Classification method and device for magnetic flux leakage signals in rail magnetic flux leakage detection

Abstract

The present invention discloses a method and device for classifying magnetic flux leakage signals in rail magnetic flux leakage detection, wherein the method comprises: converting the magnetic flux leakage signal of the detected rail into a magnetic flux leakage electric signal; the magnetic flux leakage signal is obtained by picking up the magnetic field signal of the magnetized rail being detected; extracting multiple time-domain eigenvalues of the magnetic flux leakage electric signal; using a support vector machine classifier to classify the magnetic flux leakage signal corresponding to the magnetic flux leakage electric signal based on the multiple time-domain eigenvalues of the magnetic flux leakage electric signal, thereby obtaining a classification result of the magnetic flux leakage signal; the support vector machine classifier is obtained by training the historical data of the time-domain eigenvalues of the magnetic flux leakage electric signal using a support vector machine. The present invention can improve the accuracy of rail magnetic flux leakage detection, reduce the false alarm rate of rail magnetic flux leakage detection, greatly improve the efficiency of railway inspection and maintenance, and better ensure the safety of railway lines.

Description

Technical Field

[0001] The present invention relates to the field of detection technology, in particular to flaw detection technology for in-service rails, and more particularly to a classification method and device for magnetic flux leakage signals in rail magnetic flux leakage detection. Background Art

[0002] This section is intended to provide a background or context to the embodiments of the invention that are recited in the claims. No statement herein is admitted to be prior art by virtue of its inclusion in this section.

[0003] In-service rail flaw detection is an important measure to ensure the safe operation of railways. my country generally uses high-speed rail flaw detection vehicles with a running speed of 80km / h to perform flaw detection on rails. The ultrasonic detection equipment carried by the flaw detection vehicle can only detect internal scratches on the rails, and there are blind spots on the rail surface.

[0004] By utilizing the characteristic of magnetic flux leakage detection technology that is sensitive to the surface and subsurface conditions of ferromagnetic materials, a magnetic flux leakage detection system can be installed on rail flaw detection vehicles to make up for the lack of ultrasonic testing equipment in detecting surface scratches. Magnetic flux leakage detection technology is a new non-destructive testing technology that magnetizes the rails and then uses magnetic sensors to measure the changes in the leakage magnetic field caused by defects. Figure 1 As shown in the figure, if the magnetized rail material is continuous and uniform, the magnetic flux lines in the rail are mostly confined to the rail interior. Therefore, no magnetic flux lines will pass through the rail surface or enter the workpiece from the rail surface, thus preventing the formation of a leakage magnetic field on the rail surface. However, if there are scratches on the rail surface that can cut the magnetic flux lines, the scratches will have a lower magnetic permeability and a higher magnetic resistance, thus changing the path of the magnetic flux lines. Some magnetic flux lines will overflow the rail through the defect, pass over the defect, and then enter the rail interior. This leakage of magnetic flux lines will lead to the formation of a leakage magnetic field.

[0005] Currently, rail abrasion detection is mostly performed using a rail magnetic flux leakage (MFL) detection system. This system can detect fixed markers such as rail joints and welds, as well as surface abrasions. However, the MFL signals from abrasions and welds are similar in amplitude and shape, making them difficult to distinguish. During the defect detection process, many weld signals are misidentified as abrasions, significantly impacting the accuracy of MFL detection. Summary of the Invention

[0006] The present invention provides a method for classifying magnetic flux leakage signals in rail magnetic flux leakage detection, which is used to improve the accuracy of rail magnetic flux leakage detection, reduce the false alarm rate of rail magnetic flux leakage detection, greatly improve the efficiency of railway inspection and maintenance, and better ensure railway line safety. The method includes:

[0007] Converting the magnetic flux leakage signal of the rail under test into a magnetic flux leakage electrical signal; the magnetic flux leakage signal is obtained by picking up the magnetic field signal of the magnetized rail under test;

[0008] Extract multiple time domain eigenvalues of leakage magnetic electric signal;

[0009] According to the multiple time-domain eigenvalues of the leakage magnetic electric signal, a support vector machine classifier is used to classify the leakage magnetic signal corresponding to the leakage magnetic electric signal to obtain a classification result of the leakage magnetic signal; the support vector machine classifier is obtained by using a support vector machine to train historical data of the time-domain eigenvalues of the leakage magnetic electric signal.

[0010] The present invention also provides a device for classifying magnetic flux leakage signals in rail magnetic flux leakage detection, which is used to improve the accuracy of rail magnetic flux leakage detection, reduce the false alarm rate of rail magnetic flux leakage detection, greatly improve the efficiency of railway inspection and maintenance, and better ensure railway line safety. The device includes:

[0011] A magnetic leakage electric signal conversion module is used to convert the acquired magnetic leakage signal of the measured rail into a magnetic leakage electric signal; the magnetic leakage signal is obtained by picking up the magnetic field signal of the magnetized measured rail;

[0012] Time domain eigenvalue extraction module, used to extract multiple time domain eigenvalues of leakage magnetic electric signals;

[0013] The magnetic flux leakage signal classification module is used to classify the magnetic flux leakage signal corresponding to the magnetic flux leakage electric signal using a support vector machine classifier based on multiple time domain eigenvalues of the magnetic flux leakage electric signal to obtain a classification result of the magnetic flux leakage signal; the support vector machine classifier is obtained by using a support vector machine to train historical data of the time domain eigenvalues of the magnetic flux leakage electric signal.

[0014] An embodiment of the present invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the method for classifying magnetic flux leakage signals in the rail magnetic flux leakage detection is implemented.

[0015] An embodiment of the present invention further provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the computer program implements the above-mentioned method for classifying magnetic flux leakage signals in rail magnetic flux leakage detection.

[0016] An embodiment of the present invention further provides a computer program product, which includes a computer program. When the computer program is executed by a processor, it implements the above-mentioned classification method of magnetic flux leakage signals in rail magnetic flux leakage detection.

[0017] In an embodiment of the present invention, an acquired magnetic flux leakage signal of a tested rail is converted into a magnetic flux leakage electric signal; the magnetic flux leakage signal is obtained by picking up a magnetic field signal of a magnetized tested rail; multiple time-domain eigenvalues of the magnetic flux leakage electric signal are extracted; and a support vector machine classifier is used to classify the magnetic flux leakage signal corresponding to the magnetic flux leakage electric signal based on the multiple time-domain eigenvalues of the magnetic flux leakage electric signal to obtain a classification result of the magnetic flux leakage signal. The support vector machine classifier is obtained by training a support vector machine with historical data of the time-domain eigenvalues of the magnetic flux leakage electric signal. Compared with the technical solution in the prior art that classifies magnetic flux leakage signals only based on the amplitude of the magnetic flux leakage signal, the magnetic flux leakage signal can be accurately classified by extracting multiple time-domain eigenvalues of the magnetic flux leakage electric signal and using the support vector machine classifier. This solves the problem of low accuracy of rail magnetic flux leakage detection caused by the inability to effectively classify rail scratch signals and weld signals in the prior art, thereby improving the accuracy of rail magnetic flux leakage detection and reducing the false alarm rate of rail magnetic flux leakage detection. At the same time, it can greatly improve the efficiency of railway inspection and maintenance, and better ensure railway line safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work. In the drawings:

[0019] Figure 1 Schematic diagram of a rail magnetic flux leakage detection technology principle in an embodiment of the present invention;

[0020] Figure 2 1 is a flow chart of a method for classifying magnetic flux leakage signals in rail magnetic flux leakage detection according to an embodiment of the present invention;

[0021] Figure 3 This is a specific example diagram of a rail magnetic flux leakage detection system in an embodiment of the present invention;

[0022] Figure 4 Schematic diagram of the structure of a device for classifying magnetic flux leakage signals in rail magnetic flux leakage detection according to an embodiment of the present invention;

[0023] Figure 5 A schematic diagram of a computer device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0024] To make the purpose, technical solutions and advantages of the embodiments of the present invention more clear, the embodiments of the present invention are further described in detail below with reference to the accompanying drawings. Here, the exemplary embodiments of the present invention and their descriptions are used to explain the present invention, but are not intended to limit the present invention.

[0025] The term "and / or" herein simply describes an association relationship, indicating that three relationships can exist. For example, A and / or B can represent the existence of A alone, the simultaneous existence of A and B, and the existence of B alone. In addition, the term "at least one" herein refers to any combination of at least two of any one or more of a plurality of items. For example, "at least one of A, B, and C" can represent any one or more elements selected from the set consisting of A, B, and C.

[0026] In the description of this specification, the terms "include", "including", "have", "contain", etc. are all open terms, which mean including but not limited to. The descriptions with reference to the terms "one embodiment", "a specific embodiment", "some embodiments", "for example", etc. mean that the specific features, structures or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures or characteristics described can be combined in a suitable manner in any one or more embodiments or examples. The order of steps involved in each embodiment is used to schematically illustrate the implementation of the present application, and the order of steps therein is not limited and can be appropriately adjusted as needed.

[0027] Currently, magnetic flux leakage detection systems can detect fixed markers such as rail joints and welds, as well as top surface scratches. However, the magnetic flux leakage signals from scratches and welds are similar in amplitude and shape, making them difficult to distinguish effectively. For example, the BZ direction magnetic field signals (parallel to the rail surface) of rail scratches and welds are both bimodal, with similar peak-to-peak values, making them difficult to distinguish effectively using a set amplitude threshold. As a result, many weld signals are misidentified as scratches during the defect detection process, significantly impacting the accuracy and false alarm rate of magnetic flux leakage detection.

[0028] In order to solve the above problems, the embodiment of the present invention provides a classification method for magnetic flux leakage signals in rail magnetic flux leakage detection, which is used to improve the accuracy of rail magnetic flux leakage detection, reduce the false alarm rate of rail magnetic flux leakage detection, greatly improve the efficiency of railway inspection and maintenance, and better ensure the safety of railway lines. Figure 2 , the method may include:

[0029] Step 201: converting the acquired magnetic flux leakage signal of the rail under test into a magnetic flux leakage electrical signal; the magnetic flux leakage signal is obtained by picking up the magnetic field signal of the magnetized rail under test;

[0030] Step 202: extracting multiple time-domain eigenvalues of the leakage magnetic electric signal;

[0031] Step 203: Based on the multiple time-domain eigenvalues of the leakage magnetic electric signal, a support vector machine classifier is used to classify the leakage magnetic signal corresponding to the leakage magnetic electric signal to obtain a classification result of the leakage magnetic signal; the above-mentioned support vector machine classifier is obtained by using a support vector machine to train the historical data of the time-domain eigenvalues of the leakage magnetic electric signal.

[0032] In an embodiment of the present invention, an acquired magnetic flux leakage signal of a tested rail is converted into a magnetic flux leakage electric signal; the magnetic flux leakage signal is obtained by picking up a magnetic field signal of a magnetized tested rail; multiple time-domain eigenvalues of the magnetic flux leakage electric signal are extracted; and a support vector machine classifier is used to classify the magnetic flux leakage signal corresponding to the magnetic flux leakage electric signal based on the multiple time-domain eigenvalues of the magnetic flux leakage electric signal to obtain a classification result of the magnetic flux leakage signal. The support vector machine classifier is obtained by training a support vector machine with historical data of the time-domain eigenvalues of the magnetic flux leakage electric signal. Compared with the technical solution in the prior art that classifies magnetic flux leakage signals only based on the amplitude of the magnetic flux leakage signal, the magnetic flux leakage signal can be accurately classified by extracting multiple time-domain eigenvalues of the magnetic flux leakage electric signal and using the support vector machine classifier. This solves the problem of low accuracy of rail magnetic flux leakage detection caused by the inability to effectively classify rail scratch signals and weld signals in the prior art, improves the accuracy of rail magnetic flux leakage detection, and reduces the false alarm rate of rail magnetic flux leakage detection. At the same time, it can greatly improve the efficiency of railway inspection and maintenance, and better ensure railway line safety.

[0033] In a specific implementation, the magnetic leakage signal of the measured rail is first converted into a magnetic leakage electric signal; the magnetic leakage signal is obtained by picking up the magnetic field signal of the magnetized measured rail.

[0034] In the embodiment, the magnetic flux leakage signal of the rail under test can be obtained by a rail magnetic flux leakage detection system.

[0035] For example, Figure 3 As shown, embodiments of the present invention also provide a rail magnetic flux leakage detection system capable of detecting rail head top surface scratches based on a large rail flaw detection vehicle. The rail magnetic flux leakage detection system can be composed of a computer processing system, an excitation power supply, probes 1 and 2, a signal conditioning circuit, a gate circuit, and a data acquisition card.

[0036] The excitation power supply can provide direct current, which is used to power a magnetizer for magnetizing the rail under test;

[0037] The detection probe for a single-side rail consists of two parts, probe 1 and probe 2. Each probe consists of a magnetizer and an array sensor. The magnetizer is used to magnetize the rail, and the array sensor is used to pick up the leakage magnetic field signal.

[0038] When the flaw detection vehicle is traveling upward for inspection, the magnetizer of probe 1 magnetizes the inspected rail and picks up the magnetic flux leakage signal from the rail top surface. The magnetizer of probe 2 magnetizes the inspected rail and picks up the magnetic flux leakage signal at the rail gauge angle. The magnetic flux leakage signal picked up by the array sensor in the probe is converted into an electrical signal and processed by the signal conditioning circuit. The processed magnetic flux leakage signal is then compared with the gate circuit. The gate circuit is used to monitor whether the electromagnetic signal exceeds a preset amplitude threshold. When the amplitude of the magnetic flux leakage signal exceeds the preset amplitude threshold, the gate circuit triggers the data acquisition card to collect the magnetic flux leakage signal.

[0039] The signals collected by the data acquisition card are transmitted to the computer processing system for storage and processing. With the help of the computer processing system, the classification method of the magnetic flux leakage signal in the rail magnetic flux leakage detection provided by the embodiment of the present invention can be executed, and then it can be used to determine whether the rail has abrasions.

[0040] In one embodiment, the above-mentioned detection system can be used to obtain the initial magnetic leakage signal of the tested rail by triggering acquisition with a set amplitude threshold, and a preset algorithm can be used to intercept the target magnetic leakage signal, that is, the magnetic leakage signal of suspected scratches and welds.

[0041] In a specific implementation, after the acquired magnetic leakage signal of the measured rail is converted into a magnetic leakage electric signal, a plurality of time domain characteristic values of the magnetic leakage electric signal are extracted.

[0042] In one embodiment, a plurality of time domain eigenvalues of the leakage magnetic electric signal can be freely selected according to the actual needs of the staff, and then in subsequent steps, the time domain eigenvalues of the freely selected leakage magnetic electric signal are used to classify the leakage magnetic signal corresponding to the leakage magnetic electric signal using a support vector machine classifier to obtain a classification result of the leakage magnetic signal.

[0043] It is worth noting that, since there are many types of time-domain eigenvalues of the leakage magnetic electric signal, the embodiment of the present invention is not limited to the selection method of multiple time-domain eigenvalues of the leakage magnetic electric signal, and can be freely set according to actual needs. The multiple time-domain eigenvalues of the leakage magnetic electric signal in the subsequent steps are only used as examples.

[0044] In one embodiment, the multiple time-domain characteristic values of the leakage magnetic electric signal include: one or any combination of peak-to-peak value, peak-to-peak spacing, peak-to-peak slope, peak ratio and crest factor of the leakage magnetic electric signal.

[0045] In one embodiment, multiple time-domain characteristic values of the leakage magnetic electric signal are extracted, including:

[0046] The difference between the positive peak amplitude and the negative peak amplitude of the leakage magnetic electric signal is taken as the peak-to-peak value of the leakage magnetic electric signal;

[0047] According to the positive and negative peaks of the leakage magnetic electric signal: the number of sampling points, the detection speed and the sampling speed, the peak-to-peak spacing of the leakage magnetic electric signal is converted;

[0048] The ratio of the peak-to-peak value of the leakage magnetic electric signal to the peak-to-peak distance is taken as the peak-to-peak slope of the leakage magnetic electric signal;

[0049] The ratio of the positive peak amplitude to the negative peak amplitude of the leakage magnetic electric signal is taken as the peak ratio of the leakage magnetic electric signal;

[0050] The ratio of the peak-to-peak value of the leakage magnetic electric signal to the effective value (RMS) of the leakage magnetic electric signal is used as the peak factor of the leakage magnetic electric signal.

[0051] In a specific implementation, after extracting multiple time-domain eigenvalues of the leakage magnetic electric signal, a support vector machine classifier is used to classify the leakage magnetic signal corresponding to the leakage magnetic electric signal according to the multiple time-domain eigenvalues of the leakage magnetic electric signal to obtain a classification result of the leakage magnetic signal; the above-mentioned support vector machine classifier is obtained by using a support vector machine to train historical data of the time-domain eigenvalues of the leakage magnetic electric signal.

[0052] The embodiment of the present invention can also classify these nonlinear data by importing the above-mentioned historical data into space and effectively mapping them to high dimensions. However, this mapping process consumes a large computational cost and easily reduces work efficiency. Therefore, the embodiment of the present invention specifically proposes to use a support vector machine and, with the help of kernel function techniques, process the dot product results of multiple eigenvalues (i.e., multiple time-domain eigenvalues of the leakage magnetic electric signal), which can greatly reduce the computing resource requirements. In turn, a linear classifier can be obtained, which can perform feature mapping on the leakage magnetic signal (which can be classified as abrasion signals and weld signals) and classify the leakage magnetic signal corresponding to the leakage magnetic electric signal by selecting effective eigenvectors to obtain the classification result of the leakage magnetic signal.

[0053] In one embodiment, a support vector machine classifier is constructed as follows:

[0054] Construct training and testing datasets based on historical data of time-domain eigenvalues of leakage magnetic electric signals;

[0055] Using the training data set, train and construct the initial support vector machine classifier;

[0056] According to the test data set, the initial support vector machine classifier is tested to obtain a tested support vector machine classifier.

[0057] In the above embodiment, the present invention is based on the classification method of support vector machine, which comprehensively adopts multiple time domain feature value extraction methods of leakage magnetic electric signals and support vector machine classifier to classify leakage magnetic signals. The classification performance of leakage magnetic signals is guaranteed by the support vector machine classifier. It is intuitive, easy to implement, has strong generalization ability, and has good recognition performance.

[0058] In one embodiment, a support vector machine classifier is used to classify the magnetic flux leakage signal corresponding to the magnetic flux leakage signal based on multiple time domain eigenvalues of the magnetic flux leakage signal, and the classification results of the magnetic flux leakage signal are obtained, including:

[0059] Input multiple time-domain eigenvalues of the leakage magnetic electric signal into the support vector machine classifier;

[0060] receiving an output of the support vector machine classifier: calculating function values of multiple time-domain eigenvalues of the leakage magnetic electric signal using an exponential function between time-domain eigenvalues fitted by the support vector machine classifier; the exponential function between time-domain eigenvalues being used to characterize an exponential function relationship between multiple time-domain eigenvalues;

[0061] According to the above function value, the magnetic flux leakage signal corresponding to the magnetic flux leakage electric signal is classified to obtain a classification result of the magnetic flux leakage signal.

[0062] In one embodiment, the function values of various time-domain eigenvalues of the leakage magnetic electric signal are calculated using the exponential function between the time-domain eigenvalues fitted by the support vector machine classifier according to the following formula:

[0063]

[0064] Among them, B z The function value representing various time-domain characteristic values of the leakage magnetic electric signal; a is the peak-to-peak value, in volts; b is the peak-to-peak spacing, in meters; c is the peak-to-peak slope, dimensionless; d is the peak ratio, dimensionless; f is the crest factor, dimensionless.

[0065] In the above embodiment, the magnetic flux leakage signal corresponding to the magnetic flux leakage electric signal is classified according to the above function value to obtain the classification result of the magnetic flux leakage signal, including:

[0066] When the function value is greater than a preset threshold, the classification result of the magnetic flux leakage signal is determined to be a damage signal;

[0067] When the function value is less than or equal to a preset threshold, the classification result of the magnetic flux leakage signal is determined to be a weld signal.

[0068] For example, the preset threshold value may be 44.3. When the above function value is greater than 44.3, the classification result of the magnetic flux leakage signal is determined to be a damage signal, such as determining the classification result of the magnetic flux leakage signal to be a scratch signal; when the above function value is less than or equal to 44.3, the classification result of the magnetic flux leakage signal is determined to be a weld signal.

[0069] In the above embodiment, to address the high similarity between scratch signals and weld signals in rail magnetic flux leakage top surface detection technology, a method is proposed to identify and distinguish scratch and weld signals. The BZ direction magnetic field signals (parallel to the rail surface) of rail scratches and welds are both bimodal signals with relatively close peak-to-peak values, making it difficult to effectively distinguish them by setting an amplitude threshold. By extracting characteristic values such as peak-to-peak value, peak-to-peak spacing, peak-to-peak slope, peak ratio, and peak factor of the scratch and weld signals, a classifier support vector machine is used to train a model and classify the scratch and weld signals. The embodiment of the present invention studies the algorithm of data based on the rail magnetic flux leakage detection system and obtains a method for effectively classifying rail scratches and welds. The key to this method is to extract features from a sufficient number of scratch and weld signal samples, classify and identify them, and train a classifier, and then verify the accuracy of the classifier after obtaining it. By fitting a function curve, the numerical comparison after substituting five time domain characteristic values is achieved to effectively classify rail scratches and welds.

[0070] In an embodiment of the present invention, an acquired magnetic flux leakage signal of a tested rail is converted into a magnetic flux leakage electric signal; the magnetic flux leakage signal is obtained by picking up a magnetic field signal of a magnetized tested rail; multiple time-domain eigenvalues of the magnetic flux leakage electric signal are extracted; and a support vector machine classifier is used to classify the magnetic flux leakage signal corresponding to the magnetic flux leakage electric signal based on the multiple time-domain eigenvalues of the magnetic flux leakage electric signal to obtain a classification result of the magnetic flux leakage signal. The support vector machine classifier is obtained by training a support vector machine with historical data of the time-domain eigenvalues of the magnetic flux leakage electric signal. Compared with the technical solution in the prior art that classifies magnetic flux leakage signals only based on the amplitude of the magnetic flux leakage signal, the magnetic flux leakage signal can be accurately classified by extracting multiple time-domain eigenvalues of the magnetic flux leakage electric signal and using the support vector machine classifier. This solves the problem of low accuracy of rail magnetic flux leakage detection caused by the inability to effectively classify rail scratch signals and weld signals in the prior art, improves the accuracy of rail magnetic flux leakage detection, and reduces the false alarm rate of rail magnetic flux leakage detection. At the same time, it can greatly improve the efficiency of railway inspection and maintenance, and better ensure railway line safety.

[0071] As mentioned above, the current rail magnetic leakage detection technology cannot effectively classify rail abrasions and weld signals, which has a great impact on the classification of damage after detection. A simple peak-to-peak threshold judgment will confuse most abrasion and weld signals. The present invention provides an effective classification method for rail abrasions and weld signals based on a magnetic leakage detection system, which can accurately identify and classify abrasions and welds. In the method, five characteristic values of abrasions and welds are extracted, and a classifier is generated and verified by training the model with a large amount of data. The five characteristic values can be substituted into the fitted curve to obtain the result, and the abrasions can be distinguished by numerical comparison. The magnetic leakage signals of abrasions and welds can be effectively distinguished, which greatly improves the efficiency of railway inspection and maintenance, and better ensures the safety of railway lines.

[0072] The present invention also provides a device for classifying magnetic flux leakage signals during rail magnetic flux leakage testing, as described in the following embodiments. Because the principles underlying the device are similar to those underlying the method for classifying magnetic flux leakage signals during rail magnetic flux leakage testing, the implementation of the device can be referenced to the implementation of the method for classifying magnetic flux leakage signals during rail magnetic flux leakage testing, and any repetitive details will not be repeated.

[0073] The embodiment of the present invention also provides a device for classifying magnetic flux leakage signals in rail magnetic flux leakage detection, which is used to improve the accuracy of rail magnetic flux leakage detection, reduce the false alarm rate of rail magnetic flux leakage detection, greatly improve the efficiency of railway inspection and maintenance, and better ensure the safety of railway lines. Figure 4 As shown, the device includes:

[0074] The magnetic leakage electric signal conversion module 401 is used to convert the acquired magnetic leakage signal of the measured rail into a magnetic leakage electric signal; the magnetic leakage signal is obtained by picking up the magnetic field signal of the magnetized measured rail;

[0075] A time domain eigenvalue extraction module 402 is used to extract multiple time domain eigenvalues of the leakage magnetic electric signal;

[0076] The magnetic flux leakage signal classification module 403 is used to classify the magnetic flux leakage signal corresponding to the magnetic flux leakage electric signal using a support vector machine classifier based on multiple time domain eigenvalues of the magnetic flux leakage electric signal to obtain a classification result of the magnetic flux leakage signal; the above-mentioned support vector machine classifier is obtained by using a support vector machine to train historical data of the time domain eigenvalues of the magnetic flux leakage electric signal.

[0077] In one embodiment, the multiple time-domain characteristic values of the leakage magnetic electric signal include: one or any combination of peak-to-peak value, peak-to-peak spacing, peak-to-peak slope, peak ratio and crest factor of the leakage magnetic electric signal.

[0078] In one embodiment, the time domain feature value extraction module is specifically used to:

[0079] The difference between the positive peak amplitude and the negative peak amplitude of the leakage magnetic electric signal is taken as the peak-to-peak value of the leakage magnetic electric signal;

[0080] According to the positive and negative peaks of the leakage magnetic electric signal: the number of sampling points, the detection speed and the sampling speed, the peak-to-peak spacing of the leakage magnetic electric signal is converted;

[0081] The ratio of the peak-to-peak value of the leakage magnetic electric signal to the peak-to-peak distance is taken as the peak-to-peak slope of the leakage magnetic electric signal;

[0082] The ratio of the positive peak amplitude to the negative peak amplitude of the leakage magnetic electric signal is taken as the peak ratio of the leakage magnetic electric signal;

[0083] The ratio of the peak-to-peak value of the leakage magnetic electric signal to the effective value of the leakage magnetic electric signal is used as the peak factor of the leakage magnetic electric signal.

[0084] In one embodiment, it further includes:

[0085] Support Vector Machine classifier building blocks for:

[0086] Construct a support vector machine classifier as follows:

[0087] Construct training and testing datasets based on historical data of time-domain eigenvalues of leakage magnetic electric signals;

[0088] Using the training data set, train and construct the initial support vector machine classifier;

[0089] According to the test data set, the initial support vector machine classifier is tested to obtain a tested support vector machine classifier.

[0090] In one embodiment, the magnetic flux leakage signal classification module is specifically configured to:

[0091] Input multiple time-domain eigenvalues of the leakage magnetic electric signal into the support vector machine classifier;

[0092] receiving an output of the support vector machine classifier: calculating function values of multiple time-domain eigenvalues of the leakage magnetic electric signal using an exponential function between time-domain eigenvalues fitted by the support vector machine classifier; the exponential function between time-domain eigenvalues being used to characterize an exponential function relationship between multiple time-domain eigenvalues;

[0093] According to the above function value, the magnetic flux leakage signal corresponding to the magnetic flux leakage electric signal is classified to obtain a classification result of the magnetic flux leakage signal.

[0094] In one embodiment, the magnetic flux leakage signal classification module is specifically configured to:

[0095] The function values of various time-domain eigenvalues of the leakage magnetic electric signal are calculated using the exponential function between the time-domain eigenvalues fitted by the support vector machine classifier as follows:

[0096]

[0097] Among them, B z The function value representing various time-domain characteristic values of the leakage magnetic electric signal; a is the peak-to-peak value, in volts; b is the peak-to-peak spacing, in meters; c is the peak-to-peak slope, dimensionless; d is the peak ratio, dimensionless; f is the crest factor, dimensionless.

[0098] In one embodiment, the magnetic flux leakage signal classification module is specifically configured to:

[0099] When the function value is greater than a preset threshold, the classification result of the magnetic flux leakage signal is determined to be a damage signal;

[0100] When the function value is less than or equal to a preset threshold, the classification result of the magnetic flux leakage signal is determined to be a weld signal.

[0101] Based on the above invention concept, Figure 5 As shown, the present invention also proposes a computer device 500, including a memory 510, a processor 520, and a computer program 530 stored in the memory 510 and executable on the processor 520. When the processor 520 executes the computer program 530, the classification method of the magnetic flux leakage signal in the rail magnetic flux leakage detection is implemented.

[0102] An embodiment of the present invention further provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the computer program implements the above-mentioned method for classifying magnetic flux leakage signals in rail magnetic flux leakage detection.

[0103] An embodiment of the present invention further provides a computer program product, which includes a computer program. When the computer program is executed by a processor, it implements the above-mentioned classification method of magnetic flux leakage signals in rail magnetic flux leakage detection.

[0104] In an embodiment of the present invention, an acquired magnetic flux leakage signal of a tested rail is converted into a magnetic flux leakage electric signal; the magnetic flux leakage signal is obtained by picking up a magnetic field signal of a magnetized tested rail; multiple time-domain eigenvalues of the magnetic flux leakage electric signal are extracted; and a support vector machine classifier is used to classify the magnetic flux leakage signal corresponding to the magnetic flux leakage electric signal based on the multiple time-domain eigenvalues of the magnetic flux leakage electric signal to obtain a classification result of the magnetic flux leakage signal. The support vector machine classifier is obtained by training a support vector machine with historical data of the time-domain eigenvalues of the magnetic flux leakage electric signal. Compared with the technical solution in the prior art that classifies magnetic flux leakage signals only based on the amplitude of the magnetic flux leakage signal, the magnetic flux leakage signal can be accurately classified by extracting multiple time-domain eigenvalues of the magnetic flux leakage electric signal and using the support vector machine classifier. This solves the problem of low accuracy of rail magnetic flux leakage detection caused by the inability to effectively classify rail scratch signals and weld signals in the prior art, thereby improving the accuracy of rail magnetic flux leakage detection and reducing the false alarm rate of rail magnetic flux leakage detection. At the same time, it can greatly improve the efficiency of railway inspection and maintenance, and better ensure railway line safety.

[0105] It will be understood by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0106] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0107] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0108] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 The steps for the function specified in one or more boxes.

[0109] The specific embodiments described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for classifying magnetic flux leakage signals in rail magnetic flux leakage detection, characterized in that: include: Converting the magnetic flux leakage signal of the rail under test into a magnetic flux leakage electrical signal; the magnetic flux leakage signal is obtained by picking up the magnetic field signal of the magnetized rail under test; Extract multiple time domain eigenvalues of leakage magnetic electric signal; According to the multiple time-domain eigenvalues of the leakage magnetic electric signal, the support vector machine classifier is used to classify the leakage magnetic signal corresponding to the leakage magnetic electric signal to obtain the classification result of the leakage magnetic signal; The support vector machine classifier is obtained by training the historical data of the time domain eigenvalues of the leakage magnetic electric signal using the support vector machine; The multiple time domain characteristic values of the leakage magnetic electric signal include: peak-to-peak value, peak-to-peak spacing, peak-to-peak slope, peak ratio and peak factor of the leakage magnetic electric signal; A support vector machine classifier is used to classify a magnetic flux leakage signal corresponding to the magnetic flux leakage signal according to multiple time domain eigenvalues of the magnetic flux leakage signal, thereby obtaining a classification result of the magnetic flux leakage signal, including: inputting the multiple time domain eigenvalues of the magnetic flux leakage signal into the support vector machine classifier; receiving an output of the support vector machine classifier: calculating function values of the multiple time domain eigenvalues of the magnetic flux leakage signal using an exponential function between time domain eigenvalues fitted by the support vector machine classifier; the exponential function between time domain eigenvalues is used to characterize an exponential function relationship between multiple time domain eigenvalues; and classifying the magnetic flux leakage signal corresponding to the magnetic flux leakage signal according to the function value, thereby obtaining a classification result of the magnetic flux leakage signal; The function values of various time-domain eigenvalues of the leakage magnetic electric signal are calculated using the exponential function between the time-domain eigenvalues fitted by the support vector machine classifier according to the following formula: Among them, B z The function value representing various time-domain characteristic values of the leakage magnetic electric signal; a is the peak-to-peak value, in volts; b is the peak-to-peak spacing, in meters; c is the peak-to-peak slope, dimensionless; d is the peak ratio, dimensionless; f is the crest factor, dimensionless.

2. The method according to claim 1, wherein Extract multiple time-domain eigenvalues of the leakage magnetic signal, including: The difference between the positive peak amplitude and the negative peak amplitude of the leakage magnetic electric signal is taken as the peak-to-peak value of the leakage magnetic electric signal; According to the positive and negative peaks of the leakage magnetic electric signal: the number of sampling points, the detection speed and the sampling speed, the peak-to-peak spacing of the leakage magnetic electric signal is converted; The ratio of the peak-to-peak value of the leakage magnetic electric signal to the peak-to-peak distance is taken as the peak-to-peak slope of the leakage magnetic electric signal; The ratio of the positive peak amplitude to the negative peak amplitude of the leakage magnetic electric signal is taken as the peak ratio of the leakage magnetic electric signal; The ratio of the peak-to-peak value of the leakage magnetic electric signal to the effective value of the leakage magnetic electric signal is used as the peak factor of the leakage magnetic electric signal.

3. The method according to claim 1, wherein Also includes: Construct a support vector machine classifier as follows: Construct training and testing datasets based on historical data of time-domain eigenvalues of leakage magnetic electric signals; Using the training data set, train and construct the initial support vector machine classifier; According to the test data set, the initial support vector machine classifier is tested to obtain a tested support vector machine classifier.

4. The method according to claim 1, wherein Classifying the magnetic flux leakage signal corresponding to the magnetic flux leakage electric signal according to the function value to obtain a classification result of the magnetic flux leakage signal, including: When the function value is greater than a preset threshold, determining that the classification result of the magnetic flux leakage signal is a damage signal; When the function value is less than or equal to a preset threshold, it is determined that the classification result of the magnetic flux leakage signal is a weld signal.

5. A device for classifying magnetic flux leakage signals in rail magnetic flux leakage detection, characterized in that: include: A magnetic leakage electric signal conversion module is used to convert the acquired magnetic leakage signal of the measured rail into a magnetic leakage electric signal; the magnetic leakage signal is obtained by picking up the magnetic field signal of the magnetized measured rail; Time domain eigenvalue extraction module, used to extract multiple time domain eigenvalues of leakage magnetic electric signals; The magnetic flux leakage signal classification module is used to classify the magnetic flux leakage signal corresponding to the magnetic flux leakage electric signal using a support vector machine classifier according to multiple time domain eigenvalues of the magnetic flux leakage electric signal to obtain the classification result of the magnetic flux leakage signal; The support vector machine classifier is obtained by training the historical data of the time domain eigenvalues of the leakage magnetic electric signal using the support vector machine; The multiple time domain characteristic values of the leakage magnetic electric signal include: peak-to-peak value, peak-to-peak spacing, peak-to-peak slope, peak ratio and peak factor of the leakage magnetic electric signal; The magnetic flux leakage signal classification module is specifically configured to: input multiple time-domain eigenvalues of the magnetic flux leakage electric signal into a support vector machine classifier; receive output from the support vector machine classifier: calculate function values of multiple time-domain eigenvalues of the magnetic flux leakage electric signal using an exponential function between time-domain eigenvalues fitted by the support vector machine classifier; the exponential function between time-domain eigenvalues is used to characterize the exponential function relationship between multiple time-domain eigenvalues; and classify the magnetic flux leakage signal corresponding to the magnetic flux leakage electric signal according to the function value to obtain a classification result of the magnetic flux leakage signal; The magnetic flux leakage signal classification module is specifically used to calculate the function values of multiple time domain eigenvalues of the magnetic flux leakage signal using the time domain eigenvalue inter-exponential function fitted by the support vector machine classifier according to the following formula: Among them, B z The function value representing various time-domain characteristic values of the leakage magnetic electric signal; a is the peak-to-peak value, in volts; b is the peak-to-peak spacing, in meters; c is the peak-to-peak slope, dimensionless; d is the peak ratio, dimensionless; f is the crest factor, dimensionless.

6. The device according to claim 5, characterized in that The time domain feature value extraction module is specifically used for: The difference between the positive peak amplitude and the negative peak amplitude of the leakage magnetic electric signal is taken as the peak-to-peak value of the leakage magnetic electric signal; According to the positive and negative peaks of the leakage magnetic electric signal: the number of sampling points, the detection speed and the sampling speed, the peak-to-peak spacing of the leakage magnetic electric signal is converted; The ratio of the peak-to-peak value of the leakage magnetic electric signal to the peak-to-peak distance is taken as the peak-to-peak slope of the leakage magnetic electric signal; The ratio of the positive peak amplitude to the negative peak amplitude of the leakage magnetic electric signal is taken as the peak ratio of the leakage magnetic electric signal; The ratio of the peak-to-peak value of the leakage magnetic electric signal to the effective value of the leakage magnetic electric signal is used as the peak factor of the leakage magnetic electric signal.

7. The device according to claim 5, characterized in that Also includes: Support Vector Machine classifier building blocks for: Construct a support vector machine classifier as follows: Construct training and testing datasets based on historical data of time-domain eigenvalues of leakage magnetic electric signals; Using the training data set, train and construct the initial support vector machine classifier; According to the test data set, the initial support vector machine classifier is tested to obtain a tested support vector machine classifier.

8. The device according to claim 5, wherein The magnetic flux leakage signal classification module is specifically used for: When the function value is greater than a preset threshold, determining that the classification result of the magnetic flux leakage signal is a damage signal; When the function value is less than or equal to a preset threshold, it is determined that the classification result of the magnetic flux leakage signal is a weld signal.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the method according to any one of claims 1 to 4 is implemented.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 4 is implemented.

11. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 4 is implemented.

Images