A k-means multi-working-condition clustering vehicle fault diagnosis method based on correlation analysis

Abstract

The application discloses a k-means multi-working-condition clustering vehicle fault diagnosis method based on correlation analysis, and comprises the following steps: acquiring vehicle offline working condition parameters; acquiring the optimal clustering center quantity of k-means clustering based on the elbow method and the average profile coefficient method; performing k-means clustering analysis on the vehicle offline working condition parameters to obtain the threshold range of vehicle component signal parameters corresponding to different working conditions; acquiring real-time vehicle component signal parameters, comparing and analyzing the real-time vehicle component signal parameters with the threshold range of vehicle component signal parameters corresponding to different working conditions to obtain a vehicle fault diagnosis result. The application can avoid the shortcomings that the current artificial intelligence algorithm fault diagnosis is not good in small samples and variable working conditions, and can diagnose the state of vehicle components in a complex working condition environment, thereby providing a new method for vehicle fault diagnosis in a complex working condition environment.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of vehicle fault diagnosis, and particularly relates to a k-means multi-working-condition clustering vehicle fault diagnosis method based on correlation analysis. BACKGROUND

[0002] With the increasing complexity of the traffic environment, vehicle fault diagnosis technology is crucial for improving safety, extending service life, and reducing maintenance costs. The current mainstream vehicle fault diagnosis methods are as follows:

[0003] Sensor signal diagnosis: Various sensors such as acceleration sensors, acoustic emission sensors, and temperature sensors collect vibration signals, acoustic signals, and temperature signals. For example, Qiu et al. proposed a fault diagnosis method based on vibration signal adjustment spectrogram, which is suitable for bearing fault diagnosis under different working conditions and limited sample conditions. Osman developed a new normalized Hilbert-Huang transform (NHHT) technology based on vibration signals for bearing fault detection, which proved its robustness under different experimental settings.

[0004] Feature extraction and intelligent diagnosis: Traditional signal analysis and processing methods rely on manual feature extraction, requiring a large amount of expert experience and have limited application. With the application of neural networks, feature extraction has become intelligent, and the original vibration signal can be directly used as model input to complete fault diagnosis. However, when dealing with variable working condition bearing fault diagnosis, the simple structure and limited nonlinear operation capability of traditional neural networks limit their application, and the concept of deep learning emerges as the times require, through more complex network structure and deep feature extraction, improving nonlinear fitting capability and high-dimensional data processing capability.

[0005] Domain adaptation and transfer learning: It is challenging to collect enough rolling bearing fault samples under different working conditions. Methods based on domain adaptation, such as one-dimensional multi-resolution deep domain adaptation neural network, achieve accurate diagnosis of rolling bearing faults under different working conditions by reducing the distribution difference between the source domain and the target domain, demonstrating its effectiveness in diagnosing under different working conditions.

[0006] Fault diagnosis based on classification idea: Common fault diagnosis methods based on classification ideas, such as Bayesian networks (BN), K-nearest neighbors (KNN), and support vector machines (SVM), are used to identify different working conditions and fault types. Among them, SVM is widely used in fault diagnosis due to its good generalization performance and accuracy, and the selection of its parameter optimization method is crucial for the classification effect of the model.

[0007] The existing sensor device-based diagnosis method needs to invest a large amount of funds and maintenance costs, and the artificial intelligence-based diagnosis method performs well in a big data environment, but the diagnosis effect is poor in a small sample and a complex working condition environment, which is not conducive to engineering implementation.

[0008] In view of the above problems, it is urgent to propose a k-means multi-working condition clustering vehicle fault diagnosis method based on correlation analysis, to improve the fault diagnosis effect of vehicle parts in a small sample and a complex working condition environment. SUMMARY

[0009] To solve the above technical problems, the present application provides a k-means multi-working condition clustering vehicle fault diagnosis method based on correlation analysis to solve the problems existing in the prior art.

[0010] To achieve the above purpose, the present application provides a k-means multi-working condition clustering vehicle fault diagnosis method based on correlation analysis, comprising the following steps:

[0011] Obtain vehicle offline working condition parameters;

[0012] Obtain the best cluster center number of k-means clustering based on the elbow method and the average contour coefficient method;

[0013] Perform k-means clustering analysis on the vehicle offline working condition parameters to obtain the threshold range of vehicle component signal parameters corresponding to different working conditions;

[0014] Obtain real-time vehicle component signal parameters, compare and analyze the real-time vehicle component signal parameters with the threshold range of vehicle component signal parameters corresponding to different working conditions, and obtain vehicle fault diagnosis results.

[0015] Optionally, before performing k-means clustering analysis on the vehicle offline working condition parameters, it comprises:

[0016] Perform correlation analysis on the vehicle offline working condition parameters based on the Spearman correlation coefficient, and select typical working condition parameters.

[0017] Optionally, the process of obtaining the best cluster center number of k-means clustering based on the elbow method and the average contour coefficient method comprises:

[0018] Obtain the first best cluster center number of k-means clustering based on the elbow method, obtain the second best cluster center number of k-means clustering based on the average contour coefficient method; based on engineering practice, compare and analyze the first best cluster center number and the second best cluster center number to determine the final best cluster center number of k-means clustering.

[0019] Optionally, the process of obtaining the first optimal cluster center number of k-means clustering based on the elbow method comprises:

[0020] Obtaining the within-group sum of squares under different cluster center numbers based on the elbow method, and drawing the corresponding curve graph, determining the first optimal cluster center number based on the point at which the curve graph begins to present a straight line appearance.

[0021] Optionally, the process of obtaining the second optimal cluster center number of k-means clustering based on the average silhouette coefficient method comprises:

[0022] Obtaining the silhouette coefficient under different cluster center numbers based on the average silhouette coefficient method, and drawing the corresponding curve graph, determining the second optimal cluster center number by analyzing the silhouette coefficient in the curve graph.

[0023] Optionally, the process of performing k-means clustering analysis on the off-line working condition parameters of the vehicle to obtain the threshold range of the signal parameters of the vehicle components corresponding to different working conditions comprises:

[0024] Performing k-means clustering analysis on the off-line working condition parameters of the vehicle to determine that the number of working conditions is the optimal cluster center number; after clustering is completed, analyzing the signal parameters of each component of the vehicle to obtain the threshold range of the signal parameters of the vehicle components corresponding to different working conditions.

[0025] Optionally, the process of comparing and analyzing the real-time vehicle component signal parameters with the threshold range of the signal parameters of the vehicle components corresponding to different working conditions comprises:

[0026] Obtaining the distance of the real-time vehicle component signal parameters from different cluster centers and sorting, obtaining the working condition corresponding to the real-time vehicle component signal parameters based on the cluster center with the shortest distance from the real-time vehicle component signal parameters; when the real-time vehicle component signal parameters exceed the threshold range of the signal parameters of the vehicle components under the corresponding working condition, obtaining the vehicle fault diagnosis result.

[0027] The application also provides an electronic device comprising a memory and a processor; the memory is used to store a program; the processor is used to execute the program to realize each step of the k-means multi-working condition clustering vehicle fault diagnosis method based on correlation analysis.

[0028] The application also provides a readable storage medium having a computer program stored thereon, wherein the computer program is executed by a processor to realize each step of the k-means multi-working condition clustering vehicle fault diagnosis method based on correlation analysis.

[0029] Compared with the prior art, the application has the following advantages and technical effects:

[0030] The method proposed in this invention first acquires offline vehicle operating condition parameters, performs k-means clustering analysis on these parameters to obtain threshold ranges for vehicle component signal parameters corresponding to different operating conditions, and then acquires real-time vehicle component signal parameters. These real-time parameters are then compared with the threshold ranges for different operating conditions to obtain vehicle fault diagnosis results. This method avoids the shortcomings of current AI algorithms in fault diagnosis under small sample sizes and varying operating conditions, and can diagnose the state of vehicle components under complex operating conditions, providing a new method for fault diagnosis of vehicles operating in complex environments. Attached Figure Description

[0031] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0032] Figure 1 This is a schematic diagram of the method flow according to an embodiment of the present invention;

[0033] Figure 2 This is a schematic diagram of the correlation coefficient thermodynamic matrix of four operating parameters according to an embodiment of the present invention;

[0034] Figure 3 This is a schematic diagram of the k-value curve determined by the elbow method in an embodiment of the present invention.

[0035] Figure 4 This is a schematic diagram of the k-value curve determined by the average profile coefficient method in an embodiment of the present invention.

[0036] Figure 5 This is a schematic diagram of the clustering visualization results of an embodiment of the present invention. Detailed Implementation

[0037] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0038] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0039] Example 1

[0040] like Figure 1 As shown, this embodiment provides a vehicle fault diagnosis method based on k-means multi-condition clustering with correlation analysis, including the following steps:

[0041] acquiring vehicle off-line working condition parameters;

[0042] acquiring the optimal cluster center number of k-means clustering based on the elbow method and the average silhouette coefficient method;

[0043] performing k-means clustering analysis on the vehicle off-line working condition parameters to obtain threshold ranges of vehicle component signal parameters corresponding to different working conditions;

[0044] acquiring real-time vehicle component signal parameters, comparing and analyzing the real-time vehicle component signal parameters with the threshold ranges of vehicle component signal parameters corresponding to different working conditions, and obtaining vehicle fault diagnosis results.

[0045] Before performing k-means clustering analysis on vehicle off-line working condition parameters, the method can further include performing correlation analysis on the vehicle off-line working condition parameters, and a Spearman correlation coefficient calculation formula is as follows:

[0046]

[0047] In the above formula, r i is a single time point working condition parameter data, is the average thereof in a period of time; s i is another single time point data of a working condition, is the average thereof in the same period of time; n is the number of time points in the period of time, and p is the Spearman correlation coefficient.

[0048] The correlation coefficient of each working condition parameter is calculated, the working condition parameters with large correlation coefficients are removed, and typical working condition parameters are selected according to actual engineering environment requirements, so that the redundant influence caused by too many working conditions can be solved.

[0049] The method can further include the following process of acquiring the optimal cluster center number (k value) of k-means clustering based on the elbow method and the average silhouette coefficient method:

[0050] The elbow method is used: by changing the size of k and calculating the within-cluster sum of squares (WCSS) under different k values, the optimal cluster number k, i.e., the first optimal cluster center number, is determined by finding the point at which the graph begins to present a straight line appearance, i.e., the point at which the WCSS reduction rate begins to slow down.

[0051] The silhouette coefficient method: the silhouette coefficient method is also used to determine the optimal cluster center number (k value) of k-means clustering. By calculating the silhouette coefficient (the silhouette coefficient ranges from -1 to 1) under different k values, if the silhouette coefficient of the current clustering is small, the current clustering method is not suitable; if the silhouette coefficient is large, the current clustering method is suitable, and the second optimal cluster center number is obtained.

[0052] By analyzing the first and second optimal cluster center numbers, the selection of the k value in the subsequent k-means clustering is comprehensively determined.

[0053] The process of performing k-means clustering analysis on vehicle off-line working condition parameters in Matlab to obtain the threshold range of vehicle component signal parameters corresponding to different working conditions includes:

[0054] The vehicle off-line working condition parameters are directly clustered in Matlab using the kmeans function, wherein the k value is set to the k value selected based on the elbow method and the average silhouette coefficient method, to obtain k clusters.

[0055] After clustering is completed, the signal data of each component of the vehicle is analyzed to obtain the threshold range of the component signal data corresponding to each cluster.

[0056] The above clustering clusters are visualized in Matlab to obtain k clustering centers and the threshold range of vehicle component signal parameters corresponding to different clusters (here, different clusters refer to different working conditions). The clustering centers and the obtained component signal thresholds are saved to a.mat folder.

[0057] The process of performing fault diagnosis on real-time vehicle signal data includes:

[0058] The distance of real-time vehicle component signal parameters from different clustering centers is obtained and sorted, and based on the clustering center with the shortest distance from the real-time vehicle component signal parameters, the working condition corresponding to the real-time vehicle component signal parameters is obtained. Specifically, according to the selected working condition data in real time, it is determined that the working condition at that moment is the working condition corresponding to the off-line condition.

[0059] According to the threshold range of vehicle component signal parameters under different working conditions, it is determined whether the vehicle component signal parameters at that moment under that working condition exceed the corresponding off-line threshold range. If so, it is determined that the component signal data at that moment under that working condition is abnormal.

[0060] Embodiment Two

[0061] This embodiment is based on the k-means clustering research on the off-line data of a vehicle working condition, and performs fault diagnosis on real-time vehicle components.

[0062] Extract off-line engine speed, throttle pedal opening, vehicle speed, and gear position data of a vehicle, and perform cross-correlation analysis on the four types of working condition data in Matlab, Figure 2The correlation coefficient matrix of the four working conditions is high, and the correlation of the four working conditions is high. Considering the actual driver operation, the throttle pedal and steering wheel rotation signal parameters are selected for k-means working condition clustering.

[0063] Since the working condition environment of this part of offline data is uncertain, the k value of k-means clustering is also uncertain, so the elbow method and average silhouette coefficient method are used to determine the k value of the final k-means clustering:

[0064] Elbow method: The elbow method is used to determine the optimal number of clustering centers (k value) of k-means clustering. It changes the size of k and calculates the within-cluster sum of squares (WCSS) under different k values. By finding the point where the graph starts to appear straight, that is, the point where the WCSS reduction rate starts to slow down, the optimal number of clusters k is determined.

[0065] Average silhouette coefficient method: The silhouette coefficient method is also used to determine the optimal number of clustering centers (k value) of k-means clustering. By calculating the silhouette coefficient under different k values (the silhouette coefficient ranges from -1 to 1), if the silhouette coefficient of the current clustering is small, the current clustering method is not suitable; if the silhouette coefficient is large, the current clustering method is suitable.

[0066] Figure 3 The k value curve determined by the elbow method, Figure 4 The k value curve determined by the average silhouette coefficient method, combined with engineering practice, the final k value of 9 is more in line with reality. Therefore, the number of working conditions determined by the offline data of the vehicle is 9, and k = 9 in k-means clustering.

[0067] In Matlab, the offline throttle pedal opening and steering wheel rotation are k-means clustered, k = 9, the clustering centers are printed and the clustering results are visualized, Figure 5 The clustering visualization result, Table 1 is the specific clustering center point (i.e. different working conditions).

[0068] Table 1

[0069] Cluster Cluster center value Working condition Cluster1 (0.0913，0.1274) Working condition 1 Cluster2 (36.7068，53.4096) Working condition 2 Cluster3 (30.0004，3.4425) Working condition 3 Cluster4 (25.8997，24.4463) Working condition 4 Cluster5 (5.3595，24.1276) Working condition 5 Cluster6 (7.7712，74.5556) Working condition 6 Cluster7 (81.7836，11.0680) Working condition 7 Cluster8 (8.0127，46.4623) Working condition 8 Cluster9 (12.8437，2.3641) Working condition 9

[0070] Through offline data analysis, the range of vehicle component parameter signals corresponding to different working conditions is obtained.

[0071] The following real-time signal data is used for fault diagnosis:

[0072] First, the steering wheel angle and the accelerator pedal opening at a certain moment in real time are subjected to working condition determination, and the working condition is determined according to the nearest clustering center obtained offline, and then whether the signal data of the vehicle components in the working condition exceeds the threshold range obtained offline is identified, and if it exceeds, it is determined to be abnormal, and Table 2 is the identification result of part of the real-time fault diagnosis.

[0073] Table 2

[0074]

[0075] Example Three

[0076] The embodiment provides an electronic device, comprising a memory and a processor; the memory is used for storing a program; the processor is used for executing the program, and realizing each step of the k-means multi-working condition clustering vehicle fault diagnosis method based on correlation analysis.

[0077] Example Four

[0078] The embodiment provides a readable storage medium, and a computer program is stored on the readable storage medium; when the computer program is executed by a processor, each step of the k-means multi-working condition clustering vehicle fault diagnosis method based on correlation analysis is realized.

[0079] The above is only a preferred specific embodiment of the present application, but the protection scope of the present application is not limited to this, any person skilled in the art can easily think of changes or replacements within the technical range disclosed in the present application, which should be covered in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A vehicle fault diagnosis method based on correlation analysis and k-means multi-working-condition clustering, characterized in that, The method comprises the following steps: acquiring vehicle off-line working condition parameters; acquiring the optimal cluster center number of k-means clustering based on the elbow method and the average silhouette coefficient method; performing k-means clustering analysis on the vehicle off-line working condition parameters to obtain the threshold range of vehicle component signal parameters corresponding to different working conditions; acquiring real-time vehicle component signal parameters, comparing and analyzing the real-time vehicle component signal parameters with the threshold range of vehicle component signal parameters corresponding to different working conditions, and obtaining vehicle fault diagnosis results; The process of acquiring the optimal cluster center number of k-means clustering based on the elbow method and the average silhouette coefficient method comprises: acquiring the first optimal cluster center number of k-means clustering based on the elbow method, acquiring the second optimal cluster center number of k-means clustering based on the average silhouette coefficient method, and comparing and analyzing the first optimal cluster center number and the second optimal cluster center number based on engineering practice to determine the final optimal cluster center number of k-means clustering; The process of acquiring the first optimal cluster center number of k-means clustering based on the elbow method comprises: acquiring the within-group sum of squares under different cluster center numbers based on the elbow method, drawing a corresponding curve graph, and determining the first optimal cluster center number based on the point at which the curve graph begins to present a straight line appearance.

2. The method of claim 1, wherein, before performing k-means clustering analysis on the vehicle off-line working condition parameters, the method comprises: performing correlation analysis on the vehicle off-line working condition parameters based on the Spearman correlation coefficient, and selecting typical working condition parameters.

3. The method of claim 1, wherein, the process of acquiring the second optimal cluster center number of k-means clustering based on the average silhouette coefficient method comprises: acquiring the silhouette coefficient under different cluster center numbers based on the average silhouette coefficient method, drawing a corresponding curve graph, and determining the second optimal cluster center number by analyzing the silhouette coefficient in the curve graph.

4. The method of claim 1, wherein, the process of performing k-means clustering analysis on the vehicle off-line working condition parameters to obtain the threshold range of vehicle component signal parameters corresponding to different working conditions comprises: performing k-means clustering analysis on the vehicle off-line working condition parameters to determine that the number of working conditions is the optimal cluster center number; after clustering is completed, analyzing the signal parameters of each component of the vehicle to obtain the threshold range of vehicle component signal parameters corresponding to different working conditions.

5. The method of claim 4, wherein, the process of comparing and analyzing the real-time vehicle component signal parameters with the threshold range of vehicle component signal parameters corresponding to different working conditions comprises: acquiring the distance of real-time vehicle component signal parameters from different cluster centers and sorting, obtaining the working condition corresponding to the real-time vehicle component signal parameters based on the cluster center with the shortest distance from the real-time vehicle component signal parameters, and obtaining vehicle fault diagnosis results when the real-time vehicle component signal parameters exceed the threshold range of vehicle component signal parameters under the corresponding working condition. comprise: a memory and a processor; ​ ​ 6. An electronic device, comprising: ​ ​ The memory is configured to store a program. The processor is configured to execute the program to implement each step of the k-means multi-working-condition clustering vehicle fault diagnosis method based on correlation analysis according to any one of claims 1-5.

7. A readable storage medium, having stored thereon a computer program, characterized in that, The computer program, when executed by a processor, implements each step of the k-means multi-working-condition clustering vehicle fault diagnosis method based on correlation analysis according to any one of claims 1-5.

Images