CN114355173B - Analog filter circuit fault diagnosis method based on multi-input residual error network - Google Patents
Analog filter circuit fault diagnosis method based on multi-input residual error network Download PDFInfo
- Publication number
- CN114355173B CN114355173B CN202210004437.7A CN202210004437A CN114355173B CN 114355173 B CN114355173 B CN 114355173B CN 202210004437 A CN202210004437 A CN 202210004437A CN 114355173 B CN114355173 B CN 114355173B
- Authority
- CN
- China
- Prior art keywords
- input
- filter circuit
- analog filter
- layer
- training
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/28—Testing of electronic circuits, e.g. by signal tracer
- G01R31/316—Testing of analog circuits
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
- Y04S10/52—Outage or fault management, e.g. fault detection or location
Landscapes
- Engineering & Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Testing Electric Properties And Detecting Electric Faults (AREA)
- Measurement Of Resistance Or Impedance (AREA)
Abstract
The invention discloses a fault diagnosis method of an analog filter circuit based on a multi-input residual error network, which comprises the steps of firstly setting relevant parameters of Monte Carlo analysis, and carrying out simulation analysis on the analog filter circuit by using a Monte Carlo statistical analysis method to obtain voltage signals of the output end of the analog filter circuit under different fault states; the modal components of the voltage signals are obtained through empirical wavelet transformation, the modal components are respectively formed into a training set and a testing set, the training set is used for training the multi-input ResNet, the testing set is used for testing, and a multi-input ResNet model and a parameter Parameters corresponding to the highest testing precision are stored; and finally diagnosing unknown fault states of the analog filter circuit through the trained multi-input ResNet model and Parameters.
Description
Technical Field
The invention belongs to the technical field of analog circuit fault diagnosis, and particularly relates to an analog filter circuit fault diagnosis method based on a multi-input residual error network.
Background
With the rapid development of the electronic circuit industry, the integration level of the electronic circuit is continuously improved, the functional and modularized requirements are more and more obvious, and the requirements on the operation reliability of the electronic circuit system are higher and higher. It has been shown statistically that nearly 80% of the circuits are digital circuits, but nearly 80% of the circuit faults are generated by analog circuits. Analog filter circuits are often used as filter modules in analog parts of electronic circuits because of their good performance. If the analog filter circuit malfunctions during use, the performance of the entire analog circuit is inevitably affected, and unavoidable losses are caused, so that it is very necessary to diagnose the faults of the analog filter circuit in time.
In recent years, a large number of analog filter circuit fault diagnosis methods based on combination of feature extraction and classifier are sequentially proposed, but the methods need to extract, select or fuse features, and are more complicated compared with a neural network algorithm for adaptively extracting features. However, the traditional neural network-based fault diagnosis method for the analog filter circuit only processes the original signal, and the relation between the filtering performance and the original signal component during circuit faults is not fully considered, so that the improvement of the fault diagnosis accuracy of the analog filter circuit is limited.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provide a fault diagnosis method of an analog filter circuit based on a multi-input Residual error network.
In order to achieve the above object, the invention provides a fault diagnosis method for an analog filter circuit based on a multi-input residual error network, which is characterized by comprising the following steps:
(1) Simulation analysis is carried out on the analog filter circuit to be tested by using a Monte Carlo analysis method, and voltage signals v with the number of sampling points of k multiplied by n groups of circuit output ends of k fault states being M are obtained ij (t), i=1, 2, …, k, j=1, 2, …, n, n represents the number of groups of voltage signals collected at the output of the analog filter circuit to be tested;
(2) Using conventional empirical wavelet transform algorithm for the voltage signal v ij (t) decomposing to obtain m modal components, wherein the q-th modal component is represented as ewt ijq (t), q=1, 2, …, m, and then combining the m modal components into a feature vector ewt ij ＝[ewt ij1 ；ewt ij2 ；…；ewt ijq (t)；…；ewt ijm ]；
(3) N ewt under k fault states ij According to r 1 :r 2 Is randomly allocated to form training data respectivelyCollection set
Validating a data set
(4) Will ewt ij The corresponding serial number i of the fault state is used as a label, the labels are respectively loaded for a training data set and a verification data set, and a training set of the multi-input residual error network model is constructed
And a verification set
(5) Constructing a multi-input residual error network, training a multi-input residual error network model by using the train_set, and verifying the train_set trained model by using the validation_set to obtain an optimal multi-input residual error network model max ；
(5.1) setting the training round of the multi-input residual error network model as N, and setting the input sample size to be batch_size during each round of training; the complete input of the train_set to the multi-input residual network requires t iterations, each iteration being denoted train_set to the multi-input residual network p ，p＝1…t，
(5.2)、train_set p Input to the multi-input residual network, the train_set is predicted by the Softmax function p Probability pro of the predictive label of the q-th sample in the list belonging to the loading label i qi Q=1 … batch_size, calculate train_set in this iteration p Training loss TL from training a multi-input residual network mp ，m＝1…N；
Wherein y is qi Equal to 0 or 1, when y qi When equal to 1, represents train_set p The q-th sample in (1) is loaded with a label of i, when y qi When equal to 0, represents train_set p The label loaded by the q-th sample in the (b) is not i;
(5.3) use of L Pt Performing back propagation update on the multi-input residual error network;
(5.4) repeating the steps (5.2) - (5.3) t times until all samples in the train_set are input into the multi-input residual error network, thereby completing the training of the round, and then saving the training loss after the completion of the training of the round
And a multi-input residual network model m ；
(5.5) inputting the validization_set to the multi-input residual error network model saved in this round m Predicting label value, and calculating verification accuracy Vacc m ；
Wherein e r Equal to 1 or 0, if the r-th sample in the calculation_set passes through the model m The label obtained by prediction is the same as the label loaded by the sample in the conjugation_set, and the value is 1, otherwise, the value is 0;
(5.6) repeating the steps (5.2) - (5.5) for N times to obtain N multi-input residual error network models and verification accuracy, wherein the model= [ model ] 1 …model m …model N ],Vacc＝[Vacc 1 …Vacc m …Vacc N ]；
(5.7) index using the maximum value in Vacc max Obtaining an optimal model after N rounds of training max ＝model[index max ]And is used as a fault prediction model of the analog filter circuit to be tested;
(6) And (3) obtaining a group of eigenvectors of the analog filter circuit to be tested in an unknown fault state according to the methods in the steps (1) - (2), and inputting the eigenvectors into a fault prediction model so as to obtain a fault state i of the analog filter circuit to be tested.
The invention aims at realizing the following steps:
according to the fault diagnosis method of the analog filter circuit based on the multi-input residual error network, relevant parameters of Monte Carlo analysis are set, the analog filter circuit is subjected to simulation analysis through a Monte Carlo statistical analysis method, and voltage signals of the output end of the analog filter circuit under different fault states are obtained; the modal components of the voltage signals are obtained through empirical wavelet transformation, the modal components are respectively formed into a training set and a testing set, the training set is used for training the multi-input ResNet, the testing set is used for testing, and a multi-input ResNet model and a parameter Parameters corresponding to the highest testing precision are stored; and finally diagnosing unknown fault states of the analog filter circuit through the trained multi-input ResNet model and Parameters.
The fault diagnosis method of the analog filter circuit based on the multi-input residual error network has the following beneficial effects:
when the elements in the analog filter circuit fail, the filter performance of the circuit is affected, the characteristics of each modal component of the voltage signal of the output end can be fully learned through the multi-input ResNet, and the characteristics extraction, selection and fusion methods are not needed, so that the complexity of fault diagnosis of the analog filter circuit is reduced, and the accuracy of fault diagnosis is improved.
Drawings
FIG. 1 is a flow chart of a fault diagnosis method of an analog filter circuit based on a multi-input residual error network;
FIG. 2 is a schematic diagram of a simulation of a second-order high-pass filter circuit for a four op-amp in an embodiment of the invention;
FIG. 3 is a voltage signal at the output of the four-op second-order high-pass filter circuit in a normal state;
FIG. 4 is a modal component of the output voltage signal of the four operational amplifier second-order high-pass filter circuit obtained by empirical wavelet transform decomposition in a normal state;
FIG. 5 is a specific structure of a multiple input ResNet of the present invention;
FIG. 6 is a graph of training loss and validation accuracy as a function of training round number for a multiple input ResNet in accordance with the present invention;
fig. 7 is a graph of training loss, validation accuracy as a function of training round number for a conventional single input res net.
Detailed Description
The following description of the embodiments of the invention is presented in conjunction with the accompanying drawings to provide a better understanding of the invention to those skilled in the art. It is to be expressly noted that in the description below, detailed descriptions of known functions and designs are omitted here as perhaps obscuring the present invention.
Examples
FIG. 1 is a flow chart of a fault diagnosis method of an analog filter circuit based on a multi-input residual error network.
In this embodiment, as shown in fig. 1, the fault diagnosis method of the analog filter circuit based on the multi-input residual error network of the present invention includes the following steps:
s1, performing simulation analysis on an analog filter circuit to be tested by using a Monte Carlo analysis method to obtain voltage signals v with the number of sampling points of k multiplied by n groups at the output end of the circuit under k fault states ij (t), i=1, 2, …, k, j=1, 2, …, n, n represents the number of groups of voltage signals collected at the output of the analog filter circuit to be tested;
the specific method comprises the following steps:
s1.1, setting tolerance of resistance and capacitance in an analog filter circuit to be tested and fault parameters in an ith fault state;
s1.2 setting the start time of Monte Carlo analysis
End time->
Time step TSTEP is +.>
The running times are n, f represents the to-be-detectedThe frequency of the signal input by the input end of the analog filter circuit;
s1.3, under k fault states, taking the voltage signal at the output end of the analog filter circuit to be tested as an output variable analyzed by a Monte Carlo statistical method, thereby obtaining a voltage signal v with k multiplied by n groups of sampling points being M ij (t) wherein
In the embodiment, a Monte Carlo statistical method is utilized to carry out simulation analysis on the four operational amplifier second-order high-pass filter circuit, and a fault data set is obtained.
The simulation schematic diagram of the four-operational-amplifier second-order high-pass filter circuit is shown in fig. 2, square wave signals with the frequency of 2 KHZ, the amplitude of 5V and the duty ratio of 50% are input from the input end of the circuit, and the resistance and capacitance values in the four-operational-amplifier second-order high-pass filter circuit in a normal state are shown in fig. 3; through sensitivity analysis, the four operational amplifier second-order high-pass filter circuit mainly has 12 fault modes, namely R1 ∈, R2 ∈, R3 ∈, R4 ∈, R4 ∈, C1 ∈, C2 ∈, are denoted by F1 to F12, F0 indicates that the circuit is in a normal state, and Table 1 is a specific fault setting in this embodiment. The tolerance of the resistor and the capacitor in the four operational amplifier second-order high-pass filter circuit is set to be 5%, the starting time in Monte Carlo statistical analysis is set to be 0.005s, the ending time is set to be 0.01s, and the time steps are set to be 6.25X10 -6 s, the operation times are 400, and voltage signals v with the sampling points of 800 under 13 different fault states consisting of F1 to F12 and F0 are respectively obtained ij (t)，i＝1,…,13,j＝1,2,…,400；
| Fault numbering | Fault type | Nominal value of | Fault value |
| F0 | - | - | - |
| F1 | R1↓ | 6.2kΩ | 4.34kΩ |
| F2 | R1↑ | 6.2kΩ | 8.06kΩ |
| F3 | R2↓ | 6.2kΩ | 4.34kΩ |
| F4 | R2↑ | 6.2kΩ | 8.06kΩ |
| F5 | R3↓ | 6.2kΩ | 4.34kΩ |
| F6 | R3↑ | 6.2kΩ | 8.06kΩ |
| F7 | R4↓ | 1.6kΩ | 1.12kΩ |
| F8 | R4↑ | 1.6kΩ | 2.08kΩ |
| F9 | C1↓ | 5nF | 3.5kΩ |
| F10 | C1↑ | 5nF | 6.5kΩ |
| F11 | C2↓ | 5nF | 3.5kΩ |
| F12 | C2↑ | 5nF | 6.5kΩ |
TABLE 1
S2, utilizing traditional empirical wavelet transformation to convert voltage signal v ij (t) decomposing to obtain 6 modal components ewt ijq (t), q=1, 2, …,6, respectively, will group 400 of mode components ewt ijq (t) composition feature vector ewt ij Traditional empirical wavelet transformThe modal components of the voltage signal at the output end of the four-operational-amplifier second-order high-pass filter circuit obtained through decomposition are shown in figure 4;
s3, ewt under 13 fault states ij Respectively forming training data sets according to the proportion of 3:1
Verification data set +.>
S4, will ewt ij The corresponding serial number i of the fault state is used as a label, the labels are respectively loaded for a training data set and a verification data set, and a training set of the multi-input residual error network model is constructed
And a verification set
S5, constructing a multi-input residual error network;
as shown in fig. 5, the multi-input residual network model includes an input layer, 6 residual blocks net-block with the same structure, a full connection layer and an output layer;
each net-block comprises a 1-layer convolution unit, a 4-layer residual unit with the same structure and a 1-layer average pooling layer;
the convolution unit firstly passes through a one-dimensional convolution layer with a kernel size of 5, an output channel of 16 and a step length of 2, then passes through a batch regularization layer and a Relu layer, and finally passes through a one-dimensional average pooling layer with a kernel size of 3 and a step length of 2;
the residual error unit passes through the core with the size of 3 and the input channel of channel l ＝2 3+l L=1, 2,3,4, and the output channel is channel l A one-dimensional convolution layer with the step length of 2 passes through a regularization layer and a Relu layer; then the core size is 3, and the input and output channels are all channels l A one-dimensional convolution layer with the step length of 1 passes through a regularization layer and a Relu layer, wherein l represents the layer number corresponding to the residual error unit;
the output of the 6 net-blocks are connected in series and connected to the output layer through the full connection layer;
s6, training a multi-input residual error network;
training a multi-input residual error network model by using the train_set, and verifying the train_set training model by using the validation_set to obtain the optimal multi-input residual error network model max ；
S6.1, setting training rounds of a multi-input residual error network model as N, and setting the input sample size as batch_size during each round of training; the complete input of the train_set to the multi-input residual network requires t iterations, each iteration being denoted train_set to the multi-input residual network p ，p＝1…t，
S6.2、train_set p Input to a multi-input residual network, predictive of train_set by softmax function p Probability pro of predictive label of the q-th sample in (i) belonging to loaded label i qi Q=1 … batch_size, calculate train_set in this iteration p Training loss TL from training a multi-input residual network mp ，m＝1…N；
Wherein y is qi Equal to 0 or 1, when y qi When equal to 1, represents train_set p The q-th sample in (1) is loaded with a label of i, when y qi When equal to 0, represents train_set p The label loaded by the q-th sample in the (b) is not i;
s6.3, utilize L Pt Performing back propagation update on the multi-input residual error network;
s6.4, repeating the steps S6.2-S6.3 for t times until all samples in the train_set are input into the multi-input residual error network, thereby completing the training of the round, and then saving the training loss after the training of the round is completed
And a multi-input residual network model m ；
S6.5, inputting the validization_set into the multi-input residual error network model saved in the present round m Predicting label value, and calculating verification accuracy Vacc m ；
Wherein e r Equal to 1 or 0, if the r-th sample in the calculation_set passes through the model m The label obtained by prediction is the same as the label loaded by the sample in the conjugation_set, and the value is 1, otherwise, the value is 0;
s6.6, repeating the steps S6.2-S6.5 for N rounds in total to obtain N multi-input residual error network models and verification accuracy, wherein the model= [ model ] 1 …model m …model N ],Vacc＝[Vacc 1 …Vacc m …Vacc N ]；
S6.7, utilizing index where maximum value in Vacc is located max Obtaining an optimal model after N rounds of training max ＝model[index max ]And is used as a fault prediction model of the analog filter circuit to be tested;
setting the training round number of the multi-input residual error network to 100, setting the sample size input during each round of training to 128, and completely inputting the train_set into the multi-input residual error network to perform 31 iterations, wherein each iteration inputs the train_set into the multi-input residual error network p P=1 …; the training loss TL and verification accuracy Vacc curves obtained by using the train_set and the verification_set based on the multi-input residual network in this embodiment along with the number of training rounds are shown in fig. 6, the training loss STL and the verification accuracy SVacc curves based on the conventional single-input residual network along with the number of training rounds are shown in fig. 7, and the maximum verification accuracy obtained by respectively performing 100 rounds of verification on the two residual network models is shown in table 2.
| Network type | Multi-input residual error network | Single input residual error network |
| Maximum verification accuracy | 99.46％ | 86.923％ |
TABLE 2
As can be seen from fig. 6, fig. 7 and table 2, compared with the conventional single-input residual network, the multi-input residual network has better convergence effect, higher verification precision and stronger generalization capability, can learn the fault characteristics of the analog filter circuit in different fault states more fully, and is beneficial to improving the efficiency and accuracy of fault diagnosis;
s7, obtaining a group of eigenvectors of the analog filter circuit to be tested in an unknown fault state according to the method in the steps S1-S2, and inputting the eigenvectors into a fault prediction model so as to obtain a fault state i of the analog filter circuit to be tested.
While the foregoing describes illustrative embodiments of the present invention to facilitate an understanding of the present invention by those skilled in the art, it should be understood that the present invention is not limited to the scope of the embodiments, but is to be construed as protected by the accompanying claims insofar as various changes are within the spirit and scope of the present invention as defined and defined by the appended claims.
Claims (3)
1. The fault diagnosis method of the analog filter circuit based on the multi-input residual error network is characterized by comprising the following steps of:
(1) Simulation analysis is carried out on the analog filter circuit to be tested by using a Monte Carlo analysis method, and voltage signals v with the number of sampling points of k multiplied by n groups of circuit output ends of k fault states being M are obtained ij (t), i=1, 2, …, k, j=1, 2, …, n, n represents the number of groups of voltage signals collected at the output of the analog filter circuit to be tested;
(2) Using conventional empirical wavelet transform algorithm for the voltage signal v ij (t) decomposing to obtain m modal components, wherein the q-th modal component is represented as ewt ijq (t), q=1, 2, …, m, and then combining the m modal components into a feature vector ewt ij ＝[ewt ij1 ；ewt ij2 ；…；ewt ijq (t)；…；ewt ijm ]；
(3) N ewt under k fault states ij According to r 1 :r 2 Is randomly allocated to form training data sets respectively
Validating a data set
(4) Will ewt ij The corresponding serial number i of the fault state is used as a label, the labels are respectively loaded for a training data set and a verification data set, and a training set of the multi-input residual error network model is constructed
And a verification set
(5) Constructing a multi-input residual error network, training a multi-input residual error network model by using the train_set, and verifying the train_set trained model by using the validation_set to obtain an optimal multi-input residual error network model max ；
(5.1) setting training round of the multi-input residual error network model as N, and inputting sample size during each round of trainingSet to batch_size; the complete input of the train_set to the multi-input residual network requires t iterations, each iteration being denoted train_set to the multi-input residual network p ，p＝1…t，
(5.2)、train_set p Input to the multi-input residual network, the train_set is predicted by the Softmax function p Probability pro of the predictive label of the q-th sample in the list belonging to the loading label i qi Q=1 … batch_size, calculate train_set in this iteration p Training loss TL from training a multi-input residual network mp ，m＝1…N；
Wherein y is qi Equal to 0 or 1, when y qi When equal to 1, represents train_set p The q-th sample in (1) is loaded with a label of i, when y qi When equal to 0, represents train_set p The label loaded by the q-th sample in the (b) is not i;
(5.3) use of L Pt Performing back propagation update on the multi-input residual error network;
(5.4) repeating the steps (5.2) - (5.3) t times until all samples in the train_set are input into the multi-input residual error network, thereby completing the training of the round, and then saving the training loss after the completion of the training of the round
And a multi-input residual network model m ；
(5.5) inputting the validization_set to the multi-input residual error network model saved in this round m Predicting label value, and calculating verification accuracy Vacc m ；
Wherein, the liquid crystal display device comprises a liquid crystal display device,e r equal to 1 or 0, if the r-th sample in the calculation_set passes through the model m The label obtained by prediction is the same as the label loaded by the sample in the conjugation_set, and the value is 1, otherwise, the value is 0;
(5.6) repeating the steps (5.2) - (5.5) for N times to obtain N multi-input residual error network models and verification accuracy, wherein the model= [ model ] 1 …model m …model N ],Vacc＝[Vacc 1 …Vacc m …Vacc N ]；
(5.7) index using the maximum value in Vacc max Obtaining an optimal model after N rounds of training max ＝model[index max ]And is used as a fault prediction model of the analog filter circuit to be tested;
(6) And (3) obtaining a group of eigenvectors of the analog filter circuit to be tested in an unknown fault state according to the methods in the steps (1) - (2), and inputting the eigenvectors into a fault prediction model so as to obtain a fault state i of the analog filter circuit to be tested.
2. The fault diagnosis method for analog filter circuit based on multi-input residual network as claimed in claim 1, wherein in said step (1), voltage signals v with k×n groups of sampling points being M at circuit output ends under k fault conditions are obtained ij The specific method of (t) is as follows:
(2.1) setting tolerance of resistance and capacitance in the analog filter circuit to be tested and fault parameters in the ith fault state;
(2.2) setting the start time of Monte Carlo analysis
End time->
Time step->
The running times are n;
(2.3) under k fault conditions, filtering the simulation to be testedThe voltage signal at the output end of the wave circuit is used as an output variable analyzed by a Monte Carlo statistical method, so that a k multiplied by n group voltage signal v with the sampling point number of M is obtained ij (t) wherein
3. The method for diagnosing faults of an analog filter circuit based on a multi-input residual network according to claim 1, wherein the multi-input residual network model comprises an input layer, m residual blocks with the same structure, a full connection layer and an output layer;
the outputs of m residual blocks are connected in series, and each residual block comprises a 1-layer convolution unit, a 4-layer residual unit with the same structure and a 1-layer average pooling layer;
the convolution unit passes through a one-dimensional convolution layer with a kernel size of 5, an output channel of 16 and a step length of 2, then passes through a batch regularization layer and a Relu layer, and finally passes through a one-dimensional average pooling layer with a kernel size of 3 and a step length of 2;
the residual error unit passes through the core with the size of 3 and the input channel of channel l ＝2 3+l L=1, 2,3,4, and the output channel is channel l A one-dimensional convolution layer with the step length of 2 passes through a regularization layer and a Relu layer; then the core size is 3, and the input and output channels are all channels l And a one-dimensional convolution layer with the step length of 1 passes through a regularization layer and a Relu layer, wherein l represents the number of layers corresponding to the residual unit layer.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210004437.7A CN114355173B (en) | 2022-01-04 | 2022-01-04 | Analog filter circuit fault diagnosis method based on multi-input residual error network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210004437.7A CN114355173B (en) | 2022-01-04 | 2022-01-04 | Analog filter circuit fault diagnosis method based on multi-input residual error network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114355173A CN114355173A (en) | 2022-04-15 |
| CN114355173B true CN114355173B (en) | 2023-05-30 |
Family
ID=81107229
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210004437.7A Active CN114355173B (en) | 2022-01-04 | 2022-01-04 | Analog filter circuit fault diagnosis method based on multi-input residual error network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114355173B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114325352B (en) * | 2022-01-04 | 2023-04-18 | 电子科技大学 | Analog filter circuit fault diagnosis method based on empirical wavelet transform |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6636823B1 (en) * | 1999-09-30 | 2003-10-21 | Rockwell Automation Technologies, Inc. | Method and apparatus for motor fault diagnosis |
| JP2015099047A (en) * | 2013-11-18 | 2015-05-28 | 日立オートモティブシステムズ株式会社 | Fault diagnosis rate calculation device, fault diagnosis rate calculation method, and electronic circuit device |
| CN110308384A (en) * | 2019-07-11 | 2019-10-08 | 电子科技大学 | Analog-circuit fault diagnosis method based on circle model and neural network |
| EP3651366A1 (en) * | 2018-11-09 | 2020-05-13 | STMicroelectronics Srl | Analog-to-digital converter circuit, corresponding device and method |
| CN112748368A (en) * | 2020-10-28 | 2021-05-04 | 上海交通大学 | Three-level inverter IGBT open-circuit fault diagnosis method |
| CN113111618A (en) * | 2021-03-09 | 2021-07-13 | 电子科技大学 | Analog circuit fault diagnosis method based on improved empirical wavelet transform |
| CN113409284A (en) * | 2021-06-28 | 2021-09-17 | 北京百度网讯科技有限公司 | Circuit board fault detection method, device, equipment and storage medium |
| CN115186564A (en) * | 2022-05-12 | 2022-10-14 | 山东大学 | Analog circuit fault diagnosis method based on feature fusion and improved particle swarm algorithm |
-
2022
- 2022-01-04 CN CN202210004437.7A patent/CN114355173B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6636823B1 (en) * | 1999-09-30 | 2003-10-21 | Rockwell Automation Technologies, Inc. | Method and apparatus for motor fault diagnosis |
| JP2015099047A (en) * | 2013-11-18 | 2015-05-28 | 日立オートモティブシステムズ株式会社 | Fault diagnosis rate calculation device, fault diagnosis rate calculation method, and electronic circuit device |
| EP3651366A1 (en) * | 2018-11-09 | 2020-05-13 | STMicroelectronics Srl | Analog-to-digital converter circuit, corresponding device and method |
| CN110308384A (en) * | 2019-07-11 | 2019-10-08 | 电子科技大学 | Analog-circuit fault diagnosis method based on circle model and neural network |
| CN112748368A (en) * | 2020-10-28 | 2021-05-04 | 上海交通大学 | Three-level inverter IGBT open-circuit fault diagnosis method |
| CN113111618A (en) * | 2021-03-09 | 2021-07-13 | 电子科技大学 | Analog circuit fault diagnosis method based on improved empirical wavelet transform |
| CN113409284A (en) * | 2021-06-28 | 2021-09-17 | 北京百度网讯科技有限公司 | Circuit board fault detection method, device, equipment and storage medium |
| CN115186564A (en) * | 2022-05-12 | 2022-10-14 | 山东大学 | Analog circuit fault diagnosis method based on feature fusion and improved particle swarm algorithm |
Non-Patent Citations (3)
| Title |
|---|
| LIpeng ji.Soft Fault Diagnosis of Analog Circuits Based on a ResNet With Circuit Spectrum Map.《IEEE Transactions on Circuits and Systems I: Regular Papers》.2021,第68卷(第7期),全文. * |
| 傅晨琦.基于深度残差网络的模拟电路软故障诊断方法.《飞控与探测》.2021,第4卷(第4期),全文. * |
| 刘雪梅.基于经验小波变换和残差网络的模拟电路故障诊断方法研究.刘雪梅.2023,(第1期),全文. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114355173A (en) | 2022-04-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112685961B (en) | Method and system for predicting remaining service life of analog circuit | |
| CN109635928B (en) | Voltage sag reason identification method based on deep learning model fusion | |
| CN112149373B (en) | Complex analog circuit fault identification and estimation method and system | |
| CN107894564A (en) | A kind of analog-circuit fault diagnosis method based on intersection wavelet character | |
| CN111985155B (en) | Circuit health state prediction method and system based on integrated deep neural network | |
| CN112051480A (en) | Neural network power distribution network fault diagnosis method and system based on variational modal decomposition | |
| CN114355173B (en) | Analog filter circuit fault diagnosis method based on multi-input residual error network | |
| CN113591215B (en) | Abnormal satellite component layout detection method based on uncertainty | |
| CN115081316A (en) | DC/DC converter fault diagnosis method and system based on improved sparrow search algorithm | |
| CN112485652B (en) | Analog circuit single fault diagnosis method based on improved sine and cosine algorithm | |
| CN113533945A (en) | Analog circuit fault diagnosis method based on two-dimensional convolutional neural network | |
| CN110110426A (en) | A kind of Switching Power Supply filter capacitor abatement detecting method | |
| CN113156303A (en) | Analog circuit fault diagnosis and positioning method based on output response matrix characteristic analysis | |
| CN111061708A (en) | Electric energy prediction and restoration method based on LSTM neural network | |
| CN115018133A (en) | Lithium battery capacity prediction method based on chaos theory and nonlinear system | |
| CN115902557A (en) | Switch cabinet fault diagnosis processing method and device and nonvolatile storage medium | |
| CN111935762B (en) | Distribution network fault diagnosis method and system based on EWT and CNN under 5G load-bearing network | |
| CN110309586B (en) | MA-DBN-based GIS decomposed gas content prediction method | |
| CN114781577A (en) | Buck circuit fault diagnosis method based on VMD-DCNN-SVM | |
| CN113343590A (en) | Wind speed prediction method and system based on combined model | |
| CN112651183A (en) | Reliability evaluation method for quantum distributed countermeasure unified deep hash network | |
| Mosin | A technique of analog circuits testing and diagnosis based on neuromorphic classifier | |
| CN114236365B (en) | SAR ADC circuit test optimization method based on circle model | |
| Mosin | An approach to construction the neuromorphic classifier for analog fault testing and diagnosis | |
| CN117909779A (en) | Analog circuit soft fault diagnosis method based on Koopman operator pure data driving |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |