CN101871994A - Method for diagnosing faults of analog circuit of multi-fractional order information fusion - Google Patents
Method for diagnosing faults of analog circuit of multi-fractional order information fusion Download PDFInfo
- Publication number
- CN101871994A CN101871994A CN 201010198615 CN201010198615A CN101871994A CN 101871994 A CN101871994 A CN 101871994A CN 201010198615 CN201010198615 CN 201010198615 CN 201010198615 A CN201010198615 A CN 201010198615A CN 101871994 A CN101871994 A CN 101871994A
- Authority
- CN
- China
- Prior art keywords
- mrow
- fractional order
- diagnosis
- fault
- samples
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 41
- 230000004927 fusion Effects 0.000 title claims abstract description 10
- 238000003745 diagnosis Methods 0.000 claims abstract description 64
- 238000012360 testing method Methods 0.000 claims abstract description 39
- 238000013507 mapping Methods 0.000 claims abstract description 8
- 230000005284 excitation Effects 0.000 claims abstract description 6
- 238000013145 classification model Methods 0.000 claims description 30
- 238000012549 training Methods 0.000 claims description 26
- 238000013528 artificial neural network Methods 0.000 claims description 10
- 238000007781 pre-processing Methods 0.000 claims description 6
- 238000005457 optimization Methods 0.000 claims description 4
- 230000035772 mutation Effects 0.000 claims description 3
- 238000012546 transfer Methods 0.000 claims description 3
- 230000008859 change Effects 0.000 abstract description 9
- 238000004422 calculation algorithm Methods 0.000 description 9
- 238000011160 research Methods 0.000 description 7
- 238000004458 analytical method Methods 0.000 description 6
- 230000002068 genetic effect Effects 0.000 description 6
- 238000000513 principal component analysis Methods 0.000 description 6
- 230000010354 integration Effects 0.000 description 5
- 238000000605 extraction Methods 0.000 description 4
- 230000035945 sensitivity Effects 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 2
- 230000002349 favourable effect Effects 0.000 description 2
- 238000012567 pattern recognition method Methods 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 230000007123 defense Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000013100 final test Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000009776 industrial production Methods 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000010206 sensitivity analysis Methods 0.000 description 1
Images
Landscapes
- Complex Calculations (AREA)
Abstract
The invention discloses a method for diagnosing faults of an analog circuit of multi-fractional order information fusion, belonging to the field of network fault test and diagnosis of the analog circuit. The method comprises the following steps of: A inputting N types of effective sinusoidal excitation signals with different frequencies in series at the input end of a circuit to be tested to obtain an initial dataset; B randomly dividing a sample set into k feature space subsets by adopting a random subspace method; C mapping the feature space subsets into k diverse fractional-order time-frequency spaces by using fractional-order Fourier transform; and D obtaining a final diagnosis result by utilizing the weight of a failure class and fusing the diagnosis results of k basic classifying models, and dynamically updating the weight of the failure class after completing the diagnosis on each batch of test data. The invention improves the difference degree of the feature space subsets, increases complementarity among basic classifiers, combines with the gradual change characteristics of device faults and effectively improves the precision of a diagnosing system.
Description
Technical Field
The invention relates to a fault diagnosis method for an analog circuit, in particular to a fault diagnosis method for an analog circuit with multi-fractional order information fusion.
Background
The analog circuit is an indispensable important component of an electronic circuit, the complexity and the density of the analog circuit continuously increase along with the rapid development of a large-scale analog integrated circuit, particularly an analog-digital mixed circuit, once a fault occurs at a certain position and cannot be diagnosed and recovered in time, the analog circuit is stopped at a light moment, and casualties are caused at a heavy moment, so that huge economic loss is brought. For the application field with high reliability requirement, the reliability guarantee of the device is especially important. Therefore, the technology of fault diagnosis and fault prediction of autonomous analog circuits is being developed and applied vigorously, has become an important content in modern industrial production and national defense construction, and is also one of the hot spots of research in the scientific community at present.
The research of the feature extraction and diagnosis method is an important content of the analog circuit fault diagnosis research, and the feature extraction method based on signal processing and the intelligent diagnosis method based on pattern recognition are hot spots of the research of the analog circuit fault diagnosis method at present. Such as fourier transform, wavelet transform, hilbert-yellow transform, fractional fourier transform, etc., which can acquire dynamic characteristic information of a circuit state as compared with a conventional characteristic extraction method. The pattern recognition methods such as the neural network, the expert system, the fuzzy theory, the decision tree, the Bayesian theory and the like do not need mathematical models, only specific operation rules are needed to be applied, and the measurement space is mapped to the decision space, so that complicated mathematical operations are avoided, and the fault elements in the network can be judged only through limited fault information.
The existing analog circuit fault diagnosis method mainly researches feature extraction in the same way to obtain training data sets in the same feature space, and adopts a single classifier to diagnose faults. However, practical applications show that the conventional analog circuit fault diagnosis method has the following disadvantages: (1) the training set samples obtained by extracting fault features in the same way have no too much difference, so that the generalization capability of the classification model obtained by training is poor; (2) it has been shown that any single classifier cannot completely solve the problem faced by the analog circuit fault diagnosis or meet the system requirements. When a single classification model encounters a certain fault class which is difficult to diagnose, it is difficult to improve the diagnosis precision by adjusting model parameters or adding training samples. The diversity classifier integration method can effectively solve the problems, and improves the precision of the whole classification model through the complementarity among a plurality of classifiers. In recent years, classifier integration is a research hotspot, but is still in the beginning of the field of analog circuit fault diagnosis.
In addition, the faults of components in the analog circuit have the characteristic of gradual change. Because of the existence of tolerance, the component value always changes in the normal tolerance range around the nominal value, when the component value is larger than the normal tolerance range, the fault is considered to occur, and according to the rule of component damage, the fault always changes from small soft fault to large soft fault and then to hard fault, so the fault of the component in the analog circuit has gradual change. After the large soft fault occurs to the component, the diagnostic model obtained by the training of the small soft fault sampling data does not accord with the fault characteristics of the circuit component. Therefore, according to the characteristic of gradual fault change of the analog circuit component, a fault diagnosis system needs to be updated dynamically, and research on the fault diagnosis system is still not much so far, and the fault diagnosis system is not solved completely.
Disclosure of Invention
The invention aims to provide an analog circuit fault diagnosis method which can solve the problems of poor generalization capability of a diagnosis model and neglect of fault gradual change existing in the conventional analog circuit fault diagnosis method.
The invention achieves the purpose through the following technical scheme:
in order to improve the diversity between training sample sets, an original sample set is firstly divided into a plurality of feature space subsets by adopting a random subspace method. Then, selecting different fractional order p values for the sample set of each sub-feature space, and mapping the data of the feature space subsets to different fractional order time-frequency spaces by adopting fractional order Fourier transform; in order to improve the generalization capability of the diagnosis models, the complementarity between the diagnosis models is improved by adopting a multi-classifier integration method, and the diagnosis results of a plurality of classifiers are fused to obtain a final diagnosis result; in order to solve the characteristic of gradual change of the faults of the components in the analog circuit, the weight of each fault class is dynamically updated according to the ratio of the latest fault classes in all test sample sets, and the weights of the fault classes are fused into the diagnosis result.
Specifically, the technical scheme of the invention carries out analog circuit fault diagnosis according to the following steps:
A. connecting N sinusoidal excitation signals with different frequencies within an allowable range of an input circuit in series at the input end of a circuit to be tested, wherein the value range of N is 2-5, and collecting voltage signals of measurable nodes to obtain an original data set;
B. and B, performing feature subspace preprocessing on the original data set obtained in the step A, specifically: dividing an original data set into k feature space subsets randomly by using a random subspace method, wherein the feature space subsets are not intersected with each other, and the number of feature samples of each type in each feature space subset is equal; wherein k is the number of base classification models to be adopted and is more than or equal to 2;
C. mapping the k feature subspaces obtained by the preprocessing in the step B into k different fractional order time-frequency spaces by using fractional order Fourier transform, and training feature samples in the k different fractional order time-frequency spaces to obtain k base classification models;
D. c, the k base classification models obtained in the step C are used for carrying out batch diagnosis on the dynamically acquired test data respectively, the diagnosis results of the k base classification models are fused according to the following formula by using the fault class weight, the final diagnosis result class of z is obtained, and the weight of the fault class is dynamically updated after the diagnosis of each batch of test data is finished:
wherein z represents a sample to be tested; f represents the number of fault categories; k represents the number of base classification models; outputi(z) ═ j denotes the probability value of the test sample z being diagnosed as a fault class j by the ith base classification model, j ═ 1, 2.. f, i ═ 1, 2.. k; pjThe weight assigned to the jth fault class is calculated by the following formula:
where s is the number of samples in the raw data set; m is a preset quantitative constant and represents the number of test samples in each batch of test sample set; g is the integer number of the quotient of the total number of test samples minus the number of original samples s and divided by m；numberjIs the number of samples judged as the jth fault class after diagnosisjThe number of samples containing the jth class of failure in the training sample set.
The preset quantitative constant m can be selected according to actual conditions.
Compared with the prior art, the scheme of the invention has the following technical effects:
(1) according to the method, the original data set is mapped to k different fractional order time-frequency spaces by utilizing fractional order Fourier transform, the distribution characteristics of the original data set are changed through the different fractional order time-frequency spaces, the difference between different training sample sets of the feature subspace is improved, and the method is favorable for improving the integration complementarity of the classifier.
(2) The invention adopts a neural network method as a base classification model, utilizes the fault class weight and fuses the diagnosis results of k base classification models to obtain the final diagnosis result, compared with the traditional single diagnosis method, the multi-classifier integrated diagnosis method has more generalization capability, and increases the complementarity between the diagnosis models through the diversity multi-classifier integration, thereby improving the precision of the fault diagnosis of the analog circuit.
(3) The dynamically acquired test data is diagnosed in batches, the weight of the fault class is dynamically updated according to the ratio of the latest various fault classes in all data samples, the characteristic of analog circuit fault gradual change is met, the weight is fused into the diagnosis result, and the efficiency of analog circuit fault diagnosis is improved.
Fractional order fourier transform is a uniform time-frequency transform, and by rotating factor α, the fractional order fourier transform rotates on the time-frequency plane to obtain a new signal representation, which is defined as:
wherein,
where α is p pi/2, p is the fractional order of the fractional fourier transform, and the fractional fourier domain of p is the coordinate space generated by rotating α in the counterclockwise direction on the (t, w) plane. Alpha can be any number, a counterclockwise rotation angle of 0-pi/2 is generally taken for analysis, due to the symmetry and periodicity of a fractional Fourier domain, the signal analysis result is consistent with the rotation angle of 0-pi/2, the value of p is generally 0-1, and when p is 0, fractional Fourier transform is an original signal; when p is 1, it is equal to the classical fourier transform. The fractional fourier transform smoothly changes from an original function to a normal fourier transform as p changes from 0 to 1.
From the above analysis, if a fractional Fourier transform is used, a fractional p-value must first be determined. Therefore, k random feature subspaces are mapped to different fractional order time-frequency spaces by using the uncertainty of the fractional order p value, so that the difference degree of the subspace feature samples is increased.
In order to further increase the diversity factor of the spatial feature samples, a genetic algorithm can be used for optimizing the fractional order p values, the diversity factor among the multiple classifiers is used as a fitness function, and the evolution target of the genetic algorithm is to maximize the fitness function, so that the optimal k fractional order p values are obtained; specifically, step C in the above scheme includes the following steps:
c1, setting an initial fractional order p value, wherein the value range of p is 0-1; establishing an initial population;
c2, mapping the k characteristic subspace samples to k different fractional order time-frequency domains by utilizing a fractional order Fourier transform method according to the fractional order p value;
c3, training to obtain k base classification models according to k different fractional order time frequency subspace feature samples, wherein the diagnosis model adopts a three-layer BP neural network method, a transfer function and a training function of the neural network respectively adopt 'Logsig' and 'Traingdx', and the error training precision is 0.001;
c4, calculating a difference value d (all _ classifier) between the multiple classifiers according to the following formula:
wherein,
classifierr _ n, which respectively represent the nth and mth classifiers of the k classifiers;
prob _ fail (classifier _ n, classifier _ m) is the probability of two classifiers, classifier _ n, failing the same test sample at the same time, i.e. the number of test samples misdiagnosed by two classifiers at the same time divided by the total number of test samples;
c5, judging whether the preset stop condition is met, if yes, the obtained k fractional order p values are optimal, and continuing to execute the step D; if not, performing selection, crossing, mutation and reinsertion operations on the population to obtain a new population, and then turning to the step C2; wherein the preset stop condition is that: the disparity value d (all _ classifier) obtained in step C4 is smaller than a first predetermined threshold, or the number of optimization iterations is larger than a second predetermined threshold.
The first threshold and the second threshold can be selected according to actual conditions.
Drawings
FIG. 1 is a flow chart of the method of the present invention;
FIG. 2 is a flow chart of the method for optimizing and selecting multiple fractional order p-values by using a multi-objective genetic algorithm according to an embodiment of the present invention.
Detailed Description
The technical scheme of the invention is explained in detail in the following with the accompanying drawings:
as shown in fig. 1, the analog circuit fault diagnosis is performed according to the following steps:
A. connecting N sinusoidal excitation signals with different frequencies within an allowable range of an input circuit in series at the input end of a circuit to be tested, wherein the value range of N is 2-5, and collecting voltage signals of measurable nodes to obtain an original data set;
firstly, analyzing the effective passband range of a circuit to be tested by a sweep frequency analysis method, wherein the frequency conversion range of the sweep frequency is [0.001Hz 100MHz ], in order to increase multi-frequency information in voltage signals acquired by a test node and improve the separability of a fault data sample, connecting 3 sinusoidal excitation signals with different frequencies, the same amplitude value of 1V and the phase of 0 in series at the input end of the circuit to be tested, wherein the frequencies of the sinusoidal excitation signals respectively select different frequency values representing the high, medium and low in the effective passband range of the circuit to be tested;
then, selecting a proper test node, in this embodiment, selecting an optimal test node by using a node voltage sensitivity analysis method, where the higher the node voltage sensitivity is, the more sensitive the node is to the parameter change of the element, and the more favorable the improvement of the fault identification capability is, specifically according to the following steps:
1) presetting a plurality of fault types of a circuit to be tested, and collecting voltage values of a plurality of fault type output nodes of a test node;
the method includes the steps that a fault model needs to be set manually when an original data set is collected, and as the fault types of the analog circuit are various, in order to enable the generalization capability of a diagnosis model obtained by training original data to be as large as possible, the collected fault type data need to reflect the fault types of the analog circuit as much as possible, the set fault category in the embodiment is 6, and the fault category comprises a component value increasing soft fault, a component value decreasing soft fault, a component open-circuit hard fault, a component short-circuit hard fault, a plurality of component soft faults and a plurality of component hard faults;
2) the voltage sensitivity value of each measuring point is calculated by adopting the following voltage sensitivity formula. And selecting 2-3 test nodes with the maximum voltage sensitivity value as final test nodes.
Wherein r is a component parameter, VoutA variable is output for the test node voltage.
B. And B, performing feature subspace preprocessing on the original data set obtained in the step A, specifically: dividing an original data set into k feature space subsets randomly by using a random subspace method, wherein the feature space subsets are not intersected with each other, and the number of feature samples of each type in each feature space subset is equal; wherein k is the number of base classification models to be adopted and is more than or equal to 2;
theoretically, the more the number of the base classification models is, i.e., the greater k is, the higher the accuracy of the result of the fusion diagnosis is, but the complexity of the diagnosis process is increased, so that the value of k is usually between 5 and 10, and the value of k is 8 in the present embodiment, i.e., 8 base classification models are adopted for the fusion diagnosis;
C. mapping the k feature subspaces obtained by the preprocessing in the step B into k different fractional order time-frequency spaces by using fractional order Fourier transform, and training feature samples in the k different fractional order time-frequency spaces to obtain k base classification models;
in the specific embodiment, a multi-target genetic algorithm is adopted to optimize the fractional order p values, the difference value between multiple classifiers is used as a fitness function, and the evolution target of the genetic algorithm is to maximize the fitness function, so that the optimal k fractional order p values are obtained; specifically, as shown in fig. 2, the present step includes the following steps:
c1, setting an initial fractional order p value, wherein the value range of p is 0-1, and establishing an initial population;
c2, mapping the k characteristic subspace samples to k different fractional order time-frequency domains by utilizing a fractional order Fourier transform method according to the fractional order p value;
c3, training to obtain k base classification models according to k different fractional order time frequency subspace feature samples, wherein the diagnosis model adopts a three-layer BP neural network method, a transfer function and a training function of the neural network respectively adopt 'Logsig' and 'Traingdx', and the error training precision is 0.001;
c4, calculating a difference value d (all _ classifier) between the multiple classifiers according to the following formula:
wherein,
classifierr _ n, which respectively represent the nth and mth classifiers of the k classifiers;
prob _ fail (classifier _ n, classifier _ m) is the probability of two classifiers, classifier _ n, failing the same test sample at the same time, i.e. the number of test samples misdiagnosed by two classifiers at the same time divided by the total number of test samples;
c5, judging whether the preset stop condition is met, if yes, the obtained k fractional order p values are optimal, and continuing to execute the step D; if not, performing selection, crossing, mutation and reinsertion operations on the population to obtain a new population, and then turning to the step C2; wherein the preset stop condition is that: the disparity value d (all _ classifier) obtained in the step C4 is smaller than a preset first threshold, or the number of optimization iterations is larger than a preset second threshold;
in this embodiment, the first threshold and the second threshold are 0.01 and 500, respectively, that is, when the difference value is less than 0.01 or the number of times of the optimized iterations reaches 500, the iterations are stopped, and the k fractional order p values at this time are optimal.
D. C, the k base classification models obtained in the step C are used for carrying out batch diagnosis on the dynamically acquired test data respectively, the diagnosis results of the k base classification models are fused according to the following formula by using the fault class weight, the final diagnosis result class of z is obtained, and the weight of the fault class is dynamically updated after the diagnosis of each batch of test data is finished:
wherein z represents a sample to be tested; f represents the number of fault categories; k represents the number of base classification models; outputi(z) ═ j denotes the probability value of the test sample z being diagnosed as a fault class j by the ith base classification model, j ═ 1, 2.. f, i ═ 1, 2.. k; pjThe weight assigned to the jth fault class is calculated by the following formula:
where s is the number of samples in the raw data set; m is a preset quantitative constant and represents the number of test samples in each batch of test sample set; g is the integer numerical value of the quotient obtained by subtracting the original sample number s from the total test sample number and dividing the result by m; number ofjIs the number of samples judged as the jth fault class after diagnosisjThe number of samples containing the jth class of failure in the training sample set.
In this embodiment, in order to simplify the training complexity of the neural network based classification model, after the sample set of k fractional order time-frequency feature spaces is obtained, a Principal Component Analysis (PCA) algorithm is used to extract T-dimensional main features from data samples in each fractional order time-frequency domain, and the T-dimensional main features are subjected to dimensionality reduction and used as input data of the neural network based classification model. The PCA method can reduce the original data set to any dimension by extracting main features, but when the reduced dimension of the PCA is too low, the effective feature information of the original data set is lost too much, in the specific implementation mode, the value of T is 30, namely, the feature sample in the fractional order time-frequency domain is reduced to 30 dimensions;
in this embodiment, the preset quantitative constant m is 500, that is, after 500 test samples are diagnosed, the batch of test data set is added to the original training data set, the occupancy of various fault classes is recalculated, and the weight P of various fault classes is updatedjSize; therefore, when a certain component in the analog circuit has a fault gradual change, the fault class acquired by training of the small soft fault sample does not exist, the weight of the fault class can be reduced by dynamically recalculating and updating the weight of the fault class, and the probability of diagnosing the fault class is reduced.
The genetic algorithm, the fractional fourier transform, the sweep frequency analysis method, the principal component analysis PCA algorithm and the like used in the present embodiment are all the existing technologies, and specific contents can be referred to (multi-objective intelligent optimization algorithm and application thereof, redeming et al, 2009, science publishers, fractional fourier transform and application thereof, pottery et al, 2009, qing hua university publishers, computer analysis and design of analog circuits — pspie program application, gaowen et al, 1998, qing hua university publishers, intelligent pattern recognition method, shaojian, 2006, south china university publishers).
Claims (4)
1. A multi-fractional order information fusion analog circuit fault diagnosis method is characterized by comprising the following steps:
A. connecting N sinusoidal excitation signals with different frequencies within an allowable range of an input circuit in series at the input end of a circuit to be tested, wherein the value range of N is 2-5, and collecting voltage signals of measurable nodes to obtain an original data set;
B. and B, performing feature subspace preprocessing on the original data set obtained in the step A, specifically: dividing an original data set into k feature space subsets randomly by using a random subspace method, wherein the feature space subsets are not intersected with each other, and the number of feature samples of each type in each feature space subset is equal; wherein k is the number of base classification models to be adopted and is more than or equal to 2;
C. mapping the k feature subspaces obtained by the preprocessing in the step B into k different fractional order time-frequency spaces by using fractional order Fourier transform, and training feature samples in the k different fractional order time-frequency spaces to obtain k base classification models;
D. c, the k base classification models obtained in the step C are used for carrying out batch diagnosis on the dynamically acquired test data respectively, the diagnosis results of the k base classification models are fused according to the following formula by using the fault class weight, the final diagnosis result class of z is obtained, and the weight of the fault class is dynamically updated after the diagnosis of each batch of test data is finished:
wherein z represents a sample to be tested; f represents the number of fault categories; k represents the number of base classification models; outputi(z) ═ j denotes the probability value of the test sample z being diagnosed as a fault class j by the ith base classification model, j ═ 1, 2.. f, i ═ 1, 2.. k; pjThe weight assigned to the jth fault class is calculated by the following formula:
where s is the number of samples in the raw data set; m is a preset quantitative constant and represents the number of test samples in each batch of test sample set; g is the integer numerical value of the quotient obtained by subtracting the original sample number s from the total test sample number and dividing the result by m; number ofjIs the number of samples judged as the jth fault class after diagnosisjThe number of samples containing the jth class of failure in the training sample set.
2. The method of multi-fractional order information fusion analog circuit fault diagnosis of claim 1, wherein: the step C is specifically executed according to the following steps:
c1, setting an initial fractional order p value, wherein the value range of p is 0-1, and establishing an initial population;
c2, mapping the k characteristic subspace samples to k different fractional order time-frequency domains by utilizing a fractional order Fourier transform method according to the fractional order p value;
c3, training to obtain k base classification models according to k different fractional order time frequency subspace feature samples, wherein the diagnosis model adopts a three-layer BP neural network method, a transfer function and a training function of the neural network respectively adopt 'Logsig' and 'Traingdx', and the error training precision is 0.001;
c4, calculating a difference value d (all _ classifier) between the multiple classifiers according to the following formula:
wherein,
classifierr _ n, which respectively represent the nth and mth classifiers of the k classifiers;
prob _ fail (classifier _ n, classifier _ m) is the probability of two classifiers, classifier _ n, failing the same test sample at the same time, i.e. the number of test samples misdiagnosed by two classifiers at the same time divided by the total number of test samples;
c5, judging whether the preset stop condition is met, if yes, the obtained k fractional order p values are optimal, and continuing to execute the step D; if not, performing selection, crossing, mutation and reinsertion operations on the population to obtain a new population, and then turning to the step C2; wherein the preset stop condition is that: the disparity value d (all _ classifier) obtained in step C4 is smaller than a first predetermined threshold, or the number of optimization iterations is larger than a second predetermined threshold.
3. The method of multi-fractional order information fusion analog circuit fault diagnosis of claim 2, wherein: the preset values of the first threshold and the second threshold are 0.01 and 500 respectively.
4. The method of multi-fractional order information fusion analog circuit fault diagnosis of claim 1, wherein: the predetermined quantitative constant m has a value of 500.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2010101986151A CN101871994B (en) | 2010-06-11 | 2010-06-11 | Method for diagnosing faults of analog circuit of multi-fractional order information fusion |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2010101986151A CN101871994B (en) | 2010-06-11 | 2010-06-11 | Method for diagnosing faults of analog circuit of multi-fractional order information fusion |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101871994A true CN101871994A (en) | 2010-10-27 |
| CN101871994B CN101871994B (en) | 2012-03-21 |
Family
ID=42996980
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2010101986151A Expired - Fee Related CN101871994B (en) | 2010-06-11 | 2010-06-11 | Method for diagnosing faults of analog circuit of multi-fractional order information fusion |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101871994B (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102520341A (en) * | 2011-12-05 | 2012-06-27 | 南京航空航天大学 | Analog circuit fault diagnosis method based on Bayes-KFCM (Kernelized Fuzzy C-Means) algorithm |
| CN103617557A (en) * | 2013-11-06 | 2014-03-05 | 广东电网公司电力科学研究院 | Multi-source heterogeneous power grid operation parameter analysis system |
| CN103701464A (en) * | 2013-12-13 | 2014-04-02 | 华北电力大学（保定） | Fractional order element converter |
| CN104484678A (en) * | 2015-01-07 | 2015-04-01 | 青岛理工大学 | Multi-classifier fusion fault diagnosis method based on fault type classification capability evaluation matrix |
| CN104635146A (en) * | 2015-02-06 | 2015-05-20 | 南京农业大学 | Analog circuit fault diagnosis method based on random sinusoidal signal test and HMM (Hidden Markov Model) |
| CN105654102A (en) * | 2014-11-10 | 2016-06-08 | 富士通株式会社 | Data processing device and data processing method |
| CN107192939A (en) * | 2017-06-08 | 2017-09-22 | 京东方科技集团股份有限公司 | A kind of detection method and detecting system of the working condition of circuit |
| CN108957297A (en) * | 2018-08-07 | 2018-12-07 | 北京航空航天大学 | A kind of circuit fault diagnosis extension Constructing Method for Decision introducing work information |
| CN112082639A (en) * | 2019-06-14 | 2020-12-15 | 现代自动车株式会社 | Engine state diagnosis method and diagnosis modeling method thereof |
| CN113159089A (en) * | 2021-01-18 | 2021-07-23 | 安徽建筑大学 | Pavement damage identification method, system, computer equipment and storage medium |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101533068A (en) * | 2009-04-08 | 2009-09-16 | 南京航空航天大学 | Analog-circuit fault diagnosis method based on DAGSVC |
| CN101614786A (en) * | 2009-07-07 | 2009-12-30 | 南京航空航天大学 | Power electronic circuit on-line intelligence method for diagnosing faults based on FRFT and IFSVC |
-
2010
- 2010-06-11 CN CN2010101986151A patent/CN101871994B/en not_active Expired - Fee Related
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101533068A (en) * | 2009-04-08 | 2009-09-16 | 南京航空航天大学 | Analog-circuit fault diagnosis method based on DAGSVC |
| CN101614786A (en) * | 2009-07-07 | 2009-12-30 | 南京航空航天大学 | Power electronic circuit on-line intelligence method for diagnosing faults based on FRFT and IFSVC |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102520341A (en) * | 2011-12-05 | 2012-06-27 | 南京航空航天大学 | Analog circuit fault diagnosis method based on Bayes-KFCM (Kernelized Fuzzy C-Means) algorithm |
| CN102520341B (en) * | 2011-12-05 | 2013-12-18 | 南京航空航天大学 | Analog circuit fault diagnosis method based on Bayes-KFCM (Kernelized Fuzzy C-Means) algorithm |
| CN103617557A (en) * | 2013-11-06 | 2014-03-05 | 广东电网公司电力科学研究院 | Multi-source heterogeneous power grid operation parameter analysis system |
| CN103701464A (en) * | 2013-12-13 | 2014-04-02 | 华北电力大学（保定） | Fractional order element converter |
| CN103701464B (en) * | 2013-12-13 | 2017-02-01 | 华北电力大学（保定） | Fractional order element converter |
| CN105654102A (en) * | 2014-11-10 | 2016-06-08 | 富士通株式会社 | Data processing device and data processing method |
| CN104484678A (en) * | 2015-01-07 | 2015-04-01 | 青岛理工大学 | Multi-classifier fusion fault diagnosis method based on fault type classification capability evaluation matrix |
| CN104484678B (en) * | 2015-01-07 | 2017-06-16 | 青岛理工大学 | Multi-classifier fusion fault diagnosis method based on fault type classification capability evaluation matrix |
| CN104635146B (en) * | 2015-02-06 | 2017-04-19 | 南京农业大学 | Analog circuit fault diagnosis method based on random sinusoidal signal test and HMM (Hidden Markov Model) |
| CN104635146A (en) * | 2015-02-06 | 2015-05-20 | 南京农业大学 | Analog circuit fault diagnosis method based on random sinusoidal signal test and HMM (Hidden Markov Model) |
| CN107192939A (en) * | 2017-06-08 | 2017-09-22 | 京东方科技集团股份有限公司 | A kind of detection method and detecting system of the working condition of circuit |
| US11017700B2 (en) | 2017-06-08 | 2021-05-25 | Ordos Yuansheng Optoelectronics Co., Ltd. | Circuit working state testing method and testing device |
| CN108957297A (en) * | 2018-08-07 | 2018-12-07 | 北京航空航天大学 | A kind of circuit fault diagnosis extension Constructing Method for Decision introducing work information |
| CN108957297B (en) * | 2018-08-07 | 2020-05-19 | 北京航空航天大学 | Method for constructing circuit fault diagnosis expansion decision tree by introducing working condition information |
| CN112082639A (en) * | 2019-06-14 | 2020-12-15 | 现代自动车株式会社 | Engine state diagnosis method and diagnosis modeling method thereof |
| CN113159089A (en) * | 2021-01-18 | 2021-07-23 | 安徽建筑大学 | Pavement damage identification method, system, computer equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101871994B (en) | 2012-03-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101871994B (en) | Method for diagnosing faults of analog circuit of multi-fractional order information fusion | |
| Tightiz et al. | An intelligent system based on optimized ANFIS and association rules for power transformer fault diagnosis | |
| CN110059357B (en) | Intelligent ammeter fault classification detection method and system based on self-coding network | |
| CN111914883B (en) | Spindle bearing state evaluation method and device based on deep fusion network | |
| Yin et al. | Wasserstein Generative Adversarial Network and Convolutional Neural Network (WG‐CNN) for Bearing Fault Diagnosis | |
| CN103245907B (en) | A kind of analog-circuit fault diagnosis method | |
| CN107101813A (en) | A kind of frame-type circuit breaker mechanical breakdown degree assessment method based on vibration signal | |
| CN106250442A (en) | The feature selection approach of a kind of network security data and system | |
| CN108021945A (en) | A kind of transformer state evaluation model method for building up and device | |
| CN109389325B (en) | Method for evaluating state of electronic transformer of transformer substation based on wavelet neural network | |
| Liu et al. | Multi-feature fusion for fault diagnosis of rotating machinery based on convolutional neural network | |
| Su et al. | A multi-level adaptation scheme for hierarchical bearing fault diagnosis under variable working conditions | |
| CN105572572A (en) | WKNN-LSSVM-based analog circuit fault diagnosis method | |
| CN114237046B (en) | Partial discharge pattern recognition method based on SIFT data feature extraction algorithm and BP neural network model | |
| Wang et al. | Application of an oversampling method based on GMM and boundary optimization in imbalance-bearing fault diagnosis | |
| CN112596016A (en) | Transformer fault diagnosis method based on integration of multiple one-dimensional convolutional neural networks | |
| Rajesh et al. | Influence of data balancing on transformer DGA fault classification with machine learning algorithms | |
| Han et al. | Faulted-Phase classification for transmission lines using gradient similarity visualization and cross-domain adaption-based convolutional neural network | |
| CN114266289A (en) | Complex equipment health state assessment method | |
| CN115015683A (en) | Cable production performance test method, device, equipment and storage medium | |
| CN109726770A (en) | A kind of analog circuit fault testing and diagnosing method | |
| Maruf et al. | Locating faults in distribution systems in the presence of distributed generation using machine learning techniques | |
| Ge et al. | A deep condition feature learning approach for rotating machinery based on MMSDE and optimized SAEs | |
| CN113611568A (en) | Vacuum circuit breaker based on genetic convolution depth network | |
| CN109782158A (en) | A kind of Analog circuit diagnosis method based on multiclass classification |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20120321 Termination date: 20140611 |
|
| EXPY | Termination of patent right or utility model |