A fault diagnosis method, device, system and storage medium

Abstract

The application provides a kind of fault diagnosis method, device, system and storage medium, belong to the field of fault diagnosis, method includes: the pre-processing of original vibration signal obtains the vibration characteristic parameter group to be handled;From the vibration characteristic parameter group to be handled, target vibration characteristic parameter set is screened out;From original equipment process parameter set, target equipment process parameter set is screened out;According to the fault diagnosis of target equipment process parameter set to target vibration characteristic parameter set, fault diagnosis result is obtained.The application compared with the traditional variable working condition characteristic analysis method, gives consideration to the advantage of vibration characteristic analysis, provides theoretical basis for the fault diagnosis of variable working condition industrial equipment, also realizes the optimization of variable working condition vibration diagnosis parameter of industrial equipment, improves the accuracy of diagnosis, at the same time, compared with the traditional single equipment diagnosis model method, the application has the advantage of giving consideration to multiple equipment correlation, improves the accuracy of variable working condition industrial equipment fault diagnosis model.

Description

Technical Field

[0001] This invention relates to the field of fault diagnosis technology, specifically to a fault diagnosis method, device, system, and storage medium. Background Technology

[0002] Industrial equipment, as a critical component of the petrochemical industry, operates under harsh and variable conditions year-round, constantly affected by factors such as high temperature, high pressure, and erosion, resulting in complex and difficult-to-diagnose failure modes. With the rise of big data and industrial internet technologies, data among industrial equipment has been integrated and managed, forming a unique "equipment cluster" model. However, traditional fault diagnosis models rarely consider utilizing the coupling relationships between equipment to solve the problem of fault diagnosis under variable operating conditions in industrial equipment.

[0003] Common functional faults in industrial equipment include rotor imbalance, shaft bending, bearing wear, and bearing cracks. On-site fault diagnosis typically employs diagnostic graphs, with common graph formats including time-domain analysis, shaft center trajectory, frequency-domain analysis, shaft center position, trend analysis, polar plots, Bode plots, and APHT plots. Fault diagnosis based on diagnostic graphs is the most widely used method by oil refining field personnel. However, it often faces the problem of weak fault sensitivity in diagnostic graphs for industrial equipment operating under varying conditions. In actual on-site diagnosis, if industrial equipment is operating under varying conditions, its vibration parameters will fluctuate accordingly, making it difficult to distinguish whether the vibration changes are caused by the operating conditions or by a fault. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a fault diagnosis method, device, system and storage medium to address the shortcomings of the prior art.

[0005] The technical solution of the present invention to solve the above-mentioned technical problems is as follows: A fault diagnosis method, comprising the following steps:

[0006] The original vibration signal is obtained from the sensor pre-set on the device to be diagnosed, and the original vibration signal is pre-processed to obtain multiple sets of vibration characteristic parameters to be processed;

[0007] Multiple target vibration feature parameter sets are selected from all the sets of vibration feature parameters to be processed, and all the target vibration feature parameter sets are combined to obtain the target vibration feature parameter set;

[0008] Import multiple sets of original equipment process parameters, and then filter out multiple sets of target equipment process parameters from all the sets of original equipment process parameters.

[0009] Fault diagnosis is performed on the target vibration characteristic parameter set based on all the target equipment process parameter sets to obtain fault diagnosis results.

[0010] Another technical solution of the present invention to solve the above-mentioned technical problems is as follows: A fault diagnosis device, comprising:

[0011] The preprocessing module is used to obtain raw vibration signals from sensors pre-set on the device to be diagnosed, and to preprocess the raw vibration signals to obtain multiple sets of vibration characteristic parameters to be processed.

[0012] The first screening module is used to screen out multiple target vibration feature parameter groups from all the vibration feature parameter groups to be processed, and to combine all the target vibration feature parameter groups to obtain a target vibration feature parameter set;

[0013] The second filtering module is used to import multiple sets of original equipment process parameters and filter out multiple sets of target equipment process parameters from all the sets of original equipment process parameters.

[0014] The fault diagnosis result acquisition module is used to perform fault diagnosis on the target vibration characteristic parameter set based on all the target equipment process parameter sets, and obtain the fault diagnosis result.

[0015] Based on the above-mentioned fault diagnosis method, the present invention also provides a fault diagnosis system.

[0016] Another technical solution of the present invention to solve the above-mentioned technical problems is as follows: a fault diagnosis system, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the fault diagnosis method described above is implemented.

[0017] Based on the above-mentioned fault diagnosis method, the present invention also provides a computer-readable storage medium.

[0018] Another technical solution of the present invention to solve the above-mentioned technical problems is as follows: a computer-readable storage medium storing a computer program, wherein when the computer program is executed by a processor, the fault diagnosis method described above is implemented.

[0019] The beneficial effects of this invention are as follows: By preprocessing the original vibration signal to obtain a set of vibration characteristic parameters to be processed, a target vibration characteristic parameter set is selected from the set of vibration characteristic parameters to be processed, and a target equipment process parameter set is selected from the original equipment process parameter set. Based on the fault diagnosis of the target vibration characteristic parameter set using the target equipment process parameter set, a fault diagnosis result is obtained. Compared with traditional variable operating condition characteristic analysis methods, this invention takes into account the advantages of vibration characteristic analysis, provides a theoretical basis for fault diagnosis of industrial equipment under variable operating conditions, and also achieves the optimal selection of vibration diagnosis parameters for industrial equipment under variable operating conditions, improving the accuracy of diagnosis. Furthermore, compared with traditional single-equipment diagnostic model methods, this invention has the advantage of considering the correlation of multiple equipment, improving the accuracy of industrial equipment fault diagnosis models under variable operating conditions. Attached Figure Description

[0020] Figure 1 This is a flowchart illustrating a fault diagnosis method provided in an embodiment of the present invention;

[0021] Figure 2 This is a block diagram of a fault diagnosis device provided in an embodiment of the present invention. Detailed Implementation

[0022] The principles and features of the present invention are described below with reference to the accompanying drawings. The examples given are only for explaining the present invention and are not intended to limit the scope of the present invention.

[0023] Figure 1 This is a flowchart illustrating a fault diagnosis method provided in an embodiment of the present invention.

[0024] like Figure 1 As shown, a fault diagnosis method includes the following steps:

[0025] The original vibration signal is obtained from the sensor pre-set on the device to be diagnosed, and the original vibration signal is pre-processed to obtain multiple sets of vibration characteristic parameters to be processed;

[0026] Multiple target vibration feature parameter sets are selected from all the sets of vibration feature parameters to be processed, and all the target vibration feature parameter sets are combined to obtain the target vibration feature parameter set;

[0027] Import multiple sets of original equipment process parameters, and then filter out multiple sets of target equipment process parameters from all the sets of original equipment process parameters.

[0028] Fault diagnosis is performed on the target vibration characteristic parameter set based on all the target equipment process parameter sets to obtain fault diagnosis results.

[0029] Preferably, the sensor can be a ZH3010 series eddy current displacement sensor.

[0030] It should be understood that basic parameters of key components of industrial equipment are collected to construct an analysis model of the vibration characteristics of industrial equipment under varying operating conditions based on correlation analysis.

[0031] It should be understood that the original vibration signal is obtained through a sensor.

[0032] In the above embodiments, a set of vibration characteristic parameters to be processed is obtained by preprocessing the original vibration signal. A set of target vibration characteristic parameters is selected from the set of vibration characteristic parameters to be processed, and a set of target equipment process parameters is selected from the set of original equipment process parameters. Fault diagnosis results are obtained by fault diagnosis of the target vibration characteristic parameter set based on the set of target equipment process parameters. Compared with the traditional variable operating condition characteristic analysis method, this invention takes into account the advantages of vibration characteristic analysis, provides a theoretical basis for fault diagnosis of industrial equipment under variable operating conditions, and also realizes the optimization of vibration diagnosis parameters for industrial equipment under variable operating conditions, thereby improving the accuracy of diagnosis. At the same time, compared with the traditional single-equipment diagnosis model method, this invention has the advantage of taking into account the association of multiple equipment, thereby improving the accuracy of the fault diagnosis model of industrial equipment under variable operating conditions.

[0033] Optionally, as an embodiment of the present invention, the process of preprocessing the original vibration signal to obtain a plurality of vibration characteristic parameter sets to be processed includes:

[0034] The original vibration signals are classified according to the operating condition information to obtain multiple sets of original vibration characteristic parameters;

[0035] Each of the original vibration characteristic parameter groups is normalized to obtain multiple vibration characteristic parameter groups to be processed.

[0036] It should be understood that the operating condition information includes information such as the equipment's rotational speed, load, pressure, and flow rate.

[0037] It should be understood that the operating parameters of the industrial equipment are summarized, the original vibration signal is classified according to the preset operating information, vibration characteristic parameters of different operating conditions (i.e., the original vibration characteristic parameter group) are obtained, and the vibration characteristic parameters (i.e., the original vibration characteristic parameter group) are normalized.

[0038] Specifically, there are two main types of normalization methods:

[0039] Minimum-max normalization performs a linear transformation on the original data. Let minA and maxA be the minimum and maximum values ​​of the basic data A. Minimum-max normalization maps an original data point x of A to a value X′ in the interval [0,1]. The formula is as follows:

[0040]

[0041] Z-score standardization is a data standardization method based on the mean and standard deviation of the original data. It standardizes the original data x of attribute A into x′ using Z-score. Z-score standardization is suitable when the maximum or minimum value of an attribute is unknown, or when there is discrete data outside its range.

[0042]

[0043] Where μ is the mean and σ is the standard deviation.

[0044] In the above embodiments, the original vibration signal is classified according to the operating condition information to obtain the original vibration characteristic parameter group. The normalization of the original vibration characteristic parameter group yields the vibration characteristic parameter group to be processed, laying the foundation for subsequent data processing. This realizes the optimization of vibration diagnosis parameters for industrial equipment under varying operating conditions and improves the accuracy of diagnosis.

[0045] Optionally, as an embodiment of the present invention, the process of selecting multiple target vibration feature parameter sets from all the sets of vibration feature parameters to be processed includes:

[0046] Calculate the average value of each of the groups of vibration characteristic parameters to be processed to obtain the average value of the vibration characteristic to be processed corresponding to each group of vibration characteristic parameters to be processed.

[0047] By performing a similarity calculation between the first formula, each set of vibration feature parameters to be processed, and the average value of the vibration feature to be processed, and any set of vibration feature parameters to be processed and the average value of the vibration feature to be processed corresponding to any set of vibration feature parameters to be processed, a similarity index corresponding to each set of vibration feature parameters to be processed is obtained. The first formula is:

[0048]

[0049] in,

[0050] Among them, c i Let r be the similarity index corresponding to the i-th vibration feature parameter group to be processed. ik Let x be the similarity correlation coefficient between the i-th vibration feature parameter group and the k-th vibration feature parameter group to be processed. ij Let be the j-th vibration feature parameter in the i-th vibration feature parameter group to be processed, m be the number of vibration feature parameters to be processed in the group, and n be the number of vibration feature parameter groups to be processed. x is the average value of the vibration characteristics to be processed in the i-th set of vibration characteristic parameters to be processed. kj For the j-th vibration characteristic parameter in the k-th set of vibration characteristic parameters to be processed, The average value of the vibration characteristics to be processed in the kth group of vibration characteristic parameters to be processed;

[0051] Import the working condition reference parameter group corresponding to each of the vibration characteristic parameter groups to be processed, and the multiple working condition comparison parameter groups corresponding to each of the vibration characteristic parameter groups to be processed.

[0052] An original working condition reference sequence is constructed using all the working condition reference parameter groups, and an original working condition comparison sequence corresponding to each of the working condition comparison parameter groups is constructed using multiple working condition comparison parameter groups corresponding to all the vibration characteristic parameter groups to be processed.

[0053] Normalize the original working condition reference sequence and each of the original working condition comparison sequences respectively to obtain the target working condition reference sequence and the target working condition comparison sequence corresponding to each of the working condition comparison parameter groups.

[0054] By calculating the correlation degree using the second formula, each of the aforementioned working condition reference parameter groups, the target working condition reference sequence, and each of the aforementioned target working condition comparison sequences, the correlation degree of multiple target working conditions corresponding to each of the aforementioned vibration characteristic parameter groups to be processed is obtained. The second formula is:

[0055]

[0056] in,

[0057] Among them, D′ a =|D a -D′|,

[0058] in, Let r1 be the correlation degree between the i-th working condition reference parameter group and the a-th target working condition comparison sequence. i Let r be the initial working condition correlation degree between the i-th working condition reference parameter group and the 1-th target working condition comparison sequence. a i Let D′ be the correlation degree between the i-th working condition reference parameter group and the a-th target working condition comparison sequence, where p is the number of working condition reference parameters in the working condition reference parameter group. a Let ρ be the difference sequence between the target working condition reference sequence and the comparison sequence of the a-th target working condition, where ρ is the resolution coefficient. D is the difference between the k-th reference parameter and the a-th comparison sequence in the i-th reference parameter group. a Let D be the comparison sequence for the a-th target working condition, and D′ be the reference sequence for the target working condition.

[0059] If the similarity index corresponding to the vibration feature parameter group to be processed is greater than the preset similarity threshold, and the correlation degree of multiple target working conditions corresponding to the vibration feature parameter group to be processed is less than the preset correlation threshold, then the vibration feature parameter group to be processed corresponding to the similarity index is taken as the target vibration feature parameter group, thereby obtaining multiple target vibration feature parameter groups.

[0060] It should be understood that the correlation between parameters is quantified by calculating parameter similarity indicators.

[0061] It should be understood that both the preset similarity threshold and the preset correlation threshold are determined by the 3σ principle. Taking a certain type of flue gas turbine group as an example, the calculated range of Ci (i.e., the preset similarity threshold) is [0.8, 1], and the range of Si (i.e., the preset correlation threshold) is [0, 0.45].

[0062] It should be understood that constructing an original working condition reference sequence through all the working condition reference parameter groups means forming all the working condition reference parameter groups into a sequence; constructing an original working condition comparison sequence corresponding to each of the multiple working condition comparison parameter groups corresponding to all the vibration characteristic parameter groups to be processed means forming multiple working condition comparison parameter groups into a sequence.

[0063] Specifically, assume the parameter sequence (i.e., the set of vibration characteristic parameters to be processed) is as follows:

[0064] X = [x1 x2 … x n ] T =[x ij ] n×m

[0065] In the formula, the parameter and the number of parameter samples are represented by the letters n and m, respectively. Then, the correlation coefficient r between parameter i and parameter k is... ik (That is, the similarity correlation coefficient) is:

[0066]

[0067] In the formula, i,k = 1, 2, ..., n. The correlation index c between parameters i and k is... i (That is, the similarity index) is:

[0068]

[0069] In the formula, c i The larger the value of c, the stronger the correlation between the parameters, and the more effectively it can reflect the state of industrial equipment as operating conditions change. i The smaller the value, the weaker the correlation between the parameters, and the less accurately it can reflect the changing state of the working conditions.

[0070] It should be understood that by analyzing the correlation between parameters under multiple operating conditions of industrial equipment, the sensitivity of parameters to operating conditions can be derived, thereby quantifying the degree of influence of parameters under varying operating conditions.

[0071] Specifically, assuming the parameter comparison sequence under operating condition 1 (i.e., the operating condition comparison parameter group) is D1, and the parameter comparison sequences under other operating conditions (i.e., the operating condition comparison parameter group) are {D2, D3, ..., D...} n The parameter reference sequence for operating condition one (i.e., the operating condition reference parameter set) is D1'. After normalizing the parameter sequence (i.e., the operating condition reference parameter set or the operating condition comparison parameter set), the difference sequence between operating condition one and the other operating conditions is calculated:

[0072]

[0073] The number of monitoring parameters and the number of monitoring times can be represented by n and m, respectively, D' n This represents the difference sequence between operating condition 1 and operating condition n. The expression for calculating the correlation between operating condition 1 and the other operating conditions is:

[0074]

[0075] In the formula, This indicates the degree of correlation between each parameter under operating condition 1 and the historical state of operating condition 1; This represents the correlation between each parameter under operating condition n and the historical state under operating condition 1; ρ is the resolution coefficient. Multi-condition correlation of operating state parameter i. (i.e., the correlation degree of the target working condition) is:

[0076]

[0077] From the meaning of the above formula, we can conclude that if the parameter... The smaller the parameter i is, the more significant the state differences are under varying operating conditions of industrial equipment. This further proves that the parameter can effectively reflect the influence of these varying operating conditions. Conversely, if the parameter... The larger the value, the more likely it is that parameter i is in an approximate state under varying operating conditions of industrial equipment. Therefore, this parameter cannot reflect the changes in the operating conditions of industrial equipment. Thus, in the monitoring and diagnosis of industrial equipment under varying operating conditions, the membership degree of this parameter should be minimized as much as possible.

[0078] It should be understood that similarity index c is selected. i (That is, the similarity index) is greater than the first preset value (that is, the preset similarity threshold) and the working condition correlation is The parameter whose (i.e., the target working condition correlation degree) is less than the second preset value (i.e., the preset correlation degree threshold).

[0079] In the above embodiments, multiple target vibration characteristic parameter groups are selected from all the vibration characteristic parameter groups to be processed, and the working condition sensitivity of the parameters can be derived, thereby realizing the quantification of the influence of the parameters under varying working conditions.

[0080] Optionally, as an embodiment of the present invention, the process of filtering the target equipment process parameter set from all the original equipment process parameter sets includes:

[0081] The coupling degree corresponding to each set of original equipment process parameters is obtained by calculating the coupling degree of each set of original equipment process parameters using the third equation, wherein the third equation is:

[0082]

[0083] Among them, C b δ represents the coupling degree with the process parameter set corresponding to the b-th original equipment. b This is the set of process parameters for the b-th original equipment.

[0084] If the coupling degree is greater than or equal to a preset coupling degree threshold, then the original equipment process parameter set corresponding to the coupling degree is used as the target equipment process parameter set, thereby obtaining multiple target equipment process parameter sets.

[0085] It should be understood that the preset coupling threshold can be 0.8.

[0086] Specifically, a model is used to reflect the coupling relationships between similar equipment units in an industrial equipment group. Coupling degree analysis is performed on each unit of the tobacco machine equipment group to quantify the degree of coupling between individual devices and to screen out similar equipment to the reference industrial equipment. The coupling degree of multiple industrial equipment is as follows:

[0087]

[0088] In the formula, δ n The set representing all parameters of the nth industrial device (i.e., the original device process parameter set) has a coupling degree range of 0, [0, 0.1], [0.1, 0.3], [0.3, 0.5], [0.5, 0.8], [0.8, 1.0], and 1. The coupling levels are completely uncoupled, extremely mismatched, low level, medium level, relatively high level, high level, and completely coupled, respectively. Based on actual results, high-level and completely coupled devices are generally taken as the data source for the device group (i.e., the target device process parameter set).

[0089] In the above embodiments, the process parameter set of the target equipment is selected from all the original equipment process parameter sets, which can quantify the degree of coupling between various equipment and thus select equipment similar to the reference industrial equipment.

[0090] Optionally, as an embodiment of the present invention, the process of performing fault diagnosis on the target vibration characteristic parameter set based on all the target equipment process parameter sets to obtain fault diagnosis results includes:

[0091] Import the observation vector group, and select the target common parameter feature group from all the target equipment process parameter sets according to the observation vector group, wherein the observation vector in the observation vector group corresponds one-to-one with the target equipment process parameter in the target equipment process parameter set;

[0092] The Pearson correlation coefficient is obtained by calculating the Pearson correlation coefficient between the observation vector group and the target common parameter feature group using the fourth equation. The fourth equation is:

[0093]

[0094] Where p(x) obs x, y) is the Pearson correlation coefficient between the observation vector group and the target common parameter feature group, c is the total number of observation vectors in the observation vector group, and x obs (q) is the q-th observation vector in the observation vector group, and y(q) is the q-th common parameter feature of the target in the common parameter feature group;

[0095] Based on the Pearson correlation coefficient and the target common parameter feature set, fault diagnosis is performed on the target vibration feature parameter set to obtain the fault diagnosis result.

[0096] Specifically, the common feature parameters of the reference devices are used to construct the first feature information vector x. i =[N1,N2,...,N n ] T (i.e., the observation vector group). The common parameter vectors of the industrial equipment group are used to construct the second feature information vector y. i =[M1,M2,...,M n ] T (i.e., the target common parameter feature set), and then, the Pearson correlation coefficient between the two feature vectors can be calculated as follows:

[0097]

[0098] The closer the Pearson correlation coefficient is to 0, the weaker the correlation; the closer it is to 1 or -1, the stronger the correlation. The correlation between the first eigenvector (i.e., the observed vector set) and the second eigenvector (i.e., the target common parameter characteristic set) under the operating conditions of the reference industrial equipment is calculated. A strong correlation indicates that the vibration fluctuations may be caused by the operating conditions. A weak correlation indicates that the vibration changes are largely due to a fault.

[0099] In the above embodiments, fault diagnosis results are obtained by performing fault diagnosis on the target vibration characteristic parameter set based on the process parameter set of all target equipment. Compared with the traditional variable operating condition characteristic analysis method, it takes into account the advantages of vibration characteristic analysis and provides a theoretical basis for fault diagnosis of industrial equipment under variable operating conditions.

[0100] Optionally, as an embodiment of the present invention, the process of selecting the target common parameter feature group from all the target equipment process parameter sets based on the observation vector group includes:

[0101] S411: Use the residual algorithm to calculate the weights of the target equipment process parameters in each set of target equipment process parameters to obtain the equipment process weights corresponding to each target equipment process parameter in each set of target equipment process parameters.

[0102] S412: By using the fifth formula, vector calculations are performed on each set of process parameters of the target equipment and the equipment process weights corresponding to all process parameters of the target equipment in each set of process parameters of the target equipment to obtain an estimated vector group corresponding to each set of process parameters of the target equipment. The fifth formula is:

[0103]

[0104] Among them, X s (l) represents the estimated vector set corresponding to the process parameter set of the l-th target equipment. Let q be the process weight of the target equipment corresponding to the process parameter of the l-th target equipment in the set of process parameters of the l-th target equipment. For the process parameter of the l-th target equipment in the process parameter set, it is the process parameter of the q-th target equipment.

[0105] S413: Similarity calculation is performed on each estimated vector group using the sixth equation and the observed vector group to obtain the initial similarity corresponding to each estimated vector group. The sixth equation is:

[0106]

[0107] Where S(l) is the initial similarity corresponding to the l-th estimated vector group, x obs (q) represents the q-th observation vector in the observation vector group. Let q be the q-th estimated vector in the l-th estimated vector group, and c be the total number of estimated vectors in the estimated vector group;

[0108] S414: Filter out the minimum value of all the initial similarities, and obtain the target similarity after filtering;

[0109] S415: Determine whether the target similarity is greater than the preset resolution. If not, update the parameters of all the equipment process weights according to the target similarity to obtain multiple updated equipment process weights, and return to step S412. If yes, use the estimated vector group as the original common parameter feature group to obtain multiple original common parameter feature groups, and execute S416.

[0110] S416: Calculate the average value of each common parameter feature in all the original common parameter feature groups to obtain multiple common parameter features to be processed, and construct a common parameter feature group to be processed through all the common parameter features to be processed;

[0111] S417: Normalize the common parameter feature group to be processed to obtain the target common parameter feature group.

[0112] It should be understood that, assuming a certain type of smoke machine in an industrial equipment cluster dataset has n common feature parameters, then the observation vector X of the common feature parameters of the industrial equipment under a certain operating condition can be obtained. a ={x1,x2,x3…x n} T (i.e., the observed vector group), and the estimated vector X(l) = {x} of another piece of equipment in the industrial equipment group under the same operating condition. 1l ,x 2l ,x 3l …x nl} T (i.e., the set of process parameters for the target equipment), where l represents the sequence number arrangement of similar smoke machines. For any estimated vector X(l), an n-dimensional membership weight vector is generated. The weight vector W(l) is determined using the residual method, such that the estimated vector is calculated as follows:

[0113] W(l)=[ω1ω2ω3…ω n ] T

[0114] X s (l)=D·W=ω1·X(1)+ω2·X(2)+…+ω n ·X(n)

[0115] In the formula, W(l) is the weight vector, where ω n Determined by the residual method.

[0116] Specifically, the similarity between the observed vectors (i.e., the observed vector set) and the estimated vectors (i.e., the estimated vector set) is calculated using Euclidean distance. At this time, the parameter similarity (i.e., the initial similarity) between similar devices and reference device l is:

[0117]

[0118] In the formula, x obs (i) represents X a The i-th value in the matrix, x est (i) represents X s The i-th value in (l).

[0119] Specifically, the common feature parameter with the smallest similarity (i.e., the initial similarity) is selected, indicating that the confidence of this parameter is the worst. The membership degree corresponding to it is appropriately reduced, and the similarity is recalculated. The above steps are repeated (i.e., return to step S412). The resolution δ is introduced as a measure of the similarity distribution characteristics. δ is obtained from expert experience. When S > δ, the update ends. At this time, the common parameter features obtained can best represent the common state features of the industrial equipment group.

[0120] Specifically, the common feature parameter vectors (i.e., the common parameter feature group to be processed) of similar industrial equipment under the same operating conditions are normalized to form a new normalized vector S (i.e., the target common parameter feature group), as shown in the following formula:

[0121] S = (S1, S2, ... S i ,…S n )

[0122] In the formula, n is the number of common parameters.

[0123] S i =(S i1 ,S i2 ,…S ij ,…S im )

[0124] In the formula S ij is the normalized average of common parameters, and m is the number of samples obtained for a single common parameter.

[0125] S ij =∑S ijk / N,k=1,2,3,…,n

[0126] In the formula S ijk Let S be the value of the Kth point in the sample. The normalized vector S (i.e., the target common parameter feature group) at this time fuses the average state of the industrial equipment group of the same type to obtain a state feature vector that can reflect a certain working condition of the reference equipment.

[0127] In the above embodiments, the common parameter feature group of the target is selected from the process parameter set of all target equipment based on the observation vector group. Compared with the traditional variable operating condition characteristic analysis method, it takes into account the advantages of vibration characteristic analysis and provides a theoretical basis for fault diagnosis of industrial equipment under variable operating conditions. At the same time, compared with the traditional single-equipment diagnostic model method, the present invention has the advantage of taking into account the correlation of multiple equipment, which improves the accuracy of the fault diagnosis model of industrial equipment under variable operating conditions.

[0128] Optionally, as an embodiment of the present invention, the process of performing fault diagnosis on the target vibration characteristic parameter set based on the Pearson correlation coefficient and the target common parameter characteristic group to obtain the fault diagnosis result includes:

[0129] Determine whether the Pearson correlation coefficient is greater than or equal to a preset threshold. If it is, use the preset fault cause as the fault diagnosis result; otherwise, train the SVM model based on the target common parameter feature group to obtain the fault diagnosis model.

[0130] The fault diagnosis model is used to diagnose the target vibration characteristic parameter set to obtain the fault diagnosis result.

[0131] It should be understood that the preset fault cause can be a fault caused by the operating condition.

[0132] It should be understood that the preset threshold is determined by the 3σ principle. Taking a certain type of flue gas turbine group as an example, the Pearson correlation coefficient (i.e. the preset threshold) is calculated to be in the range of [0.85, 1].

[0133] Specifically, if the vibration source is determined to be a change in fluctuation caused by a fault, a fault diagnosis model is constructed using a support vector machine (i.e., the SVM model) to determine the fault type. The common parameter datasets of different fault types of the same group of industrial equipment (i.e., the target vibration feature parameter set) are fused and input into the SVM model for training. Since this invention selects the RBF kernel function, it is necessary to select appropriate hyperparameters C and gamma to determine the fault diagnosis model of the industrial equipment. The common parameter data of the faulty equipment whose vibration source is determined are extracted and input into the fault diagnosis model of the industrial equipment to obtain the fault diagnosis result.

[0134] Specifically, this method utilizes the coupling relationships between equipment in a group of devices to address the problem of low diagnostic accuracy caused by the inability of traditional fault diagnosis methods to distinguish the source of vibration changes under varying operating conditions. First, the source of vibration is determined through multi-feature association analysis. If the vibration source is a change or fluctuation caused by a fault, a fault diagnosis model is constructed using SVM to determine the fault type. The fused common parameters are input into the SVM model for training. The common parameter data of the faulty equipment whose vibration source is determined are extracted and input into the fault diagnosis model of the industrial equipment to obtain the fault diagnosis result.

[0135] In the above embodiments, fault diagnosis results are obtained by performing fault diagnosis on the target vibration characteristic parameter set based on the Pearson correlation coefficient and the target common parameter feature group. Compared with the traditional variable working condition characteristic analysis method, it takes into account the advantages of vibration characteristic analysis and provides a theoretical basis for fault diagnosis of industrial equipment under variable working conditions. At the same time, compared with the traditional single-equipment diagnosis model method, the present invention has the advantage of taking into account the correlation of multiple equipment, which improves the accuracy of the fault diagnosis model of industrial equipment under variable working conditions.

[0136] Optionally, as another embodiment of the present invention, this invention provides a diagnostic model for industrial equipment groups based on cluster analysis. This addresses the problem that traditional graph-based diagnostic methods have weak sensitivity to some faults under varying operating conditions. By utilizing data from similar industrial equipment groups, the inherent correlations between the state characteristics of similar equipment within the group are analyzed. A common parameter feature model of the tobacco equipment group is constructed using a fuzzy clustering algorithm. Multi-parameter fusion is performed on the industrial equipment group using similarity clustering. Vibration sources are determined through feature correlation analysis. Finally, a diagnostic model is constructed using a support vector machine for fault data, achieving accurate identification of vibration sources and fault diagnosis of industrial equipment under varying operating conditions.

[0137] Optionally, as another embodiment of the present invention, this invention addresses the challenges of industrial equipment with complex structures, numerous functional components, harsh operating environments, significant susceptibility to external factors, unstable operating conditions, and highly sudden, correlated, and coupled faults. While on-site personnel often use diagnostic maps for fault diagnosis, these maps have limitations in terms of fault sensitivity and applicability to various operating conditions, making it difficult to pinpoint the vibration sources of industrial equipment under varying operating conditions. This invention utilizes data from similar industrial equipment groups to analyze the inherent correlations between the state characteristics of similar equipment within the group. A common parameter feature model of the tobacco machine equipment group is constructed using a fuzzy clustering algorithm. Similarity clustering is used to fuse multiple parameters of the industrial equipment group, and feature correlation analysis is used to determine the vibration sources. Finally, a diagnostic model is constructed using a support vector machine based on the fault data, enabling accurate identification and fault diagnosis of vibration sources under varying operating conditions in industrial equipment. This invention can be practically applied to real-world scenarios and is of great significance for improving the safety and reliability of equipment operation.

[0138] Optionally, as another embodiment of the present invention, the present invention first sets up reference equipment for industrial equipment, classifies and screens smoke machines according to a coupling model, and selects similar industrial equipment based on the coupling degree analysis method. Common parameters of the industrial equipment are extracted using a common parameter model of fuzzy clustering. To facilitate the construction of a diagnostic model dataset, the common parameters of similar equipment also need to be fused, thereby combining the common features of similar industrial equipment to more effectively represent the current state characteristics of the industrial equipment. Multiple correlation analysis is performed on the common parameters to determine the source of vibration. If the vibration source is determined to be a change or fluctuation caused by a fault, a fault diagnosis model is constructed using a support vector machine to determine the fault type. The common parameter datasets of different fault types of the screened industrial equipment group are fused and input into the SVM model for training. Since the model selects the RBF kernel function, appropriate hyperparameters C and gamma need to be selected to determine the fault diagnosis model of the industrial equipment. The common parameter data of the faulty equipment with determined vibration sources are extracted and input into the industrial equipment fault diagnosis model to obtain the fault diagnosis results.

[0139] Alternatively, as another embodiment of the present invention, compared with existing fault diagnosis methods, the method provided by the present invention has the following beneficial effects:

[0140] 1. By deeply exploring the influence relationship between vibration parameters and operating condition parameters under varying operating conditions of industrial equipment through correlation analysis, a vibration characteristic analysis model for industrial equipment based on correlation analysis is constructed. The sensitivity of vibration parameters under varying operating conditions is quantified, which can realize the optimization of vibration diagnosis parameters for industrial equipment under varying operating conditions and improve the accuracy of diagnosis.

[0141] 2. Traditional fault diagnosis methods for industrial equipment suffer from weak sensitivity to faults under varying operating conditions. In actual field diagnosis, if industrial equipment is operating under varying conditions, its vibration parameters will fluctuate accordingly, making it difficult to distinguish whether the vibration changes are caused by the operating conditions or by a fault. Compared to traditional variable operating condition characteristic analysis methods, this approach incorporates the advantages of vibration characteristic analysis, providing a theoretical basis for fault diagnosis of industrial equipment under varying operating conditions.

[0142] 3. The proposed method for constructing a cluster diagnosis model for industrial equipment introduces industrial equipment group data, classifies similar equipment through a coupling model, establishes a common parameter model for industrial equipment based on fuzzy clustering, mines common relationships among similar equipment, and constructs a feature-related cluster diagnosis model based on a multi-parameter fusion algorithm using similar clustering. This enables accurate diagnosis and identification of faults in industrial equipment under varying operating conditions. Compared to traditional single-equipment diagnosis model methods, this method has the advantage of considering the correlation of multiple equipment, improving the accuracy of industrial equipment fault diagnosis models under varying operating conditions.

[0143] Optionally, as another embodiment of the present invention, the present invention analyzes the inherent relationship between the state feature vectors of similar equipment in an industrial equipment group, constructs a common parameter feature model of the parameters of the tobacco machine equipment group by combining fuzzy clustering algorithm, performs multi-parameter fusion on the industrial equipment group by similar clustering, determines its vibration source by feature correlation analysis, and finally uses support vector machine to construct a diagnostic model to achieve accurate identification of vibration source and fault diagnosis of industrial equipment under variable operating conditions.

[0144] Optionally, as another embodiment of the present invention, the research on multi-parameter fusion method of industrial equipment based on similar clustering, for the common parameters of similar equipment, the present invention uses a similarity algorithm to measure the degree of correlation between feature vectors, filters out common parameters with poor similarity, appropriately reduces their membership values, and then performs the above comparison and adjustment until the state feature values ​​of each layer in the two feature vectors are similar before fusion.

[0145] Optionally, as another embodiment of the present invention, the present invention includes: data acquisition and collection under varying operating conditions, signal preprocessing, feature screening, classification of similar equipment, extraction of common parameters, parameter fusion, multi-feature correlation analysis, and fault diagnosis.

[0146] Alternatively, as another embodiment of the present invention, the present invention establishes a coupling model to quantitatively analyze the coupling degree between various system parameters, screens out similar devices to the reference device, and selects high-level and fully coupled devices as the data source for the device group based on actual effects.

[0147] Optionally, as another embodiment of the present invention, the present invention constructs an optimal dataset for the diagnostic model, uses fuzzy clustering to construct a common parameter model to separate the common parameters and individual parameters of industrial equipment in the data of similar equipment and the reference industrial equipment, and uses similarity clustering to fuse the common parameters of similar equipment, so that the obtained dataset more effectively represents the current state characteristics of the industrial equipment.

[0148] Optionally, as another embodiment of the present invention, the present invention summarizes the equipment parameters and monitoring parameters from different devices through sensors, classifies the collected raw vibration signal data according to different operating conditions, extracts multiple flue gas turbine vibration characteristic parameters under different operating conditions, and performs normalization processing.

[0149] Optionally, as another embodiment of the present invention, the present invention utilizes the coupling relationship between equipment in a group of equipment to solve the problem that traditional fault diagnosis methods cannot distinguish the source of equipment vibration changes under varying operating conditions, resulting in low diagnostic accuracy. First, the source of vibration is determined through multi-feature association analysis. If the source of vibration is a change or fluctuation caused by a fault, a fault diagnosis model is constructed using SVM to determine the fault type. The fused common parameters are input into the SVM model for training. The common parameter data of the faulty equipment whose vibration source is determined are extracted and input into the industrial equipment fault diagnosis model to obtain the fault diagnosis result.

[0150] Alternatively, as another embodiment of the present invention, the present invention uses historical data collected on-site at an oil refinery to verify the proposed method, which can show that the proposed method can effectively identify and diagnose abnormal vibration fluctuations in equipment under varying operating conditions, thereby achieving vibration fault diagnosis for industrial equipment under multiple operating conditions.

[0151] Figure 2 This is a block diagram of a fault diagnosis device provided in an embodiment of the present invention.

[0152] Alternatively, as another embodiment of the present invention, such as Figure 2 As shown, a fault diagnosis device includes:

[0153] The preprocessing module is used to obtain raw vibration signals from sensors pre-set on the device to be diagnosed, and to preprocess the raw vibration signals to obtain multiple sets of vibration characteristic parameters to be processed.

[0154] The first screening module is used to screen out multiple target vibration feature parameter groups from all the vibration feature parameter groups to be processed, and to combine all the target vibration feature parameter groups to obtain a target vibration feature parameter set;

[0155] The second filtering module is used to import multiple sets of original equipment process parameters and filter out multiple sets of target equipment process parameters from all the sets of original equipment process parameters.

[0156] The fault diagnosis result acquisition module is used to perform fault diagnosis on the target vibration characteristic parameter set based on all the target equipment process parameter sets, and obtain the fault diagnosis result.

[0157] Optionally, another embodiment of the present invention provides a fault diagnosis system, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the fault diagnosis method described above. This system can be a computer or similar system.

[0158] Optionally, another embodiment of the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the fault diagnosis method as described above.

[0159] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0160] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the above-described apparatus and unit can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0161] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed.

[0162] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of the embodiments of the present invention, depending on actual needs.

[0163] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0164] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0165] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A fault diagnosis method, characterized in that, Includes the following steps: The original vibration signal is obtained from the sensor pre-set on the device to be diagnosed, and the original vibration signal is pre-processed to obtain multiple sets of vibration characteristic parameters to be processed; Multiple target vibration feature parameter sets are selected from all the sets of vibration feature parameters to be processed, and all the target vibration feature parameter sets are combined to obtain the target vibration feature parameter set; Import multiple sets of original equipment process parameters, and then filter out multiple sets of target equipment process parameters from all the sets of original equipment process parameters. Based on all the target equipment process parameter sets, fault diagnosis is performed on the target vibration characteristic parameter set to obtain fault diagnosis results; The process of selecting multiple target vibration feature parameter sets from all the sets of vibration feature parameters to be processed includes: Calculate the average value of each of the groups of vibration characteristic parameters to be processed to obtain the average value of the vibration characteristic to be processed corresponding to each group of vibration characteristic parameters to be processed. By performing a similarity calculation between the first formula, each set of vibration feature parameters to be processed, and the average value of the vibration feature to be processed, and any set of vibration feature parameters to be processed and the average value of the vibration feature to be processed corresponding to any set of vibration feature parameters to be processed, a similarity index corresponding to each set of vibration feature parameters to be processed is obtained. The first formula is: ， in, , in, For the first Similarity index corresponding to each set of vibration feature parameters to be processed For the first The set of vibration characteristic parameters to be processed and the first The similarity correlation coefficient of the set of vibration characteristic parameters to be processed For the first The first set of vibration characteristic parameters to be processed One vibration characteristic parameter to be processed, This represents the number of vibration characteristic parameters to be processed in the set of vibration characteristic parameters to be processed. The number of vibration characteristic parameter sets to be processed. For the first The average value of the vibration characteristics to be processed in a set of vibration characteristic parameters to be processed. For the first The first set of vibration characteristic parameters to be processed One vibration characteristic parameter to be processed, For the first The average value of the vibration characteristics to be processed in a set of vibration characteristic parameters to be processed; Import the working condition reference parameter group corresponding to each of the vibration characteristic parameter groups to be processed, and the multiple working condition comparison parameter groups corresponding to each of the vibration characteristic parameter groups to be processed. An original working condition reference sequence is constructed using all the working condition reference parameter groups, and an original working condition comparison sequence corresponding to each of the working condition comparison parameter groups is constructed using multiple working condition comparison parameter groups corresponding to all the vibration characteristic parameter groups to be processed. Normalize the original working condition reference sequence and each of the original working condition comparison sequences respectively to obtain the target working condition reference sequence and the target working condition comparison sequence corresponding to each of the working condition comparison parameter groups. By calculating the correlation degree using the second formula, each of the aforementioned working condition reference parameter groups, the target working condition reference sequence, and each of the aforementioned target working condition comparison sequences, the correlation degree of multiple target working conditions corresponding to each of the aforementioned vibration characteristic parameter groups to be processed is obtained. The second formula is: ， in, , in, , in, For the first The reference parameter group for the first operating condition and the first The correlation degree of target working conditions in the target working condition comparison sequence For the first The initial working condition correlation degree between each set of reference parameters and the first target working condition comparison sequence. For the first The reference parameter group for the first operating condition and the first The correlation degree of target working conditions in the target working condition comparison sequence The number of operating condition reference parameters in the operating condition reference parameter group. For the target operating condition reference sequence and the first The difference sequence of the target working condition comparison sequence, The resolution coefficient, For the first The first set of reference parameters for operating conditions Reference parameters for each operating condition and the first The difference values ​​of the comparison sequences of each working condition. For the first A sequence of target operating conditions for comparison. This is the target operating condition reference sequence; If the similarity index corresponding to the vibration feature parameter group to be processed is greater than the preset similarity threshold, and the correlation degree of multiple target working conditions corresponding to the vibration feature parameter group to be processed is less than the preset correlation threshold, then the vibration feature parameter group to be processed corresponding to the similarity index is taken as the target vibration feature parameter group, thereby obtaining multiple target vibration feature parameter groups.

2. The fault diagnosis method according to claim 1, characterized in that, The process of preprocessing the original vibration signal to obtain multiple sets of vibration characteristic parameters to be processed includes: The original vibration signals are classified according to the operating conditions to obtain multiple sets of original vibration characteristic parameters; Each of the original vibration characteristic parameter groups is normalized to obtain multiple vibration characteristic parameter groups to be processed.

3. The fault diagnosis method according to claim 1, characterized in that, The process of selecting the target equipment process parameter set from all the original equipment process parameter sets includes: The coupling degree corresponding to each set of original equipment process parameters is obtained by calculating the coupling degree of each set of original equipment process parameters using the third equation, wherein the third equation is: ， in, In order to be with the first The degree of coupling corresponding to a set of original equipment process parameters For the first A set of original equipment process parameters; If the coupling degree is greater than or equal to a preset coupling degree threshold, then the original equipment process parameter set corresponding to the coupling degree is used as the target equipment process parameter set, thereby obtaining multiple target equipment process parameter sets.

4. The fault diagnosis method according to claim 1, characterized in that, The process of performing fault diagnosis on the target vibration characteristic parameter set based on all the target equipment process parameter sets to obtain fault diagnosis results includes: Import the observation vector group, and select the target common parameter feature group from all the target equipment process parameter sets according to the observation vector group, wherein the observation vector in the observation vector group corresponds one-to-one with the target equipment process parameter in the target equipment process parameter set; The Pearson correlation coefficient is obtained by calculating the Pearson correlation coefficient between the observation vector group and the target common parameter feature group using the fourth equation. The fourth equation is: ， in, The Pearson correlation coefficient is the sum of the observed vector set and the target common parameter feature set. The total number of observation vectors in the observation vector group. For the observation vector group, the first... Observation vectors, The first common parameter feature group of the target Common parameter characteristics of each target; Based on the Pearson correlation coefficient and the target common parameter feature set, fault diagnosis is performed on the target vibration feature parameter set to obtain the fault diagnosis result.

5. The fault diagnosis method according to claim 4, characterized in that, The process of selecting the target common parameter feature group from all the target equipment process parameter sets based on the observation vector group includes: S411: Use the residual algorithm to calculate the weights of the target equipment process parameters in each set of target equipment process parameters to obtain the equipment process weights corresponding to each target equipment process parameter in each set of target equipment process parameters. S412: By using the fifth formula, vector calculations are performed on each set of process parameters of the target equipment and the equipment process weights corresponding to all process parameters of the target equipment in each set of process parameters of the target equipment to obtain an estimated vector group corresponding to each set of process parameters of the target equipment. The fifth formula is: ， in, For the first The estimated vector set corresponding to the process parameters of each target device. For the first The first set of process parameters for the target equipment The equipment process weights corresponding to the process parameters of each target equipment. For the first The first set of process parameters for the target equipment Target equipment process parameters; S413: Similarity calculation is performed on each estimated vector group using the sixth equation and the observed vector group to obtain the initial similarity corresponding to each estimated vector group. The sixth equation is: ， in, For the first The initial similarity corresponding to each estimated vector group For the observation vector group, the first... Observation vectors, For the first In the estimated vector set, the th... One estimated vector, To estimate the total number of vectors in the vector group; S414: Filter out the minimum value of all the initial similarities, and obtain the target similarity after filtering; S415: Determine whether the target similarity is greater than the preset resolution. If not, update the parameters of all the equipment process weights according to the target similarity to obtain multiple updated equipment process weights, and return to step S412. If yes, use the estimated vector group as the original common parameter feature group to obtain multiple original common parameter feature groups, and execute S416. S416: Calculate the average value of each common parameter feature in all the original common parameter feature groups to obtain multiple common parameter features to be processed, and construct a common parameter feature group to be processed through all the common parameter features to be processed; S417: Normalize the common parameter feature group to be processed to obtain the target common parameter feature group.

6. The fault diagnosis method according to claim 4, characterized in that, The process of performing fault diagnosis on the target vibration characteristic parameter set based on the Pearson correlation coefficient and the target common parameter characteristic group to obtain the fault diagnosis result includes: Determine whether the Pearson correlation coefficient is greater than or equal to a preset threshold. If it is, use the preset fault cause as the fault diagnosis result; otherwise, train the SVM model based on the target common parameter feature group to obtain the fault diagnosis model. The fault diagnosis model is used to diagnose the target vibration characteristic parameter set to obtain the fault diagnosis result.

7. A fault diagnosis device, characterized in that, include: The preprocessing module is used to obtain raw vibration signals from sensors pre-set on the device to be diagnosed, and to preprocess the raw vibration signals to obtain multiple sets of vibration characteristic parameters to be processed. The first screening module is used to screen out multiple target vibration feature parameter groups from all the vibration feature parameter groups to be processed, and to combine all the target vibration feature parameter groups to obtain a target vibration feature parameter set; The second filtering module is used to import multiple sets of original equipment process parameters and filter out multiple sets of target equipment process parameters from all the sets of original equipment process parameters. The fault diagnosis result acquisition module is used to perform fault diagnosis on the target vibration characteristic parameter set based on all the target equipment process parameter sets, and obtain the fault diagnosis result; In the first screening module, the process of selecting multiple target vibration feature parameter sets from all the vibration feature parameter sets to be processed includes: Calculate the average value of each of the groups of vibration characteristic parameters to be processed to obtain the average value of the vibration characteristic to be processed corresponding to each group of vibration characteristic parameters to be processed. By performing a similarity calculation between the first formula, each set of vibration feature parameters to be processed, and the average value of the vibration feature to be processed, and any set of vibration feature parameters to be processed and the average value of the vibration feature to be processed corresponding to any set of vibration feature parameters to be processed, a similarity index corresponding to each set of vibration feature parameters to be processed is obtained. The first formula is: ， in, , in, For the first Similarity index corresponding to each set of vibration feature parameters to be processed For the first The set of vibration characteristic parameters to be processed and the first The similarity correlation coefficient of the set of vibration characteristic parameters to be processed For the first The first set of vibration characteristic parameters to be processed One vibration characteristic parameter to be processed, This represents the number of vibration characteristic parameters to be processed in the set of vibration characteristic parameters to be processed. The number of vibration characteristic parameter sets to be processed. For the first The average value of the vibration characteristics to be processed in a set of vibration characteristic parameters to be processed. For the first The first set of vibration characteristic parameters to be processed One vibration characteristic parameter to be processed, For the first The average value of the vibration characteristics to be processed in a set of vibration characteristic parameters to be processed; Import the working condition reference parameter group corresponding to each of the vibration characteristic parameter groups to be processed, and the multiple working condition comparison parameter groups corresponding to each of the vibration characteristic parameter groups to be processed. An original working condition reference sequence is constructed using all the working condition reference parameter groups, and an original working condition comparison sequence corresponding to each of the working condition comparison parameter groups is constructed using multiple working condition comparison parameter groups corresponding to all the vibration characteristic parameter groups to be processed. Normalize the original working condition reference sequence and each of the original working condition comparison sequences respectively to obtain the target working condition reference sequence and the target working condition comparison sequence corresponding to each of the working condition comparison parameter groups. By calculating the correlation degree using the second formula, each of the aforementioned working condition reference parameter groups, the target working condition reference sequence, and each of the aforementioned target working condition comparison sequences, the correlation degree of multiple target working conditions corresponding to each of the aforementioned vibration characteristic parameter groups to be processed is obtained. The second formula is: ， in, , in, , in, For the first The reference parameter group for the first operating condition and the first The correlation degree of target working conditions in the target working condition comparison sequence For the first The initial working condition correlation degree between each set of reference parameters and the first target working condition comparison sequence. For the first The reference parameter group for the first operating condition and the first The correlation degree of target working conditions in the target working condition comparison sequence The number of operating condition reference parameters in the operating condition reference parameter group. For the target operating condition reference sequence and the first The difference sequence of the target working condition comparison sequence, The resolution coefficient, For the first The first set of reference parameters for operating conditions Reference parameters for each operating condition and the first The difference values ​​of the comparison sequences of each working condition. For the first A sequence of target operating conditions for comparison. This is the target operating condition reference sequence; If the similarity index corresponding to the vibration feature parameter group to be processed is greater than the preset similarity threshold, and the correlation degree of multiple target working conditions corresponding to the vibration feature parameter group to be processed is less than the preset correlation threshold, then the vibration feature parameter group to be processed corresponding to the similarity index is taken as the target vibration feature parameter group, thereby obtaining multiple target vibration feature parameter groups.

8. A fault diagnosis system, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the fault diagnosis method as described in any one of claims 1 to 6.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the fault diagnosis method as described in any one of claims 1 to 6.

Images