Threshold setting method and system under different working conditions based on MVMD denoising, KNN classification and adaptive threshold fusion

Abstract

The present invention discloses a method and system for setting thresholds under different working conditions based on the fusion of MVMD noise reduction, KNN classification and adaptive threshold. Obtain n groups of vibration signal data of the equipment at different times; perform MVMD decomposition to obtain multiple basic modal components IMF; calculate the correlation coefficient between each IMF and the original signal; add the basic modal components IMF whose correlation coefficient is greater than the correlation coefficient threshold to obtain a reconstructed signal; calculate the effective value and kurtosis of the velocity of each group of reconstructed signals; set the KNN classification parameters, select the parameter k with the highest accuracy; calculate the thresholds under different working conditions; determine the final threshold according to the standard value. The KNN classification method is used to separate the characteristic values ​​of different operating conditions of the equipment, and then the adaptive threshold method is used to calculate the thresholds of different working conditions. Different from the defect of the traditional method that cannot set the threshold according to the working condition and cause a pre-alarm.

Description

Technical Field

[0001] The present application relates to the field of vibration monitoring technology, and in particular to a threshold setting method and system under different working conditions based on the fusion of MVMD noise reduction, KNN classification and adaptive threshold, which can improve monitoring accuracy. Background Art

[0002] During the operation of equipment, equipment is often damaged due to the inability to detect faults in time, resulting in huge economic losses. Existing vibration monitoring technology relies on vibration characteristic values ​​for fault warning, but the key lies in the setting of thresholds. Different equipment and working conditions require different threshold settings.

[0003] In existing vibration monitoring technologies, the VMD decomposition algorithm has poor performance due to the inaccurate number of decomposed modes. The MVMD decomposition algorithm extracts all IMFs in a continuous manner. Compared with VMD, this method does not require knowing the number of modes and has lower computational complexity, and removes noise interference in the signal.

[0004] However, these existing technologies have inaccurate threshold settings under different working conditions, resulting in false alarms. Summary of the invention

[0005] The purpose of this application is to provide a threshold setting method and system under different working conditions based on the fusion of MVMD noise reduction, KNN classification and adaptive threshold, so as to improve the accuracy of equipment fault warning and reduce false alarms.

[0006] The embodiment of the present application is implemented as follows:

[0007] A threshold setting method under different working conditions based on the fusion of MVMD denoising, KNN classification and adaptive threshold, characterized by comprising the following steps:

[0008] S1: Obtain n groups of vibration signal data of the device at different times;

[0009] S2: Perform MVMD decomposition on the vibration signal data of each group to obtain multiple basic modal components IMF;

[0010] S3: Calculate the correlation coefficient between each IMF and the original signal;

[0011] S4: Add the fundamental mode components IMF whose correlation coefficients are greater than the set correlation coefficient threshold to obtain a reconstructed signal;

[0012] S5: Calculate the velocity effective value and kurtosis of each reconstructed signal in n groups of data;

[0013] S6: Set the range of the classification parameter K of the KNN algorithm;

[0014] S7: taking the effective value of the speed as the first vector X1, separating the training sample train_data_vel and the test sample test_data_vel from the first vector X1, and setting the corresponding working condition labels; taking the kurtosis as the second vector X2, separating the training sample train_data_kur and the test sample test_data_kur from the second vector X2, and setting the corresponding working condition labels;

[0015] S8: Get the accuracy of the test samples under different parameters K;

[0016] S9: Select the classification parameter K with the highest accuracy;

[0017] S10: Obtain m groups of vibration signal data of the device at different times;

[0018] S11: repeat steps S2 to S5, calculate the effective value of the velocity of the m sets of data as the third vector Y1; calculate the kurtosis of the m sets of data as the fourth vector Y2;

[0019] S12: taking the third vector Y1 and the fourth vector Y2 as test samples, using KNN algorithm to classify the group of data under the condition of parameter K, and separating the effective value and kurtosis of speed under different working conditions;

[0020] S13: The speed effective value threshold and kurtosis threshold under different working conditions are calculated using an adaptive threshold algorithm based on a box plot and the 3σ principle.

[0021] In the above technology, in step S4, the correlation coefficient is greater than the set correlation coefficient threshold value of 0.3.

[0022] In the above technology, step S6 sets the classification parameters K of multiple KNN algorithms based on a grid search method or a random search method.

[0023] In the above technology, in step S6, the KNN classification parameter K is set to [3, 5, 7, 9, 11].

[0024] In the above technology, step S8: for each K value, use the training sample data to train the KNN algorithm model, and assign each test sample to the same category as its nearest K neighbors; for each K value, calculate the accuracy of the KNN algorithm model on the test sample, and obtain the accuracy of the test sample under different parameters K; the accuracy is the ratio of the number of correctly classified test samples to the total number of test samples.

[0025] In the above technology, the steps of the adaptive threshold algorithm of the box plot and the 3σ principle in step S13 are as follows: first, the effective value of the speed and the kurtosis data under different working conditions are arranged from small to large, and the upper and lower quartile data are selected using a box plot, and the average value and 3 standard deviations of the remaining data of the effective value of the speed and the kurtosis are calculated, and the respective average values ​​and standard deviations are added to obtain the thresholds under different working conditions.

[0026] In the above technology, in step S13, the average value and three times the standard deviation of the speed effective value and the kurtosis are added together to obtain the threshold value under different working conditions.

[0027] The above technology also includes step S14: after obtaining the threshold values ​​under different working conditions, determining the final threshold value according to the standard value.

[0028] In the above technology, in step S14, under the national standard value, when the threshold value of the speed effective value is less than 2.8, the alarm line is 4.5; when the threshold value of the speed effective value is greater than 2.8 and less than 4.5, the alarm line is 1.8 times the threshold value of the speed effective value; when the threshold value of the speed effective value is greater than 4.5 and less than 11.2, the maximum value of 4 and 0.6 times the threshold value of the speed effective value is compared, and the alarm line is the sum of the threshold value and the maximum value of the speed effective value.

[0029] A fault diagnosis and identification system, characterized in that it comprises:

[0030] A signal acquisition unit, used to acquire n groups of vibration signal data of the device at different times;

[0031] The signal decomposition and reconstruction unit is used to perform MVMD decomposition on each group of vibration signal data to obtain multiple basic modal components IMF; calculate the correlation coefficient between each IMF and the original signal; add the basic modal components IMF whose correlation coefficient is greater than the correlation coefficient threshold to obtain the reconstructed signal; calculate the velocity effective value and kurtosis of each group of reconstructed signals;

[0032] A classification parameter setting unit, used to set a KNN classification parameter K;

[0033] A sample separation unit is used to separate training samples and test samples and set corresponding working condition labels;

[0034] The accuracy test unit is used to test the accuracy of the test samples under different parameters K and select the parameter k with the highest accuracy;

[0035] The classification and adaptive threshold calculation unit is used to classify the test samples, separate the effective value and kurtosis of the speed under different working conditions, and calculate the thresholds under different working conditions;

[0036] The threshold determination unit is used to determine a final threshold value according to a standard value.

[0037] In the above technology, the decomposition and reconstruction signal unit can extract all IMFs in a continuous manner.

[0038] In the above technology, the sample separation unit can separate the training samples and test samples of the effective value of velocity and kurtosis.

[0039] In the above technology, the accuracy test unit can test the accuracy of the test samples under different parameters K.

[0040] In the above technology, the classification and adaptive threshold calculation unit calculates the threshold based on the box plot and the 3σ principle.

[0041] A computer-readable storage medium, characterized in that a computer program is stored thereon, and when the computer program is executed, it is used to implement the above-mentioned threshold setting method under different working conditions based on the fusion of MVMD denoising, KNN classification and adaptive threshold.

[0042] The beneficial effects of this application are:

[0043] By combining advanced signal processing technology (MVMD noise reduction), classic machine learning algorithm (KNN classification) and adaptive threshold setting method, the present invention has significantly improved accuracy, robustness, adaptability and computational efficiency compared with the existing technology, thus providing a better solution for equipment fault warning.

[0044] The KNN classification method is used to separate the characteristic values ​​of different operating conditions of the equipment, and then the adaptive threshold method is used to calculate the thresholds of different operating conditions. This is different from the traditional method that cannot set thresholds according to operating conditions and cause early alarms.

[0045] MVMD noise reduction: The MVMD (modified variational modal decomposition) algorithm can more accurately extract useful features from vibration signals while removing noise interference. Compared with the VMD (variational modal decomposition) algorithm, the MVMD algorithm can extract all intrinsic modal functions (IMFs) in a continuous manner without pre-specifying the number of modes, reducing the performance degradation caused by inaccurate number of modes. On this basis, the KNN (K-nearest neighbor) algorithm is used to classify the eigenvalues ​​under different working conditions, which can more accurately and efficiently distinguish the operating status of the equipment under different working conditions. It is simple, efficient, convenient and feasible.

[0046] At the same time, the adaptive threshold algorithm based on the box plot and 3σ principle can dynamically set the threshold according to the actual distribution of the data instead of using a fixed threshold. This method can better adapt to changes under different working conditions, dynamically adjust the threshold according to the actual operating conditions, and reduce false alarms caused by improper threshold setting.

[0047] In terms of improving computational efficiency: compared with the VMD algorithm, the MVMD algorithm has lower computational complexity, can process data faster, and improves computational efficiency.

[0048] Enhanced robustness: Combining MVMD denoising and KNN classification can better handle noise and outliers in the data and enhance the robustness of the system.

[0049] Adaptability: This method can adapt to the needs of different equipment and different working conditions, and has good versatility and adaptability.

[0050] Reduce economic losses: By timely and accurately warning of equipment failures, economic losses caused by equipment damage can be reduced.

[0051] Improve maintenance efficiency: Accurate fault warning can help maintenance personnel locate problems more quickly and improve maintenance efficiency.

[0052] Data-driven decision-making: This method is based on historical data-driven decision-making and does not need to rely on the number of modes. It provides a scientific threshold setting method, making fault warning more reliable. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0054] Figure 1 This is a flow chart of a threshold setting method under different working conditions based on the fusion of MVMD denoising, KNN classification and adaptive thresholding in an embodiment of the present application.

[0055] Figure 2 This is a set of time domain diagrams of original signals of the dephosphorization pump in Example 2 of the present application.

[0056] Figure 3 This is a time domain diagram of the reconstructed signal according to Example 2 of the present application.

[0057] Figure 4 This is an illustration of the final threshold settings for the speed effective value operating condition and the stationary condition in Example 2 of the present application.

[0058] Figure 5 This is an illustration of the final threshold setting for the kurtosis operating condition and the static condition in Example 2 of the present application. DETAILED DESCRIPTION

[0059] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and shown in the drawings here can be arranged and designed in various different configurations.

[0060] Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the present application for which protection is sought, but merely represents selected embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present application.

[0061] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0062] In the description of the present application, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc. indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, or the positions or positional relationships in which the product of the application is usually placed when in use. They are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific position, be constructed and operated in a specific position, and therefore cannot be understood as a limitation on the present application. In addition, the terms "first", "second", "third", etc. are only used to distinguish the description, and cannot be understood as indicating or implying relative importance.

[0063] In addition, the terms "horizontal", "vertical", "overhanging" and the like do not mean that the components are required to be absolutely horizontal or overhanging, but can be slightly tilted. For example, "horizontal" only means that its direction is more horizontal than "vertical", and does not mean that the structure must be completely horizontal, but can be slightly tilted.

[0064] In the description of this application, it should also be noted that, unless otherwise clearly specified and limited, the terms "set", "install", "connect", and "connect" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two elements. For ordinary technicians in this field, the specific meanings of the above terms in this application can be understood according to specific circumstances.

[0065] In the present application, unless otherwise clearly specified and limited, a first feature being "above" or "below" a second feature may include that the first and second features are in direct contact, or may include that the first and second features are not in direct contact but are in contact through another feature between them. Moreover, a first feature being "above", "above" and "above" a second feature includes that the first feature is directly above and obliquely above the second feature, or simply indicates that the first feature is higher in level than the second feature. A first feature being "below", "below" and "below" a second feature includes that the first feature is directly below and obliquely below the second feature, or simply indicates that the first feature is lower in level than the second feature.

[0066] The features and performance of the present application are further described in detail below in conjunction with the embodiments.

[0067] Example 1

[0068] Figure 1 The flowchart of the analysis algorithm of the present invention is based on the threshold setting method under different working conditions that integrates MVMD denoising, KNN classification and adaptive threshold, which specifically includes the following steps:

[0069] S1: Obtain n groups of vibration signal data of the device at different times;

[0070] S2: Perform MVMD decomposition on the vibration signal data of each group to obtain multiple basic modal components IMF;

[0071] S3: Calculate the correlation coefficient between each IMF and the original signal;

[0072] S4: Add the fundamental mode components IMF with correlation coefficient greater than 0.3 to obtain the reconstructed signal;

[0073] S5: Calculate the velocity effective value and kurtosis of each group of reconstructed signals;

[0074] S6: Set the KNN classification parameter K to 3, 5, 7, 9, 11. You can choose 2, 3, 4, or 5 of these five empirical values. You can also reduce the number based on efficiency.

[0075] S7: Take the first vector X1 of the effective value of the speed, separate the training sample train_data_vel and the test sample test_data_vel from the first vector X1, and set the corresponding labels train_label_vel and test_label_vel; take the kurtosis as the second vector X2, separate the training sample train_data_kur and the test sample test_data_kur from the second vector X2, and set the corresponding labels train_label_kur and test_label_kur;

[0076] S8: Test the accuracy of test samples under different parameters K;

[0077] S9: Select the parameter k with the highest accuracy; after the model is trained, use new data for testing. The model will automatically identify different working conditions and calculate the threshold.

[0078] S10: Obtain m groups of vibration signal data of the device at different times;

[0079] S11: repeat steps S2 to S5, calculate the effective value of the speed of this set of data as the third vector Y1; calculate the kurtosis of this set of data as the fourth vector Y2;

[0080] S12: taking the third vector Y1 and the fourth vector Y2 as test samples, using the KNN algorithm to classify the group of data under the condition of parameter k, and separating the effective value and kurtosis of speed under different working conditions;

[0081] S13: using an adaptive threshold algorithm based on a box plot and the 3σ principle to calculate the thresholds under different working conditions;

[0082] S14: Determine the final threshold value according to the national standard. For example, when the threshold value of the effective speed value is less than 2.8, the alarm line is 4.5; when the threshold value of the effective speed value is greater than 2.8 and less than 4.5, the alarm line is 1.8 times the threshold value of the effective speed value; when the threshold value of the effective speed value is greater than 4.5 and less than 11.2, compare the maximum value of 4 and 0.6 times the threshold value of the effective speed value, and the alarm line is the sum of the threshold value of the effective speed value and the maximum value.

[0083] Example 2

[0084] Obtain the vibration signal data of a phosphorus removal pump on site from May to July. Figure 2 A set of original signal time domain diagrams of the phosphorus removal pump, with the horizontal axis in seconds s， The vertical axis unit is amplitude, and the unit is m / s² .

[0085] Set the maximum balance parameter , the double ascent step size is 0, and the tolerance of the convergence criterion is 10 -7 , according to step S2, MVMD decomposition is performed on each group of vibration signals to obtain a series of basic modal components IMF.

[0086] According to step S3, the correlation coefficient between the original vibration signal and the fundamental modal component is calculated.

[0087] According to step S4, the basic modal components with correlation coefficients greater than 0.3 are added to obtain a reconstructed signal, such as Figure 3 shown.

[0088] According to step S5, the effective value and kurtosis of each group of reconstructed signals are calculated.

[0089] According to step S6, the KNN classification parameter K is set to [3, 5, 7, 9, 11].

[0090] According to step S7, the effective value of the speed is used as the first vector X1, and the training sample train_data_vel and the test sample test_data_vel are separated from the first vector X1, and the corresponding labels train_label_vel and test_label_vel are set, wherein the running state label is 1 and the static state label is 0; the kurtosis is used as the second vector X2, and the training sample train_data_kur and the test sample test_data_kur are separated from the second vector X2, and the corresponding labels train_label_kur and test_label_kur are set, wherein the running state label is 1 and the static state label is 0;

[0091] According to step S8, the accuracy of the test samples under different parameters K is tested, as shown in Table 1.

[0092] Table 1 Test sample accuracy under different parameters K

[0093]

[0094] According to step S9, K=9 is selected as the KNN classification input parameter k; after the model is trained, new data is used for testing, and the model will automatically identify different working conditions and calculate the threshold.

[0095] According to step S10, the vibration signal data of the dephosphorization pump from July to August is obtained;

[0096] According to step S11, steps S2 to S5 are repeated to calculate the effective value of the speed of this set of data as the third vector Y1; the kurtosis of this set of data is calculated as the fourth vector Y2;

[0097] According to step S12, vectors Y1 and Y2 are used as test samples, and the KNN algorithm is used to classify the group of data under the condition of parameter k, and the effective value of the speed under the running condition data_run_vel and the effective value of the speed under the stationary condition data_stop_vel are separated; the kurtosis under the running condition data_run_kur and the kurtosis under the stationary condition data_stop_kur are separated;

[0098] According to step S13, the box plot and 3 σThe principle of adaptive threshold algorithm is used to calculate the threshold under different working conditions. The specific method is: first arrange the data from small to large, use the box plot to select the upper and lower quartile data, calculate the mean and standard deviation of the remaining data, and add the mean and 3 times the standard deviation to get the threshold.

[0099] According to step S14, the international standard and the calculated threshold are compared and judged, and the final threshold of the speed effective value operating condition and the static condition is set as follows Figure 4 As shown, the horizontal axis unit is seconds, and the vertical axis unit is mm / s. Before about 795 seconds, it is a running condition, and the threshold is set to 4.5. From about 790 seconds on, it is a static condition, and the static condition threshold is set to 0.55.

[0100] The final thresholds for the kurtosis operating condition and the stationary condition are set as follows: Figure 5 As shown, the horizontal axis unit is seconds. The first 810 seconds or so are running conditions, and the threshold is set to 8.2. From about 810 seconds onwards, it is a stationary condition, and the stationary condition threshold is set to 5.

[0101] The embodiments described above are part of the embodiments of the present application, rather than all of the embodiments. The detailed description of the embodiments of the present application is not intended to limit the scope of the present application for protection, but merely represents the selected embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application.

Claims

1. A threshold setting method under different working conditions based on the fusion of MVMD denoising, KNN classification and adaptive threshold, characterized in that: The following steps are involved: S1: Obtain n groups of vibration signal data of the device at different times; S2: Perform MVMD decomposition on the vibration signal data of each group to obtain multiple basic modal components IMF; S3: Calculate the correlation coefficient between each IMF and the original signal; S4: Add the fundamental mode components IMF whose correlation coefficients are greater than the set correlation coefficient threshold to obtain a reconstructed signal; S5: Calculate the velocity effective value and kurtosis of each reconstructed signal in n groups of data; S6: Set the range of the classification parameter K of the KNN algorithm; S7: taking the effective value of the speed as the first vector X1, separating the training sample train_data_vel and the test sample test_data_vel from the first vector X1, and setting the corresponding working condition labels; taking the kurtosis as the second vector X2, separating the training sample train_data_kur and the test sample test_data_kur from the second vector X2, and setting the corresponding working condition labels; S8: Get the accuracy of the test samples under different parameters K; S9: Select the classification parameter K with the highest accuracy; S10: Obtain m groups of vibration signal data of the device at different times; S11: repeat steps S2 to S5, calculate the effective value of the velocity of the m sets of data as the third vector Y1; calculate the kurtosis of the m sets of data as the fourth vector Y2; S12: taking the third vector Y1 and the fourth vector Y2 as test samples, using KNN algorithm to classify the group of data under the condition of parameter K, and separating the effective value and kurtosis of speed under different working conditions; S13: The speed effective value threshold and kurtosis threshold under different working conditions are calculated using an adaptive threshold algorithm based on a box plot and the 3σ principle.

2. The threshold setting method under different working conditions based on the fusion of MVMD denoising, KNN classification and adaptive threshold according to claim 1 is characterized in that: Step S6 sets the classification parameters K of multiple KNN algorithms based on a grid search method or a random search method.

3. The threshold setting method under different working conditions based on the fusion of MVMD denoising, KNN classification and adaptive threshold according to claim 1 is characterized in that: In step S6, the KNN classification parameter K is set to at least three values ​​among 3, 5, 7, 9, and 11.

4. The threshold setting method under different working conditions based on the fusion of MVMD denoising, KNN classification and adaptive threshold according to claim 1 is characterized in that: Step S8: For each K value, use the training sample data to train the KNN algorithm model, and assign each test sample to the same category as its nearest K neighbors; for each K value, calculate the accuracy of the KNN algorithm model on the test sample, and obtain the accuracy of the test sample under different parameters K; the accuracy is the ratio of the number of correctly classified test samples to the total number of test samples.

5. The threshold setting method under different working conditions based on the fusion of MVMD denoising, KNN classification and adaptive threshold according to claim 1 is characterized in that: The steps of the adaptive threshold algorithm of the box plot and 3σ principle in step S13 are as follows: first, arrange the speed effective value and kurtosis data under different working conditions from small to large, use the box plot to select the upper and lower quartile data, calculate the mean and standard deviation of the remaining data of the speed effective value and kurtosis, and add up their respective mean and standard deviation to obtain the threshold under different working conditions.

6. The threshold setting method under different working conditions based on the fusion of MVMD denoising, KNN classification and adaptive threshold according to claim 5 is characterized in that: In step S13, the mean values ​​and three times the standard deviations of the speed effective value and the kurtosis are added together to obtain the threshold values ​​under different working conditions.

7. The threshold setting method under different working conditions based on the fusion of MVMD denoising, KNN classification and adaptive threshold according to claim 1 is characterized in that: The method further includes step S14: after obtaining the threshold values ​​under different working conditions, determining the final threshold value according to the standard value.

8. The threshold setting method under different working conditions based on the fusion of MVMD denoising, KNN classification and adaptive threshold according to claim 7 is characterized in that: In step S14, under the national standard value, when the threshold of the speed effective value is less than 2.8, the alarm line is 4.5; when the threshold of the speed effective value is greater than 2.8 and less than 4.5, the alarm line is 1.8 times the threshold of the speed effective value; when the threshold of the speed effective value is greater than 4.5 and less than 11.2, the maximum value of 4 and 0.6 times the threshold of the speed effective value is compared, and the alarm line is the sum of the threshold of the speed effective value and the maximum value.

9. A threshold setting system under different working conditions based on the fusion of MVMD noise reduction, KNN classification and adaptive threshold, characterized in that: include: A signal acquisition unit, used to acquire n groups of vibration signal data of the device at different times; The signal decomposition and reconstruction unit is used to perform MVMD decomposition on each group of vibration signal data to obtain multiple basic modal components IMF; calculate the correlation coefficient between each IMF and the original signal; add the basic modal components IMF whose correlation coefficient is greater than the correlation coefficient threshold to obtain the reconstructed signal; calculate the velocity effective value and kurtosis of each group of reconstructed signals; A classification parameter setting unit, used to set a KNN classification parameter K; A sample separation unit is used to separate training samples and test samples and set corresponding working condition labels; The accuracy test unit is used to test the accuracy of the test samples under different parameters K and select the parameter k with the highest accuracy; The classification and adaptive threshold calculation unit is used to classify the test samples, separate the effective value and kurtosis of the speed under different working conditions; calculate the thresholds under different working conditions; and use the adaptive threshold algorithm based on the box plot and the 3σ principle to calculate the effective value threshold and kurtosis threshold of the speed under different working conditions; The threshold determination unit is used to determine the final threshold value according to the standard value to perform fault diagnosis and identification.

10. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the computer program is executed, it is used to implement the threshold setting method under different working conditions based on the fusion of MVMD denoising, KNN classification and adaptive threshold as described in any one of claims 1 to 8.

Images