CN110308384B - 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 PDF

Info

Publication number
CN110308384B
CN110308384B CN201910622562.2A CN201910622562A CN110308384B CN 110308384 B CN110308384 B CN 110308384B CN 201910622562 A CN201910622562 A CN 201910622562A CN 110308384 B CN110308384 B CN 110308384B
Authority
CN
China
Prior art keywords
analog circuit
neural network
fault
voltage
fault diagnosis
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
Application number
CN201910622562.2A
Other languages
Chinese (zh)
Other versions
CN110308384A (en
Inventor
杨成林
周秀云
黄建国
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
University of Electronic Science and Technology of China
Original Assignee
University of Electronic Science and Technology of China
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by University of Electronic Science and Technology of China filed Critical University of Electronic Science and Technology of China
Priority to CN201910622562.2A priority Critical patent/CN110308384B/en
Publication of CN110308384A publication Critical patent/CN110308384A/en
Application granted granted Critical
Publication of CN110308384B publication Critical patent/CN110308384B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

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
    • G01R31/2851—Testing of integrated circuits [IC]

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Hardware Design (AREA)
  • Microelectronics & Electronic Packaging (AREA)
  • General Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Testing Electric Properties And Detecting Electric Faults (AREA)
  • Tests Of Electronic Circuits (AREA)

Abstract

The invention discloses a fault diagnosis method for an analog circuit based on a circle model and a neural network, which comprises the steps of firstly carrying out fuzzy group analysis on the analog circuit to obtain each fuzzy group information, then obtaining circle model parameters of different measuring points of each element in the analog circuit under different testing frequencies through circle model simulation, constructing to obtain a characteristic vector, training the constructed neural network as a training sample, obtaining the circle model parameters of each measuring point under each testing frequency according to degradation data during fault diagnosis, forming a testing characteristic vector, and inputting the testing characteristic vector into the trained neural network for fault diagnosis. The invention realizes the fault diagnosis of the analog circuit by combining the circular model parameters and the neural network, and can effectively improve the fault diagnosis rate of the analog circuit.

Description

Analog circuit fault diagnosis method based on circle model and neural network
Technical Field
The invention belongs to the technical field of analog circuit fault diagnosis, and particularly relates to an analog circuit fault diagnosis method based on a circle model and a neural network.
Background
With the rapid development of integrated circuits, digital and analog components are integrated on the same chip to improve product performance and reduce chip area and cost. It is reported that although the analog part only occupies 5% of the chip area, the failure diagnosis cost thereof occupies 95% of the total diagnosis cost, and the analog circuit failure diagnosis has been a bottleneck problem in the integrated circuit industry. At the present stage, some developed and relatively perfect analog circuit fault diagnosis theories are applied to practice, for example: the method comprises a fault dictionary method in the pre-test simulation diagnosis method and a component parameter identification method and a fault verification method in the post-test simulation diagnosis method. However, these methods can only deal with discrete parameter faults and hard faults, and cannot completely diagnose continuous parameter faults of analog components. The complex field circular model can completely model all parameter drift faults of the analog element, and is a practical fault diagnosis model. However, the specific characteristic values of the circular model under the influence of tolerance are infinite, and it is difficult to store all faults comprehensively by adopting a fault dictionary method.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provides a fault diagnosis method for an analog circuit based on a circular model and a neural network, which is used for acquiring circular model parameters of different measuring points of each element in the analog circuit under different testing frequencies, forming a feature vector corresponding to each element to train the neural network, and realizing fault diagnosis based on the neural network obtained by training so as to improve the fault diagnosis rate of the analog circuit.
In order to achieve the above object, the method for diagnosing the fault of the analog circuit based on the circular model and the neural network of the present invention comprises the following steps:
s1: recording the number of elements in the analog circuit as C, setting T measuring points as required, carrying out fuzzy group analysis on the C elements through the T measuring points to realize fault diagnosis, and recording the obtained fuzzy group number as M;
s2: setting F test frequencies according to the actual condition of the analog circuit, respectively carrying out N times of Monte Carlo simulation on each element at each test point under each test frequency, and obtaining a group of circular model parameters through each simulation
Figure GDA0002561994330000011
C1, 2, …, C, N1, 2, …, N, F1, 2, …, F, T1, 2, …, T, N
Figure GDA0002561994330000021
Wherein the corresponding circular model parameters are obtained in each simulation by adopting the following method:
s2.1: setting input voltage according to test frequency of the simulation
Figure GDA0002561994330000022
Frequency of (2) to input voltage
Figure GDA0002561994330000023
As an excitation source, the voltage sensor is used for carrying out fault-free simulation on the analog circuit to obtain the fault-free voltage of the current measuring point
Figure GDA0002561994330000024
S2.2: setting the parameter value of the element other than the element c to a value within the tolerance range, and setting the parameter value of the element c to pc1And pc2Respectively simulating to obtain the current measuring pointFault voltages, respectively
Figure GDA0002561994330000025
Calculating to obtain the output voltage of the element c acting alone
Figure GDA0002561994330000026
Figure GDA0002561994330000027
S2.3: if it is not
Figure GDA0002561994330000028
Then
Figure GDA0002561994330000029
And (3) making the circular model parameters w-1 and v-K, r-0, otherwise solving the following equations to obtain the circular model parameters w, v and r:
Figure GDA00025619943300000210
s3: constructing a neural network, wherein the number of neurons in an input layer is 3FT, and the number of neurons in an output layer is M;
s4: each feature vector obtained in the step S2
Figure GDA00025619943300000211
As an input of the neural network, taking the fuzzy group number M to which the element c belongs as an expected output, and training the neural network constructed in the step S3, wherein M is 1,2, …, M;
s5: when the analog circuit fails, the input voltage of each test frequency is used in the degradation process of the analog circuit performance
Figure GDA00025619943300000212
Inputting the voltage into an analog circuit twice, and measuring to obtain two fault voltages at each measuring point
Figure GDA00025619943300000213
Figure GDA00025619943300000214
Combined with fault-free voltage
Figure GDA00025619943300000215
Calculating to obtain the output voltage of the fault element acting alone
Figure GDA00025619943300000216
Calculating to obtain a circular model parameter when the analog circuit fails by adopting the method in the step S2.3; the center of a circle in the circle parameters obtained by measuring the point t at the f test frequency is recorded as
Figure GDA00025619943300000217
Radius of
Figure GDA00025619943300000218
Constructing test feature vectors
Figure GDA00025619943300000219
Test feature vector X*And inputting the data into the neural network trained in the step S4, wherein the classification result is the fault diagnosis result.
The invention relates to an analog circuit fault diagnosis method based on a circle model and a neural network, which comprises the steps of firstly carrying out fuzzy group analysis on an analog circuit to obtain each fuzzy group information, then obtaining circle model parameters of each element in the analog circuit at different test frequencies through circle model simulation, constructing to obtain a characteristic vector, training the constructed neural network as a training sample, obtaining the circle model parameters of each test point at each test frequency according to degradation data during fault diagnosis, and inputting the constructed neural network into the test characteristic vector to carry out fault diagnosis.
The invention realizes the fault diagnosis of the analog circuit by combining the circular model parameters and the neural network, and can effectively improve the fault diagnosis rate of the analog circuit.
Drawings
FIG. 1 is an analog circuit diagram;
FIG. 2 is an equivalent circuit diagram of the analog circuit shown in FIG. 1;
FIG. 3 is a schematic diagram of the voltage source operation of the analog circuit of FIG. 1;
FIG. 4 is a schematic diagram of the fault source effect of the analog circuit of FIG. 1;
FIG. 5 is a flow chart of an embodiment of the method for diagnosing faults of an analog circuit based on a circular model and a neural network;
FIG. 6 is a flow chart of a simulation acquisition method of the circular model of the present invention;
fig. 7 is a topology diagram of a second-order thomas analog filter circuit in the present embodiment.
Detailed Description
The following description of the embodiments of the present invention is provided in order to better understand the present invention for those skilled in the art with reference to the accompanying drawings. It is to be expressly noted that in the following description, a detailed description of known functions and designs will be omitted when it may obscure the subject matter of the present invention.
Examples
In order to better explain the technical content and the inventive point of the present invention, a theoretical derivation process of the present invention will be explained first.
Fig. 1 is an analog circuit diagram. As shown in FIG. 1, N represents a linear time invariant circuit, from an independent voltage source
Figure GDA0002561994330000031
And (4) exciting.
Figure GDA0002561994330000032
And the voltage phasor output at the selected measuring point is shown, and x is a passive element. According to an alternative theorem, the passive element x can be replaced by an independent voltage source with the same terminal voltage as it, resulting in an equivalent circuit. Fig. 2 is an equivalent circuit diagram of the analog circuit shown in fig. 1. According to the davinin theorem, any active linear time-invariant port network can be equivalently replaced externally by a series branch of a voltage source and an impedance, so that:
Figure GDA0002561994330000041
wherein the content of the first and second substances,
Figure GDA0002561994330000042
is the open circuit voltage phasor for ports a and b in fig. 2; z0Is the Davinin impedance value between a and b, ZxIs the resistance value of element x. According to the theorem of davinin,
Figure GDA0002561994330000043
and Z0Is independent of ZxAnd is determined only by the non-faulty element parameters and the location of the faulty element. Thus, in FIG. 2
Figure GDA0002561994330000044
As in FIG. 1
Figure GDA0002561994330000045
Are equal. In FIG. 2, the analog circuits N are composed of
Figure GDA0002561994330000046
And
Figure GDA0002561994330000047
and (4) jointly exciting. Voltages in FIG. 2 according to the principle of superposition
Figure GDA0002561994330000048
Is equal to
Figure GDA0002561994330000049
And
Figure GDA00025619943300000410
the algebraic sum of the output voltages when excited individually. FIG. 3 is a schematic diagram of the voltage source operation of the analog circuit of FIG. 1. Fig. 4 is a schematic diagram of the fault source operation of the analog circuit of fig. 1. As shown in fig. 3 and 4, the voltage source
Figure GDA00025619943300000411
And source of failure
Figure GDA00025619943300000412
When acting alone, the output voltages are respectively
Figure GDA00025619943300000413
And
Figure GDA00025619943300000414
it shows, as follows:
Figure GDA00025619943300000415
Figure GDA00025619943300000416
wherein H' (j ω) and H "(j ω) are transfer functions from the power port and the port of the element x to the output port, respectively, and are independent of the parameter value of the element x.
According to the principle of superposition, there are:
Figure GDA00025619943300000417
substituting formula (1) into formula (4) and eliminating
Figure GDA00025619943300000418
The impedance value Z of the output voltage to the fault source is obtained through simplificationxThe functional relationship of (a) is as follows:
Figure GDA00025619943300000419
from the above formula, thevenin equivalent impedance Z can be obtained0And ZxThe relationship of (a) to (b) is as follows:
Figure GDA00025619943300000420
wherein:
Figure GDA0002561994330000051
without loss of generality, each phasor is represented by a rectangular coordinate:
Figure GDA0002561994330000052
where j is an imaginary unit. Because of the fact that
Figure GDA0002561994330000053
H'(jω)、
Figure GDA0002561994330000054
H' (j omega) and Z0Are all independent of ZxSo that R0、X0α and β are also independent of Zx. Substituting equation (8) into equation (7) yields:
Figure GDA0002561994330000055
assuming that element x is a resistor, let Zx＝RxAnd according to the equation (9), the real part and the imaginary part of two sides are equal to obtain:
Figure GDA0002561994330000056
two equations in simultaneous (10) cancel RxObtaining the following formula:
Figure GDA0002561994330000057
the denominator in the formula (11) is eliminated, and the deduction is easy:
Figure GDA0002561994330000058
due to the fact that
Figure GDA0002561994330000059
Assuming thevenin equivalent voltageIs composed of
Figure GDA00025619943300000510
The power supply generates an output voltage of
Figure GDA00025619943300000511
The real and imaginary parts of the output voltage of the fault circuit can be expressed as follows:
Figure GDA00025619943300000512
substituting formula (13) for formula (12) yields the following formula:
Figure GDA0002561994330000061
equation (14) can be expressed as:
(Uor-w)2+(Uoj-v)2＝r2(15)
wherein the content of the first and second substances,
Figure GDA0002561994330000062
Figure GDA0002561994330000063
formula (15) represents Uor-UojThe circle center on the plane is (w, v) and the circle equation with the radius of r. Due to R0，X0α and β are independent of the value of x, so w and v are also independent of element x, i.e. whatever the value of the parameter of element x, equation (15) holds true, i.e. for each fault source the real and imaginary parts of the voltage produced at the same point under any fault source parameter satisfy the same circular equation.
Based on the theory, the invention provides an analog circuit fault diagnosis method based on a circular model and a neural network. FIG. 5 is a flow chart of an embodiment of the method for diagnosing the fault of the analog circuit based on the circular model and the neural network. As shown in fig. 5, the method for diagnosing the fault of the analog circuit based on the circular model and the neural network of the present invention specifically includes the steps of:
s501: acquiring analog circuit information:
and recording the number of elements in the analog circuit as C, setting T measuring points as required, carrying out fuzzy group analysis on the C elements through the T measuring points to realize fault diagnosis, and recording the obtained fuzzy group number as M. The fuzzy analysis is a common technical means for diagnosing the fault of the analog circuit, and the detailed process thereof is not repeated herein.
S502: and (3) simulating to obtain a feature vector:
under the condition of selecting a testing point, the testability of the analog circuit is fixed (fuzzy group is fixed), the testability cannot be improved due to different testing frequencies, but the tolerance influence can be effectively reduced due to the increase of the number of the testing frequencies. The invention sets F test frequencies according to the actual situation of the analog circuit, respectively carries out N times of Monte Carlo simulation on each representative fault element at each test point under each test frequency, and each simulation obtains a group of circular model parameters
Figure GDA0002561994330000071
C1, 2, …, C, N1, 2, …, N, F1, 2, …, F, T1, 2, …, T, N
Figure GDA0002561994330000072
FIG. 6 is a flowchart of a simulation obtaining method of the circle model in the present invention. As shown in fig. 6, the simulation obtaining method of the circular model in the present invention is as follows:
s601: fault-free simulation:
setting input voltage according to test frequency of the simulation
Figure GDA0002561994330000073
Frequency of (2) to input voltage
Figure GDA0002561994330000074
As an excitation source, the voltage sensor is used for carrying out fault-free simulation on the analog circuit to obtain the fault-free voltage of the current measuring point
Figure GDA0002561994330000075
S602: fault simulation:
setting the parameter value of the element other than the element c to a value within the tolerance range, and setting the parameter value of the element c to pc1And pc2Respectively simulating to obtain the fault voltage of the current measuring point, and respectively recording the fault voltage as
Figure GDA0002561994330000076
Calculating to obtain the output voltage of the element c acting alone
Figure GDA0002561994330000077
Figure GDA0002561994330000078
Parameter pc1And pc2Is set according to actual conditions, and generally sets pc1＜pc0，pc2＞pc0，pc0Representing the nominal value of the parameter of element c. To facilitate the operation in step S602, p may be substitutedc1Set as the minimum value p of the parameter of the element ccmin，pc2Set as the maximum value p of the parameter of the element ccmax。
S603: calculating circular model parameters
If it is not
Figure GDA0002561994330000079
Then
Figure GDA00025619943300000710
And (3) making the circular model parameters w-1 and v-K, r-0, otherwise solving the following equations to obtain the circular model parameters w, v and r:
Figure GDA00025619943300000711
s503: constructing a neural network:
and (3) constructing a neural network, wherein the number of input layer neurons is 3FT, and the number of output layer neurons is M. In the embodiment, the neural network adopts a double hidden layer neural network.
S504: training a neural network:
each feature vector obtained in step S502
Figure GDA0002561994330000081
The fuzzy group number M to which the element c belongs is used as an expected output as an input of the neural network, and the neural network constructed in step S503 is trained with M being 1,2, …, M.
Because the neural network in this embodiment adopts a double hidden layer neural network, during training, the two hidden layer internal star weights are trained by adopting an automatic coding method, the output layer internal star weight is trained by adopting a mentor training mode, and the external star weight training process adopts a layer-by-layer training mode and is divided into four steps: a first layer hidden layer outer star weight training stage, a second layer hidden layer outer star weight training stage, an output layer outer star weight training stage and an integral fine tuning stage. The double hidden layer neural network is a more common neural network, and the detailed training process thereof is not described herein.
S505: fault diagnosis:
when the analog circuit fails, the input voltage of each test frequency is used in the degradation process of the analog circuit performance
Figure GDA0002561994330000082
Inputting the voltage into an analog circuit twice, and measuring to obtain two fault voltages at each measuring point
Figure GDA0002561994330000083
Combined with fault-free voltage
Figure GDA0002561994330000084
Calculating to obtain the output voltage of the fault element acting alone
Figure GDA0002561994330000085
And (4) calculating to obtain the circular model parameters when the analog circuit fails by adopting the method in the step (S603). The center of a circle in the circle parameters obtained by measuring the point t at the f test frequency is recorded as
Figure GDA0002561994330000086
Radius of
Figure GDA0002561994330000087
Constructing test feature vectors
Figure GDA0002561994330000088
Test feature vector X*And inputting the result into the neural network trained in the step S504, wherein the classification result is the fault diagnosis result.
In order to better illustrate the implementation process and technical effects of the present invention, the present invention is illustrated by taking a second-order thomas analog filter circuit as an example.
Fig. 7 is a topology diagram of a second-order thomas analog filter circuit in the present embodiment. As shown in fig. 7, the second-order thomas analog filter circuit of the present embodiment includes 8 fault elements, denoted by VoutAs a measure point, the fuzzy set is { R1}、{R2}、{R3,C1}、{R4,R5,R6,C2And the faults of the internal elements of the fuzzy groups cannot be distinguished, and the faults between the fuzzy groups can be theoretically distinguished.
In this embodiment, 3 test frequencies are set: 500Hz, 1000Hz and 1500Hz, and the input voltage signal adopts a sine signal. And setting the values of other element parameter values within a tolerance range, and respectively performing 200 Monte Carlo simulations on each element at each measuring point under each test frequency, thereby constructing 200 feature vectors for each element. The first 100 feature vectors of each element are used as training samples, and the last 100 feature vectors of each element are used as test samples, that is, the training samples and the test samples respectively have 800 feature vectors.
In this embodiment, the neural network is a dual hidden layer neural network, and since 3 test frequencies are set in this embodiment,using only VoutAs the measuring points, the number of the measuring points is 1, the number of the neurons of the input layer of the double hidden layer neural network is 9. In this embodiment, there are 4 fuzzy groups, and the number of output layer neurons of the bilingual neural network is 4. The number of neurons in both crypt layers is 5. Training the double-hidden-layer neural network by adopting a training sample, and inputting a test sample into the double-hidden-layer neural network for fault diagnosis. Table 1 is a failure diagnosis result statistical table in the present embodiment.
Figure GDA0002561994330000091
TABLE 1
As can be seen from Table 1, when any element has a fault, the fault diagnosis rate obtained by the method can reach 96.8% at the lowest, and the average diagnosis rate reaches 99.18%. Therefore, the method can effectively realize the fault diagnosis of the analog circuit and has higher diagnosis rate.
Although illustrative embodiments of the present invention have been described above to facilitate the 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, and various changes may be made apparent to those skilled in the art as long as they are within the spirit and scope of the present invention as defined and defined by the appended claims, and all matters of the invention which utilize the inventive concepts are protected.

Claims (3)

1. A fault diagnosis method for an analog circuit based on a circle model and a neural network is characterized by comprising the following steps:
s1: recording the number of elements in the analog circuit as C, setting T measuring points as required, carrying out fuzzy group analysis on the C elements through the T measuring points to realize fault diagnosis, and recording the obtained fuzzy group number as M;
s2: setting F test frequencies according to the actual situation of the analog circuit, respectively carrying out N times of Monte Carlo simulation on each element at each test point under each test frequency, and obtaining a group of circles through each simulationModel parameters
Figure FDA0002561994320000011
C1, 2, …, C, N1, 2, …, N, F1, 2, …, F, T1, 2, …, T, N
Figure FDA0002561994320000012
Wherein the corresponding circular model parameters are obtained in each simulation by adopting the following method:
s2.1: setting input voltage according to test frequency of the simulation
Figure FDA0002561994320000013
Frequency of (2) to input voltage
Figure FDA0002561994320000014
As an excitation source, the voltage sensor is used for carrying out fault-free simulation on the analog circuit to obtain the fault-free voltage of the current measuring point
Figure FDA0002561994320000015
Wherein j is an imaginary unit;
s2.2: setting the parameter value of the element other than the element c to a value within the tolerance range, and setting the parameter value of the element c to pc1And pc2Respectively simulating to obtain the fault voltage of the current measuring point, and respectively recording the fault voltage as
Figure FDA0002561994320000016
Calculating to obtain the output voltage of the element c acting alone
Figure FDA0002561994320000017
Figure FDA0002561994320000018
S2.3: if it is not
Figure FDA0002561994320000019
Then
Figure FDA00025619943200000110
And (3) making the circular model parameters w-1 and v-K, r-0, otherwise solving the following equations to obtain the circular model parameters w, v and r:
Figure FDA00025619943200000111
s3: constructing a neural network, wherein the number of neurons in an input layer is 3FT, and the number of neurons in an output layer is M;
s4: each feature vector obtained in the step S2
Figure FDA00025619943200000112
As an input of the neural network, taking the fuzzy group number M to which the element c belongs as an expected output, and training the neural network constructed in the step S3, wherein M is 1,2, …, M;
s5: when the analog circuit fails, the input voltage of each test frequency is used in the degradation process of the analog circuit performance
Figure FDA0002561994320000021
Inputting the voltage into an analog circuit twice, and measuring to obtain two fault voltages at each measuring point
Figure FDA0002561994320000022
Figure FDA0002561994320000023
Combined with fault-free voltage
Figure FDA0002561994320000024
Calculating to obtain the output voltage of the fault element acting alone
Figure FDA0002561994320000025
Calculating to obtain a circular model parameter when the analog circuit fails by adopting the method in the step S2.3; recording the f test frequencyThe center of the circle in the circle parameters obtained by the lower measuring point t is
Figure FDA0002561994320000026
Radius of
Figure FDA0002561994320000027
Constructing test feature vectors
Figure FDA0002561994320000028
Test feature vector X*And inputting the data into the neural network trained in the step S4, wherein the classification result is the fault diagnosis result.
2. Method according to claim 1, characterized in that in step S2.2 p is selectedc1Set as the minimum value p of the parameter of the element ccmin，pc2Set as the maximum value p of the parameter of the element ccmax。
3. The analog circuit fault diagnosis method according to claim 1, wherein the neural network in step S3 employs a double hidden layer neural network.
CN201910622562.2A 2019-07-11 2019-07-11 Analog circuit fault diagnosis method based on circle model and neural network Active CN110308384B (en)

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 CN110308384A (en) 2019-10-08
CN110308384B true 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)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111474905B (en) * 2020-04-15 2021-07-06 哈尔滨工业大学 Parameter drift fault diagnosis method in manufacturing process of electromechanical product
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
CN114236365B (en) * 2021-12-21 2022-09-02 电子科技大学 SAR ADC circuit test optimization method based on circle model
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

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4914601A (en) * 1988-02-25 1990-04-03 Hewlett-Packard Company Method for profiling wafers and for locating dies thereon
CN104198922B (en) * 2014-08-15 2017-02-01 电子科技大学 Frequency selection method in early fault diagnosis of analog circuit
CN109948267B (en) * 2019-03-26 2022-06-14 电子科技大学 Linear analog circuit fault diagnosis method based on circular model parameters

Also Published As

Publication number Publication date
CN110308384A (en) 2019-10-08

Similar Documents

Publication Publication Date Title
CN110308384B (en) Analog circuit fault diagnosis method based on circle model and neural network
CN109581203B (en) Post-test simulation fault diagnosis method based on genetic algorithm
DE112008001173B4 (en) Method for determining relevance values for detection of an error on a chip, method for determining an error probability of a location on a chip and computer program
DE102007060417B4 (en) Raw chip and wafer defect classification system and method
CN111260063A (en) Analog circuit fault positioning and parameter identification method based on genetic algorithm
CN110907810B (en) Analog circuit single fault diagnosis method based on particle swarm algorithm
CN110470979B (en) Analog circuit fault diagnosis method based on fault characteristic region
CN110673017B (en) Analog circuit fault element parameter identification method based on genetic algorithm
CN112485652A (en) Analog circuit single fault diagnosis method based on improved sine and cosine algorithm
CN109948267B (en) Linear analog circuit fault diagnosis method based on circular model parameters
EP0992807B1 (en) Method and apparatus for board model correction
Grzechca Construction of an expert system based on fuzzy logic for diagnosis of analog electronic circuits
CN108828437B (en) Analog circuit fault feature extraction method based on cloud correlation coefficient matrix
CN111308327B (en) Analog circuit fault location and fault element parameter identification method
CN112505532A (en) Analog circuit single fault diagnosis method based on improved particle swarm optimization
CN113189513A (en) Ripple-based redundant power supply current sharing state identification method
Stošović et al. Oscillation-based analog diagnosis using artificial neural networks based inference mechanism
CN110308386B (en) Analog circuit fault diagnosis method based on polynomial fitting and state monitoring
CN108845247A (en) A kind of analog module method for diagnosing faults
US3102231A (en) White noise fault detection system
CN108664936A (en) A kind of diagnostic method and system based on mechanical disorder
CN108490344B (en) Analog circuit fault diagnosis method based on training sample optimization
CN109581204A (en) A kind of built-in test design method and system
CN111880073A (en) Built-in test design method and system
Zilch et al. A versatile test set generation tool for structural analog circuit testing

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