CN110308384A - Analog-circuit fault diagnosis method based on circle model and neural network - Google Patents
Analog-circuit fault diagnosis method based on circle model and neural network Download PDFInfo
- Publication number
- CN110308384A CN110308384A CN201910622562.2A CN201910622562A CN110308384A CN 110308384 A CN110308384 A CN 110308384A CN 201910622562 A CN201910622562 A CN 201910622562A CN 110308384 A CN110308384 A CN 110308384A
- Authority
- CN
- China
- Prior art keywords
- neural network
- analog circuit
- analog
- fault diagnosis
- measuring point
- 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
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
- G01R31/2851—Testing of integrated circuits [IC]
Abstract
The invention discloses a kind of analog-circuit fault diagnosis methods based on circle model and neural network, fuzzy group analysis is carried out to analog circuit first, obtain each fuzzy group information, then the circle model parameter of each element different measuring points under different test frequencies in analog circuit is obtained by circle model emulation, building obtains feature vector, constructed neural network is trained as training sample, the circle model parameter of each measuring point under each test frequency is obtained according to degraded data in fault diagnosis, it constitutes testing feature vector and inputs trained neural network progress fault diagnosis.The present invention, which passes through, combines circle model parameter and neural fusion analog circuit fault diagnosing, can effectively improve analog circuit fault diagnosing rate.
Description
Technical field
The invention belongs to Analog Circuit Fault Diagnosis Technology fields, more specifically, be related to it is a kind of based on circle model and
The analog-circuit fault diagnosis method of neural network.
Background technique
With the fast development of integrated circuit, in order to enhance product performance, reduce chip area and expense, need to by number and
Analog element is integrated on same chip.According to document announcement, although analog portion only accounts for the 5% of chip area, its failure
But Zhan always diagnoses the 95% of cost to diagnosis cost, and analog circuit fault diagnosing is always one " bottleneck " in integrated circuit industry
Problem.There is the fairly perfect analog circuit fault diagnosing theory of some development to be applied in practice at this stage, such as:
The component parameters identification method and failure proof method in fault dictionary method Simulation after test diagnosis in Simulation before test diagnosis.But this
A little methods are only capable of handling discrete parametic fault and hard fault, are unable to the continuous parameter failure of complete diagnosis analog element.Plural number
Domain circle model is capable of all parameter drift failures of complete modeling Simulation element, is a kind of more practical fault diagnosis model.
But the circle model under Effect of Tolerance also specific features value be also it is infinite more, using fault dictionary method be difficult to store comprehensively it is all therefore
Barrier.
Summary of the invention
It is an object of the invention to overcome the deficiencies of the prior art and provide a kind of simulations based on circle model and neural network
Circuit failure diagnosis method obtains the circle model parameter of each element different measuring points under different test frequencies in analog circuit, structure
Neural network is trained at each element corresponding feature vector, based on the obtained neural fusion fault diagnosis of training,
To improve analog circuit fault diagnosing rate.
For achieving the above object, the present invention is based on the analog-circuit fault diagnosis method packets of circle model and neural network
Include following steps:
S1: being denoted as C for number of elements in analog circuit, T measuring point be set as needed, and survey through T to C element
Point realizes the fuzzy group analysis of fault diagnosis, and obtained ambiguity group quantity is denoted as M；
S2: F test frequency is set according to the actual conditions of analog circuit, is existed under each test frequency for each element
Each measuring point carries out n times Monte Carlo simulation respectively, and emulation obtains one group of round model parameter every time
C=1,2 ..., C, n=1,2 ..., N, f=1,2 ..., F, t=1,2 ..., T constitute N number of feature vector of each elementWherein every time in emulation using following
Method obtains corresponding round model parameter:
S2.1: input voltage is arranged according to the test frequency of this emulationFrequency, by input voltageAs excitation
Source carries out fault-free emulation to analog circuit, obtains the faultless voltage of current measuring point
S2.2: the value in range of tolerable variance is set with the parameter value of external component by element c, sets p for the parameter value of element cc1
And pc2It is emulated respectively, obtains the false voltage of current measuring point, be denoted as respectivelyElement c independent role is calculated
Output voltage
S2.3: ifThenCircle model parameter w=1, v=-K, r=0 are enabled, is otherwise asked
It solves following equation group and obtains round model parameter w, v, r:
S3: building neural network, wherein input layer quantity is 3FT, and output layer neuron quantity is M；
S4: each feature vector that step S2 is obtainedAs the input of neural network, ambiguity group belonging to element c
Serial number m is trained the neural network of step S3 building as desired output, m=1,2 ..., M；
S5: when analog circuit breaks down, during analog circuit performance degradation, by the input electricity of each test frequency
PressureAnalog circuit is inputted in two times, and measurement obtains two false voltages under each measuring point In conjunction with fault-free electricity
PressureThe output voltage of fault element independent role is calculatedUsing the method meter in step S2.3
Calculate circle model parameter when obtaining analog circuit fault；Remember that the center of circle is in the obtained Circle Parameters of measuring point t under f-th of test frequencyRadius isConstitute testing feature vector
By testing feature vector X*It is input in the trained neural network of step S4, classification results are fault diagnosis result.
The present invention is based on the analog-circuit fault diagnosis methods of circle model and neural network, carry out mould to analog circuit first
Group analysis is pasted, each fuzzy group information is obtained, then by each element in circle model emulation acquisition analog circuit in different tests
The circle model parameter of different measuring points under frequency, building obtain feature vector, as training sample to constructed neural network into
Row training, obtains the circle model parameter of each measuring point under each test frequency according to degraded data in fault diagnosis, constitutes and surveys
It tries feature vector and inputs trained neural network progress fault diagnosis.
The present invention, which passes through, combines circle model parameter and neural fusion analog circuit fault diagnosing, can effectively improve mould
Quasi- circuit fault diagnosis rate.
Detailed description of the invention
Fig. 1 is analog circuit figure；
Fig. 2 is the equivalent circuit diagram of analog circuit shown in Fig. 1；
Fig. 3 is the voltage source effect schematic diagram of analog circuit shown in Fig. 1；
Fig. 4 is the source of trouble effect schematic diagram of analog circuit shown in Fig. 1；
Fig. 5 is the specific embodiment stream of the analog-circuit fault diagnosis method the present invention is based on circle model and neural network
Cheng Tu；
Fig. 6 is the emulation acquisition methods flow chart of circle model in the present invention；
Fig. 7 is the topological diagram of second order Thomas analogue filter circuit in the present embodiment.
Specific embodiment
A specific embodiment of the invention is described with reference to the accompanying drawing, preferably so as to those skilled in the art
Understand the present invention.Requiring particular attention is that in the following description, when known function and the detailed description of design perhaps
When can desalinate main contents of the invention, these descriptions will be ignored herein.
Embodiment
Technology contents and inventive point in order to better illustrate the present invention first carry out theory deduction process of the invention
Explanation.
Fig. 1 is analog circuit figure.As shown in Figure 1, N indicates a linearly invariant circuit, by independent voltage sourceSwash
It encourages.Indicate the voltage phasor exported on selected measuring point, x is passive element.According to Substitution Theoren, passive element x can be replaced
It is changed to independent voltage source identical with its end voltage, obtains equivalent circuit.Fig. 2 is the equivalent circuit diagram of analog circuit shown in Fig. 1.
According to Thevenin's theorem, when any one active linear, the string of a voltage source and an impedance was externally can be used in constant port network
Join branch equally to be replaced, then obtain:
Wherein,It is a and b port open voltage phasor in Fig. 2；Z0It is the Dai Weining impedance value between a and b, ZxFor member
The impedance value of part x.According to Thevenin's theorem,And Z0Value independently of Zx, and only by fault-free component parameters and fault element
Position determine.Then, in Fig. 2In Fig. 1It is equal.In Fig. 2, analog circuit N byWithCommon activation.
Voltage according to principle of stacking, in Fig. 2It is equal toWithThe algebraical sum of output voltage when independent drive.Fig. 3 is shown in Fig. 1
The voltage source of analog circuit acts on schematic diagram.Fig. 4 is the source of trouble effect schematic diagram of analog circuit shown in Fig. 1.Such as Fig. 3 and Fig. 4
It is shown, voltage sourceAnd the source of troubleWhen independent role, output voltage is used respectivelyWithIt indicates, has:
Wherein, H'(j ω) and H " (j ω) be respectively port where power port and element x to the transmitting letter of output port
Number, and it is unrelated with the parameter value of element x.
According to principle of stacking, have:
Formula (1) is substituted into formula (4), is eliminatedOutput voltage is obtained to source of trouble impedance value Z by abbreviationxLetter
Number relationship is as follows:
From the available Dai Weining equivalent impedance Z of above formula0With ZxRelationship it is as follows:
Wherein:
Without loss of generality, each phasor is indicated with rectangular co-ordinate:
Wherein, j is imaginary unit.BecauseH'(jω)、H " (j ω) and Z0It is independently of each Zx, so R0、X0、α
Z is also independent from βx.(8) formula substitution (7) formula is obtained:
It is assumed that element x is resistance, Z is rememberedx=Rx, it is equal according to formula (9) both sides real and imaginary parts, it obtains:
Two equations in simultaneous (10) disappear Rx, it obtains such as following formula formula:
Denominator in (11) formula that disappears, is not difficult to release:
Due toIt is assumed that Dai Weining equivalent voltage isPower supply generates
Output voltage beThe output voltage real and imaginary parts of faulty circuit can indicate as follows:
Formula (13) are substituted into formula (12), obtain following formula:
Formula (14) can indicate are as follows:
(Uor-w)2+(Uoj-v)2=r2 (15)
Wherein,
Formula (15) indicates Uor-UojThe center of circle is the equation of a circle that (w, v) radius is r in plane.Due to R0, X0, α and β are independently of x
Value, therefore w and v are also independent from element x.I.e. no matter what value is the parameter of element x take, and formula (15) is always set up, i.e., for each
The real and imaginary parts of the source of trouble, the voltage generated under any source of trouble parameter in same observation station are all satisfied the same equation of a circle.
Therefore, equation of a circle (15) is the fault model that can be applied to soft fault and hard fault simultaneously.And it is unrelated with test method.More than
Conclusion assumes that (element x) is what resistance obtained to the source of trouble, if the source of trouble is capacitor or inductance, can be derived by identical
Conclusion.
Based on the above theory, the invention proposes a kind of analog circuit fault diagnosing sides based on circle model and neural network
Method.Fig. 5 is the specific embodiment flow chart of the analog-circuit fault diagnosis method the present invention is based on circle model and neural network.
As shown in figure 5, the present invention is based on the specific steps of analog-circuit fault diagnosis method of circle model and neural network to include:
S501: analog circuit information is obtained:
Number of elements in analog circuit is denoted as C, T measuring point is set as needed, C element is carried out through T measuring point
The fuzzy group analysis for realizing fault diagnosis, is denoted as M for obtained ambiguity group quantity.Fuzzy group analysis is analog circuit fault diagnosing
Common technology means, details are not described herein for detailed process.
S502: emulation obtains feature vector:
In selected measuring point, the measurability of analog circuit is fixed (ambiguity group is fixed), will not be because of test frequency
The difference of rate and improve testability, but Effect of Tolerance can be effectively reduced in the increase of test frequency quantity.The present invention is according to mould
F test frequency is arranged in the actual conditions of quasi- circuit, for each representing fault element in each survey under each test frequency
Point carries out n times Monte Carlo simulation respectively, and emulation obtains one group of round model parameter every timeC=1,
2 ..., C, n=1,2 ..., N, f=1,2 ..., F, t=1,2 ..., T constitute N number of feature vector of each element
Fig. 6 is the emulation acquisition methods flow chart of circle model in the present invention.As shown in fig. 6, justifying the imitative of model in the present invention
True acquisition methods are as follows:
S601: fault-free emulation:
According to the test frequency of this emulation, input voltage is setFrequency, by input voltageAs driving source pair
Analog circuit carries out fault-free emulation, obtains the faultless voltage of current measuring point
S602: fault simulation:
The value in range of tolerable variance is set with the parameter value of external component by element c, sets p for the parameter value of element cc1With
pc2It is emulated respectively, obtains the false voltage of current measuring point, be denoted as respectivelyElement c independent role is calculated
Output voltage
Parameter pc1And pc2It is arranged according to the actual situation, generally setting pc1< pc0, pc2> pc0, pc0Indicate element c ginseng
Several nominal values.It, can be by p for the ease of the operation in step S602c1It is set as the minimum value p of element c parametercmin, pc2If
It is set to the maximum value p of element c parametercmax。
S603: circle model parameter is calculated
IfThenCircle model parameter w=1, v=-K, r=0 are enabled, is otherwise solved as follows
Equation group obtains round model parameter w, v, r:
S503: building neural network:
Neural network is constructed, wherein input layer quantity is 3FT, and output layer neuron quantity is M.In the present embodiment
Neural network uses two hidden-layer neural network.
S504: neural metwork training:
Each feature vector that step S502 is obtainedAs the input of neural network, ambiguity group sequence belonging to element c
Number m is trained the neural network of step S503 building as desired output, m=1,2 ..., M.
By neural network in this present embodiment using two hidden-layer neural network, in training, star power in two hidden layers
It is trained using automatic coding, star power using there is the progress of tutor's training method, adopt by extraterrestrial power training process in output layer
With layer-by-layer training method, be divided into four steps: the extraterrestrial power training of first layer hidden layer, the extraterrestrial power of second layer hidden layer are trained, and output layer is extraterrestrial
The power trained and whole fine tuning stage.Two hidden-layer neural network is a kind of more common neural network, and detailed training process exists
This is repeated no more.
S505: fault diagnosis:
When analog circuit breaks down, during analog circuit performance degradation, by the input voltage of each test frequencyAnalog circuit is inputted in two times, and measurement obtains two false voltages under each measuring pointIn conjunction with faultless voltageThe output voltage of fault element independent role is calculatedUsing the method meter in step S603
Calculate circle model parameter when obtaining analog circuit fault.Remember that the center of circle is in the obtained Circle Parameters of measuring point t under f-th of test frequencyRadius isConstitute testing feature vector
By testing feature vector X*It is input in the trained neural network of step S504, classification results are fault diagnosis result.
Implementation process and technical effect in order to better illustrate the present invention, by taking second order Thomas's analogue filter circuit as an example
The present invention will be described.
Fig. 7 is the topological diagram of second order Thomas analogue filter circuit in the present embodiment.As shown in fig. 7, the two of the present embodiment
Rank Thomas's analogue filter circuit includes 8 fault elements, with VoutAs measuring point, ambiguity group is { R1}、{R2}、{R3,C1}、
{R4,R5,R6,C2, the failure undistinguishable of ambiguity group internal element can be distinguished on the failure theory between ambiguity group.
3 test frequencies: 500Hz, 1000Hz and 1500Hz are set in the present embodiment, and input voltage signal is using sinusoidal letter
Number.Other elements parameter value value in range of tolerable variance is set, for each element in each measuring point under each test frequency
200 Monte Carlo simulations are carried out respectively, to construct to obtain 200 feature vectors to each element.Before each element
100 feature vectors are as training sample, using rear 100 feature vectors of each element as test sample, i.e. training sample
Respectively there are 800 feature vectors with test sample.
It is two hidden-layer neural network that neural network is arranged in the present embodiment, by being provided with 3 test frequencies in this present embodiment
Rate, only with VoutAs measuring point, i.e., measuring point quantity is 1, then the input layer quantity of two hidden-layer neural network is 9.This
There are 4 ambiguity groups in embodiment, then the output layer neuron quantity of two hidden-layer neural network is 4.Two hidden layers are set
Neuronal quantity is 5.Two hidden-layer neural network is trained using training sample, test sample is then inputted into two hidden-layer
Neural network carries out fault diagnosis.Table 1 is the fault diagnosis result statistical form in the present embodiment.
Table 1
As it can be seen from table 1 when any element failure, using the obtained fault diagnosis rate of the present invention is minimum can also
Reach 96.8%, average diagnosis reaches 99.18%.As it can be seen that the present invention can effectively realize the fault diagnosis of analog circuit, and
With higher diagnosis.
Although the illustrative specific embodiment of the present invention is described above, in order to the technology of the art
Personnel understand the present invention, it should be apparent that the present invention is not limited to the range of specific embodiment, to the common skill of the art
For art personnel, if various change the attached claims limit and determine the spirit and scope of the present invention in, these
Variation is it will be apparent that all utilize the innovation and creation of present inventive concept in the column of protection.
Claims (3)
1. a kind of analog-circuit fault diagnosis method based on circle model and neural network, which comprises the following steps:
S1: being denoted as C for number of elements in analog circuit, T measuring point be set as needed, and carries out to C element real through T measuring point
The fuzzy group analysis of existing fault diagnosis, is denoted as M for obtained ambiguity group quantity；
S2: F test frequency is set according to the actual conditions of analog circuit, for each element every under each test frequency
A measuring point carries out n times Monte Carlo simulation respectively, and emulation obtains one group of round model parameter every timeStructure
At N number of feature vector of each element
Wherein corresponding round model parameter is obtained using following methods in emulation every time:
S2.1: input voltage is arranged according to the test frequency of this emulationFrequency, by input voltageAs driving source pair
Analog circuit carries out fault-free emulation, obtains the faultless voltage of current measuring point
S2.2: the value in range of tolerable variance is set with the parameter value of external component by element c, sets p for the parameter value of element cc1
And pc2It is emulated respectively, obtains the false voltage of current measuring point, be denoted as respectivelyElement c is calculated individually to make
Output voltage
S2.3: ifThenCircle model parameter w=1, v=-K, r=0 are enabled, is otherwise solved as follows
Equation group obtains round model parameter w, v, r:
S3: building neural network, wherein input layer quantity is 3FT, and output layer neuron quantity is M；
S4: each feature vector that step S2 is obtainedAs the input of neural network, ambiguity group serial number m belonging to element c
As desired output, m=1,2 ..., M are trained the neural network of step S3 building；
S5: when analog circuit breaks down, during analog circuit performance degradation, by the input voltage of each test frequency
Analog circuit is inputted in two times, and measurement obtains two false voltages under each measuring point In conjunction with faultless voltageThe output voltage of fault element independent role is calculatedIt is calculated using the method in step S2.3
Circle model parameter when to analog circuit fault；Remember that the center of circle is in the obtained Circle Parameters of measuring point t under f-th of test frequencyRadius isConstitute testing feature vector
By testing feature vector X*It is input in the trained neural network of step S4, classification results are fault diagnosis result.
2. analog-circuit fault diagnosis method according to claim 1, which is characterized in that p in the step S2.2c1Setting
For the minimum value p of element c parametercmin, pc2It is set as the maximum value p of element c parametercmax。
3. analog-circuit fault diagnosis method according to claim 1, which is characterized in that neural network in the step S3
Using two hidden-layer neural network.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910622562.2A CN110308384B (en) | 2019-07-11 | 2019-07-11 | Analog circuit fault diagnosis method based on circle model and neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910622562.2A CN110308384B (en) | 2019-07-11 | 2019-07-11 | Analog circuit fault diagnosis method based on circle model and neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110308384A true CN110308384A (en) | 2019-10-08 |
| CN110308384B CN110308384B (en) | 2020-09-18 |
Family
ID=68081123
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910622562.2A Active CN110308384B (en) | 2019-07-11 | 2019-07-11 | Analog circuit fault diagnosis method based on circle model and neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110308384B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111474905A (en) * | 2020-04-15 | 2020-07-31 | 哈尔滨工业大学 | Parameter drift fault diagnosis method in manufacturing process of electromechanical product |
| CN113687995A (en) * | 2021-10-27 | 2021-11-23 | 成都嘉纳海威科技有限责任公司 | Chip screening method based on neural network |
| CN114236365A (en) * | 2021-12-21 | 2022-03-25 | 电子科技大学 | SAR ADC circuit test optimization method based on circle model |
| CN114355173A (en) * | 2022-01-04 | 2022-04-15 | 电子科技大学 | Analog filter circuit fault diagnosis method based on multi-input residual error network |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH01253247A (en) * | 1988-02-25 | 1989-10-09 | Hewlett Packard Co <Hp> | Method and apparatus for positioning die |
| CN104198922A (en) * | 2014-08-15 | 2014-12-10 | 电子科技大学 | Frequency selection method in early fault diagnosis of analog circuit |
| CN109948267A (en) * | 2019-03-26 | 2019-06-28 | 电子科技大学 | Diagnosis for Linear Analog Circuit method based on circle model parameter |
-
2019
- 2019-07-11 CN CN201910622562.2A patent/CN110308384B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH01253247A (en) * | 1988-02-25 | 1989-10-09 | Hewlett Packard Co <Hp> | Method and apparatus for positioning die |
| CN104198922A (en) * | 2014-08-15 | 2014-12-10 | 电子科技大学 | Frequency selection method in early fault diagnosis of analog circuit |
| CN109948267A (en) * | 2019-03-26 | 2019-06-28 | 电子科技大学 | Diagnosis for Linear Analog Circuit method based on circle model parameter |
Non-Patent Citations (2)
| Title |
|---|
| SHULIN TIAN等: "Circle Equation-Based Fault Modeling Method for Linear Analog Circuits", 《IEEE》 * |
| 胡鸿志等: "面向三维空间的模拟电路故障建模与故障检测", 《仪表技术与传感器》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111474905A (en) * | 2020-04-15 | 2020-07-31 | 哈尔滨工业大学 | Parameter drift fault diagnosis method in manufacturing process of electromechanical product |
| CN113687995A (en) * | 2021-10-27 | 2021-11-23 | 成都嘉纳海威科技有限责任公司 | Chip screening method based on neural network |
| CN113687995B (en) * | 2021-10-27 | 2022-01-04 | 成都嘉纳海威科技有限责任公司 | Chip screening method based on neural network |
| CN114236365A (en) * | 2021-12-21 | 2022-03-25 | 电子科技大学 | SAR ADC circuit test optimization method based on circle model |
| CN114236365B (en) * | 2021-12-21 | 2022-09-02 | 电子科技大学 | SAR ADC circuit test optimization method based on circle model |
| CN114355173A (en) * | 2022-01-04 | 2022-04-15 | 电子科技大学 | Analog filter circuit fault diagnosis method based on multi-input residual error network |
| CN114355173B (en) * | 2022-01-04 | 2023-05-30 | 电子科技大学 | Analog filter circuit fault diagnosis method based on multi-input residual error network |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110308384B (en) | 2020-09-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110308384A (en) | Analog-circuit fault diagnosis method based on circle model and neural network | |
| TWI395959B (en) | Method and apparatus for determining relevance values for a detection of a fault on a chip and for determining a fault probability of a location on a chip | |
| CN102721941B (en) | Method for fusing and diagnosing fault information of circuit of electric meter on basis of SOM (self-organized mapping) and D-S (Dempster-Shafer) theories | |
| CN110470980B (en) | Analog circuit fault characteristic range determining method based on genetic algorithm | |
| CN108732528A (en) | A kind of digitalized electrical energy meter method for diagnosing faults based on depth confidence network | |
| CN110470979B (en) | Analog circuit fault diagnosis method based on fault characteristic region | |
| CN102636742A (en) | Large-scale analogue circuit fault diagnosis method based on wavelet neural network | |
| CN104237770B (en) | A kind of analog-circuit fault diagnosis method | |
| CN110907810B (en) | Analog circuit single fault diagnosis method based on particle swarm algorithm | |
| CN111260063A (en) | Analog circuit fault positioning and parameter identification method based on genetic algorithm | |
| CN107576904B (en) | Analog-circuit fault diagnosis method based on sawtooth wave | |
| CN101251574B (en) | Method for recognition of failure diagnosis parameter of real time electric circuit | |
| CN109948267A (en) | Diagnosis for Linear Analog Circuit method based on circle model parameter | |
| CN111308327B (en) | Analog circuit fault location and fault element parameter identification method | |
| CN104090228A (en) | Analog circuit fuzzy group identification method | |
| CN113189513A (en) | Ripple-based redundant power supply current sharing state identification method | |
| CN108845247A (en) | A kind of analog module method for diagnosing faults | |
| CN106226637A (en) | A kind of Short method of testing | |
| US3102231A (en) | White noise fault detection system | |
| CN103605063B (en) | The port error value feedback system of test multiport device and method | |
| CN109814026A (en) | A kind of built-in test design method and system | |
| CN109782157A (en) | A kind of built-in test design method and system | |
| CN110308386B (en) | Analog circuit fault diagnosis method based on polynomial fitting and state monitoring | |
| CN105158679B (en) | Filter circuit method for diagnosing faults based on transfer function | |
| US5550951A (en) | Metrics for specifying and/or testing neural networks |
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 |