Turnout equipment health degree diagnosis method, system, equipment and medium
By filtering the current data of the switch equipment, building feature data sets and clustering analysis, combined with the XGboost model, the problem of low data processing complexity and prediction accuracy in the health diagnosis of switch equipment is solved, achieving more accurate and efficient Health assessment.
Patent Information
- Application Number
- CN202410994781.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-24
- Publication Date
- 2024-11-12
AI Technical Summary
In the health diagnosis of switch equipment, there is a problem of high complexity in operation data processing and low prediction accuracy.
By acquiring current data, filtering and correction, building feature data sets, performing clustering calculations, and analyzing clustering results using the XGboost model to evaluate the health of the switch equipment.
It improves the accuracy of current data analysis, automatically discovers data clustering structures, reduces the complexity of complex features and diversified data processing, and improves the accuracy and efficiency of health diagnosis.
Smart Images

Figure CN118940062A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of turnout diagnosis and prediction, and in particular to a turnout equipment health diagnosis method, system, equipment and medium. Background Art
[0002] Turnout equipment health diagnosis is a key technical means in the field of railway transportation. It aims to evaluate the working status and health of the equipment by real-time monitoring and analysis of the operating data of the turnout system.
[0003] By collecting the operating data of the turnout system, such as electrical signals, mechanical vibrations, and temperature changes, the diagnostic system can analyze this data and identify potential fault characteristics. As an important mode of public transportation, railway transportation has extremely high requirements for safety and operating efficiency. Turnouts are key equipment in railway lines. Their failure will not only affect the normal operation of trains, but may also cause safety accidents and transportation delays. Therefore, it is particularly important to monitor and diagnose the health status of turnout equipment in a timely and effective manner. The introduction of health diagnosis technology can realize real-time monitoring and intelligent analysis of the turnout system, help railway operating units to discover and solve problems in a timely manner, and improve the overall safety, reliability and efficiency of the transportation system.
[0004] However, the existing technology for health diagnosis of turnout equipment still has the problems of high complexity in processing turnout equipment operation data and low accuracy in health diagnosis prediction. Summary of the invention
[0005] The present invention provides a method, system, electronic device and computer-readable storage medium for diagnosing the health of a turnout device, the main purpose of which is to solve the problems of high complexity in processing the operating data of the turnout device and low accuracy in health diagnosis prediction.
[0006] To achieve the above object, the present invention provides a method for diagnosing the health of a turnout device, comprising:
[0007] Obtain current data of target turnout equipment within a preset time;
[0008] filtering and correcting the current data to obtain current analysis data;
[0009] constructing a characteristic data set of the current analysis data;
[0010] Performing clustering calculation on the feature data set to obtain a clustering result;
[0011] The clustering results are analyzed using a preset XGboost model to obtain the health of the target turnout equipment.
[0012] In one embodiment of the present invention, filtering and correcting the current data to obtain current analysis data includes:
[0013] Filtering out a single turnout action current data segment in the current data;
[0014] Counting the number of occurrences of null value current in the single turnout action current data segment;
[0015] If the occurrence number is greater than or equal to the preset number, filtering the null value current in the single turnout action current data segment to obtain current analysis data;
[0016] If the occurrence number is less than the preset number, the null value current in the single turnout action current data segment is interpolated to obtain current analysis data.
[0017] In one embodiment of the present invention, the interpolation of the null value current in the single turnout action current data segment to obtain current analysis data includes:
[0018] Calculating the first-order difference quotient of the adjacent data points according to the adjacent data points of the null current;
[0019] Constructing a Newton interpolation polynomial according to the calculation result of the first-order difference quotient;
[0020] Calculating the function value of the null current using the Newton interpolation polynomial;
[0021] The null current is interpolated using the function value to obtain current analysis data.
[0022] In one embodiment of the present invention, the step of constructing a characteristic data set of the current analysis data includes:
[0023] Calculating the time domain characteristics of the current analysis data;
[0024] Calculating sample entropy and sort entropy of the current analysis data;
[0025] The time domain feature, the sample entropy and the sorting entropy are used as a feature data set of the current analysis data.
[0026] In one embodiment of the present invention, performing clustering calculation on the feature data set to obtain a clustering result includes:
[0027] Taking each feature data in the feature data set as a cluster;
[0028] Calculating the cluster distance between each cluster and converting the cluster distance into a distance matrix;
[0029] Merging the two clusters with the smallest cluster distance in the distance matrix to obtain an updated cluster;
[0030] Calculating the update distance between the update cluster and other clusters, and iterating the distance matrix using the update distance;
[0031] Determine whether the minimum value of the updated distance is less than a preset distance threshold;
[0032] When the minimum value of the updated distance is greater than or equal to the distance threshold, returning to the step of merging two clusters with the smallest cluster distance in the distance matrix;
[0033] When the minimum value of the updated distance is less than the distance threshold, the updated cluster is taken as the clustering result.
[0034] In one embodiment of the present invention, converting the cluster distance into a distance matrix includes:
[0035] Construct an initialized distance matrix whose rows and columns are equal to the number of feature data in the feature data set;
[0036] The initialization distance matrix is filled using the cluster distances to obtain a distance matrix.
[0037] In one embodiment of the present invention, the clustering results are analyzed using a preset XGboost model to obtain the health of the target turnout equipment, including:
[0038] Calculate the predicted label of each feature data in the clustering result according to the XGboost model;
[0039] Grouping each feature data in the clustering result according to the predicted label, and calculating the cluster center of each group;
[0040] Calculate the eigenvalues of the preset features of each cluster center;
[0041] The characteristic values are weighted averaged, and the health of the target turnout equipment is determined according to the result of the weighted average.
[0042] In order to solve the above problems, the present invention also provides a turnout equipment health diagnosis system, the system comprising:
[0043] A data acquisition module, used to acquire current data of a target turnout device within a preset time;
[0044] A data cleaning module, used for filtering and correcting the current data to obtain current analysis data;
[0045] A feature data set construction module, constructing a feature data set of the current analysis data;
[0046] A clustering calculation module, used to perform clustering calculation on the feature data set to obtain a clustering result;
[0047] The data analysis module is used to analyze the clustering results using a preset XGboost model to obtain the health of the target turnout equipment.
[0048] In order to solve the above problem, the present invention further provides an electronic device, the electronic device comprising:
[0049] at least one processor; and,
[0050] a memory communicatively connected to the at least one processor; wherein,
[0051] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can perform the steps of the above-mentioned method for diagnosing the health of turnout equipment.
[0052] In order to solve the above problem, the present invention also provides a computer-readable storage medium, in which at least one computer program is stored. The at least one computer program is executed by a processor in an electronic device to implement the steps of the above-mentioned method for diagnosing the health of a turnout equipment.
[0053] The current data of the target turnout equipment within the preset time in the present invention can effectively realize the real-time monitoring and evaluation of the health status of the equipment, filter and correct the current data, ensure that the analysis results are more accurate, help to timely discover current abnormalities, and optimize data in time; capture the key features in the current data, and display the current data of the target turnout equipment in a more macroscopic way; perform hierarchical clustering calculations on the feature data set based on the shortest distance, and automatically discover the clustering structure in the data without pre-specifying the number of clusters. Use the preset XGBoost model to analyze the clustering results of the target turnout equipment, effectively process high-dimensional features and large-scale data, reduce the complexity of processing complex features and diversified data of turnout equipment, and capture the potential similarities and differences between different devices. Further optimize and adjust the clustering algorithm to improve the accuracy of the clustering results. Therefore, the present invention proposes a turnout equipment health diagnosis method, system, electronic device and computer-readable storage medium, the main purpose of which is to solve the problem of high complexity in processing turnout equipment operation data and low accuracy in health diagnosis prediction. BRIEF DESCRIPTION OF THE DRAWINGS
[0054] Figure 1 A schematic flow chart of a method for diagnosing the health of a turnout device provided in one embodiment of the present invention;
[0055] Figure 2 A functional module diagram of a turnout equipment health diagnosis system provided by one embodiment of the present invention;
[0056] Figure 3 A schematic diagram of the structure of an electronic device for implementing a method for diagnosing the health status of a turnout device provided in one embodiment of the present invention.
[0057] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION
[0058] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.
[0059] The embodiment of the present application provides a method for diagnosing the health of a turnout device. The execution subject of the method for diagnosing the health of a turnout device includes but is not limited to at least one of the electronic devices such as a server and a terminal that can be configured to execute the method provided by the embodiment of the present application. In other words, the method for diagnosing the health of a turnout device can be executed by software or hardware installed in a terminal device or a server device.
[0060] Reference Figure 1 FIG. 1 is a flow chart of a method for diagnosing the health of a turnout device according to an embodiment of the present invention. In this embodiment, the method for diagnosing the health of a turnout device includes:
[0061] S1. Obtain the current data of the target turnout equipment within a preset time.
[0062] In the embodiment of the present invention, the current data of the stored target turnout equipment can be acquired through the database.
[0063] By acquiring the current data of the target turnout equipment within a preset time, the present invention can effectively realize real-time monitoring and evaluation of the health status of the equipment, thereby predicting potential faults in advance, improving the reliability and safety of the equipment, and providing analytical data for the health diagnosis of the turnout equipment.
[0064] S2. Filter and correct the current data to obtain current analysis data.
[0065] In an embodiment of the present invention, the filtering and correcting of the current data to obtain current analysis data includes: screening out a single turnout action current data segment in the current data; counting the number of occurrences of null-value current in the single turnout action current data segment; if the number of occurrences is greater than or equal to the preset number of times, filtering the null-value current in the single turnout action current data segment to obtain current analysis data; if the number of occurrences is less than the preset number of times, interpolating the null-value current in the single turnout action current data segment to obtain current analysis data.
[0066] In detail, the single turnout action current data segment refers to the complete current data for completing a turnout action, wherein the complete current data for a turnout action includes a complete cycle of a turnout from before unlocking - unlocking - switching - locking - after locking.
[0067] In detail, if the number of occurrences is greater than or equal to the preset number of times, the null value current in the single turnout action current data segment is deleted, and the current data other than the null value current is used as the current analysis data. For example, if the preset number of times is 2 times, if the number of occurrences is less than the preset number of times, the null value current that has appeared once in the single turnout action current data segment is interpolated; if the number of occurrences is greater than or equal to the preset number of times, the null value current that has appeared 2 times or more in the single turnout action current data segment is deleted.
[0068] In an embodiment of the present invention, the interpolation of the null current in the single turnout action current data segment to obtain current analysis data includes: calculating the first-order difference quotient of the adjacent data points of the null current according to the adjacent data points of the null current; constructing a Newton interpolation polynomial according to the calculation result of the first-order difference quotient; calculating the function value of the null current using the Newton interpolation polynomial; and interpolating the null current using the function value to obtain current analysis data.
[0069] In the embodiment of the present invention, the null current in the single turnout action current data segment can be interpolated by Newton, assuming that the known adjacent data points of the null current are A (a1, b1) and B (a2, b2), and the first-order difference quotient is calculated. The first-order difference quotient result is used to construct the polynomial p(x)=b1+F[a1,a2](a-a1) according to the Newton interpolation formula, and the horizontal coordinate corresponding to the null current is substituted into the constructed Newton interpolation polynomial to obtain the function value, which is used to interpolate the null current.
[0070] Among them, Newton interpolation is a method of recursively constructing interpolation polynomials, so that when adding new data points, it is not necessary to recalculate the entire interpolation polynomial. Instead, the polynomial is directly updated using the difference quotient of known data points, thereby saving computing resources and improving computing efficiency. Newton interpolation is applicable to various types of data sets. Whether it is uniformly spaced or non-uniformly spaced data points, it can accurately construct interpolation polynomials with high numerical stability. Compared with other interpolation methods such as Lagrange interpolation, it can better guarantee numerical accuracy and stability during the calculation process.
[0071] The present invention filters and corrects the current data, which can improve the accuracy of current analysis. Through effective data processing, abnormal data can be corrected to ensure more accurate analysis results, which helps to timely discover current abnormalities and optimize data in a timely manner.
[0072] S3. Construct a feature data set of the current analysis data.
[0073] In an embodiment of the present invention, constructing a feature data set of the current analysis data includes: calculating time domain features of the current analysis data; calculating sample entropy and sorting entropy of the current analysis data; and using the time domain features, the sample entropy and the sorting entropy as the feature data set of the current analysis data.
[0074] In detail, the time domain features include: mean value, standard deviation, root mean square, maximum value, difference between maximum and minimum values, waveform factor, impulse factor, skewness, kurtosis and mode, etc. Time domain features are characteristic parameters obtained by statistically analyzing the changes of signals or data on the time axis, which can intuitively reflect the overall characteristics and fluctuations of data.
[0075] Sample entropy and sort entropy are indicators used to measure the complexity and irregularity of signals, which can provide deeper data analysis. Sample entropy evaluates the complexity of a signal by calculating the frequency distribution of adjacent data points in the signal sequence, while sort entropy reflects the irregularity and amount of information in the data based on the changes in the order of data points.
[0076] These characteristic parameters calculated by the present invention not only help to accurately describe and understand the characteristics of the signal, but also provide a reliable basis and method for data prediction, fault diagnosis, equipment health monitoring and system optimization, thereby improving the reliability and performance efficiency of the equipment.
[0077] S4. Perform clustering calculation on the feature data set to obtain a clustering result.
[0078] In an embodiment of the present invention, the clustering calculation is performed on the feature data set to obtain a clustering result, including: taking each feature data in the feature data set as a cluster; calculating the cluster distance between each cluster, and converting the cluster distance into a distance matrix; merging two clusters with the smallest cluster distance in the distance matrix to obtain an updated cluster; calculating the updated distance between the updated cluster and other clusters, and iterating the distance matrix using the updated distance; determining whether the minimum value of the updated distance is less than a preset distance threshold; when the minimum value of the updated distance is greater than or equal to the distance threshold, returning to the step of merging the two clusters with the smallest cluster distance in the distance matrix; when the minimum value of the updated distance is less than the distance threshold, taking the updated cluster as the clustering result.
[0079] In the embodiment of the present invention, when the update distance between the update cluster and other clusters is less than a preset distance threshold, the update cluster containing a number of feature data at this time is taken as the clustering result.
[0080] In an embodiment of the present invention, converting the cluster distance into a distance matrix includes: constructing an initialization distance matrix whose rows and columns are the number of feature data in the feature data set; and filling the initialization distance matrix with the cluster distance to obtain a distance matrix.
[0081] Specifically, an initialization distance matrix is constructed, where the rows and columns are the number of feature data in the feature data set. According to the calculated cluster distances between each cluster, the cluster distances are filled into the corresponding positions in the initialization distance matrix to obtain a distance matrix. Assume that the distance between cluster u and cluster v is d uv , because the distance matrix is symmetric, the positions D[u,v] and D[v,u] in the distance matrix should be set to d uv , when filling the distance, the values at the symmetrical positions should be updated at the same time.
[0082] For the remaining unfilled positions D[u,v] (where u≠v), set an initial value as needed. Typically, these elements can be set to infinity, indicating that the clusters are initially considered independent of each other.
[0083] Clustering ends when the update distance between the update cluster and other clusters is less than the preset distance threshold through the dynamic time warping algorithm, and the obtained update cluster is the clustering result. When calculating the distance of each feature data in the feature data set, the dynamic time warping algorithm dynamically adjusts the alignment path between time series and considers the changes in length and speed, so as to effectively solve the offset and deformation problems on the time axis and provide more accurate clustering results.
[0084] The distance between each cluster can be calculated using three eigenvalues: current length difference, difference in average value of non-null current curve after removing the maximum value, and peak difference in the locking stage, to balance the defects of the dynamic time warping algorithm, and the distance and eigenvalues are weighted to obtain a new distance.
[0085] For example, the current length difference with a weight of 0.2 is calculated as follows:
[0086] eps1=abs(len(line1)-len(line2))*0.2;
[0087] Among them, line1 and line2 represent the current curve, and abs(len(line1)-len(line2)) represents the absolute value of the current length difference.
[0088] The difference between the average values of the non-null current curve after removing the maximum value, with a weight of 8, is calculated as follows:
[0089]
[0090] Where line1 and line2 represent the current curves, sum(line2) represents the total length of current curve 2, max(line2) represents the maximum value of current curve 2, len(np.nonzero(line2)[0] represents the length of current curve 2 excluding null values, sum(line1) represents the total length of current curve 1, max(line1) represents the maximum value of current curve 1, len(np.nonzero(line1)[0] represents the length of current curve 1 excluding null values.
[0091] The peak difference in the locking phase with a weight of 5 is calculated as follows:
[0092] eps3=abs(max(line1_t3)-max(line2_t3))*5;
[0093] Among them, line1 and line2 represent the current curve, line1_t3 and line2_t3 represent the locking stage, max(line1_t3) represents the maximum length of line1 in the locking stage, max(line2_t3) represents the maximum length of line2 in the locking stage, and abs(max(line1_t3)-max(line2_t3)) represents the absolute value of the peak difference in the locking stage.
[0094] The current length difference can evaluate the duration change of the current waveform and reflect the continuity and periodicity of the current activity. The difference between the average values of the non-null current curve after removing the maximum value can eliminate the influence of the current peak on the overall mean, and more accurately reveal the true change trend of the current data, which is helpful to identify and analyze abnormal conditions or changes. The peak difference in the locking stage is to evaluate the peak change of the current in a specific operation stage, which can reflect the characteristics of the current fluctuation of the turnout equipment in the locking stage, and provides an important reference for the health diagnosis of the turnout equipment.
[0095] Hierarchical clustering of feature data sets based on the shortest distance can automatically discover clustering structures in the data without specifying the number of clusters in advance. By gradually merging the closest data points or clusters, the algorithm can provide a detailed clustering hierarchy and is generally robust to noisy data.
[0096] S5. Analyze the clustering results using a preset XGboost model to obtain the health of the target turnout equipment.
[0097] In an embodiment of the present invention, the clustering results are analyzed using a preset XGboost model to obtain the health of the target turnout equipment, including: calculating the predicted label of each feature data in the clustering results according to the XGboost model; grouping each feature data in the clustering results according to the predicted label, and calculating the cluster center of each group; calculating the feature value of the preset feature of each cluster center; weighted averaging the feature values, and determining the health of the target turnout equipment according to the result of the weighted average.
[0098] Specifically, each feature data in the clustering result is grouped according to the predicted label, and all feature data points with the predicted label of the same category are put into the same group.
[0099] In the embodiment of the present invention, for each group, the minimum value of the sum of distances between the current curve corresponding to each group and all other curves is calculated as the cluster center, and the cluster center is the actual current curve.
[0100] In the embodiment of the present invention, the preset features include the maximum value, the minimum value, the difference between the maximum and minimum values, the running time, the non-zero average value and the like in the conversion stage.
[0101] In detail, the eigenvalues calculated for each cluster center and the weights corresponding to the eigenvalues are weighted averaged, and the weighted average results are given health levels in turn according to the 13 types of labels. For example, the weights of the preset features are as follows: the maximum value weight of the transition stage is 1.1, the minimum value weight of the transition stage is 0.9, the maximum and minimum value difference weight is 1.5, the running time weight is 0.1, and the non-zero average value weight is 0.5, where the transition stage refers to the transition state in the complete cycle of a single turnout action.
[0102] The preset features are weighted with the corresponding weights to obtain the weighted results of the preset features, the weighted results are averaged, and the weighted average results are given health levels in sequence according to the labels. The turnout equipment health diagnosis system divides the healthy state and degraded state into 13 categories, with labels of 0, 1, 2, ..., 11, 12. Among them, label 0 means that it is basically very healthy and has no degradation; label 12 means that it is unhealthy and has very serious degradation.
[0103] The preset XGboost model predicts each feature data in the clustering result by iterating multiple trees, continuously optimizing the objective function of the XGboost model and improving the accuracy of the XGboost model prediction. Assume that the tree model to be trained in the tth iteration is f t (x), the tree model of the XGboost algorithm can be expressed as:
[0104]
[0105] in, Represents the feature data x in the clustering result after the tth iteration i The predicted label of represents the predicted label after the first t-1 iterations, f t (x i ) represents the tree model of the tth iteration.
[0106] The objective function of the tree model at the tth iteration consists of the loss function of the XGboost model and the regularization term, and the formula is as follows:
[0107]
[0108] in, represents the loss function in the XGboost model, It means to sum the complexity of all t trees as the regularization term of the objective function.
[0109] set up Represents the clustering result x after the tth iteration i Prediction label:
[0110]
[0111] in, represents the predicted label after the first t-1 iterations, f t (x i ) represents the tree model of the tth iteration.
[0112] Will Substitute into the objective function:
[0113]
[0114] in,
[0115]
[0116] Will Substitute into Obj (t) :
[0117]
[0118] in, represents the loss function in the XGboost model, represents the predicted label after the first t-1 iterations, f t (x i ) represents the tree model of the tth iteration, Ω(f t ) represents the complexity of the tree model at the tth iteration.
[0119] right Perform Taylor expansion and obtain the following formula:
[0120]
[0121] Among them, g i is the first-order derivative of the loss function, h i is the second-order derivative of the loss function, f t (x i ) represents the tree model of the tth iteration, Ω(f t ) represents the complexity of the tree model at the tth iteration.
[0122] The regularization term of the objective function is determined by the number of leaf nodes of all generated trees and the L2 norm of the vector composed of all node weights:
[0123]
[0124] Substituting into the objective function we get:
[0125]
[0126] Where T represents the number of leaves, represents the L2 norm of the jth leaf node weight vector, g i is the first-order derivative of the loss function, h i It is the second-order derivative of the loss function, γ represents the leaf node coefficient, and λ represents the regularization term penalty coefficient.
[0127] The objective function is ω j Find the first-order derivative and set the derivative result equal to 0, then we get the weight corresponding to the leaf node j. Substitution And define and The final expression of the objective function is:
[0128]
[0129] Among them, ω j represents the weight of the jth leaf node, T represents the number of leaves, γ represents the leaf node coefficient, λ represents the regularization term penalty coefficient, G j represents the cumulative sum of the first-order partial derivatives of the samples contained in the j-th leaf node, H j Obj represents the cumulative sum of the second-order partial derivatives of the samples contained in the jth leaf node. (t) Represents the objective function.
[0130] The XGboost model can handle complex nonlinear relationships and interaction effects, effectively capture the trend of current data through ensemble learning of multiple trees, improve the accuracy and generalization ability of the prediction of the clustering results, and has the characteristics of strong interpretability, easy understanding and adjustment.
[0131] The preset XGboost model is used to analyze the clustering results of the target turnout equipment, effectively processing high-dimensional features and large-scale data, reducing the complexity of processing the complex features and diversified data of the turnout equipment, capturing the potential similarities and differences between different devices, further optimizing and adjusting the clustering algorithm, and improving the accuracy of the clustering results.
[0132] like Figure 2 As shown, it is a functional module diagram of a turnout equipment health diagnosis system provided by an embodiment of the present invention. According to the functions to be implemented, the turnout equipment health diagnosis system 400 may include a data acquisition module 401, a data cleaning module 402, a feature data set construction module 403, a clustering calculation module 404 and a data analysis module 405. The module described in the present invention may also be referred to as a unit, which refers to a series of computer program segments that can be executed by the processor of the turnout equipment health diagnosis system and can complete fixed functions, which are stored in the memory of the turnout equipment health diagnosis system.
[0133] In this embodiment, the functions of each module / unit are as follows:
[0134] A data acquisition module, used to acquire current data of a target turnout device within a preset time;
[0135] A data cleaning module, used for filtering and correcting the current data to obtain current analysis data;
[0136] A feature data set construction module, constructing a feature data set of the current analysis data;
[0137] A clustering calculation module, used to perform clustering calculation on the feature data set to obtain a clustering result;
[0138] The data analysis module is used to analyze the clustering results using a preset XGboost model to obtain the health of the target turnout equipment.
[0139] In detail, each module described in a turnout equipment health diagnosis system 400 described in an embodiment of the present invention adopts the same technical means as a turnout equipment health diagnosis method described in the accompanying drawings when used, and can produce the same technical effects, which will not be repeated here.
[0140] like Figure 3 FIG. 1 is a schematic diagram of the structure of an electronic device for implementing a method for diagnosing the health of a turnout device provided by an embodiment of the present invention.
[0141] The electronic device 1000 may include a processor 1001, a memory 1002, a communication bus 1003 and a communication interface 1004, and may also include a computer program stored in the memory 1002 and executable on the processor 1001, such as a switch equipment health diagnosis program.
[0142] Among them, the processor 1001 is the control core (Control Unit) of the electronic device, which uses various interfaces and lines to connect the various components of the entire electronic device, and executes various functions of the electronic device and processes data by running or executing programs or modules stored in the memory 1002, and calling data stored in the memory 1002.
[0143] The memory 1002 includes at least one type of readable storage medium. In some embodiments, the memory 1002 may be an internal storage unit of the electronic device, such as a mobile hard disk of the electronic device. The memory 1002 may be used to store not only application software installed in the electronic device but also various data.
[0144] The communication bus 1003 is configured to implement connection and communication between the memory 1002 and at least one processor 1001, etc.
[0145] The communication interface 1004 is used for communication between the electronic device and other devices, including a network interface and a user interface.
[0146] The figure only shows an electronic device with components. Those skilled in the art will understand that the structure shown in the figure does not constitute a limitation on the electronic device 1000, and may include fewer or more components than shown in the figure, or combine certain components, or arrange the components differently.
[0147] It should be understood that the embodiment is for illustration only and the scope of the patent application is not limited to this structure.
[0148] The turnout equipment health diagnosis program stored in the memory 1002 in the electronic device 1000 is a combination of multiple instructions. When running in the processor 1001, the steps in the above-mentioned turnout equipment health diagnosis method can be implemented.
[0149] Specifically, the specific implementation system of the processor 1001 for the above instructions can refer to the description of the relevant steps in the corresponding embodiment of the accompanying drawings, which will not be repeated here.
[0150] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor of an electronic device, the computer program can implement the steps in the above-mentioned method for diagnosing the health of a turnout device.
[0151] It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.
[0152] Therefore, no matter from which point of view, the embodiments should be regarded as illustrative and non-restrictive, and the scope of the present invention is limited by the appended claims rather than the above description, so it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present invention. Any attached figure mark in the claims should not be regarded as limiting the claims involved.
[0153] In addition, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices stated in the system claim can also be implemented by one unit or device through software or hardware. The words first, second, etc. are used to indicate names, and do not indicate any particular order.
[0154] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution 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 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.
Claims
1. A method for diagnosing the health of a turnout equipment, characterized in that: The method comprises: Obtain current data of target turnout equipment within a preset time; filtering and correcting the current data to obtain current analysis data; constructing a characteristic data set of the current analysis data; Performing clustering calculation on the feature data set to obtain a clustering result; The clustering results are analyzed using a preset XGboost model to obtain the health of the target turnout equipment.
2. A method for diagnosing the health of a turnout equipment according to claim 1, characterized in that: The filtering and correcting of the current data to obtain current analysis data includes: Filtering out a single turnout action current data segment in the current data; Counting the number of occurrences of null value current in the single turnout action current data segment; If the occurrence number is greater than or equal to the preset number, filtering the null value current in the single turnout action current data segment to obtain current analysis data; If the number of occurrences is less than a preset number, the null value current in the single turnout action current data segment is interpolated to obtain current analysis data.
3. A method for diagnosing the health of a turnout equipment according to claim 2, characterized in that: The interpolation of the null value current in the single turnout action current data segment to obtain current analysis data includes: Calculating the first-order difference quotient of the adjacent data points according to the adjacent data points of the null current; Constructing a Newton interpolation polynomial according to the calculation result of the first-order difference quotient; Calculating the function value of the null current using the Newton interpolation polynomial; The null current is interpolated using the function value to obtain current analysis data.
4. A method for diagnosing the health of a turnout equipment according to claim 1, characterized in that: The step of constructing a characteristic data set of the current analysis data comprises: Calculating the time domain characteristics of the current analysis data; Calculating sample entropy and sort entropy of the current analysis data; The time domain feature, the sample entropy and the sorting entropy are used as a feature data set of the current analysis data.
5. A method for diagnosing the health of a turnout equipment according to claim 4, characterized in that: The performing clustering calculation on the feature data set to obtain a clustering result includes: Taking each feature data in the feature data set as a cluster; Calculating the cluster distance between each cluster and converting the cluster distance into a distance matrix; Merging the two clusters with the smallest cluster distance in the distance matrix to obtain an updated cluster; Calculating the update distance between the update cluster and other clusters, and iterating the distance matrix using the update distance; Determine whether the minimum value of the updated distance is less than a preset distance threshold; When the minimum value of the updated distance is greater than or equal to the distance threshold, returning to the step of merging two clusters with the smallest cluster distance in the distance matrix; When the minimum value of the updated distance is less than the distance threshold, the updated cluster is taken as the clustering result.
6. A method for diagnosing the health of a turnout equipment according to claim 5, characterized in that: The step of converting the cluster distance into a distance matrix comprises: Construct an initialized distance matrix whose rows and columns are equal to the number of feature data in the feature data set; The initialization distance matrix is filled using the cluster distances to obtain a distance matrix.
7. A method for diagnosing the health of a turnout equipment according to claim 5, characterized in that: The clustering results are analyzed by using a preset XGboost model to obtain the health of the target turnout equipment, including: Calculate the predicted label of each feature data in the clustering result according to the XGboost model; Grouping each feature data in the clustering result according to the predicted label, and calculating the cluster center of each group; Calculate the eigenvalues of the preset features of each cluster center; The characteristic values are weighted averaged, and the health of the target turnout equipment is determined according to the result of the weighted average.
8. A turnout equipment health diagnosis system, characterized in that: The system comprises: A data acquisition module, used to acquire current data of a target turnout device within a preset time; A data cleaning module, used for filtering and correcting the current data to obtain current analysis data; A feature data set construction module, constructing a feature data set of the current analysis data; A clustering calculation module, used to perform clustering calculation on the feature data set to obtain a clustering result; The data analysis module is used to analyze the clustering results using a preset XGboost model to obtain the health of the target turnout equipment.
9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and, a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute a method for diagnosing the health of a turnout equipment as described in any one of claims 1 to 7.
10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, a method for diagnosing the health of a turnout device as described in any one of claims 1 to 7 is implemented.
Citation Information
Patent Citations
Turnout management system and method
CN110930119A
Intelligent diagnosis method for turnout switching equipment
CN118070188A
