CN109828168A - Converter method for diagnosing faults based on Density Estimator - Google Patents
Converter method for diagnosing faults based on Density Estimator Download PDFInfo
- Publication number
- CN109828168A CN109828168A CN201910097620.4A CN201910097620A CN109828168A CN 109828168 A CN109828168 A CN 109828168A CN 201910097620 A CN201910097620 A CN 201910097620A CN 109828168 A CN109828168 A CN 109828168A
- Authority
- CN
- China
- Prior art keywords
- fault
- sample
- training
- data
- density estimator
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Landscapes
- Test And Diagnosis Of Digital Computers (AREA)
Abstract
The present invention relates to the converter method for diagnosing faults unanimously based on Density Estimator.It is pre-processed by the analysis of cubic B-Spline interpolation based on mallat algorithm to collecting data, obtains the sample with fault signature;By KDE fault grader, the more excellent parameter of fault grader is chosen after off-line training, the normal condition for including in training sample and all types of fault conditions is accurately divided, and more excellent parameter is used for classifier network, obtains optimized parameter;By in the classifier network implantation in-circuit emulation with optimized parameter, the fault diagnosis of real time on-line monitoring actual circuit is done;The classifier network for having chosen optimized parameter can distinguish known fault type sample and normal sample and complete the fault location of known fault type, moreover it is possible in the case where UNKNOWN TYPE failure occurs, can identify that unknown failure realizes circuit protection.More acurrate, the more reliable health status for judging converter of present invention energy, also improves the efficiency of converter fault diagnosis.
Description
Technical field
The present invention relates to power electronics field, especially a kind of converter fault diagnosis side based on Density Estimator
Method.
Background technique
With the development of power electronics technology, device of the converters as power electronics AC-DC conversion, it is complete
The high efficiency conversion and control of pairs of electric energy.Since converters play indispensable angle in electric field
Color is particularly important so can converters operate normally.Conventional electric power electronic converter fault diagnosis side
Method has artificial neural network, support vector machines, fault dictionary method etc..
Artificial Neural Network is established input feature vector due to being interconnected using artificial neuron and exports result
Mapping relations, by the continuous amendment of neuron and its corresponding construction (weight, deviation), when each back transfer updates weight,
Network ownership value requires to update, and convergence rate is slow, and in training process, has when learning new samples and forgets old sample
Trend, be unfavorable for the exact classification to failure.Support vector machines is computationally relatively easy, but due to vulnerable to sampled signal
Influence of noise can cause the erroneous judgement to output result.The strong antijamming capability of fault dictionary, but the fault sample needed for it is big,
It can be only achieved good effect.In the case that method based on Density Estimator can be realized small sample, to known fault and just
Reason condition is distinguished and is positioned, and can be distinguished to unknown failure and normal condition.
Summary of the invention
The purpose of the present invention is to provide a kind of converter method for diagnosing faults based on Density Estimator, can it is more acurrate,
The more reliable health status for judging converter, and can identify unknown failure, also improve the accurate of converter accident analysis
Rate.
To achieve the above object, the technical scheme is that a kind of converter fault diagnosis based on Density Estimator
Method, comprising the following steps:
Step S1, by data acquisition and noise reduction process, the sample with fault data is obtained;
Step S2, time domain in fault data, frequency domain failure spy are extracted using the cubic B-Spline interpolation analysis based on mallat
Sign, and by Data Dimensionality Reduction, gained fault signature and fault type one-to-one correspondence are established data sample library by secondary noise reduction process,
For training, testing characteristics of network;
Step S3, the fault grader based on Density Estimator algorithm is constructed, will include in training sample after off-line training
Normal condition accurately divided with all types of failures, extract the more excellent parameter of fault grader, and more excellent parameter is directly assigned
Classifier network carries out classifier test job, selects optimized parameter by test;
Step S4, the classifier network for being endowed optimized parameter is implanted into simulink, does the failure of actual circuit
Converter circuit quick self-checking is realized in diagnosis.
In an embodiment of the present invention, the specific implementation steps are as follows by the step S1:
Step S11, measurement point electric signal is acquired using capture card PCI-6229;
Step S12, sampling electric signal removes noise by simulink module, obtains raw sample data, is had
The sample of fault data
In an embodiment of the present invention, the specific implementation steps are as follows by the step S3:
Step S31, the data sample library that step S2 is obtained is divided into training sample and test sample;
Step S32, the input using training sample as the fault grader based on Density Estimator algorithm carries out it
Training;
Step S33, training sample is trained using different kernel functions, passes through the good core of training result using effect
Function;
Step S34, whether training of judgement error meets default, if then entering step S35, otherwise changes bandwidth h size
It is for further adjustments;
Step S35, more excellent kernel function and bandwidth h are obtained, and is assigned the failure modes based on Density Estimator algorithm
Device is endowed the fault grader based on Density Estimator algorithm of more excellent parameter using test sample test;
Step S36, judge to test whether accuracy meets preset requirement, if meeting, the more excellent parameter that will finally select
As optimized parameter and terminate process, otherwise return step S33.
That is step S3 is trained the failure training sample inputted using Density Estimator algorithm, i.e., will be trained defeated
Result selects optimal most suitable kernel function to keep training error small as far as possible compared with expected result to meet expected require out.If
Error deviation is larger, then changes kernel function, if error meets expected standard, changing the smoothing parameter of kernel function, (bandwidth h) is right
Training network is modified, until (bandwidth h) can make error convergence to meeting required precision to the smoothing parameter of selection.If meeting
Error requirements, then it is assumed that training process is completed in algorithm, and as the optimal ginseng of Density Estimator algorithm fault grader
Number, to test test sample.
The present invention uses NI company capture card PCI-6229 collecting sample, and speed is fast, and accuracy rate is high, and data are accurate;As
A kind of Nonparametric Estimation, Density Estimator do not do the data distribution form of probability function default special only in accordance with data itself
Estimated probability distributed model is levied, this has the ability that well adapts to the classification of failure, and similar fault signature will generate similar event
Hinder arithmetic result, accurate division can be made between different faults;The principle of Density Estimator algorithm is simple, that is, utilizes kernel function will
The data of bandwidth and each data point go multiple probability density of the data to be observed of fitting, then linear superposition as parameter
It is formed the estimation function of cuclear density, the probability density function of cuclear density is formed after normalization.So using KDE conduct
Fault grader, which detects converters health status, has apparent advantage, and judgement that can be more acurrate, more reliable converts
The health status of device also improves the efficiency of converter accident analysis.
Compared to the prior art, the invention has the following advantages: present invention employs Density Estimator algorithm (KDE)
Analysis on Fault Diagnosis is made to converters, since it is non-parametric estmation, the priori in relation to data distribution is not utilized to know
Know, it is only necessary to training is normal and fault sample can be achieved with the positioning of failure, the failure that indiscipline is crossed after testing can with just
Reason condition distinguishes, to detect converters failure situation;The present invention uses it as fault grader,
The advantage that KDE algorithm is simple, calculating speed is fast is not only played, but also studies its distribution character from data sample, is made
It can complete that parameter is preferred with less sample, enhance the ability of converters fault diagnosis Nonlinear Classification, more
The deficiency of traditional neural network is mended;Fault signature is extracted using the method that time domain, frequency-domain analysis combine simultaneously, can simplify
Classifier network structure just can carry out converter failure with less input value to accurately identify positioning.
Detailed description of the invention
Fig. 1 is the step S1 flow diagram of the embodiment of the present invention.
Fig. 2 is the method flow schematic diagram of the embodiment of the present invention.
Fig. 3 is the Selection of kernel function schematic diagram of the embodiment of the present invention.
Fig. 4 is that the bandwidth parameter h of the embodiment of the present invention selects schematic diagram.
Specific embodiment
With reference to the accompanying drawing, technical solution of the present invention is specifically described.
It is noted that following detailed description is all illustrative, it is intended to provide further instruction to the application.Unless another
It indicates, all technical and scientific terms used herein has usual with the application person of an ordinary skill in the technical field
The identical meanings of understanding.
It should be noted that term used herein above is merely to describe specific embodiment, and be not intended to restricted root
According to the illustrative embodiments of the application.As used herein, unless the context clearly indicates otherwise, otherwise singular
Also it is intended to include plural form, additionally, it should be understood that, when in the present specification using term "comprising" and/or " packet
Include " when, indicate existing characteristics, step, operation, device, component and/or their combination.
It is examined as shown in Fig. 2, present embodiments providing a kind of converters failure based on Density Estimator algorithm
Disconnected method, specifically includes the following steps:
Step S1, by data acquisition and noise reduction process, the sample with fault message is obtained;
Step S2, time domain in fault data, frequency domain failure spy are extracted using the cubic B-Spline interpolation analysis based on mallat
Sign, and by Data Dimensionality Reduction, gained fault signature and fault type one-to-one correspondence are established data sample library by secondary noise reduction process,
For training, testing characteristics of network;
Step S3, building is based on Density Estimator algorithm fault grader, will include in training sample after off-line training
Normal condition is accurately divided with all types of failures, extracts the optimized parameter of fault grader, and optimized parameter is directly assigned and is divided
Class device network carries out classifier test job;
Step S4, the classifier network for being endowed optimized parameter is implanted into simulink, does the failure of actual circuit
Converter circuit quick self-checking is realized in diagnosis.
As shown in Figure 1, in the present embodiment, step S1 specifically includes the following steps:
Step S11, according to physical fault, a situation arises, applies failure, circuit under simulating actual conditions to circuit components
Failure generates output waveform;
Step S12, Usage data collection card acquires measurement point electric signal;
Step S13, sampled signal removes noise by simulink module, obtains raw sample data, obtains with event
Hinder the sample of information.
In special step S12, converter circuit measurement point telecommunications to be measured is acquired using NI company capture card PCI-6229
Number.
Preferably, in the present embodiment, when diagnosing to actual motion circuit, the sample for also extracting record is believed
Breath may participate in when sample is accumulated to certain quantity and construct new sample database, enrich the data of sample database.
As shown in Fig. 2, acquiring converter circuit measurement point electric signal to be measured, including collection voltages by data acquisition device
Signal (input voltage, output voltage), current signal (output electric current, inductive current), with simulink module to the letter of acquisition
Number removal noise, obtain the faulty information of grandfather tape sample, reuse time domain, frequency-domain analysis, extracted from raw information
Feature vector { the I of energy response transform device health status1,I2,…,Ii,…,Iq, wherein q is the dimension of feature vector, and uses this
Input signal of the feature vector as Density Estimator algorithm, while establishing fault signature and being converted correspondingly with fault type
Device data sample library, for training, testing characteristics of network, network after training test saves optimum network structure parameter,
And (KDE) fault grader is reconstructed using optimized parameter, simulink then is written into the classifier network with optimized parameter
In, real-time fault diagnosis and positioning are done to the converters in actual motion, if breaking down and alarm, and informed
Fault type and abort situation realize converter circuit quick self-checking function.The present embodiment can record simultaneously mentions in actual work
The sample got, the sample that will build up on when sample accumulation is to certain quantity and the reconstruct of old sample form new training sample,
Off-line training processing further is carried out to network, this operates the failure modes ability that can preferably improve classifier, improves failure
Recognition capability realizes the Accurate Diagnosis to converters failure.
In the present embodiment, Density Estimator algorithm (KDE) fault grader, it is so-called to be carried out using nonparametric model
Density Estimator, i.e., the data point observed is fitted using smooth peak function, thus bent to true probability distribution
Line is simulated.
The general type of probability density function are as follows:
In formula, k is training sample (k=1,2 ..., c), nkFor the sample number of kth class training set;K (*) is kernel function, h(k)For the bandwidth parameter of kernel function K (*).Above formula indicates that kth class sample is the probability density function of r with a distance from central point.Herein
Assuming that sample point obeys exponential distribution with its affiliated class central point distance r, i.e. K (*) fetching number core is shown below:
K (x)=exp (- x), x >=0
The selection of kernel function to estimation effect there are certain influence, and bandwidth h(k)Selection to Density Estimator function
Play a significant role.Bandwidth h(k)Size reflect the intensive and sparse degree of kth class sample distribution, therefore can be according to sample
Distribution situation realize bandwidth calculating, be shown below:
In formula,Indicate sample point farthest with a distance from central point;α is the confidence level of the training set sample, that is, indicates one
A new sample point is fallen inProbability size on section.
When there are c class training sample, every one kind sample has corresponding probability density letter The probability density function for constructing all training samples isIt can be defined as:
Define the minimum value that β is probability density function:
β value is normal and fault sample line of demarcation in formula.WhenWhen, test sample belongs to normal condition;WhenWhen, test sample is fault sample.
According to the probability density function of all training samples obtainedThen for the definition of single failure specific gravity ρ
It is as follows:
As specific gravity ρ(k)For specific gravity ρ(1), ρ(2)..., ρ(c)In maximum value when, sample to be tested then belongs to kth class failure classes
Type.In this way, classification identification can be carried out to fault sample using KDE algorithm.
In the present embodiment, step S3 specifically includes the following steps:
Step S31, the data sample library that step S2 is obtained is divided into training sample and test sample;
Step S32, the input using training sample as Density Estimator algorithm fault grader is trained it;
Step S33, training sample is trained using different kernel functions, passes through the good core of training result using effect
Function.
Step S34, whether training of judgement error meets default, if then entering step S35, otherwise changes bandwidth h size
It is for further adjustments;
Step S35, more excellent kernel function and bandwidth h are obtained, and is assigned based on Density Estimator algorithm classification device, is used
Test sample tests the Density Estimator algorithm fault grader for being endowed more excellent parameter;
Step S36, judge to test whether accuracy meets preset requirement, if meeting, the more excellent parameter that will finally select
As optimized parameter and terminate process, otherwise return step S33.
That is step S3 is trained the failure training sample inputted using Density Estimator algorithm, by trained output
As a result compared with expected result, if error deviation is larger, change kernel function, as shown in Figure 3.If error meets expected mark
Standard, then change kernel function smoothing parameter (bandwidth h) to training network be modified, as shown in figure 4, then by change kernel function
Smoothing parameter (bandwidth h) to training network be modified, be allowed to along error reduce direction carry out, until selection smooth ginseng
(bandwidth h) can make error convergence to meeting required precision to number.If meeting error requirements, then it is assumed that algorithm, which is completed, trained
Journey, and as the optimized parameter of Density Estimator algorithm fault grader, to test test sample.
Particularly, in the present embodiment, it can be distinguished using Density Estimator algorithm when unknown failure occurs for circuit
The difference of malfunction and normal condition, to realize self-test, and algorithm can reduce time and the complexity of sample training
Degree, also reduces the dependence to fault sample.
The critically important step of the algorithm, is the selection of bandwidth parameter.Bandwidth reflects the flat journey of KDE curve entirety
Degree, that is, the data point observed specific gravity shared during the formation of KDE curve.Bandwidth is bigger, and the data point observed is most
End form at curve shape in proportion it is smaller, KDE integrated curved is more flat;Bandwidth is smaller, and the data point observed exists
Proportion is bigger in finally formed curve shape, and KDE integrated curved is more precipitous.As principle, the choosing of bandwidth is carried out
It selects, the construction of kernel function is made to be able to satisfy the requirement of error precision.
In actual application, a large amount of fault data is generally difficult to obtain converter, and its training pattern when
Between it is long, process is cumbersome.For this problem, modified way is required compared to all weights of other traditional neural networks, this
What text was proposed only needs individual representative fault samples based on Density Estimator algorithm, can be achieved with the positioning of failure,
So that the classification approximation capability of KDE classifier is stronger, it is used to detect converters health status as fault grader,
Identification of defective type has apparent advantage.
It is above-mentioned to construct KDE fault grader network, and derive that classifier parameters select rule, off-line training is tested
In optimum classifier network write-in simulink afterwards, the real-time self-test to converters circuit, efficiently, Gao Ke are realized
By the identification of defective of property, positioning failure.
The present embodiment uses NI company capture card PCI-6229 collecting sample, and speed is fast, and accuracy rate is high, and data are accurate;Make
For a kind of Nonparametric Estimation, Density Estimator does not do the data distribution form of probability function default only in accordance with data itself
Feature assessment probability Distribution Model, this, which has the classification of failure, well adapts to ability, and similar fault signature will generate similar
Fail result can make accurate division between different faults;The principle of Density Estimator algorithm is simple, that is, utilizes kernel function by band
The data of wide and each data point go multiple probability density of the data to be observed of fitting as parameter, then linear superposition is just
The estimation function of cuclear density is formd, the probability density function of cuclear density is formed after normalization.So using KDE as event
Hindering detection of classifier converters health status has apparent advantage, more acurrate, the more reliable judgement converter of energy
Health status, moreover it is possible to identify unknown failure, also improve the efficiency of converter accident analysis.
The above are preferred embodiments of the present invention, all any changes made according to the technical solution of the present invention, and generated function is made
When with range without departing from technical solution of the present invention, all belong to the scope of protection of the present invention.
Claims (3)
1. a kind of converter method for diagnosing faults based on Density Estimator, which comprises the following steps:
Step S1, by data acquisition and noise reduction process, the sample with fault data is obtained;
Step S2, time domain, frequency domain fault signature in fault data are extracted using the cubic B-Spline interpolation analysis based on mallat,
And by Data Dimensionality Reduction, gained fault signature and fault type one-to-one correspondence are established data sample library, are used for by secondary noise reduction process
Training, testing characteristics of network;
Step S3, the fault grader based on Density Estimator algorithm is constructed, will include just in training sample after off-line training
Normal state is accurately divided with all types of failures, extracts the more excellent parameter of fault grader, and directly assign more excellent parameter to classification
Device network carries out classifier test job, selects optimized parameter by test;
Step S4, the classifier network for being endowed optimized parameter is implanted into simulink, does the fault diagnosis of actual circuit,
Realize converter circuit quick self-checking.
2. the converter method for diagnosing faults according to claim 1 based on Density Estimator, which is characterized in that the step
The specific implementation steps are as follows by rapid S1:
Step S11, measurement point electric signal is acquired using capture card PCI-6229;
Step S12, sampling electric signal removes noise by simulink module, obtains raw sample data, obtains with faulty
The sample of data.
3. the converter method for diagnosing faults according to claim 1 based on Density Estimator, which is characterized in that the step
The specific implementation steps are as follows by rapid S3:
Step S31, the data sample library that step S2 is obtained is divided into training sample and test sample;
Step S32, the input using training sample as the fault grader based on Density Estimator algorithm is trained it;
Step S33, training sample is trained using different kernel functions, passes through the good core letter of training result using effect
Number;
Step S34, whether training of judgement error meets default, if then entering step S35, otherwise change bandwidth h size make into
One successive step;
Step S35, more excellent kernel function and bandwidth h are obtained, and is assigned the fault grader based on Density Estimator algorithm, is adopted
The fault grader based on Density Estimator algorithm of more excellent parameter is endowed with test sample test;
Step S36, judge test accuracy whether meet preset requirement, if meeting, using the more excellent parameter finally selected as
Optimized parameter simultaneously terminates process, otherwise return step S33.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910097620.4A CN109828168A (en) | 2019-01-31 | 2019-01-31 | Converter method for diagnosing faults based on Density Estimator |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910097620.4A CN109828168A (en) | 2019-01-31 | 2019-01-31 | Converter method for diagnosing faults based on Density Estimator |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109828168A true CN109828168A (en) | 2019-05-31 |
Family
ID=66863254
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910097620.4A Pending CN109828168A (en) | 2019-01-31 | 2019-01-31 | Converter method for diagnosing faults based on Density Estimator |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109828168A (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110596506A (en) * | 2019-10-30 | 2019-12-20 | 福州大学 | Converter fault diagnosis method based on time convolution network |
CN111626821A (en) * | 2020-05-26 | 2020-09-04 | 山东大学 | Product recommendation method and system for realizing customer classification based on integrated feature selection |
CN113051092A (en) * | 2021-02-04 | 2021-06-29 | 中国人民解放军国防科技大学 | Fault diagnosis method based on optimized kernel density estimation and JS divergence |
CN114325480A (en) * | 2021-11-19 | 2022-04-12 | 广东核电合营有限公司 | Diode open-circuit fault detection method and device for multiphase brushless exciter |
CN114429235A (en) * | 2020-10-29 | 2022-05-03 | 新智数字科技有限公司 | Equipment fault prediction method and device, readable medium and electronic equipment |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103868692A (en) * | 2014-03-18 | 2014-06-18 | 电子科技大学 | Rotary machine fault diagnosis method based on kernel density estimation and K-L divergence |
CN107144430A (en) * | 2017-06-27 | 2017-09-08 | 电子科技大学 | A kind of Method for Bearing Fault Diagnosis based on incremental learning |
US20170276571A1 (en) * | 2016-03-24 | 2017-09-28 | Johnson Controls Technology Company | Systems and methods for fault detection and handling by assessing building equipment performance |
JP2018155507A (en) * | 2017-03-15 | 2018-10-04 | 日新電機株式会社 | Partial discharge detection device and partial discharge detection method |
CN109116150A (en) * | 2018-08-03 | 2019-01-01 | 福州大学 | A kind of converters method for diagnosing faults based on Cerebellar Model Articulation Controller |
-
2019
- 2019-01-31 CN CN201910097620.4A patent/CN109828168A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103868692A (en) * | 2014-03-18 | 2014-06-18 | 电子科技大学 | Rotary machine fault diagnosis method based on kernel density estimation and K-L divergence |
US20170276571A1 (en) * | 2016-03-24 | 2017-09-28 | Johnson Controls Technology Company | Systems and methods for fault detection and handling by assessing building equipment performance |
JP2018155507A (en) * | 2017-03-15 | 2018-10-04 | 日新電機株式会社 | Partial discharge detection device and partial discharge detection method |
CN107144430A (en) * | 2017-06-27 | 2017-09-08 | 电子科技大学 | A kind of Method for Bearing Fault Diagnosis based on incremental learning |
CN109116150A (en) * | 2018-08-03 | 2019-01-01 | 福州大学 | A kind of converters method for diagnosing faults based on Cerebellar Model Articulation Controller |
Non-Patent Citations (2)
Title |
---|
MINGYANG LI等: "A novel seizure diagnostic model based on kernel density estimationand least squares support vector machine", 《BIOMEDICAL SIGNAL PROCESSING AND CONTROL》 * |
吴海燕等: "ELMD 及核密度估计的滚动轴承故障诊断", 《机械强度》 * |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110596506A (en) * | 2019-10-30 | 2019-12-20 | 福州大学 | Converter fault diagnosis method based on time convolution network |
CN111626821A (en) * | 2020-05-26 | 2020-09-04 | 山东大学 | Product recommendation method and system for realizing customer classification based on integrated feature selection |
CN111626821B (en) * | 2020-05-26 | 2024-03-12 | 山东大学 | Product recommendation method and system for realizing customer classification based on integrated feature selection |
CN114429235A (en) * | 2020-10-29 | 2022-05-03 | 新智数字科技有限公司 | Equipment fault prediction method and device, readable medium and electronic equipment |
CN113051092A (en) * | 2021-02-04 | 2021-06-29 | 中国人民解放军国防科技大学 | Fault diagnosis method based on optimized kernel density estimation and JS divergence |
CN114325480A (en) * | 2021-11-19 | 2022-04-12 | 广东核电合营有限公司 | Diode open-circuit fault detection method and device for multiphase brushless exciter |
CN114325480B (en) * | 2021-11-19 | 2023-09-29 | 广东核电合营有限公司 | Diode open-circuit fault detection method and device for multiphase brushless exciter |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109828168A (en) | Converter method for diagnosing faults based on Density Estimator | |
Liu et al. | Classifying transformer winding deformation fault types and degrees using FRA based on support vector machine | |
CN106646096B (en) | Transformer fault classification and recognition methods based on vibration analysis method | |
CN111459700B (en) | Equipment fault diagnosis method, diagnosis device, diagnosis equipment and storage medium | |
CN106443310B (en) | A kind of transformer fault detection method based on SOM neural network | |
CN110596506A (en) | Converter fault diagnosis method based on time convolution network | |
CN108535635B (en) | EEMD and HMM based analog circuit intermittent fault diagnosis method | |
CN108520357A (en) | A kind of method of discrimination, device and the server of line loss abnormal cause | |
CN106443316A (en) | Power transformer winding deformation state multi-information detection method and device | |
CN112200032A (en) | Attention mechanism-based high-voltage circuit breaker mechanical state online monitoring method | |
CN107909118A (en) | A kind of power distribution network operating mode recording sorting technique based on deep neural network | |
CN112464995A (en) | Power grid distribution transformer fault diagnosis method and system based on decision tree algorithm | |
CN112485747A (en) | Error state evaluation method for capacitor voltage transformer | |
CN104635146B (en) | Analog circuit fault diagnosis method based on random sinusoidal signal test and HMM (Hidden Markov Model) | |
CN112149569B (en) | Voiceprint fault diagnosis method of transformer based on fuzzy C-means clustering algorithm | |
CN116520068B (en) | Diagnostic method, device, equipment and storage medium for electric power data | |
CN114255784A (en) | Substation equipment fault diagnosis method based on voiceprint recognition and related device | |
CN105375449A (en) | Method and system for setting direct-current line traveling wave protection voltage change rate criterion | |
CN109726770A (en) | A kind of analog circuit fault testing and diagnosing method | |
CN113406439B (en) | Fault positioning method for power distribution network | |
CN112600659B (en) | Method and system for detecting side channel leakage of safety chip based on neural network | |
CN113946920A (en) | Rolling bearing fault diagnosis method with unbalanced data and data set deviation | |
CN108052953A (en) | The relevant sample extended method of feature based | |
CN110888011B (en) | In-situ test method and test device based on testability modeling | |
CN112880726A (en) | Sensor fault diagnosis method and device based on variational modal decomposition sample entropy |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190531 |
|
RJ01 | Rejection of invention patent application after publication |