Industrial equipment fault diagnosis method, system and storage medium based on multi-view expansion statistical features

Abstract

The present invention provides an industrial equipment fault diagnosis method, system, and storage medium based on multi-view expansion statistical features, which relate to the technical field of industrial equipment fault diagnosis. The method aims to solve the problems that existing methods are easily contaminated by noise, and that the extracted features contain redundant and irrelevant features, resulting in long calculation time, low calculation efficiency, and high hardware computing power requirements. The method comprises the following steps: step 1, obtaining a time series data set of the operating fault status of the industrial equipment; step 2, preprocessing the time series data using an unsupervised differential expansion mapping preprocessor; step 3, constructing multiple filtering feature selectors at different storage ratios to obtain multi-view feature subset vectors and a combined feature selector; step 4, constructing multiple classifiers, training the multiple classifiers based on the multi-view feature subsets, and constructing a combined classifier; and step 5, using the combined feature selector and the combined classifier to perform fault diagnosis on the equipment operating status.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial equipment fault diagnosis, and in particular to an industrial equipment fault diagnosis method, system and storage medium based on multi-view expansion statistical features. Background Art

[0002] With the development of contemporary industry and science and technology, the overall trend in modern equipment development is towards complexity, intelligence, and automation. The value of equipment is increasing, and companies are investing more and more in it. To create sufficient economic value, many companies will put their equipment into full-load and continuous operation. As a result, many equipment suffers from problems such as frequent use, harsh operating conditions, high power, heavy loads, and excessive wear. If timely inspections and appropriate maintenance plans are not implemented, not only will the service life of the equipment be shortened, but it may also affect its normal operation, cause damage to the equipment, and even lead to serious safety accidents.

[0003] In order to monitor the status of industrial production and diagnose faults in industrial equipment, it is an urgent industrial need to use a detection system with multiple sensors to collect and store a large amount of equipment status sequence data. Improving equipment reliability by analyzing status sequence data is an urgent industrial need.

[0004] Feature-based time series classification methods, due to their simple architecture and ease of parallelization, can be applied to process a large number of state time series collected and stored by various sensor systems to diagnose faults in industrial equipment. However, existing technologies have many shortcomings, the most important of which is that they only increase feature diversity in a single way. When the feature scale reaches a certain level, it is difficult to obtain new features due to the limitations of professional and statistical knowledge. In order to flexibly improve feature diversity, differential mapping is widely used, but differential transformations are easily contaminated by noise, resulting in the annihilation of some useful features of the signal. On the other hand, the massive amount of features extracted contain redundant and irrelevant features, resulting in many shortcomings such as long calculation time, low computational efficiency, and high hardware computing power requirements. Summary of the Invention

[0005] The technical problems to be solved by the present invention are:

[0006] Existing methods are easily contaminated by noise, which causes some useful features of the signal to be annihilated, and the extracted features contain redundant and irrelevant features, resulting in long calculation time, low computational efficiency, and high requirements on hardware computing power.

[0007] The present invention is to solve the above technical problems using the following technical solutions:

[0008] The present invention provides an industrial equipment fault diagnosis method based on multi-view expansion statistical features, comprising the following steps:

[0009] Step 1: Obtain a time series dataset of the fault status of industrial equipment;

[0010] Step 2: Use the unsupervised differential expansion mapping preprocessor to preprocess the time series data, including the following steps:

[0011] Step 2-1, using a tracking differentiator with an adjustable filter factor to perform a first-order differential transformation on the time series data to obtain a first-order differential time series;

[0012] Step 2-2: Use a tracking differentiator with an adjustable filter factor to perform a first-order differential transformation on the time series data again to obtain a second-order differential time series;

[0013] Step 2-3: introducing an expansion mechanism to perform expansion mapping of multiple proportions on the time series data, the first-order differential time series, and the second-order differential time series;

[0014] Steps 2-4: extract the global statistical features of the sequences after expansion mapping and perform full splicing to obtain a diversity feature vector dataset;

[0015] Step 3: Construct multiple filtering feature selectors under different storage ratios, train the feature selectors, obtain multi-view feature subset vectors and feature selectors containing multiple feature indexes, and construct a combined feature selector based on the multiple feature selectors;

[0016] Step 4: Build multiple classifiers, train multiple classifiers based on multi-view feature subsets and build a combined classifier;

[0017] Step 5: Preprocess the operating status data of the industrial equipment to be diagnosed, use the trained combined feature selector for feature selection, and use the trained combined classifier for fault diagnosis. By introducing a voting mechanism, the output results of the classifier on each view are integrated to finally obtain the fault diagnosis results of the equipment operating status.

[0018] Furthermore, the industrial equipment described in step 1 includes one or more general equipment, special equipment or special equipment in the fields of machinery, electrical electronics, computer networks, automobiles, chemicals or energy, and the equipment operating status information includes one or more of temperature, pressure, speed, torque, sound, vibration or electrical signals.

[0019] Furthermore, the functional implementation process of the tracking differentiator with adjustable filter factor is as follows:

[0020] The calculation formula for the first column x1[1] of the matrix x1 and the first column x2[1] of the matrix x2 is:

[0021]

[0022] For i=1,2,…,m, increasing in sequence:

[0023] y i =x1[i]-s[i]+khx2[i];

[0024] Where s is the time series of the input tracking differentiator;

[0025] If |y i |>k 2 h 2 r, then otherwise,

[0026] If |a|>rhk, then fh i = -r·sign(a), otherwise,

[0027] The formula for calculating the i+1th column x1[i+1] of the matrix x1 and the i+1th column x2[i+1] of the matrix x2 is:

[0028]

[0029] The differential time series is transformed by a tracking differentiator with an adjustable filter factor to obtain the following differential time series: Where r is the speed factor and h is the step size.

[0030] Furthermore, the implementation process of the expansion mechanism is as follows:

[0031] For the filter factor k=i, for the time series s i Perform downsampling to obtain the downsampling time subsequence at that moment in It represents the largest integer less than (mk) / d, where m is the length of the time series and d is the sampling interval, with the expansion ratio as the sampling interval;

[0032] Complete the iterative cycle process of k=1,2,…,d in sequence, and splice the time subsequences obtained by downsampling each moment in sequence to obtain the final expansion mapping sequence

[0033] Furthermore, the global statistical features described in steps 2-4 include one or more of mean, sum of squares, range, standard deviation, arithmetic mean, median, coefficient of variation, variance, standard deviation, approximate entropy, autoregressive model coefficient, Fourier coefficient of one-dimensional discrete Fourier transform, kurtosis of absolute Fourier transform spectrum or autocorrelation coefficient.

[0034] Furthermore, in the training process of the feature selector in step 3, the correlation measure between the feature vector and the label is used as the feature quality score, and the top-ranked features under different storage ratios are used as multi-view features; the correlation measure between the feature vector and the label adopts the F statistic in variance analysis.

[0035] Furthermore, the classifier adopts a ridge regression classifier with cross validation.

[0036] Furthermore, the voting mechanism adopts a hard voting strategy.

[0037] An industrial equipment fault diagnosis system, comprising:

[0038] The data acquisition module is used to obtain the operating status information of the industrial equipment to be diagnosed and arrange it according to the acquisition time sequence to construct a time series data set;

[0039] A data processing module is used to preprocess the time series data using an unsupervised differential expansion mapping preprocessor to obtain a feature vector dataset with diversity;

[0040] Feature selector construction and training module, used to construct multiple filtering feature selectors under different storage ratios, train the feature selectors, obtain multi-view feature subset vectors and feature selectors containing multiple feature indexes, and construct a combined feature selector based on multiple feature selectors;

[0041] Classifier construction and training module, used to build multiple classifiers, train multiple classifiers based on multi-view feature subsets and build a combined classifier;

[0042] The fault diagnosis module is used to preprocess the operating status data of the industrial equipment to be diagnosed, use a combined feature selector for feature selection, and use a combined classifier for fault diagnosis. By introducing a voting mechanism to integrate the output results of the classifiers on each view, the fault diagnosis results of the equipment operating status are finally obtained;

[0043] The communication module is used to connect the above modules and transmit data.

[0044] A storage medium, wherein the computer-readable storage medium stores a computer program, wherein the computer program is configured to implement the steps of the industrial equipment fault diagnosis method based on multi-view expansion statistical features described in any one of the above technical solutions when called by a processor.

[0045] Compared with the prior art, the present invention has the following beneficial effects:

[0046] This invention processes collected time series data on equipment operating status by combining a tracking differentiator with an expansion mechanism to efficiently enhance feature diversity. Furthermore, a tracking differentiator with an adjustable filter factor performs differential transformation on the time series, resulting in improved noise suppression. The invention employs multiple filtering feature selection methods at varying storage ratios to generate multi-view features, improving feature stability and removing redundant and irrelevant features. Multi-view features are then used with a linear classifier to achieve efficient and high-performance equipment fault diagnosis. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 Flowchart of a method for diagnosing industrial equipment faults based on multi-view expansion statistical features in an embodiment of the present invention;

[0048] Figure 2 Flowchart of the training process of the feature selector in an embodiment of the present invention;

[0049] Figure 3 This is a flow chart of the functional implementation process of the adjustable filter factor tracking differentiator in an embodiment of the present invention;

[0050] Figure 4 Schematic diagram of a fault diagnosis system in an embodiment of the present invention;

[0051] Figure 5 is a scatter plot of accuracy in an embodiment of the present invention;

[0052] Figure 6 This is a radar chart comparing the scope of problems applicable to the fault diagnosis method of the present invention in an embodiment of the present invention. DETAILED DESCRIPTION

[0053] In order to enable those skilled in the art to better understand the present invention, exemplary embodiments or examples of the present invention will be described below with reference to the accompanying drawings. Obviously, the described embodiments or examples are only some of the embodiments or examples of the present invention, and not all of them. Based on the embodiments or examples of the present invention, all other embodiments or examples obtained by those skilled in the art without creative work should fall within the scope of protection of the present invention.

[0054] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0055] Example 1

[0056] like Figure 1 As shown, the present invention provides an industrial equipment fault diagnosis method based on multi-view expansion statistical features, comprising the following steps:

[0057] Step 1.1: Obtain the data of the operating status information of the industrial equipment to be diagnosed, and arrange it according to the acquisition time sequence to construct a real-valued unary time series set of equal length and equal interval Where y is the number of time series samples to be diagnosed;

[0058] The industrial equipment includes one or more general equipment, special equipment or specialized equipment in the fields of machinery, electrical electronics, computer networks, automobiles, chemicals or energy, and the equipment operating status information includes one or more of temperature, pressure, speed, torque, sound, vibration or electrical signals.

[0059] Step 1.2: Use the unsupervised differential expansion mapping preprocessor to preprocess the time series data, including the following steps:

[0060] Step 1.2.1: Construct a tracking differentiator with adjustable filter factor, such as Figure 3 As shown, the function implementation process is:

[0061] The calculation formula for the first column x1[1] of the matrix x1 and the first column x2[1] of the matrix x2 is:

[0062]

[0063] For i=1,2,…,m, increasing in sequence:

[0064] y i =x1[i]-s[i]+khx2[i];

[0065] Where s is the time series of the input tracking differentiator;

[0066] If |y i |>k 2 h 2 r, then otherwise,

[0067] If |a|>rhk, then fh i = -r·sign(a), otherwise,

[0068] The formula for calculating the i+1th column x1[i+1] of the matrix x1 and the i+1th column x2[i+1] of the matrix x2 is:

[0069]

[0070] A tracking differentiator with an adjustable filter factor is used to perform a first-order differential transformation on the univariate time series to obtain a first-order differential time series set. Where k1 is the filter factor of the first-order differential map tracking differentiator.

[0071] Step 1.2.2: Use a tracking differentiator with an adjustable filter factor to perform a first-order differential transformation on the first-order differential time series set to obtain a second-order differential time series set. Where k2 is the filter factor of the second-order differential map tracking differentiator;

[0072] Step 1.2.3: Introduce the expansion mechanism to the unary time series set First-order differential time series set and the second-order differential time series set Perform expansion mapping of multiple scales separately:

[0073]

[0074]

[0075]

[0076] Among them, d List =[2 0 ,2 1 ,…,d ceil ] is the expansion parameter list, d max is the expansion ratio of the maximum expansion map, m is the length of the time series;

[0077] The functional implementation process of the expansion mechanism is as follows:

[0078] For the filter factor k=i, for the time series s i Perform downsampling to obtain the downsampling time subsequence at that moment in It represents the largest integer less than (mk) / d, where m is the length of the time series and d is the sampling interval, with the expansion ratio as the sampling interval;

[0079] Complete the iterative cycle process of k=1,2,…,d in sequence, and splice the time subsequences obtained by downsampling each moment in sequence to obtain the final expansion mapping sequence

[0080] Step 1.2.4: Extract the global statistical features of the sequence after expansion mapping And perform full splicing of the global statistical features to obtain the diversity feature vector FV U ;

[0081] The global statistical features include but are not limited to one or more of mean, sum of squares, range, standard deviation, arithmetic mean, median, coefficient of variation, variance, standard deviation, approximate entropy, autoregressive model coefficients, Fourier coefficients of one-dimensional discrete Fourier transform, kurtosis of absolute Fourier transform spectrum or autocorrelation coefficient.

[0082] Step 1.3: Use the trained combined feature selector to select the diversity feature vector FV U Perform feature selection to obtain multi-view feature subset vectors Where v is the number of views;

[0083] Step 1.4: Use the trained combined classifier to perform fault diagnosis. By introducing a voting mechanism, the output results of the classifiers on each view are integrated to finally obtain the fault diagnosis result of the equipment operation status.

[0084] Example 2

[0085] like Figure 2 As shown, the training process of multiple filtering feature selectors and classifiers under different storage ratios of the present invention includes:

[0086] Step 2.1: Obtain the fault time series set containing the fault state of industrial equipment operation and its label set L, where x is the number of time series samples used for feature index training;

[0087] Step 2.2: Use the unsupervised differential expansion mapping preprocessor to preprocess the time series data, including the following steps:

[0088] Step 2.2.1: Construct a tracking differentiator with an adjustable filter factor as in Example 1, and use the tracking differentiator with an adjustable filter factor to perform a first-order differential transformation on the fault time series to obtain a first-order differential time series set.

[0089] Step 2.2.2: Use a tracking differentiator with an adjustable filter factor to perform a first-order differential transformation on the first-order differential time series set to obtain a second-order differential time series set.

[0090] Step 2.2.3: Introduce the same expansion mechanism as in Example 1 to the fault time series set First-order differential time series set and the second-order differential time series set Perform expansion mapping of multiple scales separately:

[0091]

[0092]

[0093]

[0094] Step 2.2.4: Extract the global statistical features of the sequence after expansion mapping And perform full splicing of the global statistical features to obtain the diversity feature vector FV for training L Dataset;

[0095] Step 2.3, construct multiple filter feature selectors with different storage ratios, train each filter feature selector based on the diversity feature vector dataset used for training, and during the training process, the feature vector FV L The correlation measure with the label set L is used as the feature quality score. The correlation measure uses the F statistic in variance analysis. The top-ranked features under different storage ratios are used as multi-view features. The multi-view feature subset vector and the feature selector containing multiple feature indexes are obtained. A combined feature selector is constructed based on multiple feature selectors.

[0096] Step 2.4: Build multiple ridge regression classifiers with cross-validation, train multiple classifiers based on multi-view feature subsets, and build a combined classifier.

[0097] Example 3

[0098] like Figure 4 As shown, the present invention provides an industrial equipment fault diagnosis system, comprising:

[0099] The data acquisition module 401 is used to acquire the data of the operating status information of the industrial equipment to be diagnosed, and to construct a time series data set by arranging the data according to the acquisition time sequence;

[0100] The data processing module 402 is used to preprocess the time series data using an unsupervised differential dilation mapping preprocessor to obtain a feature vector dataset with diversity;

[0101] A feature selector construction and training module 403 is used to construct multiple filtering feature selectors under different storage ratios, train the feature selectors, obtain multi-view feature subset vectors and feature selectors containing multiple feature indexes, and construct a combined feature selector based on the multiple feature selectors;

[0102] A classifier construction and training module 404 is used to construct multiple classifiers, train multiple classifiers based on multi-view feature subsets, and construct a combined classifier;

[0103] The fault diagnosis module 405 is used to pre-process the operating status data of the industrial equipment to be diagnosed, use a combined feature selector to perform feature selection, use a combined classifier to perform fault diagnosis, and integrate the output results of the classifiers on each view through the introduction of a voting mechanism to ultimately obtain the fault diagnosis results of the equipment operating status;

[0104] The communication module 406 is used to connect the above modules and perform data transmission.

[0105] The developed system (software) computer program is stored on a computer-readable storage medium. When called by a processor, the computer program is configured to implement the steps of the aforementioned method for diagnosing industrial equipment faults based on multi-view expansion statistical features. This materializes the present invention on a carrier, becoming a computer program product.

[0106] Example 4

[0107] To validate the effectiveness of our method, we conducted computer simulation classification comparisons between our method and existing methods on 85 datasets from the UCR open-source dataset. Each dataset used the default training / testing split, and each sample was normalized. Classification accuracy was used as the performance metric, and the Wilcoxon signed-rank test with a Holm correction at a p-value of 0.02 was used to measure the significance of the differences between the algorithms. The results are shown in Table 1.

[0108] Existing methods include the baseline distance-based Dynamic Time Warping using a one nearest neighbor classifier (DTW-1NN), the classification algorithm based on 22 canonical time-series characteristics (catch22), a generalized signature method for multivariate time series feature extraction (Signatures), a timeseries feature extraction method on basis of scalable hypothesis tests method (TSFresh), and a simple transformation based pipeline time series classifier (FreshPRINCE).

[0109] Table 1

[0110]

[0111]

[0112]

[0113] like Figure 5 In the accuracy scatter plot shown, points falling above / above / below the diagonal line respectively represent that the classification accuracy of the method of the present invention on the data set is worse than / the same as / better than the accuracy of the prior art method.

[0114] From Table 1 and Figure 5 The results show that the classification accuracy of the method of the present invention is not only better than that of the existing method on most data sets, but also significantly better on some data sets.

[0115] In order to test the applicable problem range of the method of the present invention, the method of the present invention is tested on 7 types of tasks on the UCR open source dataset together with the existing technical methods of TSFresh and FreshPRINCE. The performance radar chart is shown in the attached figure. Figure 6 As shown in the figure, it can be seen that the method of the present invention has the best performance in the four types of tasks: image, sensor, spectrum, and electrocardiogram data.

[0116] Although the present invention is disclosed as above, the scope of protection disclosed by the present invention is not limited thereto. Those skilled in the art of the present invention may make various changes and modifications without departing from the spirit and scope of the present invention, and these changes and modifications will fall within the scope of protection of the present invention.

Claims

1. A method for industrial equipment fault diagnosis based on multi-view expansion statistical features, characterized in that: The following steps are involved: Step 1: Obtain a time series dataset of the fault status of industrial equipment; Step 2: Use the unsupervised differential expansion mapping preprocessor to preprocess the time series data, including the following steps: Step 2-1, using a tracking differentiator with an adjustable filter factor to perform a first-order differential transformation on the time series data to obtain a first-order differential time series; Step 2-2: Use a tracking differentiator with an adjustable filter factor to perform a first-order differential transformation on the first-order differential time series to obtain a second-order differential time series; Step 2-3: introducing an expansion mechanism to perform expansion mapping of multiple proportions on the time series data, the first-order differential time series, and the second-order differential time series; Steps 2-4: extract the global statistical features of the sequences after expansion mapping and perform full splicing to obtain a diversity feature vector dataset; Step 3: Construct multiple filtering feature selectors under different storage ratios, train the feature selectors, obtain multi-view feature subset vectors and feature selectors containing multiple feature indexes, and construct a combined feature selector based on the multiple feature selectors; Step 4: Build multiple classifiers, train multiple classifiers based on multi-view feature subsets and build a combined classifier; Step 5: Preprocess the operating status data of the industrial equipment to be diagnosed, use a combined feature selector for feature selection, and use a combined classifier for fault diagnosis. By introducing a voting mechanism to integrate the output results of the classifier on each view, the fault diagnosis result of the equipment operating status is finally obtained; The functional implementation process of the tracking differentiator with adjustable filter factor is as follows: matrix The first column With the matrix The first column The calculation formula is: ； For i=1,2,…,m, increasing in sequence: ； in, Tracking the time series of the differentiator for input; like ,but ,otherwise, ; like ,but ,otherwise, ; matrix The i+1th column of With the matrix The i+1th column of The calculation formula is: ； The differential time series is transformed by a tracking differentiator with an adjustable filter factor to obtain the following differential time series: , where r is the speed factor and h is the step size; The implementation process of the expansion mechanism is as follows: For the filter factor k=i, for the time series Perform downsampling to obtain the downsampling time subsequence at that moment ,in Indicates less than The maximum integer of , where m is the length of the time series, d is the sampling interval, and the expansion ratio is used as the sampling interval; Complete the iterative cycle process of k=1,2,…,d in sequence, and downsample each moment to obtain the time subsequences and splice them in sequence to obtain the final expansion mapping sequence .

2. The industrial equipment fault diagnosis method based on multi-view expansion statistical features according to claim 1 is characterized in that: The industrial equipment described in step 1 includes one or more general equipment, special equipment, or specialized equipment in the fields of machinery, electrical electronics, computer networks, automobiles, chemicals, or energy. The equipment operating status information includes one or more of temperature, pressure, speed, torque, sound, vibration, or electrical signals.

3. The industrial equipment fault diagnosis method based on multi-view expansion statistical features according to claim 1 is characterized in that: The global statistical features described in steps 2-4 include one or more of mean, sum of squares, range, standard deviation, arithmetic mean, median, coefficient of variation, variance, standard deviation, approximate entropy, autoregressive model coefficient, Fourier coefficient of one-dimensional discrete Fourier transform, kurtosis of absolute Fourier transform spectrum or autocorrelation coefficient.

4. The industrial equipment fault diagnosis method based on multi-view expansion statistical features according to claim 1 is characterized in that: In the feature selector training process in step 3, the correlation between the feature vector and the label is measured as the feature quality score, and the top-ranked features under different storage ratios are used as multi-view features; The correlation measure between the feature vector and the label adopts the F statistic in variance analysis.

5. The industrial equipment fault diagnosis method based on multi-view expansion statistical features according to claim 1 is characterized in that: The classifier adopts a ridge regression classifier with cross validation.

6. The industrial equipment fault diagnosis method based on multi-view expansion statistical features according to claim 5 is characterized in that: The voting mechanism adopts a hard voting strategy.

7. An industrial equipment fault diagnosis system, characterized in that: include: A data acquisition module (401) is used to acquire data on the operating status of the industrial equipment to be diagnosed, and to arrange the data in accordance with the acquisition time sequence to construct a time series data set; The data processing module (402) is used to preprocess the time series data using an unsupervised differential expansion mapping preprocessor to obtain a feature vector data set with diversity; the steps include: Step 2-1, using a tracking differentiator with an adjustable filter factor to perform a first-order differential transformation on the time series data to obtain a first-order differential time series; Step 2-2: Use a tracking differentiator with an adjustable filter factor to perform a first-order differential transformation on the first-order differential time series data to obtain a second-order differential time series; Step 2-3: introducing an expansion mechanism to perform expansion mapping of multiple proportions on the time series data, the first-order differential time series, and the second-order differential time series; Steps 2-4: extract the global statistical features of the sequences after expansion mapping and perform full splicing to obtain a diversity feature vector dataset; A feature selector construction and training module (403) is used to construct multiple filtering feature selectors under different storage ratios, train the feature selectors, obtain multi-view feature subset vectors and feature selectors containing multiple feature indexes, and construct a combined feature selector based on the multiple feature selectors; A classifier construction and training module (404) is used to construct multiple classifiers, train multiple classifiers based on multi-view feature subsets and construct a combined classifier; A fault diagnosis module (405) is used to pre-process the operating status data of the industrial equipment to be diagnosed, use a combined feature selector to perform feature selection, use a combined classifier to perform fault diagnosis, and integrate the output results of the classifier on each view by introducing a voting mechanism to finally obtain the fault diagnosis result of the equipment operating status; A communication module (406) is used to connect the above modules and perform data transmission; The functional implementation process of the tracking differentiator with adjustable filter factor is as follows: matrix The first column With the matrix The first column The calculation formula is: ； For i=1,2,…,m, increasing in sequence: ； in, Tracking the time series of the differentiator for input; like ,but ,otherwise, ; like ,but ,otherwise, ; matrix The i+1th column of With the matrix The i+1th column of The calculation formula is: ； The differential time series is transformed by a tracking differentiator with an adjustable filter factor to obtain the following differential time series: , where r is the speed factor and h is the step size; The implementation process of the expansion mechanism is as follows: For the filter factor k=i, for the time series Perform downsampling to obtain the downsampling time subsequence at that moment ,in Indicates less than The maximum integer of , where m is the length of the time series, d is the sampling interval, and the expansion ratio is used as the sampling interval; Complete the iterative cycle process of k=1,2,…,d in sequence, and downsample each moment to obtain the time subsequences and splice them in sequence to obtain the final expansion mapping sequence .

8. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and the computer program is configured to implement the steps of the industrial equipment fault diagnosis method based on multi-view expansion statistical features according to any one of claims 1 to 6 when called by a processor.

Images