A method, device, medium and product for tracing the factors affecting the health status of equipment

Abstract

The present application discloses a method, equipment, medium and product for tracing the factors affecting the health status of equipment, and relates to the field of equipment health assessment and management. The method includes: forming a sample matrix based on the original variables, extracting the main component features and the main component statistics features; using the Gaussian membership function to convert the main component statistics features into evidence; using the test sample to activate the evidence, and fusing all the evidence of feature activation based on the evidence reasoning algorithm to obtain the confidence of the evidence support state; constructing the relationship between the confidence of the evidence support state and the main component statistics features and the relationship between the original variables and the main component statistics features, and based on the two constructed relationships, obtaining the relationship between the confidence of the evidence support state and the original variables; based on the relationship between the confidence of the evidence support state and the original variables, the equipment health status influencing factors can be traced. The present application can evaluate the equipment status and trace the key factors affecting the status change in the case of high-dimensional correlated variables.

Description

Technical Field

[0001] The present application relates to the field of equipment health assessment and management, and in particular to a method, equipment, medium and product for tracing the factors affecting the health status of equipment. Background Art

[0002] Conducting status assessment on complex equipment is a key link to master its operating status. At present, multivariate information fusion methods, such as evidence reasoning methods, can fuse data with uncertainty to obtain the comprehensive status of the equipment. However, the challenges faced by complex equipment status assessment include: a large number of monitoring variables, high-dimensional correlation characteristics, and inaccurate evaluation results. Therefore, establishing a health assessment model that considers variable correlation is the basis for giving accurate evaluation results. Variable feature extraction is a method that can decorrelate and meet the requirements of evidence independence. However, the transformation from features to evidence has always been a difficult point. In addition, in order to enable equipment managers to carry out maintenance based on the evaluation results, it is necessary to trace the factors that affect the status degradation, which helps to eliminate risks in a timely manner. The fusion process of the evidence reasoning method is clear and transparent, the parameter settings can be adjusted, and it has human-computer interaction capabilities. It can provide decision makers with understandable and credible equipment status assessment results, and the results can be traced, meeting the decision makers' requirements for interpretability of the assessment model. However, the complex correspondence between features and variables and the conversion of evidence involved make it challenging to trace the evaluation results. Summary of the invention

[0003] The purpose of this application is to provide a method, device, medium and product for tracing the factors affecting the health status of equipment, which can evaluate the equipment status under high-dimensional correlated variables and trace the key factors affecting the status change.

[0004] To achieve the above objectives, this application provides the following solutions:

[0005] In a first aspect, the present application provides a method for tracing the factors affecting the health status of a device, including:

[0006] Acquire original variables of the device in different states, and form a sample matrix based on the original variables; the different states of the device include state and fault state;

[0007] Extracting principal component features from the sample matrix to obtain principal component statistical features;

[0008] The principal component statistic characteristics are converted into evidence using a Gaussian membership function;

[0009] Activate the evidence using a test sample, and fuse all feature-activated evidence based on an evidence reasoning algorithm to obtain a confidence level of the evidence support state;

[0010] Construct the relationship between the confidence of the evidence support status and the characteristics of the principal component statistics, and the relationship between the original variables and the characteristics of the principal component statistics;

[0011] Based on the relationship between the confidence of the evidence support status and the characteristics of the principal component statistics and the relationship between the original variables and the characteristics of the principal component statistics, the relationship between the confidence of the evidence support status and the original variables is obtained;

[0012] The factors affecting the health status of equipment can be traced back based on the relationship between the confidence of the evidence-supported status and the original variables.

[0013] Optionally, extracting principal component features from the sample matrix to obtain principal component statistical features includes:

[0014] Performing standardization on the sample matrix;

[0015] Determine the sample covariance matrix based on the standardized sample matrix;

[0016] Performing principal component eigenvalue decomposition on the sample covariance matrix, and arranging the eigenvectors obtained by the decomposition in descending order to obtain a vector sequence;

[0017] Selecting in sequence from the vector sequence the eigenvectors whose principal component cumulative contribution rates reach a predetermined value to form a load matrix;

[0018] Obtaining a score matrix based on the load matrix and the standardized sample matrix;

[0019] Determining a principal component covariance matrix based on the score matrix;

[0020] Obtaining statistics based on the load matrix, the standardized sample matrix, and the principal component covariance matrix;

[0021] The statistics are filtered to obtain the principal component statistics feature.

[0022] Optionally, filtering the statistics to obtain the principal component statistics feature includes:

[0023] A median filter with a sliding window length of a set value is added to the statistic to obtain the main component statistic characteristics.

[0024] Optionally, the Gaussian membership function is used to convert the principal component statistic feature into an expression of evidence:

[0025] ;

[0026] ;

[0027] In the formula, Indicates The evidence obtained by feature transformation of principal component statistics is Indicates The principal component statistics feature, Indicates the first state, Indicates The first state principal components, , Indicates The confidence of the sample characterized by the principal component statistics relative to the first state, The first state The principal component statistics feature, Indicates the second state, Indicates The confidence of the sample characterized by the principal component statistics relative to the second state, The second state The principal component statistics feature, represents the Hth state, Indicates The confidence of the sample characterized by the principal component statistics relative to the Hth state, The Hth state The principal component statistics feature, Indicates The confidence of the sample characterized by the principal component statistics relative to all states, represents the Gaussian membership in all states, Indicates status Gaussian membership under .

[0028] Optionally, a sensitivity coefficient is used to represent the relationship between the confidence of the evidence support status and the original variable; wherein the sensitivity coefficient is expressed as:

[0029] ;

[0030] In the formula, represents the sensitivity coefficient, represents the original variable, Indicates the confidence level of the evidence supporting the status, Indicates the maximum confidence value of the evidence support state, max() represents the maximum value function, represents the number of original variables, Indicates the number of states.

[0031] In a second aspect, the present application provides a computer device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the method for tracing the factors affecting the health status of an equipment as described in any one of the above.

[0032] In a third aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method for tracing the factors affecting the health status of an equipment as described in any one of the above.

[0033] In a fourth aspect, the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the method for tracing the factors affecting the health status of an equipment as described in any one of the above.

[0034] According to the specific embodiments provided in this application, this application has the following technical effects:

[0035] The present application provides a method, equipment, medium and product for tracing the factors affecting the health status of equipment. The relationship between the confidence of the evidence-supported status and the original variables is derived from the relationship between the original variables and the statistical characteristics of the principal components. The evaluation results can be analyzed, and the equipment status can be evaluated in the case of high-dimensional correlated variables to trace the key factors that affect the changes in the health status. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0037] Figure 1 This is an application environment diagram of a method for tracing the influencing factors of the health status of a device in an embodiment of the present application;

[0038] Figure 2 A flowchart of a method for tracing the factors affecting the health status of equipment provided in one embodiment of the present application;

[0039] Figure 3 A schematic diagram of a Gaussian membership function model curve of a first principal component statistic feature provided in an embodiment of the present application;

[0040] Figure 4 A schematic diagram of a Gaussian membership function model curve of a second principal component statistic feature provided in an embodiment of the present application;

[0041] Figure 5A schematic diagram of a motor rotor state assessment result based on multi-feature fusion provided in an embodiment of the present application;

[0042] Figure 6 A schematic diagram of the sensitivity coefficient of the motor rotor health state affected by variable 1 provided in an embodiment of the present application;

[0043] Figure 7 A schematic diagram of the sensitivity coefficient of the motor rotor health state affected by variable 2 provided in an embodiment of the present application;

[0044] Figure 8 A schematic diagram of the sensitivity coefficient of the motor rotor health state affected by variable 3 provided in an embodiment of the present application;

[0045] Fig. 9 A schematic diagram of the sensitivity coefficient of the motor rotor health state affected by variable 4 provided in an embodiment of the present application;

[0046] Fig.10 A schematic diagram of a Gaussian membership function model curve of original variables under different vibration conditions provided by another embodiment of the present application;

[0047] Fig.11 A schematic diagram of a motor rotor state assessment result based on multivariable fusion provided by another embodiment of the present application;

[0048] Fig.12 A schematic diagram of the structure of a computer device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0049] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0050] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0051] The device health status influencing factor tracing method provided in the embodiment of the present application can be applied to Figure 1In the application environment shown, the terminal 102 communicates with the server 104 through a network. The data storage system can store data that the server 104 needs to process. The data storage system can be set up separately, integrated on the server 104, or placed on the cloud or other servers. The terminal 102 can send the original variables of the device in different states to the server 104. After the server 104 receives the original variables of the device in different states, for the original variables of the device in different states, the server 104 forms a sample matrix based on the original variables; extracts the principal component features of the sample matrix to obtain the principal component statistic features; uses the Gaussian membership function to convert the principal component statistic features into evidence; uses the test sample to activate the evidence, and fuses all the evidence of feature activation based on the evidence reasoning algorithm to obtain the confidence of the evidence support state; constructs the relationship between the confidence of the evidence support state and the principal component statistic features and the relationship between the original variables and the principal component statistic features; obtains the relationship between the confidence of the evidence support state and the original variables based on the relationship between the confidence of the evidence support state and the principal component statistic features and the relationship between the original variables and the principal component statistic features; realizes the tracing of the factors affecting the health status of the device based on the relationship between the confidence of the evidence support state and the original variables. The server 104 can feed back the obtained tracing results of the factors affecting the health status of the device to the terminal 102. In addition, in some embodiments, the method for tracing the factors affecting the health status of the device can also be implemented independently by the server 104 or the terminal 102. For example, the terminal 102 can directly trace the factors affecting the health status of the device for the original variables in different states of the device, or the server 104 can obtain the original variables in different states of the device from the data storage system, and trace the factors affecting the health status of the device for the original variables in different states of the device.

[0052] The terminal 102 may be, but is not limited to, various desktop computers, laptop computers, smart phones, tablet computers, IoT devices, and portable wearable devices. The IoT devices may be smart speakers, smart TVs, smart air conditioners, smart vehicle-mounted devices, etc. The portable wearable devices may be smart watches, smart bracelets, head-mounted devices, etc. The server 104 may be implemented as an independent server or a server cluster consisting of multiple servers, or may be a cloud server.

[0053] In an exemplary embodiment, Figure 2 As shown, a method for tracing the factors affecting the health status of an equipment is provided. The method is executed by a computer device, and can be executed by a computer device such as a terminal or a server alone, or can be executed by a terminal and a server together. In the embodiment of the present application, the method is applied to Figure 1 The server 104 in the example is used as an example to illustrate the method, which includes the following steps 200 to 206. Among them:

[0054] Step 200: Acquire original variables of the device in different states, and form a sample matrix based on the original variables. The different states of the device include a state and a fault state.

[0055] Step 201: extract the principal component features of the sample matrix to obtain the principal component statistical features.

[0056] Step 202: Use Gaussian membership function to convert the principal component statistic features into evidence.

[0057] Step 203: Use the test sample to activate evidence, and fuse all feature activated evidence based on the evidence reasoning algorithm to obtain the confidence of the evidence support status.

[0058] Step 204: construct the relationship between the confidence of the evidence support state and the characteristics of the principal component statistics, and the relationship between the original variables and the characteristics of the principal component statistics.

[0059] Step 205: Based on the relationship between the confidence of the evidence support state and the characteristics of the principal component statistics and the relationship between the original variables and the characteristics of the principal component statistics, the relationship between the confidence of the evidence support state and the original variables is obtained.

[0060] Step 206: Tracing the influencing factors of the equipment health status based on the relationship between the confidence of the evidence-supported status and the original variables.

[0061] By implementing the above steps 200 to 206, the present application can analyze the traceability results and determine which variable has a more obvious impact on the traceability assessment. The monitoring of this variable should be strengthened to facilitate early detection of changes in the health status of the equipment.

[0062] In another exemplary embodiment of the present application, in order to ensure that the generated evidence is irrelevant, principal component features are extracted from the sample matrix formed by the original variables so as to generate evidence from the obtained principal component statistical features. Based on this, the implementation process of the above step 201 may include:

[0063] (1) Standardize the sample matrix.

[0064] (2) Determine the sample covariance matrix based on the standardized sample matrix.

[0065] (3) Perform principal component eigenvalue decomposition on the sample covariance matrix and arrange the eigenvectors obtained by the decomposition in descending order to obtain a vector sequence.

[0066] (4) Select the eigenvectors whose cumulative contribution rate of the principal component reaches a predetermined value in the vector sequence to form a load matrix.

[0067] (5) The score matrix is ​​obtained based on the load matrix and the standardized sample matrix.

[0068] (6) Determine the principal component covariance matrix based on the score matrix.

[0069] (7) Obtain statistics based on the loading matrix, the standardized sample matrix, and the principal component covariance matrix.

[0070] (8) Filter the statistics to obtain the main component statistics characteristics. For example, add a median filter with a sliding window length of a set value to the statistics to obtain the main component statistics characteristics.

[0071] In another exemplary embodiment of the present application, the method for tracing the factors affecting the health status of equipment provided in the present application can be divided into two execution processes: training and testing. Among them, the training process mainly uses the measurement data of each feature in a known state to build an evidence matrix. The testing process is to activate the corresponding evidence through the membership of each feature, and then integrate the evidence of all feature activations to obtain a comprehensive health status assessment result. Based on this, in this embodiment, the specific implementation method of the training process and the testing process can be described as:

[0072] Step 1: Evidence generation based on principal component statistics and Gaussian membership function.

[0073] In this step, the principal component features can be extracted from the original variables, and evidence can be generated from the extracted principal component statistic features to ensure that the generated evidence is irrelevant. The reference values ​​of the original variables related to the state have physical meaning and can be clearly given. However, when the variables are converted into features, the physical meaning of the features gradually becomes unclear, which makes it difficult to give reference values ​​indicating state changes, affecting the process of constructing evidence from features. Therefore, evidence is generated based on the principal component statistic features and the Gaussian membership function.

[0074] In this embodiment, the known device states in the training data constitute the identification framework for evaluation, denoted as :

[0075] (1)

[0076] in, It is kind of state. It is the first state, which also indicates a healthy state.

[0077] In the first state The sample matrix below ( ) as a benchmark, extract the load matrix :

[0078] First, the sample matrix After standardization, the standardized sample matrix is ​​recorded as Based on this, the covariance matrix of all samples is for:

[0079] (2)

[0080] Based on the principal component model, the standardized sample matrix Decomposed, we have:

[0081] (3)

[0082] In the formula, represents the residual matrix. represents the score matrix, , each column in the score matrix is ​​called a principal component. Represents the different principal components in the first state. represents the load matrix, , Both represent the load vector in the first state, and T represents the transpose of the matrix.

[0083] Covariance matrix Perform eigenvalue decomposition and arrange the eigenvalues ​​in descending order ,have , and obtain a vector sequence. According to whether the cumulative contribution rate of the principal component reaches the predetermined value , the number of principal components in the retained vector sequence is The eigenvector of is used to obtain the loading matrix.

[0084] Based on the score matrix, the principal component covariance matrix can be obtained. Among them, the principal component covariance matrix is :

[0085] (4)

[0086] in, is the reserved vector sequence feature vectors.

[0087] Considering that each principal component has different sensitivity to state changes and the principal components are unrelated, each principal component is used as an indicator for state assessment separately, and then the results are fused.

[0088] Statistics are indicators that reflect the spatial information of the principal components. Statistics are calculated for each component separately. principal components For example, The statistics are calculated as follows:

[0089] (5)

[0090] In the formula, The first state A statistic. Indicates principal components The principal component covariance matrix of represents the intermediate matrix, , represents the kth load vector in the first state. In order to make the statistics more distinct from the states, a sliding window with a length of The median filter of the first state after the median filter The statistic is used as the first state The principal component statistics are characterized by .

[0091] Unlike the original variables that have clear physical meanings, it is difficult to give a reference value of the principal component statistic feature when converting the principal component statistic feature to evidence. Therefore, this embodiment proposes to describe the reference value of the principal component statistic feature conversion in the form of a Gaussian membership function. In addition, the single-value reference value can be regarded as a triangular fuzzy number conversion of the principal component statistic feature to generate evidence. Compared with the reference value given by a single value, the Gaussian reference value can better describe the characteristics of the principal component statistic feature under different states. Unlike establishing a Gaussian membership function model for the training data of a single variable, the principal component statistic feature integrates multiple variables, and evidence based on the Gaussian membership function can be constructed using a few components less than the number of variables.

[0092] For example, in the first state Next, Principal component statistics The Gaussian membership function of is:

[0093] (6)

[0094] in, Indicates status Next The Gaussian membership of the principal component statistics features, Indicates the first state Next The mean of the principal component statistics features, Indicates the first state Next The standard deviation of the principal component statistics features.

[0095] Based on the above description, The Gaussian membership of the principal component statistics feature in all states is :

[0096] (7)

[0097] In the formula, The second state The principal component statistics feature, The Hth state principal component statistics features.

[0098] Similarly, the statistical features of each principal component in all states can be constructed A Gaussian membership function, represents the number of states known in the training data, Represents the number of principal components.

[0099] Based on the above steps, the principal component statistics characteristics are converted into evidence based on the Gaussian membership function as follows:

[0100] (8)

[0101] (9)

[0102] In the formula, Indicates The evidence obtained by feature transformation of principal component statistics is Indicates The principal component statistics feature, Indicates The first state principal components, , Indicates The confidence of the sample characterized by the principal component statistics relative to the first state, Indicates The confidence of the sample characterized by the principal component statistics relative to the second state, represents the Hth state, Indicates The confidence of the sample characterized by the principal component statistics relative to the Hth state, Indicates The confidence of the sample characterized by the principal component statistics relative to all states, represents the Gaussian membership in all states, Indicates Gaussian membership in this state.

[0103] Step 2: Evidence activation based on the membership of the test sample.

[0104] First, the test data (Status to be evaluated) Standardize uniformly according to the training data. Secondly, according to the load matrix Finally, the principal component statistics of the test data are calculated. Principal component statistics For example, there are:

[0105] (10)

[0106] From equations (8) and (9), we can get the test data No. Principal component statistics The evidence is expressed in the form of :

[0107] (11)

[0108] in, Represents test data No. The confidence level of the principal component statistics features, Represents test data No. The confidence of the sample characterized by the principal component statistics relative to the first state, Represents test data The first state The principal component statistics feature, Represents test data No. The first state principal components, Represents test data No. The confidence of the sample characterized by the principal component statistics relative to the second state, Represents test data The second state The principal component statistics feature, Represents test data No. The confidence of the sample characterized by the principal component statistics relative to the Hth state, Represents test data The Hth state principal component statistics features. Represents test data No. principal component statistics features.

[0109] Similarly, test data can be obtained of The evidence of the principal component statistics , for:

[0110] (12)

[0111] in, Represents test data No. The confidence of the sample characterized by the principal component statistics relative to all states. Represents test data No. principal component statistics features.

[0112] Step 3: Multi-feature fusion based on evidence reasoning method.

[0113] In order to synthesize the judgment results of the evidence composed of each principal component, the evidence composed of each principal component feature can be fused based on the evidence reasoning algorithm. The fusion of two pieces of evidence under a sample is taken as an example to demonstrate the multi-feature fusion process. The three elements of evidence include: evidence representation, evidence weight and evidence reliability.

[0114] From equations (11) and (12), we can get the two pieces of evidence as follows:

[0115] (13)

[0116] (14)

[0117] in, Represents test data The first principal component statistic features The evidence consists of a sample. Represents test data The second principal component statistic features The evidence consists of a sample. Represents test data The first principal component statistic features Samples relative to the state confidence level. Represents test data The first principal component statistic features Samples relative to the state confidence level. Represents test data In the The first principal component statistic feature of the state The principal components of the samples. Represents test data In the The second principal component statistics feature of the state The principal components of the samples.

[0118] Since the evidence used in this embodiment is composed of the statistical features of the principal components, the contribution of the principal components can be selected as the weight of evidence, and the weight of evidence is recorded as The reliability of evidence can be calculated from data or given by expert knowledge. The reliability of evidence is recorded as .

[0119] The fusion process of two pieces of evidence for a certain sample includes:

[0120] The two pieces of evidence merged to support the assessment as a recognition framework Medium status The confidence level is :

[0121] (15)

[0122] (16)

[0123] In the formula, Represents two pieces of evidence for the state The unnormalized joint probability mass supported. It represents the sum of the unnormalized joint probability masses of the two pieces of evidence supporting all states in the identification framework. Represents two pieces of evidence for the state The unnormalized joint probability mass supported, represents a state that belongs to the identification framework (also called a proposition in evidential reasoning), Indicates belonging to.

[0124] Therefore, from equations (8), (9), (15) and (16), we can get the evidence support state The confidence level is :

[0125] (17)

[0126] In the formula, .

[0127] From formula (17), we can see that establishing the evaluation result (i.e., the evidence support status Confidence ) and principal component statistics characteristics relationship.

[0128] Step 4: Trace the factors affecting health status based on variable contribution.

[0129] The above steps 1 to 3 can be integrated to form an overall model, which is called the equipment health status assessment model based on multi-feature fusion. The equipment health status assessment model based on multi-feature fusion can be formed by integrating health indicators to obtain comprehensive health status assessment results. In this process, sensitivity analysis can be used to quantitatively analyze the extent to which the health status assessment results are affected by input variables. Statistics are a combination of variable contributions, based on which each variable can be established. And the principal component statistics characteristics The relationship between evaluation results and variables can be further established The relationship between the two can be realized, thereby tracing the factors affecting the health status.

[0130] Among them, the test data No. Principal component statistics It can be expressed as the sum of the contributions of the variables:

[0131] (18)

[0132] In the formula, Indicates Sample variables, . is the identity matrix Column, used to indicate the variable .

[0133] Taking the evaluation results of two pieces of evidence fusion as an example, the process of tracing the influencing factors is demonstrated. The relationship between the evaluation results and the variables obtained from equations (17) and (18) can be described as:

[0134] (19)

[0135] In the formula, Represents test data The first principal component statistic of the feature Samples relative to the state The confidence level of the test data The first principal component statistic of the feature The samples can be expressed by formula (18) as The sum of the original variables, so in formula (17) It can be written as , It represents the first principal component statistic characteristic represented by all variables and for the state Similarly, It represents the second principal component statistic characteristics represented by all variables and for the state confidence level. Represents the intermediate matrix for calculating the first principal component statistic in the first state; Represents the intermediate matrix for calculating the second principal component statistics under the first state.

[0136] In order to trace the results affected by variables, the sensitivity coefficient is used It indicates the relationship between the confidence of the evidence support status and the original variable. Among them, the sensitivity coefficient is expressed as:

[0137] (20)

[0138] In the formula, Indicates the maximum confidence value of the evidence support state, max() represents the maximum value function, represents the number of original variables, Indicates the number of states.

[0139] The implementation process of step 4 above can be integrated into another model, which can be called a health status influencing factor tracing model based on variable contribution.

[0140] In summary, this application derives the relationship between variables and evaluation results from the relationship between variables and features. Therefore, the evaluation results can be analyzed to determine which variable has a more obvious impact on the traceability evaluation, and the monitoring of this variable should be strengthened to detect changes in the health status of the equipment in a timely manner at an early stage.

[0141] In another exemplary embodiment of the present application, in order to verify the effectiveness of the method provided by the present application, the method provided by the present application is used to evaluate the health status of the motor rotor, and the influencing factors of the change of the health status of the motor rotor are traced according to the evaluation results. The specific implementation steps are as follows:

[0142] 1) Establishment of indicator system and setting of health status assessment framework.

[0143] The indicators that can reflect the changes in the health status of the motor rotor are: 1-fold vibration amplitude (variable 1), 2-fold vibration amplitude (variable 2), 3-fold vibration amplitude (variable 3) and vibration displacement (variable 4). Four health status assessment levels are set according to the health status labels of the motor rotor training data. , respectively, indicating the deterioration of health status. In this embodiment, Indicates health status, Indicates a fault condition.

[0144] 2) Motor rotor health status assessment based on multi-feature fusion.

[0145] (1) Principal component feature extraction and evidence construction based on Gaussian membership function.

[0146] The health index data of the motor rotor is collected by the acceleration sensor and the vibration displacement sensor. The training data is 500 samples in each healthy state, and the test data is 300 samples in each state. The variables of the three vibration amplitudes will have a certain degree of correlation. Therefore, the application of the method provided in this application can solve the information fusion problem under variable correlation and make the evaluation results more reasonable. Based on principal component analysis, two components are retained and the principal component statistical features are calculated. According to the known health status category, a Gaussian membership function model is constructed for each principal component statistical feature, such as Figure 3 and Figure 4 Compared with constructing 16 Gaussian membership functions using four indicators, the method proposed in this application can describe the characteristics of the training data with only 8 Gaussian membership functions constructed using two features.

[0147] (2) Membership-based test sample evidence generation and fusion.

[0148] After the test data is standardized, the two principal component statistics are calculated and the evidence of the test data is constructed based on the Gaussian membership model constructed based on the training data, and the evidence forms shown in equations (11) and (12) are obtained. Based on the evidence reasoning method, the evidence constructed by the two principal component statistics is fused, the reliability of the evidence is set to 1, and the weight of the evidence is given by normalizing the contribution of the principal component to obtain the health status assessment result of the motor rotor, as shown in Figure 5 The accuracy of the health status assessment results of the motor rotor test data is 96.17%, which can accurately assess the health status and fault status However, there is some ambiguity in the evaluation of intermediate state changes, which is the same as the conclusion shown by the training data. There is ambiguity between the two states and it is difficult to distinguish them completely.

[0149] 3) Tracing the factors affecting the health status of the motor rotor based on variable contribution.

[0150] After the evaluation results are obtained, the sensitivity coefficient shown in formula (20) is used to quantitatively analyze the degree to which the motor rotor health status is affected by the variables, thus tracing the influencing factors. Figures 6 to 9 As shown in the figure, the sensitivity coefficient of the motor rotor health status affected by the original four variables can be obtained. Figures 6 to 9It can be seen that in the process of the change of the motor rotor health status, the sensitivity coefficients of variable 1 and variable 2 are roughly the same, with average sensitivity coefficients of 12.4697 and 12.4932 respectively, which are much larger than the sensitivity coefficients of variable 3 (2.4381) and variable 4 (1.3420). This shows that in the evaluation, more attention should be paid to the changes of these two variables, which have a greater impact on the health status. In addition, when the motor rotor is in the state of When compared with other states, the sensitivity coefficients of each variable are increased, which are 35.3374, 35.4040, 6.9095 and 4.1406 respectively, indicating that The state is a key state and is difficult to distinguish from other states. It is also the most complex to be affected by variables.

[0151] 4) Comparative experiment.

[0152] In order to verify that the method proposed in this application can accurately evaluate the health status changes of the motor rotor and trace the influencing factors, the Gaussian membership function model is directly constructed using the original variables, such as Fig.10 As shown, the Gaussian membership function model of the original variable can be obtained. Among them, Fig.10 Part (a) shows the Gaussian membership function model of the original variables under 1-fold frequency vibration conditions, Fig.10 Part (b) shows the Gaussian membership function model of the original variables under the 2-fold frequency vibration condition, Fig.10 Part (c) shows the Gaussian membership function model of the original variables under the 3rd frequency vibration condition, Fig.10 Part (d) shows the Gaussian membership function model of the original variables under vibration displacement conditions. Fig.10 As can be seen from the figure, there are four variables and four states, and 16 Gaussian membership functions need to be established. All variables are used to distinguish the states. , , There is always a certain degree of ambiguity when fusing four variables, and the influence of potential variable correlation is not taken into account.

[0153] Based on the Gaussian membership function model of each variable given above, the evidence corresponding to the test sample is generated, and then the evidence of all variables is integrated based on the evidence reasoning method to obtain the data health status assessment result of the test data, such as Fig.11 The accuracy of the evaluation result of the motor rotor health status is 93.67%. Fig.11 It can be seen that for the state and The judgment accuracy is low.

[0154] In an exemplary embodiment, a computer device is provided. The computer device may be a server or a terminal. The internal structure diagram thereof may be as follows: Fig.12 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store traceability data of factors affecting the device status. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for tracing the factors affecting the health status of the device is implemented.

[0155] Those skilled in the art will understand that Fig.12 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components. In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0156] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0157] In an exemplary embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0158] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0159] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0160] The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. The non-relational database may include a distributed database based on blockchain, etc., but is not limited thereto. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but is not limited thereto.

[0161] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0162] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A method for tracing the factors affecting the health status of equipment, characterized in that: The method for tracing the factors affecting the health status of the equipment includes: Acquire original variables of the device in different states, and form a sample matrix based on the original variables; the different states of the device include a healthy state and a fault state; Extracting principal component features from the sample matrix to obtain principal component statistical features; The principal component statistic characteristics are converted into evidence using a Gaussian membership function; Activate the evidence using a test sample, and fuse all feature-activated evidence based on an evidence reasoning algorithm to obtain a confidence level of the evidence support state; Construct the relationship between the confidence of the evidence support status and the characteristics of the principal component statistics, and the relationship between the original variables and the characteristics of the principal component statistics; Based on the relationship between the confidence of the evidence support status and the characteristics of the principal component statistics and the relationship between the original variables and the characteristics of the principal component statistics, the relationship between the confidence of the evidence support status and the original variables is obtained; Based on the relationship between the confidence of the evidence-supported status and the original variables, the influencing factors of the equipment health status can be traced back; The principal component feature extraction is performed on the sample matrix to obtain the principal component statistic feature, including: Performing standardization on the sample matrix; Determine the sample covariance matrix based on the standardized sample matrix; Performing principal component eigenvalue decomposition on the sample covariance matrix, and arranging the eigenvectors obtained by the decomposition in descending order to obtain a vector sequence; Selecting in sequence from the vector sequence the eigenvectors whose principal component cumulative contribution rates reach a predetermined value to form a load matrix; Obtaining a score matrix based on the load matrix and the standardized sample matrix; Determining a principal component covariance matrix based on the score matrix; Obtaining statistics based on the load matrix, the standardized sample matrix, and the principal component covariance matrix; The statistics are filtered to obtain the principal component statistics feature.

2. The method for tracing the factors affecting the health status of equipment according to claim 1 is characterized in that: The statistics are filtered to obtain the principal component statistics feature, including: A median filter with a sliding window length of a set value is added to the statistic to obtain the main component statistic characteristics.

3. The method for tracing the factors affecting the health status of equipment according to claim 1 is characterized in that: The Gaussian membership function is used to convert the principal component statistic characteristics into the expression of evidence: ； ； In the formula, Indicates The evidence obtained by feature transformation of principal component statistics is Indicates The principal component statistics feature, Indicates the first state, Indicates The first state principal components, , Indicates The confidence of the sample characterized by the principal component statistics relative to the first state, The first state The principal component statistics feature, Indicates the second state, Indicates The confidence of the sample characterized by the principal component statistics relative to the second state, The second state The principal component statistics feature, represents the Hth state, Indicates The confidence of the sample characterized by the principal component statistics relative to the Hth state, The Hth state The principal component statistics feature, Indicates The confidence of the sample characterized by the principal component statistics relative to all states, represents the Gaussian membership in all states, Indicates status Gaussian membership under .

4. The method for tracing the factors affecting the health status of equipment according to claim 1, characterized in that: The sensitivity coefficient is used to represent the relationship between the confidence of the evidence support status and the original variable; wherein the sensitivity coefficient is expressed as: ； In the formula, represents the sensitivity coefficient, represents the original variable, Indicates the confidence level of the evidence supporting the status, Indicates the maximum confidence value of the evidence support state, max() represents the maximum value function, represents the number of original variables, Indicates the number of states.

5. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method for tracing the factors affecting the health status of an equipment as described in any one of claims 1 to 4.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for tracing the factors affecting the health status of equipment described in any one of claims 1 to 4 is implemented.

7. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for tracing the factors affecting the health status of equipment described in any one of claims 1 to 4 is implemented.

Images