life prediction model includes taking the characteristics of the railway wheel diameter value, circumferential wear, rim thickness and rim height as input features, and the running mileage as the output label, and dividing the data set into a training set and a test set.

[0029] The loss function of ridge regression is expressed as:

[0030]

[0031] Among them, J(θ) is the loss function of ridge regression, h θ (x i ) is the model’s predicted value for the i-th sample, y i is the true label value, λ is the regularization parameter, which controls the regularization strength, and θ j are the parameters of the model, n is the number of features, and m is the number of samples.

[0032] As a preferred solution of the railway wheel life prediction method based on sample expansion described in the present invention, wherein: the construction of the railway wheel life prediction model also includes training the ridge regression model using the training set, adjusting the parameters to minimize the loss function, and evaluating the trained ridge regression model using the test set, and the evaluation indicators include mean square error and root mean square error. According to the evaluation results, the model is tuned, the regularization parameter λ is adjusted, or cross-validation is performed to select the best regularization parameter.

[0033] As a preferred solution of the railway wheel life prediction method based on sample expansion described in the present invention, the combined prediction results and other railway wheel fault diagnosis information are used to formulate a maintenance plan, including using a trained ridge regression model to predict the railway wheel data to be predicted to obtain the predicted running mileage.

[0034] The predicted running mileage is compared with the actual running mileage to calculate the remaining life information of the railway wheels.

[0035] Assuming that the life of the railway wheel is proportional to the mileage traveled, the remaining life is calculated based on the difference between the predicted mileage and the actual mileage. If the predicted mileage is less than the actual mileage, it means that the remaining life of the railway wheel is a negative value.

[0036] Another object of the present invention is to provide a railway wheel life prediction system based on sample expansion, which can increase the data sample volume and diversity through data expansion, so that the prediction model has better generalization ability, and the model can better adapt to data under various operating conditions, so that the prediction result is more reliable and stable. The problem that the current railway wheel life prediction technology requires a large amount of data is solved.

[0037] As a preferred solution of the railway wheel life prediction system based on sample expansion described in the present invention, it includes: a data acquisition module, a data expansion module, and a life prediction module. The data acquisition module is used to collect the railway wheel wear data set within a preset period of time of the railway line. The data expansion module is used to solve the data set model based on the horizontal and vertical expansion of the railway wheel wear data set. The life prediction module is used to construct a railway wheel life prediction model, and formulate a maintenance plan by combining the prediction results with other railway wheel fault diagnosis information.

[0038] A computer device comprises a memory and a processor. The memory stores a computer program. The processor executes the computer program to implement the steps of a railway wheel life prediction method based on sample expansion.

[0039] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of a railway wheel life prediction method based on sample expansion.

[0040] Beneficial effects of the present invention: The railway wheel life prediction method based on sample expansion provided by the present invention collects railway wheel wear data within a certain period of time on a railway line, adopts the method of horizontal expansion and vertical expansion, and uses synthetic data, data enhancement and sample replication and other technologies to increase the data sample size and diversity. By increasing the data sample size and diversity through data expansion, the prediction model has better generalization ability, and the model can better adapt to data under various operating conditions, so the prediction result is more reliable and stable. The present invention achieves better results in terms of stability, accuracy and data demand. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0042] Figure 1 An overall flow chart of a railway wheel life prediction method based on sample expansion provided in the first embodiment of the present invention.

[0043] Figure 2 A lateral data expansion flow chart of a railway wheel life prediction method based on sample expansion is provided for the first embodiment of the present invention.

[0044] Figure 3 A longitudinal data expansion flow chart of a railway wheel life prediction method based on sample expansion provided in the first embodiment of the present invention. DETAILED DESCRIPTION

[0045] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0046] Example 1

[0047] Reference Figure 1-Figure 3 , which is an embodiment of the present invention, provides a railway wheel life prediction method based on sample expansion, comprising:

[0048] S1: Collect railway wheel wear datasets within a preset period of time on railway lines.

[0049] Furthermore, collecting the railway wheel wear data set within a preset period of time of the railway line includes collecting the running mileage of the railway vehicles, the wheel diameter value of the railway wheels, the circumferential wear, the flange thickness and the flange height.

[0050] S2: Expand the dataset model horizontally and vertically based on the railway wheel wear dataset.

[0051] Further, refer to Figure 2For horizontal expansion, the model of the horizontal and vertical expansion of the railway wheel wear dataset is solved. The model of the horizontal expansion of the railway wheel wear dataset is solved, the test data collected by the vehicle type to be modeled is input, and the test data of the same car type is extracted. For each feature, a polynomial function is used to fit the existing data points. Suppose a feature x and the corresponding observation value y, and an n-order polynomial is used to fit these data points, which is expressed as:

[0052] f(x)=a0+a1x+a2x 2 +…+a n x n

[0053] Among them, a0, a1, a2…a n are the coefficients of the polynomial and n is the order of the polynomial.

[0054] New data points are inserted between existing data points. The data points are evenly selected within the feature range. For each interpolation point, the corresponding eigenvalue is calculated using the fitted polynomial function. The x value of the interpolation point is substituted into the polynomial function to obtain the y value corresponding to the interpolation point.

[0055] The polynomial interpolation steps are repeated for each feature to generate corresponding new data points.

[0056] Use statistical methods to verify the newly generated data points to ensure the rationality and accuracy of the data points, including:

[0057] Calculate the mean and variance between the newly generated data points and the original data points to ensure that they are within a reasonable range.

[0058] Perform a hypothesis test, such as a t-test or z-test, to ensure that the new data points are not statistically significantly different from the original data points.

[0059] Use confidence intervals to assess whether new data points fall within the expected range.

[0060] Perform comparative analysis on data distribution to ensure that new data points conform to the distribution characteristics of the original data points.

[0061] If the verification is successful, the new data points that have passed the verification are merged with the original data set to form an expanded data set. If the verification fails, the parameters of the polynomial are adjusted, the data points are regenerated, and the polynomial function is refitted and interpolated for each feature to generate new data points until the verification passes.

[0062] It should be noted that, referring to Figure 3For vertical expansion, the multivariate Gaussian distribution model is a probabilistic model used to model multidimensional data, which can represent the complex relationship between multiple features. The model solution of the railway wheel wear data set is expanded horizontally and vertically, and the model solution of the railway wheel wear data set is also included. The horizontally expanded data set is used as input to fit the multivariate Gaussian distribution model, and the parameters of the model are estimated by maximum likelihood estimation, including the mean vector and covariance matrix. The probability density function of the multivariate Gaussian distribution is:

[0063]

[0064] where x is a d-dimensional feature vector, u is the mean vector, and Σ is the covariance matrix.

[0065] Based on the fitted multivariate Gaussian distribution model, new data points are generated using the model. For each eigenvector, new data points are randomly sampled from the corresponding multivariate Gaussian distribution, and then new data points are generated according to the required amount of data.

[0066] After generating new data, perform the same data validation as horizontal expansion to ensure its rationality and accuracy, including:

[0067] Calculate the mean and variance between the newly generated data points and the original data points to ensure that they are within a reasonable range.

[0068] Perform a hypothesis test, such as a t-test or z-test, to ensure that the new data points are not statistically significantly different from the original data points.

[0069] Use confidence intervals to assess whether new data points fall within the expected range.

[0070] Perform comparative analysis on data distribution to ensure that new data points conform to the distribution characteristics of the original data points.

[0071] If the generated data deviates from the actual situation, the parameters of the multivariate Gaussian distribution model are adjusted. After the verification is completed, the generated new data are merged with the existing data set to obtain the expanded railway wheel wear data set. It should also be noted that first, by collecting railway wheel wear data on a certain railway line within a certain period of time, and sorting and summarizing the data, a complete railway wheel wear data set is established. Subsequently, the method of horizontal expansion and vertical expansion is adopted, and the data sample size and diversity are increased by using technologies such as synthetic data, data enhancement and sample replication. Not only does it make full use of the existing railway wheel wear data, but it also enables the prediction model to better cover the railway wheel wear conditions under various operating conditions and time periods, thereby improving the accuracy and reliability of the prediction.

[0072] S3: Construct a railway wheel life prediction model.

[0073] Furthermore, ridge regression is an extension of linear regression that prevents overfitting by penalizing the coefficients. It adds a regularization term to the loss function, which helps deal with multicollinearity and noise issues. Building a railway wheel life prediction model includes taking the railway wheel diameter value, circumferential wear, rim thickness, and rim height as input features, running kilometres as output labels, and dividing the data set into a training set and a test set.

[0074] The loss function of ridge regression is expressed as:

[0075]

[0076] Among them, J(θ) is the loss function of ridge regression, h θ (x i ) is the model’s predicted value for the i-th sample, y i is the true label value, λ is the regularization parameter, which controls the regularization strength, and θ j are the parameters of the model, n is the number of features, and m is the number of samples.

[0077] It should be noted that building a railway wheel life prediction model also includes training the ridge regression model using the training set, adjusting the parameters to minimize the loss function, and evaluating the trained ridge regression model using the test set. The evaluation indicators include mean square error and root mean square error. According to the evaluation results, the model is tuned, the regularization parameter λ is adjusted, or cross-validation is performed to select the best regularization parameter.

[0078] S4: Combine the prediction results with other railway wheel fault diagnosis information to develop a maintenance plan.

[0079] Furthermore, the prediction results are combined with other railway wheel fault diagnosis information to formulate a maintenance plan, including using the trained ridge regression model to predict the railway wheel data to obtain the predicted running mileage.

[0080] The predicted running mileage is compared with the actual running mileage to calculate the remaining life information of the railway wheels.

[0081] Assuming that the life of the railway wheel is proportional to the mileage traveled, the remaining life is calculated based on the difference between the predicted mileage and the actual mileage. If the predicted mileage is less than the actual mileage, it means that the remaining life of the railway wheel is a negative value.

[0082] Example 2

[0083] An embodiment 2 of the present invention provides a railway wheel life prediction method based on sample expansion. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0084] First, railway wheel wear data from the past five years were collected from a specific railway line, including key features such as running mileage, wheel diameter, circumferential wear, wheel rim thickness and wheel rim height. Subsequently, the data were expanded horizontally and vertically using the method of the present invention to generate a larger data set. Based on the expanded data set, a wheel life prediction model based on the ridge regression model was established, and the expanded data set was used for training and verification.

[0085] Referring to Table 1, it can be seen from the economic benefit analysis that compared with the traditional regular maintenance strategy, the intelligent maintenance plan predicted and formulated by the method of the present invention significantly reduces the maintenance cost. Specifically, it can be observed that the frequency of wheel replacement is reduced by about 20%, the total maintenance cost is reduced by about 15%, and the operational reliability is improved by about 10%. These results clearly show that the method of the present invention not only effectively extends the service life of railway wheels, but also significantly reduces related maintenance costs, thereby improving transportation efficiency and economic benefits. The following are specific economic benefit analysis results:

[0086] Table 1 Experimental data table

[0087] project Traditional methods Method of the present invention change(%) Replacement frequency (times / year) 10 8 -20％ Total maintenance cost (10,000 yuan) 500 425 -15％ Operation reliability (%) 90 99 +10％

[0088] In terms of simulation experiments, the effectiveness and reliability of the method of the present invention are further verified. Experiments conducted on python and Simulink platforms show that reference table 2 is a comparative test table.

[0089] Table 2 Comparative test table

[0090] Evaluation Metrics Traditional methods Method of the present invention change(%) Mean Square Error (MSE) 0.025 0.015 -40％ Root mean square error (RMSE) 0.158 0.122 -23％ Prediction accuracy (%) 78 92 +18％ Recall rate (%) 75 90 +20％

[0091] Specifically, the traditional method and the method of the present invention were trained and tested using the original data set, and their respective evaluation indicators were calculated, including mean square error (MSE), root mean square error (RMSE), prediction accuracy and recall. The experimental results show that the MSE and RMSE of the method of the present invention are reduced by about 40% and 23%, respectively, while the prediction accuracy and recall are also significantly improved. These data verify the effectiveness of the method of the present invention in improving prediction accuracy and reliability, and provide a scientific basis and practical solution for the optimization of railway transportation systems.

[0092] In addition, the application of the method of the present invention also helps to improve the overall safety of the railway transportation system. By accurately predicting the life of the wheel and formulating a maintenance plan in a timely manner, the safety hazards caused by wheel failure can be effectively reduced, ensuring the smoothness of railway transportation and the safety of passengers. Therefore, the present invention not only has significant advantages in economic benefits, but also makes important contributions to improving the safety and reliability of the transportation system.

[0093] In summary, Example 2 of the present invention has not only been verified in theory through economic benefit analysis and simulation experiments, but also has demonstrated significant advantages in practical applications, providing innovative solutions and scientific basis for the development and management of the railway transportation industry.

[0094] Example 3

[0095] An embodiment of the present invention provides a railway wheel life prediction system based on sample expansion, including a data acquisition module, a data expansion module, and a life prediction module.

[0096] The data acquisition module is used to collect the railway wheel wear data set within the preset years of the railway line. The data expansion module is used to expand the data set model horizontally and vertically according to the railway wheel wear data set. The life prediction module is used to build a railway wheel life prediction model, combine the prediction results with other railway wheel fault diagnosis information, and formulate a maintenance plan.

[0097] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0098] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0099] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering, or processing in another suitable manner as necessary, and then stored in a computer memory.

[0100] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc. It should be noted that the above embodiments are only used to illustrate the technical solution of the present invention and are not limited. Although the present invention is described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention, which should be included in the scope of the claims of the present invention.

[0101] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A railway wheel life prediction method based on sample expansion, characterized in that: include: Collect railway wheel wear data sets within a preset period of time on the railway line; The model solution is obtained by expanding the data set horizontally and vertically based on the railway wheel wear data set; Construct a railway wheel life prediction model; The prediction results are combined with other railway wheel fault diagnosis information to formulate a maintenance plan.

2. The railway wheel life prediction method based on sample expansion according to claim 1 is characterized in that: The collecting of the railway wheel wear data set within a preset number of years of the railway line includes collecting the running mileage of railway vehicles, the wheel diameter value, the circumferential wear, the wheel flange thickness and the wheel flange height.

3. The railway wheel life prediction method based on sample expansion according to claim 2 is characterized in that: The method of solving the model of the horizontally and vertically expanded data set according to the railway wheel wear data set includes solving the model of the horizontally expanded data set of the railway wheel wear data set, inputting the detection data collected by the vehicle type to be modeled, extracting the detection data of the same vehicle type, and fitting the existing data points with a polynomial function for each feature. Assuming a feature x and a corresponding observation value y, an n-order polynomial is used to fit these data points, which is expressed as: f(x)=a0+a1x+a2x 2 +…+a n x n Among them, a0, a1, a2…a n are the coefficients of the polynomial, and n is the order of the polynomial; Insert new data points between existing data points. The data points are evenly selected within the feature range. For each interpolation point, the corresponding eigenvalue is calculated using the fitted polynomial function. The x value of the interpolation point is substituted into the polynomial function to obtain the y value corresponding to the interpolation point. Repeat the polynomial interpolation steps for each feature to generate corresponding new data points; Use statistical methods to validate the newly generated data points, Calculate the mean and variance between the newly generated data points and the original data points. The change of the mean is controlled within ±5%, and the change of the variance is controlled within ±10%. Conduct hypothesis testing, using t-test or z-test, to ensure that there is no statistically significant difference between the new data points and the original data points. The test result requires a p-value greater than 0.05, indicating that at a 95% confidence level, there is no significant difference between the new data points and the original data points. Use confidence intervals to assess whether new data points fall within ±1.96 times the standard error of the mean of the original data points; By drawing a data distribution graph to compare the distribution characteristics of the new data points and the original data points, ensure that the new data points conform to the distribution characteristics of the original data points. The distribution shape of the new data points is required to be highly consistent with the distribution shape of the original data points, and the changes in skewness and kurtosis should be within ±0.1; If the verification is successful, the new data points that have passed the verification are merged with the original data set to form an expanded data set. If the verification fails, the parameters of the polynomial are adjusted, the data points are regenerated, and the polynomial function is refitted and interpolated for each feature to generate new data points until the verification passes.

4. The railway wheel life prediction method based on sample expansion according to claim 3 is characterized in that: The method of solving the model of the horizontally and vertically expanded data sets of the railway wheel wear data set also includes solving the model of the vertically expanded data set of the railway wheel wear data set, using the horizontally expanded data set as input, fitting the multivariate Gaussian distribution model, and estimating the parameters of the model by maximum likelihood estimation, including the mean vector and the covariance matrix. The probability density function of the multivariate Gaussian distribution is: Where x is a d-dimensional feature vector, u is the mean vector, and Σ is the covariance matrix; Based on the fitted multivariate Gaussian distribution model, new data points are generated using the model. For each eigenvector, new data points are randomly sampled from the corresponding multivariate Gaussian distribution, and then new data points are generated according to the required data volume. After generating new data, perform the same data validation as horizontal expansion to ensure its rationality and accuracy. Calculate the mean and variance between the newly generated data points and the original data points. The change in the mean is controlled within ±5% of the original data mean, and the change in the variance is controlled within ±10%; Conduct hypothesis testing, using t-test or z-test to ensure that there is no statistically significant difference between the new data points and the original data points. Set the significance level to 0.

05. When the p-value between the new data points and the original data points is greater than 0.05, it means that there is no statistically significant difference between the two. Use confidence intervals to assess whether new data points fall within the 95% confidence interval of the original data mean to ensure that the statistical characteristics of the new data points meet expectations; By drawing a data distribution graph to compare the distribution characteristics of the new data points and the original data points, ensure that the new data points conform to the distribution characteristics of the original data points, and the distribution shape of the new data points is highly consistent with the distribution shape of the original data points, so that the changes in skewness and kurtosis should be within ±0.1; If the generated data deviates from the actual situation, the parameters of the multivariate Gaussian distribution model are adjusted. After verification, the generated new data are merged with the existing data set to obtain the expanded railway wheel wear data set.

5. The railway wheel life prediction method based on sample expansion according to claim 4 is characterized in that: The construction of the railway wheel life prediction model includes taking the railway wheel diameter value, circumferential wear, rim thickness and rim height as input features, the running mileage as output labels, and dividing the data set into a training set and a test set; The loss function of ridge regression is expressed as: Among them, J(θ) is the loss function of ridge regression, h θ (x i ) is the model’s predicted value for the i-th sample, y i is the true label value, λ is the regularization parameter, which controls the regularization strength, and θ j are the parameters of the model, n is the number of features, and m is the number of samples.

6. The railway wheel life prediction method based on sample expansion according to claim 5 is characterized in that: The construction of the railway wheel life prediction model also includes training the ridge regression model using the training set, adjusting the parameters to minimize the loss function, and evaluating the trained ridge regression model using the test set, wherein the evaluation indicators include mean square error and root mean square error. According to the evaluation results, the model is tuned, the regularization parameter λ is adjusted, or cross-validation is performed to select the best regularization parameter.

7. The railway wheel life prediction method based on sample expansion according to claim 6 is characterized in that: The combined prediction results and other railway wheel fault diagnosis information are used to formulate a maintenance plan, including using a trained ridge regression model to predict the railway wheel data to be predicted to obtain a predicted running mileage; Compare the predicted running mileage with the actual running mileage to calculate the remaining life information of the railway wheels; Assuming that the life of the railway wheel is proportional to the mileage traveled, the remaining life is calculated based on the difference between the predicted mileage and the actual mileage. If the predicted mileage is less than the actual mileage, it means that the remaining life of the railway wheel is a negative value.

8. A system using the railway wheel life prediction method based on sample expansion as claimed in any one of claims 1 to 7, characterized in that: Including data acquisition module, data expansion module, life prediction module; The data acquisition module is used to collect railway wheel wear data sets within a preset period of time on the railway line; The data expansion module is used to expand the data set model horizontally and vertically according to the railway wheel wear data set to solve; The life prediction module is used to construct a railway wheel life prediction model, and to formulate a maintenance plan by combining the prediction results with other railway wheel fault diagnosis information.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the railway wheel life prediction method based on sample expansion according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the railway wheel life prediction method based on sample expansion according to any one of claims 1 to 7 are implemented.

Images