CN112163189A - Equipment fault diagnosis method based on test sequence probability matrix - Google Patents
Equipment fault diagnosis method based on test sequence probability matrix Download PDFInfo
- Publication number
- CN112163189A CN112163189A CN202011208271.8A CN202011208271A CN112163189A CN 112163189 A CN112163189 A CN 112163189A CN 202011208271 A CN202011208271 A CN 202011208271A CN 112163189 A CN112163189 A CN 112163189A
- Authority
- CN
- China
- Prior art keywords
- fault
- test
- matrix
- correlation matrix
- double
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000012360 testing method Methods 0.000 title claims abstract description 143
- 239000011159 matrix material Substances 0.000 title claims abstract description 132
- 238000000034 method Methods 0.000 title claims abstract description 43
- 238000003745 diagnosis Methods 0.000 title claims abstract description 28
- 239000013598 vector Substances 0.000 claims abstract description 56
- 230000000694 effects Effects 0.000 claims abstract description 10
- 238000012163 sequencing technique Methods 0.000 claims abstract description 4
- 230000002146 bilateral effect Effects 0.000 claims description 23
- 230000005284 excitation Effects 0.000 claims description 6
- 238000010276 construction Methods 0.000 claims 2
- 230000000875 corresponding effect Effects 0.000 description 13
- 238000011156 evaluation Methods 0.000 description 10
- 230000002596 correlated effect Effects 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 230000006870 function Effects 0.000 description 4
- 230000002068 genetic effect Effects 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 230000003321 amplification Effects 0.000 description 2
- 238000013459 approach Methods 0.000 description 2
- 230000001364 causal effect Effects 0.000 description 2
- 230000008878 coupling Effects 0.000 description 2
- 238000010168 coupling process Methods 0.000 description 2
- 238000005859 coupling reaction Methods 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- BTCSSZJGUNDROE-UHFFFAOYSA-N gamma-aminobutyric acid Chemical compound NCCCC(O)=O BTCSSZJGUNDROE-UHFFFAOYSA-N 0.000 description 2
- 238000002955 isolation Methods 0.000 description 2
- 238000003199 nucleic acid amplification method Methods 0.000 description 2
- 238000013024 troubleshooting Methods 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 239000000470 constituent Substances 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000002405 diagnostic procedure Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000004451 qualitative analysis Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/15—Correlation function computation including computation of convolution operations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Analysis (AREA)
- Theoretical Computer Science (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Optimization (AREA)
- Computational Mathematics (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Algebra (AREA)
- Computing Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Operations Research (AREA)
- Probability & Statistics with Applications (AREA)
- Test And Diagnosis Of Digital Computers (AREA)
Abstract
The invention provides a device fault diagnosis method based on a test sequence probability matrix, which comprises the following steps: constructing a double-edge correlation matrix representing cause and effect relationships determined among equipment fault reasons, fault modes and test project nodes; acquiring equipment test data, sequencing the test data according to the row vector sequence of the double-edge correlation matrix test items and converting the test data into a vector, wherein the value of an element in the vector depends on whether the test data is in accordance with the specification requirement compared with the range of the test items and the preset value of the parameter index to be tested; and reporting faults of the test items with the element values in the vectors which do not meet the specification requirements, matching the row vectors of the test items in the double-edge correlation matrix to obtain a fault mode set related to the fault reporting test items, and obtaining a fault reason set related to the fault mode set through the double-edge correlation matrix. The invention does not need a large amount of data samples, and can solve the problem of fault diagnosis under the condition of incomplete data in a complex system.
Description
Technical Field
The invention relates to the technical field of fault diagnosis, in particular to a device fault diagnosis method based on a test sequence probability matrix.
Background
As is known, methods for fault diagnosis of equipment can be divided into three main categories: mathematical model-based methods, signal processing-based methods, and knowledge-based methods. In a knowledge-based approach, for example, a failure mode-test correlation matrix may be generated using a multi-signal flow graph model, and then fault isolation may be performed by matching test data to the correlation matrix. Or a method for carrying out qualitative analysis on the fault tree by using a binary decision diagram organically integrates a fault tree principle and an expert system so as to realize fault diagnosis. By using the method, more accurate fault isolation can be realized, but complete test data is required, otherwise, the diagnostic result has higher ambiguity, namely, the method has weaker processing capability on uncertain information.
In order to solve the problem of fault diagnosis under the condition of incomplete test data, an intelligent fault diagnosis method combining a directed graph, a fuzzy theory and a genetic algorithm is proposed, a system model is constructed by using the directed graph, the problem of uncertainty in the model is solved by using a fuzzy set, and a possible fault propagation path is searched by using the genetic algorithm. In addition, a rough set and a genetic algorithm are used for reducing redundant attributes and conflict samples, and then the simplest fault attribute is extracted to be used as a training sample of the SVM and used as a classifier for rapidly isolating faults. The methods well solve the ambiguity problem caused by information uncertainty, but a large amount of sample data is needed to train the model.
Disclosure of Invention
The invention aims to provide a test sequence-based equipment fault diagnosis method to solve the problem of uncertain fault diagnosis under the condition of incomplete test data in a complex equipment system.
In order to achieve the purpose, the invention adopts the following technical scheme:
according to the 1 st aspect of the present invention, there is provided an equipment fault diagnosis model building method, the fault diagnosis model being capable of representing cause and effect relationships determined among fault causes, fault patterns, and test item nodes, by which diagnosis of equipment faults can be achieved, the method comprising:
constructing a first correlation matrix C between fault reasons and fault modes, wherein the matrix C is a binary matrix, row vectors of the matrix C represent the influence conditions of all the fault reasons on the fault modes, and column vectors represent the influence conditions of all the fault modes corresponding to the fault reasons;
constructing a second correlation matrix D between the fault mode and the test items, wherein the matrix D is a binary matrix, row vectors of the matrix D represent the response of all tests to the fault mode excitation signals, and column vectors of the matrix D represent the response of all fault mode excitation signals corresponding to the tests;
and combining the first correlation matrix C and the second correlation matrix D to generate a double-edge correlation matrix B, wherein in the double-edge correlation matrix B, a fault mode-test item correlation matrix element represents the correlation between a fault mode and a test item, and a fault reason-fault mode correlation matrix element represents the correlation between the fault mode and the fault reason.
In some embodiments, the method further comprises the step of constructing a bilateral vector probability matrix capable of representing the correlation degree of the test timing sequence, the fault reason and the fault mode based on the bilateral correlation matrix and expert knowledge
Further, constructing the bilateral vector probability matrix
The method comprises the following steps: the sequence of the test items in the double-side correlation matrix is arranged according to the equipment test sequence, the fault mode sequence is sequenced according to the test items and the correlation degree of the test items to obtain a double-side vector matrix, and expert knowledge is adopted to obtain the correlation degree between each test item and the corresponding fault mode and between each fault mode and the corresponding fault reason.
Further, the most confident application of expert knowledge is performed using triangular fuzzy number analytic hierarchy process.
According to the 2 nd aspect of the present invention, there is provided an equipment failure diagnosis method including:
constructing a double-edge correlation matrix representing cause and effect relationships determined among equipment fault reasons, fault modes and test project nodes;
acquiring equipment test data, sequencing the test data according to the row vector sequence of the double-edge correlation matrix test items and converting the test data into a vector, wherein the value of an element in the vector depends on whether the test data is in accordance with the specification requirement compared with the range of the test items and the preset value of the parameter index to be tested;
and reporting faults of the test items with the element values in the vectors which do not meet the specification requirements, matching the row vectors of the test items in the double-edge correlation matrix to obtain a fault mode set related to the fault reporting test items, and obtaining a fault reason set related to the fault mode set through the double-edge correlation matrix.
Compared with the prior art, the method does not need a large number of data samples, and can solve the problem of uncertain diagnosis under the condition of incomplete data in a complex system.
Drawings
FIG. 1 is a flow chart of failure mode-test correlation matrix generation;
FIG. 2 is a diagnostic process based on a bilateral correlation matrix model;
FIG. 3 is a flow chart of generating a bilateral vector probability matrix based on a bilateral correlation matrix;
FIG. 4(a) is a diagram of membership functions of triangular fuzzy numbers, and FIG. 4(b) is a diagram of membership functions corresponding to triangular fuzzy numbers of different linguistic values;
FIG. 5 is a flow chart of an analytic hierarchy process;
FIG. 6 is a schematic diagram of a filter amplifier;
FIG. 7 is a failure cause-failure mode correlation matrix;
FIG. 8 is a failure mode-test item correlation matrix;
FIG. 9 is a fault cause-fault mode-test item double-sided correlation matrix.
Detailed Description
The technical solution of the present invention is described in more detail below by way of examples and figures.
In the following description:
"related" means two elements XiAnd YiAre related, inter-related, i.e., represent a shared or causal relationship between two elements.
The correlation matrix represents two groups of elements X ═ X1,X2,...,Xi,...,Xn}、Y＝{Y1,Y2,...,Yi,...,YnCause and effect relationships between. For example, if test point XiAnd constituent unit YiIs correlated, then YiIf a fault occurs, it means XiShould be abnormal; conversely, XiIf the test is passed, Y is provediIs normal.
The logical relationship between a certain test point and its input component unit (1 or n) and any test point (1 or n) directly input into the component unit is only indicated, and is called first-order correlation.
Example 1:
1. correlation matrix between failure causes and failure modes
Let the set of failure cause nodes c ═ c1,c2,...,ci...,crIn the failure mode node set m ═ m1,m2,...,mi...,mmFifthly, the specific definition of the correlation matrix of the fault cause and the fault mode is as follows:
in the formula (1), the row vector α ═ ci1 ci2,…,cir]Indicates all the causes of failure c1,c2,...,ci...,crFor failure mode miThe resulting impact conditions; column vector betaj＝[c1j,c2j,...,cij...,crj]TIndicates the cause of the failure ciCorresponding all failure modes m1,m2,...,mmThe resulting impact situation.
Cm*rIs a binary matrix, i.e. any cell c in the matrixijThe value of (A) is only 0 or 1, 1 represents the fault reason ciCan lead to failure mode mjWhereas 0 means that it cannot be generated.
2. Correlation matrix between failure modes and test items
According to the correlation between the fault mode and the signal and the correlation between the signal and the test, a fault-signal-test model can be obtained and used for representing the dependency relationship between the test and the fault mode. Thus, the failure mode-test correlation matrix is defined as:
in the formula (2), the row vector αi＝[di1 di2,…,din]Represents all the tests t1,t2,...,tnFor failure mode miA response of the excitation signal; column vector betaj＝[d1j d2j,…,dmj]TRepresents the test tjCorresponding all failure modes m1,m2,...,mmA response of the excitation signal.
Dm×nIs a binary matrix, i.e. any cell d in the matrixjkCan only take 0 or 1, 1 represents the failure mode mjCan be tested tkOn the contrary, 0 means that it cannot be detected.
As shown in FIG. 1, the correlation matrix D is generated bym×n：
(1) Determining a correlation matrix M of a failure mode signalm×k；
(2) Determining a correlation matrix T for a signal testk×n；
(3) According to matrix Mm×kAnd matrix Tk×nObtaining a correlation matrix D of failure mode-testsm×nI.e. the matrix Mm×kM and the matrix Tk×nThe number of columns n forming a zero matrix Dm×n(0) By analysing the fault pattern-signal correlation matrix Mm×kRow vector of
The value of each element in (a) if theta ij1, it indicates the occurrence of failure mode miWill influence the signal soThen in the signal-test correlation matrix Tk×nFind signal soThe line vector of
If it is
1 denotes test tjCapable of detecting a signal soThen test tjCapable of detecting failure mode miI.e.:
3. double-sided correlation matrix
The above-mentioned failure cause-failure mode correlation matrix Cm*nAnd failure mode-test correlation matrix Dm×nRespectively representing the cause and effect relationship between the fault reason and the fault mode and between the fault mode and the test item, and combining the two matrixes to generate a double-edge correlation matrix Bm×(n+r)Namely:
at the double-edge correlation matrix Bm×(n+r)Middle, failure mode-test item correlation matrix element di-jIndicates a failure mode miAnd test items tjCorrelation between, failure cause-failure mode correlation matrix element ei-jIndicates a failure mode miAnd cause of failure cjThe correlation between them. Therefore, the double-edge correlation matrix realizes the expression of the causal relationship among the fault reason, the fault mode and the test items.
Due to the double-edge correlation matrix Bm×(n+r)And the cause and effect relationship determined among the fault reason, the fault mode and the test item node is shown, and the fault diagnosis can be realized through the cause and effect relationship.
According to one embodiment of the invention, a fault diagnosis method based on a double-edge correlation matrix is provided.
As shown in fig. 2, based on a double-sided correlation matrix Bm×(n+r)The fault diagnosis comprises the following steps:
step 101, testing results according to a double-edge correlation matrix Bm×(n+r)The test item row vector sequence is ordered and converted into a vector
In the vector, wiThe value of (d) depends on the value and set range of the respective test data. In particular toIn other words, the test result value is compared with the range of the preset values of the test item and the index of the parameter to be tested, whether the test result and the index meet the standard requirement is judged, and if the test result and the index meet the standard requirement, the w isiSet to 0, if not meeting the standard requirement, then wiSet to 1, with other indicia, e.g., "? "means.
102, aiming at the fault reporting test items, matching the test item row vectors in the double-edge correlation matrix to obtain a fault mode set related to the fault reporting test items.
For example, if the vector
In which a certain element w appearsk1, the test item t in the test is representedkThe fault sign appears, and for the test item tkReporting faults and applying a double-edge correlation matrix Bm×(n+r)The test item row vectors in (1) are matched to obtain and report failure test item tkSet of relevant failure modes ml,mj,...,mk}。
Further, from the set of failure modes { m }l,mj,...,mkGet rid of and test pass item set t1,t2,...,tk-1Related failure modes, obtaining the failure mode set m causing the failure symptomj,...,mk}。
And 103, searching a fault reason set related to the fault mode set in the double-edge correlation matrix to realize fault reason positioning.
By a double-sided correlation matrix Bm×(n+r)Set of search and failure modes { mj,...,mkRelated set of causes of failure cj,...,ckAnd further, the positioning of the fault reason can be realized.
Example 2:
in some cases, some equipments have high coupling characteristics, and the testing process has strict operation flow and specified detection means, so that the temporary test points cannot be freely disassembled and arranged like automobiles in the test troubleshooting process, and the equipment is provided with a redundant system like an airplane, so that the fault test cannot be carried out. In addition, the equipment test may terminate the test flow due to the error reporting phenomenon of part of the key test items, so that the data of part of the test items is unknown.
For such equipment, when a double-edge correlation matrix is used for diagnosing and troubleshooting, a group of fault fuzzy groups may be obtained, so that a tester cannot select the most probable fault mode from the group of fault fuzzy groups, and therefore the fault fuzzy groups need to be decoupled.
In a complex equipment system, since most fault modes are strongly correlated with some specific tests and weakly correlated or uncorrelated with others, the presence of an element "1" in most row vectors of a bilateral correlation matrix is first order correlated with some tests, and other "1" s are higher order correlated or uncorrelated.
Based on the above, according to another embodiment of the invention, based on the bilateral correlation matrix and expert knowledge, a bilateral vector probability matrix capable of representing the test time sequence, the fault reason and the correlation degree of the fault mode is constructed
Using the bilateral vector probability matrix
And the fault mode and the fault reason can be accurately positioned.
As shown in fig. 3, the order of tests in the bilateral correlation matrix is arranged according to the equipment test order, and the order of the failure modes is ordered according to the test items and their correlation degrees (e.g. represented by probabilities), so as to obtain the bilateral vector matrix
Data on the degree (e.g., in terms of probability) of correlation between each test item and the corresponding failure mode and between each failure mode and the corresponding failure cause is also missing due to the bilateral vector matrix. Expert knowledge may be employed to obtain the degree to which each test item is associated with a corresponding failure mode and each failure mode is associated with a corresponding failure cause. The expert knowledge may be from a knowledge base based on an expert evaluation index system constructed using known methods, however such results may have incompleteness, uncertainty and ambiguity.
To this end, according to an alternative approach, the present invention employs fuzzy numbers, such as triangular fuzzy numbers, to represent the degree of correlation. As shown in fig. 4, the membership function of a triangular fuzzy number has the following form:
for example, in order to associate the judgment result of the expert knowledge on the degree of correlation with the blur number, nine linguistic variables such as "very high", "higher", "medium upper", "medium lower", "low", and "very low" may be introduced, and the correspondence relationship between each linguistic variable and the triangular blur number is shown in table 1:
| serial number | Language value | Triangular |
| 1 | Is very high | (08，09，10) |
| 2 | Height of | (07，08，09) |
| 3 | Is higher than the original | (06，07，08) |
| 4 | On the middle upper side | (05，06，07) |
| 5 | Medium and high grade | (04，05，06) |
| 6 | Moderate partial downward | (03，04，05) |
| 7 | Is on the low side | (02，03，04) |
| 8 | Is low in | (01，02，03) |
| 9 | Is very low | (0，01，02) |
According to another optional method, a judgment matrix A is constructed by using the relative importance of an Analytic Hierarchy Process (AHP) to the pairwise comparison of factors, so that the problem that when an expert evaluates the influence factors of a certain system, the influence of each factor on the system cannot be directly given due to the particularity of some factors, especially the strong coupling condition of multiple systems and multiple factors occurs, and the influence degree of each sub-factor on each subsystem cannot be given comprehensively and objectively by using expert knowledge can be solved.
Set n sub-factor sets X ═ X1，x2，...，xnGet two sub-factors X from the set XiAnd xjI is 1, …, n; j is 1, …, n; then element aijRepresents a sub-factor xiAnd xjThe ratio of the weights of the influence on the subsystem Z can be obtained by taking n × n values and constructing the matrix a ═ aij)n×nA is called the comparison matrix or decision matrix of Z-X.
Can be measured by evaluating the scale pair aijFor example, the evaluation scale is divided into five grades of "important", "more important", "less important" and "as important", and the evaluation scale 9, 7, 5, 3, 1 is used as the measure value, and the measure values between the five scales 8, 6, 4, 2 are used as the fold-median value of each important meaning represented by each scale.
After the determination matrix a is established, the weights of the elements in each layer in the matrix can be obtained by, for example, a eigenvector value method. Among the obtained feature values, the feature vector W corresponding to the largest feature value is used as each factor weight. Further, the final weight may be obtained through verification of the consistency index.
Because the analytic hierarchy process only takes the factor weight estimation value as the decision basis, the uncertainty of expert knowledge is not considered in the processing process, and the precision problem exists. The invention integrates the triangle fuzzy theory and the analytic hierarchy process to reduce the inaccuracy. Aiming at the conditions of limitation, uncertainty and the like of expert knowledge, a Triangular Fuzzy Analytic Hierarchy Process (TFAHP) is adopted, the Triangular Fuzzy analytic Hierarchy Process is used for replacing the evaluation scale of a surface analytic Hierarchy Process to factors, and the evaluation standard is shown in a table 2:
| evaluating the scale | Definition of | Language value | Triangular |
| 9 | Are of great importance | Is very high | (08，09，10) |
| 8 | Between a significant number and a significant number | Height of | (07，08，09) |
| 7 | Most importantly | Is higher than the original | (06，07，08) |
| 6 | Between more and more important | On the middle upper side | (05，06，07) |
| 5 | Of greater importance | Medium and high grade | (04，05，06) |
| 4 | Between more and less important | Moderate partial downward | (03，04，05) |
| 3 | Of little importance | Is on the low side | (02，03，04) |
| 2 | Between slightly and equally important | Is low in | (01，02，03) |
| 1 | Of equal importance | Is very low | (0，01，02) |
For example, if the factor i is slightly more important than the factor j, the corresponding element a in the decision matrix AijA value of 3, a triangular blur number b can be usedij＝(sij，3，uij) Is shown, wherein sijAnd uijThe left-right spread is represented, and the difference between the left-right spread and the left-right spread represents the uncertainty of the judgment result according to expert knowledge, i.e. the larger the difference is, the larger the uncertainty of the result is. Using b in triangular fuzzy numbersij -1Representing the relative importance of factor j and factor i, then
Therefore, all elements of the judgment matrix A can be finally expressed by triangular fuzzy numbers, and the fuzzy judgment matrix B is obtained.
And (4) applying a triangular fuzzy number analytic hierarchy process to perform most confidence application on expert knowledge. For example, the result of scoring N times the degree of correlation between a certain failure mode and a test item is:
wherein s isi，mi，uiIs the lower, middle, upper limit of the triangular blur number, θiAnd representing the reference degree of the ith grade to the correlation degree of the fault mode and the test item, and determining the grade weight by using the average as follows:
the mean score ambiguity number is:
the sequence of the fault modes is arranged according to the level of the grading probability value of the correlation degree of the fault modes and the test items and the test sequence, and then a bilateral vector probability matrix containing the correlation degree of each test item and the corresponding fault mode can be obtained
Arranged longitudinally with the degree of correlation from high to low, the closer to the test item the failure mode is, the higher the degree of correlation, e.g. assuming failure mode m12And m1All with the test item t10Associated, then failure mode m12And test item t10Is greater than the failure mode m1And test item t10Of (2), i.e. P (d)12-10)＞P(d1-10)。
Similarly, the fault reasons are arranged according to the level of the grading fuzzy value of the fault reasons and the correlation degree of the fault modes and the sequence of the fault modes, and a bilateral vector matrix containing the correlation degree of each fault mode and the corresponding fault reason can be obtained
By the above-mentioned bilateral vector matrix
The fault reason, the fault mode, the test item and the equipment test sequence can be effectively and organically combined, thereby solving the problem based on the double-edge correlation matrix Bm×(n+r)The decoupling problem of the fault fuzzy group in the diagnosis realizes the precise fault positioning of a complex and coupled system.
The method provided by the invention is adopted to analyze and obtain the bilateral vector probability model of the filter amplifying circuit, which is composed of fault reasons, fault modes and test item nodes.
Fig. 6 shows a schematic diagram of a filter amplifying circuit, which is a filter amplifying circuit composed of an inverting proportional amplifier, an RC low-pass filter and an inverter, wherein the signal related to the inverting proportional amplifier is s1(gain), s2(degree of linearity), s4(slew rate) and s5(DC offset) the filter dependent signal is s3(cut-off frequency) the signal associated with the inverter is s1(gain), s2(degree of linearity), s4(slew rate) and s5(DC offset).
As can be seen from the figure, when R is1||R2≠R3In the first stage, DC offset is generated, so R1、R2And R3All influence signal s5. In the same way as R4And C1Will influence s3And R is1、R2And the open loop gain of the operational amplifier, affect the gain of the first stage of the amplification filter. In addition, the system is provided with 3 measuring points which are respectively TP1、TP2And J1. Each station has an associated test, and each test has an associated signal. Wherein the measuring point TP1With 4 tests t1、t2、t3、t4And t10They separately detect the signal s1、s2、s4And s5(ii) a Measuring point TP2Containing test t5、t11It detects the signal s3(ii) a Measurement point J1Of which there are 4 tests t6、t7、t8And t9Separately detecting the signal s1、s2、s4And s5。
Reliability related information of the circuit shown in fig. 6 is collected, analysis is performed according to aspects such as functional signals, fault reasons, fault modes, fault influences, detection means and fault reason occurrence probability, and the fault modes are divided into the following parts according to multi-signal model requirements: the two types of functional faults and global faults, denoted by the symbols F and G, respectively, have the following results:
analyzing the data in the table above of the filter amplifying circuit, constructing a fault cause-fault mode correlation matrix C as shown in FIG. 714×17And a failure mode-test item correlation matrix D as shown in FIG. 814×11Further, a double-edge correlation matrix B as shown in FIG. 9 can be calculated14×28。
Assuming that the amplification filter is evaluated 3 times with expert knowledge:
1) if in the bilateral vector matrix
The test items and failure modes in (1) are in a one-to-many relationship, e.g. cause t10Failure mode of reporting failure is m1、m2Then, according to 3 evaluations, obtain fuzzy judgment matrix B1、B2、B3。
The maximum eigenvalue of the fuzzy judgment matrix given by the 3 evaluations is respectively:
fuzzy judgment matrix B1、B2、B3And performing consistency check to obtain a consistency ratio as follows:
as can be seen from equation (14), the consistency satisfies the requirement, and 3 evaluation pairs can be obtained to cause t10Failure mode m of reporting failure1、m2Estimation of the likelihood of occurrence.
The weights taken for 3 evaluations are equal, i.e. θiThe fused estimate is finally obtained 1/3.
V＝[0.9928,0.1195] (16)
2) If in the bilateral vector matrix
The test items and failure modes in (1) are in a one-to-one relationship, e.g., t11And m12Are uniquely correlated, the probability between them is directly estimated by the triangular fuzzy function, and equal weighting fusion is carried out.
V'＝0.889 (18)
3) If in the bilateral vector matrix
There is no correlation between the cause of the failure and the failure mode or between the test item and the failure mode, and the probability thereof is directly represented by 0.
Thereby obtaining the vector probability matrix of the amplifying filter circuit
As shown in the table below, 3 evaluations were obtained from the table
An estimate of the probability of an event occurring.
Therefore, if the test points report faults, the probability sequencing of the relevant fault modes and fault reasons can be realized through the table so as to position the faults.
The above examples are only intended to illustrate the technical solution of the present invention and not to limit it; although the present invention has been described in detail with reference to preferred embodiments, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention; without departing from the spirit of the present invention, it is intended to cover all aspects of the invention as defined by the appended claims.
Claims (5)
1. An equipment fault diagnosis model building method is characterized in that the fault diagnosis model can represent cause and effect relationships determined among fault causes, fault modes and test item nodes, and diagnosis of equipment faults can be realized through the cause and effect relationships, and the method comprises the following steps:
constructing a first correlation matrix C between fault reasons and fault modes, wherein the matrix C is a binary matrix, row vectors of the matrix C represent the influence conditions of all the fault reasons on the fault modes, and column vectors represent the influence conditions of all the fault modes corresponding to the fault reasons;
constructing a second correlation matrix D between the fault mode and the test items, wherein the matrix D is a binary matrix, row vectors of the matrix D represent the response of all tests to the fault mode excitation signals, and column vectors of the matrix D represent the response of all fault mode excitation signals corresponding to the tests;
and combining the first correlation matrix C and the second correlation matrix D to generate a double-edge correlation matrix B, wherein in the double-edge correlation matrix B, a fault mode-test item correlation matrix element represents the correlation between a fault mode and a test item, and a fault reason-fault mode correlation matrix element represents the correlation between the fault mode and the fault reason.
2. The method for constructing the equipment fault diagnosis model according to claim 1, further comprising constructing a bilateral vector probability matrix capable of representing the correlation degree of the test timing sequence, the fault reason and the fault mode based on the bilateral correlation matrix and expert knowledge
3. The equipment fault diagnosis model construction method according to claim 2, characterized by constructing the bilateral vector probability matrix
The method comprises the following steps: the sequence of the test items in the double-side correlation matrix is arranged according to the equipment test sequence, the fault mode sequence is sequenced according to the test items and the correlation degree of the test items to obtain a double-side vector matrix, and expert knowledge is adopted to obtain the correlation degree between each test item and the corresponding fault mode and between each fault mode and the corresponding fault reason.
4. The equipment fault diagnosis model construction method according to claim 3, characterized in that the most confident application of expert knowledge is performed by applying triangular fuzzy number analytic hierarchy process.
5. An equipment fault diagnosis method characterized by comprising:
constructing a double-edge correlation matrix representing cause and effect relationships determined among equipment fault reasons, fault modes and test project nodes;
acquiring equipment test data, sequencing the test data according to the row vector sequence of the double-edge correlation matrix test items and converting the test data into a vector, wherein the value of an element in the vector depends on whether the test data is in accordance with the specification requirement compared with the range of the test items and the preset value of the parameter index to be tested;
and reporting faults of the test items with the element values in the vectors which do not meet the specification requirements, matching the row vectors of the test items in the double-edge correlation matrix to obtain a fault mode set related to the fault reporting test items, and obtaining a fault reason set related to the fault mode set through the double-edge correlation matrix.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011208271.8A CN112163189A (en) | 2020-11-03 | 2020-11-03 | Equipment fault diagnosis method based on test sequence probability matrix |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011208271.8A CN112163189A (en) | 2020-11-03 | 2020-11-03 | Equipment fault diagnosis method based on test sequence probability matrix |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112163189A true CN112163189A (en) | 2021-01-01 |
Family
ID=73865526
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011208271.8A Pending CN112163189A (en) | 2020-11-03 | 2020-11-03 | Equipment fault diagnosis method based on test sequence probability matrix |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112163189A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115712840A (en) * | 2022-11-15 | 2023-02-24 | 中国人民解放军陆军工程大学 | Multi-fault diagnosis method and system for electronic information system |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105021275A (en) * | 2015-07-15 | 2015-11-04 | 吉林大学 | Time domain dynamic characteristic analysis-based intra-vehicle sound quality prediction method |
| CN108304661A (en) * | 2018-02-05 | 2018-07-20 | 南京航空航天大学 | Diagnosis prediction method based on TDP models |
| CN109325293A (en) * | 2018-09-25 | 2019-02-12 | 中国人民解放军海军航空大学 | A kind of Multiple faults diagnosis approach based on correlation models |
| CN109670610A (en) * | 2018-12-24 | 2019-04-23 | 中国人民解放军32181部队 | Fault diagnosis optimization method based on fault propagation analysis |
-
2020
- 2020-11-03 CN CN202011208271.8A patent/CN112163189A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105021275A (en) * | 2015-07-15 | 2015-11-04 | 吉林大学 | Time domain dynamic characteristic analysis-based intra-vehicle sound quality prediction method |
| CN108304661A (en) * | 2018-02-05 | 2018-07-20 | 南京航空航天大学 | Diagnosis prediction method based on TDP models |
| CN109325293A (en) * | 2018-09-25 | 2019-02-12 | 中国人民解放军海军航空大学 | A kind of Multiple faults diagnosis approach based on correlation models |
| CN109670610A (en) * | 2018-12-24 | 2019-04-23 | 中国人民解放军32181部队 | Fault diagnosis optimization method based on fault propagation analysis |
Non-Patent Citations (2)
| Title |
|---|
| 张昭 等: "基于双边矢量概率矩阵的故障诊断方法研究", 《计算机测量与控制》, vol. 28, no. 07, 25 July 2020 (2020-07-25), pages 5 - 8 * |
| 李彬赫: "远程物联健康监测系统设计", 《万方数据》, 2 November 2022 (2022-11-02), pages 1 - 69 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115712840A (en) * | 2022-11-15 | 2023-02-24 | 中国人民解放军陆军工程大学 | Multi-fault diagnosis method and system for electronic information system |
| CN115712840B (en) * | 2022-11-15 | 2023-06-13 | 中国人民解放军陆军工程大学 | Multi-fault diagnosis method and system for electronic information system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102818948B (en) | Synthetic diagnosing method based on fuzzy fault diagnosis and relevancy model diagnosis | |
| US7844873B2 (en) | Fault location estimation system, fault location estimation method, and fault location estimation program for multiple faults in logic circuit | |
| CN103970997B (en) | A kind of depopulated helicopter sensor fault fast diagnosis method | |
| CN109657797B (en) | Fault diagnosis capability analysis method based on hybrid diagnosis Bayesian network | |
| US20100131800A1 (en) | Diagnostic device | |
| CN105137324B (en) | A kind of more detection point failure component localization methods based on emulation disaggregated model | |
| JPH08234832A (en) | Device and method for monitoring and diagnostic plant | |
| CN107490758B (en) | Modularization BP neural network circuit failure diagnosis method based on fault propagation | |
| CN106021671B (en) | In conjunction with the circuit health grading evaluation method of correlative relationship and grey cluster technology | |
| CN116450399A (en) | Fault diagnosis and root cause positioning method for micro service system | |
| CN113626267A (en) | Method for evaluating uncertainty fault diagnosis efficiency of complex electronic system | |
| CN109726770A (en) | A kind of analog circuit fault testing and diagnosing method | |
| CN112163189A (en) | Equipment fault diagnosis method based on test sequence probability matrix | |
| CN105182219A (en) | Power converter fault classification method based on Hamming error correcting code support vector machine | |
| CN110441669A (en) | The gradual failure diagnosis of uncertain sophisticated circuitry system and life-span prediction method | |
| Laidani et al. | Analog Circuit Fault Classification and Data Reduction Using PCA-ANFIS Technique Aided by K-means Clustering Approach. | |
| CN116658433A (en) | Intelligent detection method and system for particle-resistant magnetic pump | |
| CN107103134A (en) | Low-speed wireless sensor network testability analysis method based on Bayesian network | |
| CN115184055A (en) | Method and system for determining test set with optimized hierarchical testability | |
| D'Antona et al. | Bad data detection and identification in power system state estimation with network parameters uncertainty | |
| Gaita et al. | A machine learning based wafer test ranking for root cause analysis | |
| KR102182678B1 (en) | Method and appratus for predicting fault pattern using multi-classifier based on feature selection method in semiconductor manufacturing process | |
| Guerreiro et al. | Fault list compression for efficient analogue and mixed-signal production test preparation | |
| CN114441698A (en) | Gas chromatograph fault diagnosis method based on Bayesian network | |
| Mosin | Quality improvement of analog circuits fault diagnosis based on ANN using clusterization as preprocessing |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |