A multi-source information integration method and system for complex equipment

Abstract

The present application relates to the technical field of multi-source data fusion, and a multi-source information unified measurement method and system for complex equipment, comprising: calculating the sample overall variance of a data sample set of the complex equipment as a parameter of a Gaussian kernel function to obtain uncertainty modeling; using the uncertainty modeling, calculating the contribution degree of each data sample to each preset focus element to obtain a contribution degree set, and performing evidence modeling on uncertain parameters to obtain a univariate evidence structure; calculating an evidence distance sequence between each evidence body in test information and expert information, constructing a credibility index sequence according to the evidence distance sequence, and then obtaining a weighted multi-source data set through weight configuration; fusing the weighted multi-source data set to obtain multi-source uncertainty fusion data; and constructing a multi-source information unified measurement model using the univariate evidence structure and the multi-source uncertainty fusion data. The present application can realize unified measurement of incomplete, multi-source and correlated uncertain parameters of a complex system.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of multi-source data fusion, and particularly relates to a multi-source information unified measurement method and system for complex equipment. BACKGROUND

[0002] In the analysis and design of complex equipment systems, fully considering the reliability of the structure and system plays an indispensable role in ensuring the continuous reliable operation and safe service of complex equipment. The traditional reliability analysis technology is based on a probability model, which requires a large amount of test data and test samples to construct an accurate reliability analysis model.

[0003] Establishing a quantitative model representing complex uncertain factors is the basis of reliability analysis. In complex equipment systems, due to the limitations of cognition of complex things, the variability of service environment, the error of processing and measurement, the lack of test data and expert judgment, etc., the uncertainty in materials, structures and loads presents the characteristics of incomplete sample information, multi-source and correlation. The traditional evidence theory model is mostly provided by experts or experience, and can only handle independent uncertain parameters. SUMMARY

[0004] The present application provides a multi-source information unified measurement method for complex equipment, which mainly aims to realize the unified measurement of incomplete, multi-source and correlated uncertain parameters of complex systems.

[0005] To achieve the above purpose, the present application provides a multi-source information unified measurement method for complex equipment, comprising:

[0006] Obtaining a data sample set based on complex equipment, calculating the sample overall variance of the data sample set, and configuring a preset bandwidth parameter using the sample overall variance;

[0007] According to the pre-constructed Gaussian kernel function and the bandwidth parameter, an uncertainty modeling based on the data sample set is obtained;

[0008] Obtaining a focus element set, calculating the contribution degree of each data sample in the data sample set to each focus element in the focus element set using the uncertainty modeling, obtaining a contribution degree set, and performing evidence modeling of uncertain parameters on the contribution degree set to obtain a univariate evidence structure;

[0009] Acquire test information and expert information, calculate the evidence distance between each piece of evidence in the test information and expert information according to the pre-constructed Jousselme evidence distance calculation formula, obtain the evidence distance sequence, construct a credibility index based on each piece of evidence in the test information and expert information according to the evidence distance sequence, obtain a credibility index sequence, and configure the weights of each piece of evidence in the test information and expert information according to the credibility index sequence to obtain a weighted multi-source data set;

[0010] By using pre-constructed evidence fusion rules, the weighted multi-source dataset is fused to obtain multi-source uncertainty fused data;

[0011] Based on the pre-built unified measurement model construction strategy, a multi-source information unified measurement model is constructed using the univariate evidence structure and multi-source uncertainty fusion data.

[0012] Optionally, calculating the overall variance of the data sample set includes:

[0013] Obtain the total number of samples in the data sample set, identify the number of data sources in the data sample set, obtain the number of data sources, and identify the number of samples contained in each data source to obtain the data source sample number sequence;

[0014] Calculate the variance and mean of each data source in the data sample set to obtain the data source variance sequence and the data source mean sequence.

[0015] Based on the mean sequence of the data source, the sample number sequence of the data source, and the total number of samples, the global mean is calculated, wherein the global mean is expressed as:

[0016]

[0017] In the formula, This represents the global mean. Indicates the number of samples in the data source sequence. Number of samples in each data source Indicates the first value in the mean sequence of the data source. The mean of each data source. Indicates the total number of samples;

[0018] Based on the total number of samples, the number of data sources, the sequence of sample numbers from the data sources, the sequence of variances from the data sources, the sequence of means from the data sources, and the global mean, the overall sample variance is calculated, wherein the overall sample variance is expressed as:

[0019]

[0020] In the formula, Indicates the overall variance of the sample. Indicates the number of data sources. Indicates the th variance sequence of the data source The variance corresponding to each data source Indicates the first The square of the mean of each data source This represents the square of the global mean. Indicates the size of a number.

[0021] Optionally, obtaining the uncertainty modeling based on the data sample set according to the pre-constructed Gaussian kernel function and the bandwidth parameter includes:

[0022] Randomly select two data samples from the data sample set as the first data sample and the second data sample;

[0023] Calculate the sample distance between the first data sample and the second data sample;

[0024] Based on the sample distance and bandwidth parameters, a pre-constructed Gaussian kernel function is assigned a value to obtain an uncertainty model, wherein the uncertainty model is expressed as:

[0025]

[0026] In the formula, express Uncertainty modeling between two data samples This indicates that the bandwidth parameter is the same as the overall variance of the sample. express and The sample distance between two data samples is the square of the Euclidean distance between the two data samples. and Both represent the size of the number.

[0027] Optionally, the uncertainty modeling is used to calculate the contribution of each data sample in the data sample set to each focal element in the focal element set, resulting in a contribution set, including:

[0028] Obtain the number of focal elements in the focal element set to get the total number of focal elements, and obtain the local variance of each data sample in the data sample set;

[0029] Based on the local variance of each data sample and the bandwidth parameter, a sample confidence weight sequence is obtained, wherein a sample confidence weight in the sample confidence weight sequence is represented as:

[0030]

[0031] wherein, denotes the sample confidence weight of the i-th data sample in the sample confidence weight sequence, denotes the local variance of the i-th data sample, denotes the square of the bandwidth parameter;

[0032] According to the total number of focal elements, the sample confidence weight sequence, a contribution degree function is constructed, wherein the contribution degree function is expressed as:

[0033]

[0034] wherein, denotes any one data sample, such as the first data sample , the contribution degree of the i-th focal element, denotes the uncertainty modeling, denotes the typical feature vector of the i-th focal element, denotes the total number of focal elements, denotes the feature vector of the i-th focal element, denotes the size of the numerical value; According to the contribution degree function, a contribution degree set is obtained. Optionally, the contribution degree set is subjected to evidence modeling of an uncertainty parameter to obtain a univariate evidence structure, including:

[0035] A variable set corresponding to the focal element set is obtained, and a preset composite focal element set in the focal element set is obtained, wherein a preset composite focal element in the preset composite focal element set is a subset in the focal element set;

[0036] The contribution degree set is subjected to contribution degree normalization operation based on a target variable in the variable set to obtain a contribution degree normalization sequence, wherein one contribution degree normalization in the contribution degree normalization sequence is expressed as:

[0037]

[0038] wherein,

[0039] denotes the contribution degree normalization value of the i-th data sample to the i-th variable in the i-th focal element, denotes the contribution degree function, denotes the i-th variable in the i-th focal element,

[0040] ​​​​​​​​​a focal element, denotes a focal element set comprising a focal element set comprising denotes a variable measurement confidence of a first variable, , , and all denote the size of a number;

[0041] According to a pre-constructed aggregation rule, the contribution degree normalization sequence is subjected to an aggregation operation based on each preset composite focal element in the preset composite focal element set, to obtain an evidence body set;

[0042] The evidence body set is obtained, and the effective focal element set is obtained from the focal element set corresponding to the effective evidence body, and the effective focal element set is obtained by merging the effective focal element set, to obtain a simple focal point set;

[0043] The effective evidence body, the simple focal point set, and the variable set are summarized to obtain a single variable evidence structure.

[0044] Optionally, the Jousselme evidence distance calculation formula is represented as:

[0045]

[0046] In the formula, denotes the evidence distance between the evidence body and the evidence body , denotes Jousselme, denotes the basic probability assignment of the evidence body , denotes the basic probability assignment of the evidence body , denotes the transpose, denotes the similarity matrix.

[0047] Optionally, the credibility index sequence is obtained based on the test information and the expert information, comprising:

[0048] The number of each evidence body in the test information and the expert information is obtained to obtain the total number of fusion evidence bodies;

[0049] According to the total number of fusion evidence bodies and the evidence distance sequence, a conflict matrix is constructed;

[0050] The conflict matrix is mapped using a pre-constructed exponential decay function to obtain a credibility sequence;

[0051] Normalize each confidence level in the confidence level sequence to obtain a confidence level index sequence.

[0052] Optionally, before fusing the weighted multi-source dataset using pre-constructed evidence fusion rules to obtain multi-source uncertainty fused data, the method further includes:

[0053] Obtain traditional conflict factors, and construct adjustment coefficients using these factors, wherein the adjustment coefficients are expressed as follows:

[0054]

[0055] In the formula, This represents the adjustment coefficient. Represents a hyperbolic function. Indicates traditional conflict factors;

[0056] Based on the adjustment coefficient, a conflict compensation term is constructed, wherein the conflict compensation term is expressed as:

[0057]

[0058] In the formula, Representing a subset Conflict compensation items, Represents traditional conflict compensation items, where the subset For set subset of Represents a set of recognition frames;

[0059] Based on the conflict compensation term and adjustment coefficient, an evidence fusion rule is constructed, wherein the evidence fusion rule is expressed as:

[0060]

[0061] In the formula, Representing a subset The merged data Indicates the first element in a weighted multi-source dataset. A subset of data samples Normalized value of contribution express The corresponding credibility index.

[0062] Optionally, the step of constructing a multi-source information unified measurement model based on a pre-built unified measurement model construction strategy, utilizing the univariate evidence structure and multi-source uncertainty fusion data, includes:

[0063] According to a pre-constructed unified dimension model construction strategy, feature extraction is performed on the single variable evidence structure and the multi-source uncertainty fusion data to obtain a variable feature set, and correlation analysis is performed on the variable feature set to obtain a correlation coefficient matrix;

[0064] The correlation coefficient matrix is converted into a parallel polyhedral structure framework, and a shape matrix is constructed according to the parallel polyhedral structure framework.

[0065] The shape matrix is used to perform independent space mapping on the variable feature set to obtain independent space variable data, and each variable data in the independent space variable data is assigned a basic credibility to obtain credibility assignment data.

[0066] According to the credibility assignment data, a joint focus element combination is generated by using a pre-constructed Cartesian product algorithm.

[0067] According to the joint focus element combination and the multi-source uncertainty fusion data, a multi-source information unified dimension model is constructed.

[0068] To achieve the above object, the application further provides a multi-source information unified dimension system for complex equipment, comprising:

[0069] A sample kernel density estimation module is configured to obtain a data sample set based on complex equipment, calculate a sample overall variance of the data sample set, configure a preset bandwidth parameter by using the sample overall variance, obtain an uncertainty modeling based on the data sample set according to a pre-constructed Gaussian kernel function and the bandwidth parameter, obtain a focus element set, calculate a contribution degree of each data sample in the data sample set to each focus element in the focus element set by using the uncertainty modeling, obtain a contribution degree set, and perform evidence modeling on the contribution degree set to obtain a single variable evidence structure.

[0070] A multi-source uncertainty information fusion module is configured to obtain test information and expert information, calculate an evidence distance between each evidence body in the test information and the expert information according to a pre-constructed Jousselme evidence distance calculation formula to obtain an evidence distance sequence, construct a credibility index based on each evidence body in the test information and the expert information according to the evidence distance sequence to obtain a credibility index sequence, configure a weight for each evidence body in the test information and the expert information according to the credibility index sequence to obtain a weighted multi-source data set, and fuse the weighted multi-source data set by using a pre-constructed evidence fusion rule to obtain multi-source uncertainty fusion data.

[0071] A single variable evidence structure is constructed, and a multi-source uncertainty fusion data is fused according to a pre-constructed single variable evidence structure construction strategy.

[0072] To solve the above problems, the application further provides an electronic device, which comprises:

[0073] a memory, which stores at least one instruction;

[0074] a processor, which executes the instruction stored in the memory to implement the multi-source information unified measurement method for complex equipment.

[0075] To solve the above problems, the application further provides a computer readable storage medium, which stores at least one instruction, and the at least one instruction is executed by a processor in an electronic device to implement the multi-source information unified measurement method for complex equipment.

[0076] To solve the problems in the background art, the application firstly takes the overall variance of a data sample set as a bandwidth parameter, uses a Gaussian kernel function to model the uncertainty of the data sample set, to obtain the contribution degree of each data sample to a focus element, and then constructs a single variable evidence structure according to the contribution degree set, wherein the single variable evidence structure can quantify the cognitive uncertainty caused by incomplete information; then, the application calculates the evidence distance of test information and expert information, and then establishes a credibility index representing the conflict between multi-source evidence information according to the evidence distance, to obtain a credibility index sequence, and further realizes the weight distribution of multi-source information; then, the application uses an evidence fusion rule to fuse the weighted multi-source evidence information, to obtain multi-source uncertainty fusion data, wherein the multi-source uncertainty fusion data is used to realize the comprehensive measurement of multi-source uncertain information in complex equipment; finally, to realize the unified measurement of high-dimensional and correlated uncertain information in complex equipment, the application develops an evidence theory model based on a parallel polyhedron, such as: the correlation coefficient of an evidence variable is obtained through correlation analysis, an evidence recognition framework in the form of a parallel polyhedron that measures correlation is established, then sample information is converted into an independent standard space through a shape matrix, to realize the basic credibility distribution of each evidence variable and construct a joint focus element, and based on this, the consistency between the recognition framework and the joint focus element is used for space conversion, to realize the unified quantification of high-dimensional and correlated uncertain information. Therefore, the application can realize the unified measurement of incomplete, multi-source and correlated uncertain parameters in a complex system. BRIEF DESCRIPTION OF DRAWINGS

[0077] Figure 1 A flowchart of a multi-source information unified measurement method for complex equipment provided by an embodiment of the application;

[0078] Figure 2 A structural flowchart of a multi-source information unified measurement method for complex equipment is provided for an embodiment of the present application.

[0079] Figure 3 A functional module diagram of a multi-source information unified measurement system for complex equipment is provided for an embodiment of the present application.

[0080] Figure 4 A structural schematic diagram of an electronic device for implementing the multi-source information unified measurement method for complex equipment is provided for an embodiment of the present application.

[0081] Legend of reference signs:

[0082] 1, electronic device; 10, processor; 11, memory; 12, bus.

[0083] The implementation, functional features and advantages of the present application will be further described with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION

[0084] It should be understood that the specific embodiments described herein are merely intended to explain the present application and not to limit the present application.

[0085] An embodiment of the present application provides a multi-source information unified measurement method for complex equipment. The execution subject of the multi-source information unified measurement method for complex equipment includes but is not limited to at least one of electronic devices such as a server and a terminal, which can be configured to execute the method provided by the embodiment of the present application. In other words, the multi-source information unified measurement method for complex equipment can be executed by software or hardware installed in a terminal device or a server device, and the software can be a blockchain platform. The server includes but is not limited to a single server, a server cluster, a cloud server or a cloud server cluster, etc.

[0086] Referring to Figure 1 Fig. 1 shows a flowchart of a multi-source information unified measurement method for complex equipment provided by an embodiment of the present application. In this embodiment, the multi-source information unified measurement method for complex equipment includes:

[0087] S1, obtaining a data sample set based on complex equipment, calculating a sample overall variance of the data sample set, and configuring a preset bandwidth parameter by using the sample overall variance.

[0088] The complex equipment refers to equipment in the fields of aerospace, navigation, high-speed rail, etc., which has a harsh working environment, complex and variable load, complex structure of parts and assembly process.

[0089] The data sample set includes data from various information sources of complex equipment, such as vibration sensor data, temperature sensor data, and maintenance log files. Due to the long manufacturing cycle and high trial-and-error costs of complex equipment, the data in the data sample set is limited.

[0090] The overall sample variance, within the framework of evidence theory, requires consideration of the heterogeneity and incompleteness of multi-source data in the data sample set during its calculation.

[0091] The bandwidth parameter controls the decay rate of the Gaussian kernel function: the larger the variance, the larger the value of h, and the slower the weight decay of samples at the same distance. The Gaussian kernel function quantifies the uncertainty of data distribution by calculating the ratio of the distance between samples to the variance.

[0092] Specifically, in this embodiment of the invention, to achieve basic confidence allocation based on kernel density estimation for a limited set of data samples, it is first necessary to configure an appropriate bandwidth parameter, and then use a Gaussian kernel function to model the uncertainty of the data sample set. However, to configure the bandwidth parameter, it is necessary to first calculate the overall variance of the data sample set.

[0093] In detail, in this embodiment of the invention, calculating the overall variance of the data sample set includes:

[0094] Obtain the total number of samples in the data sample set, identify the number of data sources in the data sample set, obtain the number of data sources, and identify the number of samples contained in each data source to obtain the data source sample number sequence;

[0095] Calculate the variance and mean of each data source in the data sample set to obtain the data source variance sequence and the data source mean sequence.

[0096] Based on the mean sequence of the data source, the sample number sequence of the data source, and the total number of samples, the global mean is calculated, wherein the global mean is expressed as:

[0097]

[0098] In the formula, This represents the global mean. Indicates the number of samples in the data source sequence. Number of samples in each data source Indicates the first value in the mean sequence of the data source. The mean of each data source. Indicates the total number of samples;

[0099] According to the total number of samples, the number of data sources, the data source sample number sequence, the data source variance sequence, the data source mean value sequence and the global mean value, the sample overall variance is calculated, wherein the sample overall variance is represented as:

[0100]

[0101] In the formula, the sample overall variance is represented as, the number of data sources is represented as, the variance corresponding to the i-th data source in the data source variance sequence is represented as, the mean value corresponding to the i-th data source is represented as, the square of the mean value corresponding to the i-th data source is represented as, the square of the global mean value is represented as, the size of the number is represented as.

[0102] The total number of samples is the total number of elements in the data sample set. The data source is a subset in the data sample set derived from different data interfaces (or sensors). The number of data sources is the number of subsets in the data sample set. The sample number is the number of elements in the subset. The data source sample number sequence is a record of the number of elements corresponding to each subset.

[0103] The variance is an index for measuring the dispersion degree between each data point in the data source and the mean value. The greater the variance, the greater the difference between the data points. The mean value is the average value of each data point in the data source. The data source variance sequence is a record of the variance of each data source. The data source mean value sequence is a record of the mean value of each data source.

[0104] The global mean value is obtained by calculating the average value of the data sample set according to the mean value of each data source.

[0105] Specifically, in the embodiment of the application, the basic data information of the data sample set is first obtained, such as the total number of samples, the number of data sources and the number of samples contained in each data source. Then, the mean value and the variance of each data source are calculated by using the mean value and the variance formula, and the data source variance sequence and the data source mean value sequence are obtained.

[0106] ​Specifically, according to the global mean value formula, the global mean value is obtained by substituting the data source mean value sequence, the data source sample number sequence and the total number of samples. For example, the basic data information of the data sample set is as follows: 50 samples of vibration source data, local variance σ1 2 = 0.15, mean value μ1 = 0.33 Hz; 30 samples of temperature source data, σ2 2 = 4.2 ℃, μ2 = 65 ℃; 2 samples of maintenance log, and σ3 2 = 0.08 is obtained through semantic analysis.

[0107] Then, according to the sample overall variance formula, the sample overall variance of the data sample set is obtained by substituting the total number of samples, the number of data sources, the data source sample number sequence, the data source variance sequence, the data source mean value sequence and the global mean value.

[0108] S2, according to the pre-constructed Gaussian kernel function and the bandwidth parameter, the uncertainty modeling based on the data sample set is obtained.

[0109] The Gaussian kernel function is used to quantify the uncertainty of data distribution by calculating the ratio of the distance between each data sample in the data sample set and the variance.

[0110] The uncertainty modeling is a Gaussian kernel function with specific values.

[0111] In detail, in the embodiment of the present application, the uncertainty modeling based on the data sample set is obtained according to the pre-constructed Gaussian kernel function and the bandwidth parameter, which comprises:

[0112] Any two data samples are selected from the data sample set as the first data sample and the second data sample;

[0113] The sample distance between the first data sample and the second data sample is calculated.

[0114] The pre-constructed Gaussian kernel function is valued according to the sample distance and the bandwidth parameter, and the uncertainty modeling is obtained, wherein the uncertainty modeling is represented as:

[0115]

[0116] In the formula, represents The uncertainty modeling between the two data samples, The bandwidth parameter is the same as the sample overall variance, represents and The sample distance between the two data samples is the square of the Euclidean distance between the two data samples, and All represent the size of the number.

[0117] Wherein, the Euclidean distance is used for calculating the straight line distance between two points, and can be used in a multi-dimensional space.

[0118] Wherein, the first data sample and the second data sample are any two data samples in the data sample set, which can be obtained by The first data sample and the second data sample are represented, but it cannot be said that the first data sample and the second data sample are .

[0119] Specifically, in the embodiment of the application, two data samples are extracted from the data sample set, and are named as the first data sample and the second data sample, respectively. Then, the sample distance between the first data sample and the second data sample is calculated by using the Euclidean formula, and then the sample distance and the bandwidth parameter calculated in S1 are substituted into the Gaussian kernel function to obtain the uncertainty modeling.

[0120] S3, obtaining a focus element set, calculating the contribution degree of each data sample in the data sample set to each focus element in the focus element set by using the uncertainty modeling, obtaining a contribution degree set, and performing evidence modeling of the uncertainty parameter on the contribution degree set to obtain a univariate evidence structure.

[0121] Wherein, the focus element set is a set of storing each focus element. The focus element in the equipment health monitoring scene usually corresponds to a specific fault mode assumption, for example,

single focus element: bearing fault, gear wear

compound focus element: bearing fault ∪ lubrication deficiency

[0122] Wherein, the contribution degree represents the influence or importance. The contribution degree set is a record of the influence of each data sample on each focus element.

[0123] Wherein, the evidence modeling of the uncertainty parameter refers to the operation of converting each contribution degree into an evidence body. The evidence body is used to represent the uncertainty information from a single or multiple data sources, and can quantify the support degree of multiple hypothesis focus elements through a mathematical structure.

[0124] Wherein, the evidence modeling refers to the process of performing basic credibility allocation processing based on kernel density estimation on each data uncertainty modeling. The univariate evidence structure refers to the generation result of the evidence body.

[0125] In detail, in the embodiment of the application, the contribution degree of each data sample in the data sample set to each focus element in the focus element set is calculated by using the uncertainty modeling to obtain a contribution degree set, including:

[0126] Obtain the number of focal elements in the focal element set to get the total number of focal elements, and obtain the local variance of each data sample in the data sample set;

[0127] Based on the local variance of each data sample and the bandwidth parameter, a sample confidence weight sequence is obtained, wherein a sample confidence weight in the sample confidence weight sequence is represented as:

[0128]

[0129] In the formula, Indicates the first [number] in the sample confidence weight sequence. Sample confidence weights for each data sample Indicates the first Local variance of a data sample Represents the square of the bandwidth parameter;

[0130] Based on the total number of focal elements and the sample confidence weight sequence, a contribution function is constructed, wherein the contribution function is expressed as:

[0131]

[0132] In the formula, This refers to any data sample, such as the first data sample. For the first Each jiao yuan Contribution This indicates the modeling of the uncertainty. Indicates the first Typical eigenvectors of each focal element This indicates the total number of the focal elements. Indicates the first The feature vectors of each focal element Indicates the magnitude of a numerical value;

[0133] Based on the contribution function, obtain the contribution set.

[0134] The total number of focal elements is the number of elements in the focal element set. The calculation process for the local variance is the same as the variance calculation method for a single data source in S1 above.

[0135] The sample confidence weight sequence is a set of confidence weights for each sample. These sample confidence weights are used to measure the reliability of the data samples.

[0136] The contribution function is used to quantify each data sample. For each focal element ( The contribution of the data sample reflects the data sample The degree of influence of the uncertainty modeling.

[0137] Specifically, in the embodiment of the present application, firstly, the basic data information of the focus element set is acquired to obtain the total number of focus elements.

[0138] Specifically, the present application takes each sample confidence weight as a weight coefficient, and according to the contribution degree function, the uncertainty of each data sample to each focus element is modeled. The additional weight is normalized to obtain the contribution value of each data sample to each focus element, and a contribution degree set is obtained.

[0139] In detail, in the embodiment of the present application, the evidence modeling of the uncertainty parameter of the contribution degree set to obtain a univariate evidence structure comprises:

[0140] The variable set corresponding to the focus element set is acquired, and a preset composite focus element set in the focus element set is acquired, wherein the preset composite focus element in the preset composite focus element set is a subset of the focus element set.

[0141] The contribution degree normalization operation based on the target variable in the variable set is performed on the contribution degree set to obtain a contribution degree normalization sequence, wherein one contribution degree normalization in the contribution degree normalization sequence is represented as:

[0142]

[0143] In the formula, the contribution degree normalization value of the i th data sample to the j th variable in the i th focus element is represented as the contribution degree function is represented as the i th focus element is represented as the focus element set containing i focus elements is represented as the variable measurement confidence of the i th variable is represented as , , , , , and all represent the size of the number.

[0144] According to the pre-constructed aggregation rule, the aggregation operation based on each preset composite focus element in the preset composite focus element set is performed on the contribution degree normalization sequence to obtain an evidence body set.

[0145] ​​​Obtaining the evidence body set in the value less than the preset focus threshold, obtaining the effective evidence body, and obtaining the effective focus element set corresponding to the effective evidence body from the focus element set, obtaining the effective focus element set, merging the effective focus element set, and obtaining the simple focus set;

[0146] The effective evidence body, the simple focus set, and the variable set are summarized to obtain a single variable evidence structure.

[0147] The variable set refers to the set of the sub-variable set corresponding to each focus element in the focus element set. The composite focus element in the composite focus element set is the intersection or union of the sub-variable sets corresponding to multiple focus elements. For example, a single focus element is "bearing failure", "insufficient lubrication", etc., and a composite focus element is "bearing failure and insufficient lubrication", "bearing failure or insufficient lubrication", etc.

[0148] The contribution degree normalization sequence is record information for storing the contribution degree normalization. The contribution degree normalization indicates that the contribution degree of the data sample to the focus element is normalized, and the product of the normalization result and the variable measurement confidence corresponding to each specific variable is obtained to obtain the contribution degree normalization. Thus, the contribution degree of the data sample to the focus element is converted into the contribution value of the data sample to each specific variable in the focus element.

[0149] The variable measurement confidence needs to be evaluated according to the working environment of the complex equipment. For example, when the complex equipment is in a high-temperature, stormy working environment, the value of the variable measurement confidence is lower than that of the variable measurement confidence of the measured data in the room temperature, windless condition.

[0150] The aggregation rule is the Dempster-Pawlak aggregation rule.

[0151] The focus threshold is configured as 0.05. The effective focus element set is a set of focus elements with a value greater than 0.05.

[0152] The merging operation refers to the union operation of each set, which is equivalent to deleting the invalid focus element (compared with the effective focus element) from the focus element set. The simple focus set is the name of the merging operation result.

[0153] The single variable evidence structure is a composite information structure body containing the effective evidence body, the simple focus set, and the variable set.

[0154] Specifically, in the embodiment of the application, first, according to the contribution degree normalization sequence formula, the contribution degree of the data sample on the variable is converted into a conditional probability to obtain a contribution degree normalization value, and then the contribution degree normalization sequence is obtained.

[0155] Then the application divides the data samples of different data sources according to the division range of the composite focus elements in the new preset composite focus element set according to the Dempster-Pawlak aggregation rule, obtains the evidence bodies corresponding to each composite focus element, and further obtains the evidence body set.

[0156] Specifically, in the embodiment of the application, the composite focus element set is obtained by a technician subjectively configuring according to the complex equipment adaptation scene. Therefore, there is a problem of improper configuration.

[0157] In the embodiment of the application, when the value of a certain evidence body in the evidence body set is less than the focus threshold 0.05, the evidence body can be regarded as redundant and needs to be deleted from the evidence body set to obtain the effective evidence body, and further obtain the effective focus element. Moreover, the focus element corresponding to each effective evidence body is a composite focus element including multiple single focus elements. The application performs a union operation on the effective focus element set to obtain the simple focus set, which can better identify each single focus element.

[0158] Specifically, in the embodiment of the application, the single-variable evidence structure is an information complex, and various information needs to be fused to form it. Therefore, the information of the effective evidence body, the simple focus set and the variable set needs to be summarized to realize the evidence modeling of uncertain parameters. Further, the cognitive uncertainty caused by incomplete information in the data sample set is quantified.

[0159] Specifically, referring to FIG. 1, Figure 2 As shown in the figure, the application intends to input the data combined by the single-variable evidence structure and multi-source uncertain information fusion into the parallel polyhedral structure to construct an evidence theory model. Through the steps S1-S3, the construction of the single-variable evidence structure is completed. Next, multi-source uncertain information fusion is needed.

[0160] S4, acquiring test information and expert information, calculating the evidence distance between each evidence body in the test information and the expert information according to the pre-constructed Jousselme evidence distance calculation formula, obtaining an evidence distance sequence, constructing a credibility index based on each evidence body in the test information and the expert information according to the evidence distance sequence, obtaining a credibility index sequence, and configuring weights for each evidence body in the test information and the expert information according to the credibility index sequence, obtaining a weighted multi-source data set.

[0161] Among them, the test information is the data information obtained by the sensor. Taking the detection of an aero-engine as an example, the test information includes vibration sensor information and oil analysis data. The expert information is data collected subjectively by a person. Still taking the detection of an aero-engine as an example, the expert information includes expert evaluation data and historical case data.

[0162] The Jousselme evidence distance calculation formula is a quantitative method for measuring the difference between two pieces of evidence. The evidence distance is the calculation result of the Jousselme evidence distance calculation formula. The evidence distance sequence is a set of all evidence distances.

[0163] In detail, in this embodiment of the invention, the Jousselme evidence distance calculation formula is expressed as:

[0164]

[0165] In the formula, Indicating evidence With evidence Evidence distance between them Jousselme, Indicating evidence The basic probability distribution, Indicating evidence The basic probability distribution, Indicates transpose. This represents the similarity matrix.

[0166] The evidence body is a mathematical object in Dempster-Shafer used to represent uncertain information. In this invention, it is a collection of data sample objects and uncertain information (such as source credibility).

[0167] The credibility index is used to quantify the reliability or credibility of each piece of evidence under specific circumstances.

[0168] The total number of fused evidence bodies is the sum of the number of evidence bodies from test information and the number of evidence bodies from expert information.

[0169] The conflict matrix is ​​a two-dimensional arrangement representation of the evidence distance sequence. For example, if the total number of fused evidence bodies is W, then the conflict matrix is ​​a W×W evidence distance numerical distribution matrix.

[0170] The exponential decay function is used to map the evidence distance to confidence level, and is expressed as:

[0171]

[0172] In the formula, Indicates the credibility of the i-th piece of evidence. This represents the evidentiary distance between evidence pieces i and j. This represents the sensitivity factor, with a value of 3.

[0173] The credibility sequence is a one-dimensional mapping result of the conflict matrix, and the credibility index sequence is a normalized result of the credibility sequence.

[0174] The basic probability assignment is used to define the distribution of the evidence body.

[0175] In detail, in the embodiment of the present application, the credibility index based on each evidence body in the test information and the expert information is constructed according to the evidence distance sequence, and a credibility index sequence is obtained, which comprises:

[0176] The number of each evidence body in the test information and the expert information is obtained, and a total number of fusion evidence bodies is obtained;

[0177] A conflict matrix is constructed according to the total number of fusion evidence bodies and the evidence distance sequence;

[0178] The conflict matrix is mapped by using a pre-constructed exponential decay function, and a credibility sequence is obtained;

[0179] Each credibility in the credibility sequence is normalized, and a credibility index sequence is obtained.

[0180] Specifically, in the embodiment of the present application, the basic data information of the test information and the expert information is identified to obtain the total number of fusion evidence bodies. Then, a W×W matrix space is constructed according to the total number of fusion evidence bodies W, and each evidence distance is filled according to the corresponding position to obtain the conflict matrix. Then, the present application uses a pre-constructed exponential decay function to perform a mapping operation on the conflict matrix to obtain the credibility sequence. The credibility sequence is then standardized by a normalization operation to obtain the credibility index sequence.

[0181] Further, in the embodiment of the present application, each evidence body in the test information and the expert information is configured with a weight by using the credibility index sequence, and a weighted multi-source data set is obtained. The weighted multi-source data set can preliminarily configure the credibility of multi-source information of uncertain information.

[0182] S5, using a pre-constructed evidence fusion rule to fuse the weighted multi-source data set to obtain multi-source uncertainty fusion data.

[0183] The evidence fusion rule is used to comprehensively integrate information from multiple evidence sources to form a more comprehensive and reliable result. The fusion operation refers to the execution process of the evidence fusion rule. The multi-source uncertainty fusion data is the fusion result.

[0184] Specifically, in the embodiment of the present application, the traditional evidence fusion rule is improved before the weighted multi-source data set is fused.

[0185] In detail, before the multi-source uncertainty fusion data is obtained by fusing the weighted multi-source data set using the pre-constructed evidence fusion rule, the method further comprises:

[0186] obtaining a traditional conflict factor, and constructing a regulation coefficient using the traditional conflict factor, wherein the regulation coefficient is expressed as:

[0187]

[0188] wherein, the regulation coefficient, the hyperbolic function, the traditional conflict factor;

[0189] constructing a conflict compensation term according to the regulation coefficient, wherein the conflict compensation term is expressed as:

[0190]

[0191] wherein, the conflict compensation term of the subset, the traditional conflict compensation term, wherein the subset is a subset in the set the set of identification frameworks; constructing an evidence fusion rule according to the conflict compensation term and the regulation coefficient, wherein the evidence fusion rule is expressed as:

[0192]

[0193] wherein,

[0194] the fused data, the contribution degree normalized value of the i-th data sample in the weighted multi-source data set to the subset the corresponding confidence index. wherein, the traditional conflict factor is:

[0195] However, under the traditional Dempster rule, the conflict factor K>0.5 is invalid, so the present application adopts adaptive weighted fusion.

[0196]

[0197] However, under the traditional Dempster rule, the conflict factor K>0.5 is invalid, so the present application adopts adaptive weighted fusion.

[0198] wherein, the regulation coefficient is the result of adaptive adjustment of the traditional conflict factor.​​​​​

[0199] The conflict compensation term is a compensation mechanism for adjusting conflict information in the evidence theory.

[0200] Specifically, in the embodiment of the present application, the adjustment coefficient is used to replace the traditional conflict factor in the traditional evidence fusion rule, and a conflict compensation term constructed by the adjustment coefficient is added to the traditional evidence fusion rule, so that an improved evidence fusion rule is obtained.

[0201] After obtaining the evidence fusion rule, the present application fuses the weighted multi-source data set to obtain multi-source uncertainty fusion data, so as to realize the fusion of multi-source uncertainty information of different spaces and different forms, and realize the comprehensive measurement of multi-source uncertainty information in complex equipment.

[0202] Specifically, referring to FIG. 4, the process of S4-S5 is used to complete the processing of the multi-source uncertainty fusion data. Figure 2 Thus, the phase of data structure processing of the present application is completed, and the construction process of the multi-source information unified measurement model is described as follows, so that the multi-source information unified measurement model can process the single-variable evidence structure and the multi-source uncertainty fusion data generated by the present application.

[0203] S6, according to the pre-constructed unified measurement model construction strategy, using the single-variable evidence structure and the multi-source uncertainty fusion data, constructing a multi-source information unified measurement model.

[0204] The unified measurement model construction strategy refers to a process guide constructed based on the parallel polyhedron evidence theory model for realizing the unified measurement of high-dimensional and related uncertainty information in complex equipment, which is used to realize the unified quantification of high-dimensional and related uncertainty information by using the consistency of the recognition framework and the joint focal element for spatial transformation.

[0205] The multi-source information unified measurement model is the construction result of the unified measurement model construction strategy, and is used to realize the unified measurement of multi-source and uncertainty information.

[0206] In detail, according to the pre-constructed unified measurement model construction strategy, using the single-variable evidence structure and the multi-source uncertainty fusion data, constructing a multi-source information unified measurement model, including:

[0207] According to the pre-constructed unified measurement model construction strategy, performing feature extraction on the single-variable evidence structure and the multi-source uncertainty fusion data to obtain a variable feature set, and performing correlation analysis on the variable feature set to obtain a correlation coefficient matrix;

[0208] transforming the correlation coefficient matrix into a parallel polyhedral structure framework, and constructing a shape matrix according to the parallel polyhedral structure framework;

[0209] performing independent space mapping on the variable feature set by using the shape matrix to obtain independent space variable data, and assigning a basic credibility to each variable data in the independent space variable data to obtain credibility assignment data;

[0210] generating a joint focus element combination by using a pre-constructed Cartesian product algorithm according to the credibility assignment data;

[0211] constructing a multi-source information system quantitative model according to the joint focus element combination and the multi-source uncertainty fusion data.

[0212] The feature extraction is a common structure in a neural network, including a convolution layer, a pooling layer, etc. The variable feature set is a feature extraction result.

[0213] The correlation analysis refers to a process of calculating a nonlinear correlation coefficient by using a Copula function. The correlation coefficient matrix is a two-dimensional representation of a correlation analysis result.

[0214] The parallel polyhedral structure framework is used to model a multi-dimensional recognition framework into a hypergeometric body.

[0215] The shape matrix is configured as:

[0216]

[0217] In the formula, denotes a shape matrix, denotes a correlation analysis result between evidence variable 2 and evidence variable 1, denotes a scaling factor of evidence variable 1 in the shape matrix. The scaling factor is used to adjust the proportion of each dimension, so that the shape matrix can correctly reflect the scaling relationship of the evidence variable in different dimensions.

[0218] The independent space mapping operation is represented as:

[0219]

[0220] In the formula, denotes a vector of independent space mapping, denotes a correlation coefficient matrix, denotes an average value of all elements in the correlation coefficient matrix.

[0221] The independent space variable data is a result of independent space mapping of the shape matrix.

[0222] The operation of assigning the basic credibility refers to corresponding assignment of the credibility index sequence to each data in the independent spatial variable data.

[0223] The Cartesian product algorithm refers to all possible ordered pairs composed of one element from multiple sets. The combined focus group refers to a focus that meets a specific condition among numerous focuses obtained by the Cartesian product algorithm.

[0224] Specifically, in the embodiments of the present application, the correlation coefficient between evidence variables is obtained through correlation analysis to obtain a correlation coefficient matrix. Then, a parallel polyhedral form of the evidence recognition framework is established according to the correlation coefficient matrix. Then, information is converted into an independent standard space through a shape matrix to realize basic credibility assignment of each evidence variable.

[0225] Specifically, the Cartesian product algorithm is used to calculate all possible focuses of the credibility assignment data. Then, according to the preconfigured threshold 0.7, the correlation coefficient greater than 0.7 between each evidence variable in the focus is determined as a strong correlation variable (evidence variable 2 and evidence variable 1 are summarized as a strong correlation variable), and the strong correlation variable is bound as a combined focus. (For example ) the variable relationship with the correlation coefficient less than 0.7 is classified as an independent focus. Thus, the combined focus group containing the combined focus and the independent focus is obtained.

[0226] When the above evidence recognition framework and the combined focus group are constructed, the evidence recognition framework and the combined focus group can be used as the backbone of the multi-source information unified measurement model.

[0227] The multi-source information unified measurement model realizes unified quantification of high-dimensional and correlated uncertain information through spatial conversion according to the consistency of the evidence recognition framework and the combined focus group.

[0228] In order to solve the problems in the background art, the present application firstly takes the overall variance of the data sample set as the bandwidth parameter, uses the Gaussian kernel function to model the uncertainty of the data sample set, so as to obtain the contribution degree of each data sample to the focus element, and then constructs a univariate evidence structure according to the contribution degree set, wherein the univariate evidence structure can quantify the cognitive uncertainty caused by incomplete information; then, the present application calculates the evidence distance of the test information and the expert information, and then establishes a credibility index representing the conflict between the multi-source evidence information according to the evidence distance, so as to obtain a credibility index sequence, and further realizes the weight distribution of the multi-source information; then, the weighted multi-source evidence information is fused by using the evidence fusion rule, so as to obtain multi-source uncertainty fusion data, wherein the multi-source uncertainty fusion data are used to realize the comprehensive measurement of multi-source uncertain information in complex equipment; finally, in order to realize the unified measurement of high-dimensional and related uncertain information in complex equipment, the present application develops an evidence theory model based on parallel polyhedron, such as: the correlation coefficient of the evidence variable can be obtained by correlation analysis, the evidence recognition framework in the form of parallel polyhedron is established to measure the correlation, then the sample information is converted into an independent standard space by using the shape matrix, the basic credibility of each evidence variable is distributed, and the joint focus element is constructed, and based on this, the consistency between the recognition framework and the joint focus element is used for space conversion, so as to realize the unified quantification of high-dimensional and related uncertain information. Therefore, the present application can realize the unified measurement of incomplete, multi-source and related uncertain parameters in a complex system.

[0229] As Figure 3 shown, it is a functional module diagram of the multi-source information unified measurement system for complex equipment provided by an embodiment of the present application.

[0230] The multi-source information unified measurement system for complex equipment 100 can be installed in an electronic device. According to the realized functions, the multi-source information unified measurement system for complex equipment 100 can include a sample kernel density estimation module 101, a multi-source uncertain information fusion module 102 and a unified measurement model construction 103. The modules of the present application can also be called units, which refer to a series of computer program segments that can be executed by an electronic device processor and can complete fixed functions, and are stored in the memory of the electronic device.

[0231] The sample kernel density estimation module 101 is configured to obtain a data sample set based on complex equipment, calculate sample overall variance of the data sample set, configure a preset bandwidth parameter by using the sample overall variance, obtain uncertainty modeling based on the data sample set according to a pre-constructed Gaussian kernel function and the bandwidth parameter, and obtain a focus element set, calculate contribution degrees of each data sample in the data sample set to each focus element in the focus element set by using the uncertainty modeling, obtain a contribution degree set, and perform evidence modeling of an uncertainty parameter on the contribution degree set to obtain a univariate evidence structure.

[0232] The multi-source uncertainty information fusion module 102 is configured to obtain test information and expert information, calculate evidence distances between each evidence body in the test information and the expert information according to a pre-constructed Jousselme evidence distance calculation formula, obtain an evidence distance sequence, construct a reliability index based on each evidence body in the test information and the expert information according to the evidence distance sequence, obtain a reliability index sequence, configure weights for each evidence body in the test information and the expert information according to the reliability index sequence, obtain a weighted multi-source data set, and fuse the weighted multi-source data set by using a pre-constructed evidence fusion rule to obtain multi-source uncertainty fusion data.

[0233] The unified measurement model construction 103 is configured to construct a multi-source information unified measurement model by using the univariate evidence structure and the multi-source uncertainty fusion data according to a pre-constructed unified measurement model construction strategy.

[0234] In detail, the modules in the multi-source information unified measurement system 100 for complex equipment in the embodiments of the present application adopt the same technical means as the multi-source information unified measurement method for complex equipment in the embodiments of the present application described above when in use, and can produce the same technical effects, which will not be described herein again. Figure 1

[0235] As shown in Figure 4 FIG. 1, it is a structural schematic diagram of an electronic device for implementing the multi-source information unified measurement method for complex equipment according to an embodiment of the present application.

[0236] The electronic device 1 can include a processor 10, a memory 11 and a bus 12, and can further include a computer program stored in the memory 11 and executable on the processor 10, such as a multi-source information unified measurement method program for complex equipment.

[0237] ​The memory 11 includes at least one type of readable storage medium, such as a flash memory, a mobile hard disk, a multimedia card, a card-type memory (e.g., an SD or DX memory, etc.), a magnetic memory, a disk, an optical disk, etc. In some embodiments, the memory 11 can be an internal storage unit of the electronic device 1, such as a mobile hard disk of the electronic device 1. In other embodiments, the memory 11 can also be an external storage device of the electronic device 1, such as a plug-in mobile hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the electronic device 1. Further, the memory 11 includes both an internal storage unit and an external storage device of the electronic device 1. The memory 11 can be used to store application software and various data installed on the electronic device 1, such as a code of a multi-source information unified measurement method program for complex equipment, etc., and can also be used to temporarily store data that has been output or will be output.

[0238] The processor 10 can be composed of an integrated circuit in some embodiments, such as a single packaged integrated circuit, or a plurality of packaged integrated circuits with the same or different functions, including one or more combinations of a central processing unit (CPU), a microprocessor, a digital processing chip, a graphics processor, and various control chips, etc. The processor 10 is a control unit of the electronic device, which connects various components of the entire electronic device through various interfaces and lines, executes programs or modules stored in the memory 11 (such as a multi-source information unified measurement method program for complex equipment, etc.), and calls data stored in the memory 11 to perform various functions and process data of the electronic device 1.

[0239] The bus 12 can be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus 12 can be divided into an address bus, a data bus, a control bus, etc. The bus 12 is configured to realize the connection and communication between the memory 11 and at least one processor 10, etc.

[0240] Figure 4 Only the electronic device with components is shown, and those skilled in the art can understand that, Figure 4The illustrated structure does not constitute a limitation on the electronic device 1, and can include fewer or more components than illustrated, or combine certain components, or arrange different components.

[0241] For example, although not shown, the electronic device 1 can also include a power supply (such as a battery) to power the various components. Preferably, the power supply can be logically connected to the at least one processor 10 through a power management device, so that the power management device can implement functions such as charge management, discharge management, and power consumption management. The power supply can also include one or more direct current or alternating current power supplies, recharging devices, power supply fault detection circuits, power supply converters or inverters, power supply status indicators, and any other components. The electronic device 1 can also include various sensors, Bluetooth modules, Wi-Fi modules, and the like, which are not described here.

[0242] Further, the electronic device 1 can also include a network interface, which can optionally include a wired interface and / or a wireless interface (such as a WI-FI interface, a Bluetooth interface, etc.), and is typically used to establish a communication connection between the electronic device 1 and other electronic devices.

[0243] Optionally, the electronic device 1 can also include a user interface, which can be a display (Display), an input unit (such as a keyboard (Keyboard)), and optionally a standard wired interface, a wireless interface. Optionally, in some embodiments, the display can be an LED display, a liquid crystal display, a touch liquid crystal display, an OLED (Organic Light-Emitting Diode) touch, etc. The display can also be appropriately referred to as a display screen or a display unit, and is used to display information processed in the electronic device 1 and to display a visualized user interface.

[0244] The multi-source information unified measurement method program for complex equipment stored in the memory 11 in the electronic device 1 is a combination of multiple instructions, which, when executed in the processor 10, can achieve:

[0245] Obtain a data sample set based on complex equipment, calculate the overall sample variance of the data sample set, and configure a preset bandwidth parameter using the overall sample variance;

[0246] According to a pre-constructed Gaussian kernel function and the bandwidth parameter, an uncertainty modeling based on the data sample set is obtained;

[0247] Obtaining a focus element set, calculating the contribution degree of each data sample in the data sample set to each focus element in the focus element set by using the uncertainty modeling, obtaining a contribution degree set, and performing evidence modeling of the uncertainty parameters on the contribution degree set to obtain a univariate evidence structure;

[0248] Obtaining test information and expert information, calculating the evidence distance between each evidence body in the test information and the expert information according to a pre-constructed Jousselme evidence distance calculation formula, obtaining an evidence distance sequence, constructing a credibility index based on each evidence body in the test information and the expert information according to the evidence distance sequence, obtaining a credibility index sequence, and configuring weights for each evidence body in the test information and the expert information according to the credibility index sequence, obtaining a weighted multi-source data set;

[0249] Fusing the weighted multi-source data set by using a pre-constructed evidence fusion rule to obtain multi-source uncertainty fusion data;

[0250] Constructing a multi-source information unified dimension model by using the univariate evidence structure and the multi-source uncertainty fusion data according to a pre-constructed unified dimension model construction strategy.

[0251] Specifically, the specific implementation method of the processor 10 to the above instructions can refer to Figures 1 to 4 The description of related steps in the corresponding embodiments will not be repeated here.

[0252] Further, the modules / units integrated in the electronic device 1 can be stored in a computer readable storage medium if they are realized in the form of software function units and sold or used as independent products. The computer readable storage medium can be volatile or non-volatile. For example, the computer readable medium can include any entity or device capable of carrying the computer program code, a recording medium, a U disk, a mobile hard disk, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM, Read-Only Memory).

[0253] The application also provides a computer readable storage medium, which stores a computer program, and the computer program can realize the following when executed by a processor of an electronic device:

[0254] Obtaining a data sample set based on complex equipment, calculating the sample overall variance of the data sample set, and configuring a preset bandwidth parameter by using the sample overall variance;

[0255] Obtaining an uncertainty modeling based on the data sample set according to a pre-constructed Gaussian kernel function and the bandwidth parameter;

[0256] Obtain a focus element set, use the uncertainty modeling to calculate the contribution degree of each data sample in the data sample set to each focus element in the focus element set, obtain a contribution degree set, and perform evidence modeling on the uncertainty parameters of the contribution degree set to obtain a univariate evidence structure;

[0257] Obtain test information and expert information, calculate the evidence distance between each evidence body in the test information and the expert information according to a pre-constructed Jousselme evidence distance calculation formula, obtain an evidence distance sequence, construct a credibility index based on each evidence body in the test information and the expert information according to the evidence distance sequence, obtain a credibility index sequence, and perform weight configuration on each evidence body in the test information and the expert information according to the credibility index sequence, and obtain a weighted multi-source data set;

[0258] Fuse the weighted multi-source data set by using a pre-constructed evidence fusion rule to obtain multi-source uncertainty fusion data;

[0259] Construct a multi-source information unified dimension model by using the univariate evidence structure and the multi-source uncertainty fusion data according to a pre-constructed unified dimension model construction strategy.

[0260] In several embodiments provided in the present application, it should be understood that the disclosed devices, systems and methods can be implemented in other ways. For example, the above-described system embodiments are only illustrative, and actual implementation can have another division manner.

[0261] The modules described as separate components can or can not be physically separated, and the components shown as modules can or can not be physical units, i.e., can be located in one place or distributed on multiple network units. Part or all of the modules can be selected to achieve the purpose of the embodiment according to actual needs.

[0262] In addition, each functional module in each embodiment of the present application can be integrated in one processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of hardware plus software functional modules.

[0263] It is obvious for those skilled in the art that the present application is not limited to the details of the above exemplary embodiments, and the present application can be implemented in other specific forms without departing from the spirit or essential characteristics of the present application.

[0264] Finally, it should be noted that the above examples are merely intended to illustrate the technical solutions of the present application and not to limit the present application. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or equivalently replaced without departing from the spirit and scope of the present application.

Claims

1. A multi-source information integration measurement method for complex equipment, characterized in that, The method comprises: acquiring a data sample set based on complex equipment, calculating sample overall variance of the data sample set, and configuring a preset bandwidth parameter by using the sample overall variance, wherein the complex equipment refers to equipment in the fields of aerospace, navigation and high-speed rail, and the data sample set comprises vibration sensor data, temperature sensor data and maintenance log files; obtaining uncertainty modeling based on the data sample set according to a pre-constructed Gaussian kernel function and the bandwidth parameter; acquiring a focus element set, calculating contribution degrees of each data sample in the data sample set to each focus element in the focus element set by using the uncertainty modeling, obtaining a contribution degree set, and performing evidence modeling of an uncertainty parameter on the contribution degree set to obtain a univariate evidence structure; acquiring test information and expert information, calculating evidence distances between each evidence body in the test information and the expert information according to a pre-constructed Jousselme evidence distance calculation formula, obtaining an evidence distance sequence, constructing a credibility index based on each evidence body in the test information and the expert information according to the evidence distance sequence, obtaining a credibility index sequence, and configuring weights of each evidence body in the test information and the expert information according to the credibility index sequence to obtain a weighted multi-source data set; acquiring a traditional conflict factor, and constructing an adjustment coefficient by using the traditional conflict factor, wherein the adjustment coefficient is expressed as: ； wherein denotes a regulation coefficient, denotes a hyperbolic function, denotes a traditional conflict factor; constructing a conflict compensation term according to the adjustment coefficient, wherein the conflict compensation term is expressed as: ； In the formula, Representing a subset Conflict compensation items, Represents traditional conflict compensation items, where the subset For set subset of Represents a set of recognition frames; constructing an evidence fusion rule according to the conflict compensation term and the adjustment coefficient, wherein the evidence fusion rule is expressed as: ； In the formula, Representing a subset The merged data Indicates the first element in a weighted multi-source dataset. A subset of data samples Normalized value of contribution express Corresponding credibility metrics; fusing the weighted multi-source data set by using the pre-constructed evidence fusion rule to obtain multi-source uncertainty fusion data; constructing a multi-source information unified dimension model by using the univariate evidence structure and the multi-source uncertainty fusion data according to a pre-constructed unified dimension model construction strategy.

2. The multi-source information integration measurement method for complex equipment according to claim 1, characterized in that, The calculation of the sample overall variance of the data sample set comprises: acquiring a total number of samples in the data sample set, identifying a number of data sources in the data sample set to obtain a data source number, and identifying a number of samples contained in each data source to obtain a data source sample number sequence; calculating variances and means corresponding to each data source in the data sample set to obtain a data source variance sequence and a data source mean sequence; calculating a global mean according to the data source mean sequence, the data source sample number sequence and the total number of samples, wherein the global mean is expressed as: ； In the formula, represents the global mean, represents the number of samples in the i-th data source in the data source sample number sequence, represents the number of samples in the i-th data source in the data source sample number sequence, represents the mean corresponding to the i-th data source in the data source mean sequence, represents the mean corresponding to the i-th data source in the data source mean sequence, represents the total number of samples; calculating a sample overall variance according to the total number of samples, the number of data sources, the data source sample number sequence, the data source variance sequence, the data source mean sequence and the global mean, wherein the sample overall variance is expressed as: ； In the formula, denotes the overall variance of the sample, denotes the number of data sources, denotes the variance corresponding to the th data source in the variance sequence of the data sources, denotes the mean value of the th data source, denotes the square of the global mean value, denotes the size of the number.

3. The multi-source information unified measurement method for complex equipment according to claim 2, characterized in that, The uncertainty modeling based on the data sample set according to the pre-constructed Gaussian kernel function and the bandwidth parameter comprises: arbitrarily selecting two data samples from the data sample set as a first data sample and a second data sample; calculating a sample distance between the first data sample and the second data sample; According to the sample distance and the bandwidth parameter, a pre-constructed Gaussian kernel function is valued to obtain uncertainty modeling, wherein the uncertainty modeling is expressed as: ； wherein denotes uncertainty modeling between two data samples, denotes the bandwidth parameter, same as the sample overall variance, denotes and sample distance between two data samples, square of the Euclidean distance between two data samples, and denotes the size of a number.

4. The multi-source information integration measurement method for complex equipment according to claim 3, characterized in that, According to the sample distance and the bandwidth parameter, a pre-constructed Gaussian kernel function is valued to obtain uncertainty modeling, wherein the uncertainty modeling is expressed as: The contribution degree set is obtained by calculating the contribution degree of each data sample in the data sample set to each focus element in the focus element set by using the uncertainty modeling, and the contribution degree set comprises: The total number of focus elements is obtained by acquiring the number of focus elements in the focus element set, and the local variance of each data sample is obtained; ； wherein denotes the sample confidence weight of the i-th data sample in the sequence of sample confidence weights, denotes the local variance of the i-th data sample, denotes the square of the bandwidth parameter;​​ According to the local variance of each data sample and the bandwidth parameter, a sample confidence weight sequence is obtained, wherein one sample confidence weight in the sample confidence weight sequence is expressed as: ； In the formula, This refers to any data sample, such as the first data sample. For the first Each jiao yuan Contribution This indicates the modeling of the uncertainty. Indicates the first Typical eigenvectors of each focal element This indicates the total number of the focal elements. Indicates the first The feature vectors of each focal element Indicates the magnitude of a numerical value; According to the total number of focus elements and the sample confidence weight sequence, a contribution degree function is constructed, wherein the contribution degree function is expressed as:

5. The multi-source information integration measurement method for complex equipment according to claim 4, characterized in that, According to the contribution degree function, the contribution degree set is obtained. The contribution degree set is subjected to evidence modeling of the uncertainty parameter to obtain a univariate evidence structure, comprising: A variable set corresponding to the focus element set is obtained, and a pre-set composite focus element set in the focus element set is obtained, wherein a pre-set composite focus element in the pre-set composite focus element set is a subset of the focus element set; ； wherein represents the contribution of the i-th data sample to the j-th variable in the k-th focus element normalized value, represents the contribution of the i-th data sample to the j-th variable in the k-th focus element normalized value, represents the contribution of the i-th data sample to the j-th variable in the k-th focus element normalized value, represents the contribution of the i-th data sample to the j-th variable in the k-th focus element normalized value, represents the contribution function, represents the k-th focus element, represents the k-th focus element, represents the focus element set containing k focus elements, represents the k-th focus element, represents the variable measurement confidence of the j-th variable, , , , and all represent the magnitude of a number; The contribution degree set is subjected to contribution degree normalization operation based on a target variable in the variable set to obtain a contribution degree normalization sequence, wherein one contribution degree normalization in the contribution degree normalization sequence is expressed as: According to a pre-constructed aggregation rule, the contribution degree normalization sequence is subjected to aggregation operation based on each pre-set composite focus element in the pre-set composite focus element set to obtain an evidence body set; An effective evidence body is obtained by acquiring an evidence body with a value less than a pre-set focus threshold in the evidence body set, and an effective focus element set is obtained by acquiring a focus element corresponding to the effective evidence body from the focus element set, and the effective focus element set is merged to obtain a simple focus point set; 6. The multi-source information integration measurement method for complex equipment according to claim 5, characterized in that, The effective evidence body, the simple focus point set and the variable set are summarized to obtain a univariate evidence structure. ； wherein denotes the evidence distance between the evidence bodies denotes the evidence distance between the evidence bodies denotes the evidence distance between the evidence bodies denotes Jousselme, denotes the basic probability assignment of the evidence body denotes the basic probability assignment of the evidence body denotes the basic probability assignment of the evidence body denotes the basic probability assignment of the evidence body denotes the transpose, denotes the similarity matrix.

7. The multi-source information integration measurement method for complex equipment according to claim 6, characterized in that, The Jousselme evidence distance calculation formula is expressed as: According to the evidence distance sequence, a credibility index sequence is obtained by constructing a credibility index based on each evidence body in the test information and the expert information, comprising: The total number of fusion evidence bodies is obtained by acquiring the number of each evidence body in the test information and the expert information; A conflict matrix is constructed according to the total number of fusion evidence bodies and the evidence distance sequence; A pre-constructed exponential decay function is used to map the conflict matrix to obtain a credibility sequence; 8. The multi-source information integration measurement method for complex equipment according to claim 7, characterized in that, Each credibility in the credibility sequence is normalized to obtain a credibility index sequence. According to a pre-constructed unified measurement model construction strategy, a multi-source information unified measurement model is constructed by using the univariate evidence structure and the multi-source uncertainty fusion data, comprising: According to a pre-constructed unified measurement model construction strategy, a variable feature set is obtained by performing feature extraction on the univariate evidence structure and the multi-source uncertainty fusion data, and a correlation coefficient matrix is obtained by performing correlation analysis on the variable feature set; The correlation coefficient matrix is converted into a parallel polyhedral structure framework, and a shape matrix is constructed according to the parallel polyhedral structure framework; Independent space variable data are obtained by performing independent space mapping on the variable feature set by using the shape matrix, and basic credibility is assigned to each variable data in the independent space variable data to obtain credibility assignment data; According to the credibility assignment data, a joint focus element combination is generated by using a pre-constructed Cartesian product algorithm; A multi-source information unified measurement model is constructed according to the joint focus element combination and the multi-source uncertainty fusion data.

9. A multi-source information integration measurement system for complex equipment for performing the multi-source information integration measurement method for complex equipment according to claim 1, characterized in that, The system comprises: A sample kernel density estimation module is configured to obtain a data sample set based on complex equipment, calculate sample overall variance of the data sample set, configure a preset bandwidth parameter by using the sample overall variance, and obtain uncertainty modeling based on the data sample set according to a pre-constructed Gaussian kernel function and the bandwidth parameter, and obtain a focus element set, calculate contribution degrees of each data sample in the data sample set to each focus element in the focus element set by using the uncertainty modeling, obtain a contribution degree set, and perform evidence modeling of an uncertainty parameter on the contribution degree set to obtain a single-variable evidence structure; A multi-source uncertainty information fusion module is configured to obtain test information and expert information, calculate evidence distances between each evidence body in the test information and the expert information according to a pre-constructed Jousselme evidence distance calculation formula, obtain an evidence distance sequence, construct a credibility index based on each evidence body in the test information and the expert information according to the evidence distance sequence, obtain a credibility index sequence, configure weights for each evidence body in the test information and the expert information according to the credibility index sequence, obtain a weighted multi-source data set, and fuse the weighted multi-source data set by using a pre-constructed evidence fusion rule to obtain multi-source uncertainty fusion data; A unified measurement model construction is configured to construct a multi-source information unified measurement model by using the single-variable evidence structure and the multi-source uncertainty fusion data according to a pre-constructed unified measurement model construction strategy.

Images