CN109085495A - A kind of monitoring and diagnosis method and system of breaker Internet of Things information system - Google Patents
A kind of monitoring and diagnosis method and system of breaker Internet of Things information system Download PDFInfo
- Publication number
- CN109085495A CN109085495A CN201810758937.3A CN201810758937A CN109085495A CN 109085495 A CN109085495 A CN 109085495A CN 201810758937 A CN201810758937 A CN 201810758937A CN 109085495 A CN109085495 A CN 109085495A
- Authority
- CN
- China
- Prior art keywords
- breaker
- sample
- vibration signal
- internet
- monitoring
- 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
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
- G01R31/327—Testing of circuit interrupters, switches or circuit-breakers
- G01R31/3277—Testing of circuit interrupters, switches or circuit-breakers of low voltage devices, e.g. domestic or industrial devices, such as motor protections, relays, rotation switches
Abstract
The present invention is a kind of monitoring and diagnosis method and system of breaker Internet of Things information system, is monitored transmission to the electric current, voltage and vibration signal of breaker.Rank scores are carried out to electric current and voltage signal according to transmission data, it is scored using limited Boltzmann machine vibration signal, and the influence for comprehensively considering three kinds of signals is classified the fault degree of breaker, completes the monitoring and diagnosis process of breaker, to avoid all kinds of fault harms.The invention has the advantages that the electric current, voltage and vibration signal to breaker are monitored, the failure of circuit breaker voltage, electric current and vibration aspect can be found in time;The comprehensive diagnos circuit-breaker status in terms of three, diagnostic result are accurate.
Description
Technical field
The present invention relates to equipment running status evaluation areas more particularly to a kind of monitoring of breaker Internet of Things information system to examine
Disconnected method and system.
Background technique
Industry 4.0 has been arrived, and informationization has become the new direction of industrial development.In IT wave, Internet of Things
Net becomes the new coordinate in epoch.Internet of Things is the information carriers such as internet, conventional telecommunications net, and allowing all can exercise standalone feature
Common objects realize the network that interconnects, it is the extensive use of various cognition technologies.Magnanimity is deployed on Internet of Things
Multiple types sensor, each sensor are an information source, the information content and letter that different classes of sensor is captured
It is different to cease format.The data that sensor obtains have real-time, by the acquisition environmental information of certain frequency periodicity, constantly more
New data.
With the development of Internet of Things, breaker inline diagnosis is the important prerequisite of breaker safe operation, finds breaker
Security risk present in operation carries out repair and maintenance to fault point in time, stablizes maintenance power system security with important
Meaning.Existing high voltage fault diagnosis is all to carry out correlation analysis to the vibration signal of breaker mostly, frequently with each quasi-tradition
Method and machine learning algorithm carry out state recognition and diagnosis, and diagnostic method is more unilateral and cannot find breaker mistake in time
The failures such as load, short circuit, single-phase earthing, under-voltage.
Summary of the invention
Only detecting breaker according to breaker vibrational state present invention mainly solves traditional diagnosis method, whether failure is deposited
There is the problem of can not diagnosing the failures such as breaker overload, short circuit, single-phase earthing, under-voltage, open circuit can be diagnosed by providing one kind
Device electric current, voltage failure, the monitoring and diagnosis method and system of the accurate breaker Internet of Things information system of diagnostic result.
The technical solution adopted by the present invention to solve the technical problems is: a kind of monitoring of breaker Internet of Things information system is examined
Disconnected method, comprising the following steps:
S100: current status assessment, voltage status assessment and vibrational state assessment are carried out respectively to breaker, is obtained corresponding
Current status scoring, voltage status scoring and vibrational state scoring;
S200: each condition grading obtains comprehensive score multiplied by summing after respective weight;
S300: circuit-breaker status grade is obtained according to comprehensive score;
S400: breaker is performed corresponding processing according to circuit-breaker status grade.
Scheme as a preference of the above scheme, the described vibrational state assessment the following steps are included:
S101: Circuit breaker vibration signal is obtained;
S102: intercepted samples segment;
S103: albefaction, standardization and naturalization processing are successively carried out to sample segment;
S104: will treated sample segment input-bound Boltzmann machine;
S105: sample characteristics are obtained;
S106: step S102-S105 is repeated, different sample segments is intercepted, obtains different sample characteristics, obtain sample
Feature set;
S107: the median of each sample characteristics is taken to obtain vibration signal characteristics;
S108: vibration signal characteristics are inputted into Softmax classifier;
S109:Softmax classifier is to assume the probability numbers of each state of function representation breaker;
S110: taking maximum probability numerical value corresponding states is current breaker vibrational state.
As a kind of further preferred embodiment of above scheme, the limited Boltzmann machine and Softmax classification
Device needs to be trained before use, comprising the following steps:
S201: the vibration signal set of each state of breaker is obtained;
S202: intercepted samples segment constitutes sample data set in vibration signal;
S203: albefaction, standardization and naturalization processing are successively carried out to sample data set, obtains sample training data set;
S204: pass through sample training data set and the limited Boltzmann machine of contrast divergence algorithm training;
S205: the multiple segment input-bound Boltzmann machines for intercepting same state vibration signal obtain sample training feature
Collection;
S206: take the median of each sample training feature as the state vibration signal characteristics;
S207: vibration signal characteristics are inputted into Softmax classifier;
S208: the tag value of the maximum probability for the hypothesis function for taking Softmax classifier to export is vibrated as the state
The corresponding equipment state of signal.Using the vibration signal of known state to limited Boltzmann machine and Softmax classifier, can obtain
The final data that the vibration signal of different conditions obtains is obtained, convenient for diagnosing to breaker.
Scheme as a preference of the above scheme, the described current status assessment according to breaker actual current with it is specified
Circuit breaker electric stream mode is divided into safety by the value of electric current, small amount transfinites, wholesale transfinites and is shorted.Four kinds of current status are equipped with phase
Corresponding condition grading, condition grading is for calculating comprehensive score.
Scheme as a preference of the above scheme, the described voltage status assessment according to breaker virtual voltage with it is specified
Circuit breaker electric stream mode is divided into normal, low pressure, ultralow pressure and high pressure by the value of voltage.Four kinds of voltage status are equipped with corresponding
Condition grading, condition grading is for calculating comprehensive score.
Scheme as a preference of the above scheme, the hypothesis function expression are as follows:
Wherein, θ=[θ1,θ2,…,θK]TIt is the parameter of classifier Softmax,It is entraining agent, χiIndicate sample
Eigen or sample training feature, βiIndicate all kinds of state, that is, tag values of breaker.Vibrator state shares eight kinds, so
The value of φ is 8, eight probability numbers and be 1.
Scheme as a preference of the above scheme, the vibrator state include normal, control power supply air switch
Initial failure, control power supply air switch later period failure, coil initial failure, coil later period failure, tank circuit initial failure,
Tank circuit later period failure, resultant fault.Each vibrator state has a corresponding state classification number, i.e. reference numerals
Value, each vibrator state are additionally provided with a corresponding condition grading for calculating comprehensive score.
The present invention also provides a kind of monitoring and diagnosis systems of breaker Internet of Things information system, including cloud server, data
Library, current sensor, voltage sensor and the vibrating sensor of processor and setting on-board the circuit breaker, the current sensor,
Breaker data is transferred to cloud server by network by voltage sensor and vibrating sensor, and cloud server utilizes data
Library storing data, processor pass through network call database data.
Scheme as a preference of the above scheme, the processor are PC machine.
The invention has the advantages that the electric current, voltage and vibration signal to breaker are monitored, open circuit can be found in time
The failure of device voltage, electric current and vibration aspect;The comprehensive diagnos circuit-breaker status in terms of three, diagnostic result are accurate.
Detailed description of the invention
Fig. 1 is a kind of flow chart of the monitoring and diagnosis method of breaker Internet of Things information system.
Fig. 2 is a kind of flow chart figure of vibrational state assessment.
Fig. 3 is a kind of flow chart of limited Boltzmann machine and Softmax classifier training.
Fig. 4 is a kind of structural schematic diagram of the monitoring and diagnosis system of breaker Internet of Things information system.
Fig. 5 is a kind of condition grading table of circuit breaker current status assessment.
Fig. 6 is a kind of condition grading table of circuit breaker voltage status assessment.
Fig. 7 is a kind of condition grading table of breaker vibrational state assessment.
1- current sensor 2- voltage sensor 3- vibrating sensor 4- database 5- cloud server 6- processor.
Specific embodiment
Below with reference to the embodiments and with reference to the accompanying drawing further description of the technical solution of the present invention.
Embodiment:
The monitoring and diagnosis method of the present embodiment breaker Internet of Things information system, as shown in Figure 1, comprising the following steps:
S100: current status assessment, voltage status assessment and vibrational state assessment are carried out respectively to breaker, is obtained corresponding
Current status scoring, voltage status scoring and vibrational state scoring;
S200: each condition grading obtains comprehensive score multiplied by summing after respective weight;
S300: circuit-breaker status grade is obtained according to comprehensive score;
S400: breaker is performed corresponding processing according to circuit-breaker status grade.
As shown in Fig. 2, the described vibrational state assessment the following steps are included:
S101: Circuit breaker vibration signal is obtained;
S102: intercepted samples segment;
S103: albefaction, standardization and naturalization processing are successively carried out to sample segment;
S104: will treated sample segment input-bound Boltzmann machine;
S105: sample characteristics are obtained;
S106: step S102-S105 is repeated several times, intercepts different sample segments, obtains different sample characteristics, obtain
Sample characteristics collection;
S107: the median of each sample characteristics is taken to obtain vibration signal characteristics;
S108: vibration signal characteristics are inputted into Softmax classifier;
S109:Softmax classifier is to assume the probability numbers of each state of function representation breaker;
S110: taking maximum probability numerical value corresponding states is current breaker vibrational state.
As shown in figure 3, the limited Boltzmann machine and Softmax classifier needs to be trained before use, packet
Include following steps:
S201: the vibration signal set of each state of breaker is obtained;
S202: intercepted samples segment constitutes sample data set in vibration signal;
S203: albefaction, standardization and naturalization processing are successively carried out to sample data set, obtains sample training data set;
S204: pass through sample training data set and the limited Boltzmann machine of contrast divergence algorithm training;
S205: the multiple segment input-bound Boltzmann machines for intercepting same state vibration signal obtain sample training feature
Collection;
S206: take the median of each sample training feature as the state vibration signal characteristics;
S207: vibration signal characteristics are inputted into Softmax classifier;
S208: the tag value of the maximum probability for the hypothesis function for taking Softmax classifier to export is vibrated as the state
The corresponding equipment state of signal.
The circuit breaker failure based on limited Boltzmann machine is established according to all kinds of fault-signals of known breaker first
Diagnostic model, then to the limited Boltzmann machine of training, using trained limited Boltzmann machine after each offset
Vibration signal in extract characteristic value, finally using extract fault signature, training Softmax classifier, after training
Classifier assesses fault state of circuit breaker.
The vibration signal of all kinds of malfunctions of breaker is obtained, vibration signal data set is constitutedBy each
The vibration signal α of breakeriIt is classified as training sample, constitutes a new sample setα in sample seti∈δI×DIt indicates
I-th of vibration signal, there are D data points in each signal.βiIndicate all kinds of shapes of the corresponding breaker of i-th of vibration signal
State, ε are expressed as number of signals.
Limited Boltzmann machine is trained, the selection of characteristic point needs to set length as Xin, output dimension is XoutWith
The step-length for intercepting vibration signal sample is Ytrain, construct length XinFor visible layer, output dimension is XoutLimited Boltzmann
Machine, from the vibration signal α of known breakeri∈δI×DIn pass through step-length YtrainThe sample segment X of intercepted lengthin, thus structure
At sample data set It is j-th of sample segment, includes XinA data point.First by sample number
Albefaction is carried out according to collection, Z-score standardization then is carried out to processed sample data set, later whitening and standardized number
According to naturalization processing is carried out, zooms between [0,1], finally obtain sample training data set
Random selection initialization weight E, visible biasing r and initial hidden layer biasing t are initialized, it will be in sample training data set
A sampleInput-bound Boltzmann machine, so that it may it is direct to obtain hidden layer output probability, i.e.,Sigm (x) is sigmoid activation primitive, available by carrying out two-value sampling to hidden layer
Hidden layer stateThen hidden layer is reconstructed can be obtained visible layer output probabilityWith
Afterwards willThe middle factor carries out data sampling as mean values, 1 for standard deviation, obtains visible layer stateFurther according to can
The state of layer is seen as input, and reverse goes out the new output probability of hidden layerE, r, t are updated
Cheng Wei η indicates learning rate.
Pass through step-length YtrainUse length XinIn the vibration signal α of breakeriIn moved, after interception jth time translation
Sample segmentInput-bound Boltzmann machine extracts corresponding state feature after albefaction, standardization and naturalization processing,
Obtain sample characteristicsOn the vibration signal of breaker after repeatedly translating, sample characteristics collection can be obtainedThe feature of the medians of each column as the vibration signal of breaker is obtained again
ForMEIAN (x) is that the median row vector of each column of matrix x is constituted.
The characteristic value that limited Boltzmann machine is extracted inputs Softmax classifier, obtains sample characteristics χi, utilize ε
Sample forms sample characteristics collectionRespectively correspond all kinds of states of breakerCircuit-breaker status is indicated with φ
?The sample characteristics χ of each inputiUsing Softmax classifier, the φ for assuming that function gives is used
Dimensional vector indicates that φ probability numbers of breaker different faults situation, the adduction of φ probability numbers are 1, it is assumed that function call table
It is as follows up to formula:
θ=[θ1,θ2,…,θK]TIt is the parameter of classifier Softmax,It is entraining agent, model passes through minimum
Following cost function M (θ) reaches.
λ is weight decay factor.It is calculated by background program, acquires the probability vector of output, and choose maximum probability
Numeric indicia is signal alphaiThe equipment state of corresponding breaker, to assess vibrational state.
The data set of breaker is divided into 8 subsets, the different conditions of corresponding breaker, and vibrator state includes normal, control
Power supply air switch initial failure processed, control power supply air switch later period failure, coil initial failure, coil later period failure, storage
It can circuit initial failure, tank circuit later period failure, resultant fault.State classification number carries out analysis and assessment for appealing method,
Condition grading is used for last comprehensive score, and the scoring of breaker vibrational state is as shown in Figure 7.
After the completion of limited Boltzmann machine and Softmax classifier training, detected using above-mentioned identical method interception
Vibration signal sample set, finally show that breaker is in each shape probability of state, the state for choosing maximum probability is worked as breaker
Preceding state obtains breaker current vibration condition grading by grade form shown in Fig. 7.
Current status assessment according to the value of breaker actual current and rated current by circuit breaker electric stream mode be divided into safety,
Small amount transfinites, wholesale transfinites and is shorted, and for circuit breaker current condition grading as shown in figure 5, ω is rated current, ξ is current sense
Device obtains numerical value.
Voltage status assessment according to the value of breaker virtual voltage and voltage rating by circuit breaker electric stream mode be divided into it is normal,
Low pressure, ultralow pressure and high pressure, for circuit breaker voltage condition grading as shown in fig. 6, ψ is voltage rating, ζ is that voltage sensor obtains number
Value.
The present embodiment also provides a kind of monitoring and diagnosis system of breaker Internet of Things information system, as shown in figure 4, including cloud
Server 5, database 4, current sensor 1, voltage sensor 2 and the vibrating sensor of processor 6 and setting on-board the circuit breaker
3, breaker data is transferred to cloud server by network by the current sensor, voltage sensor and vibrating sensor,
Cloud server utilizes data database storing, and processor passes through network call database data.Processor 6 is PC machine.Electric current
Sensor, voltage sensor and vibrating sensor setting on-board the circuit breaker, select sim mode that breaker evidence is transferred to cloud,
Cloud server uses Ali's cloud Internet of Things external member, in addition uses mysql database, and PC machine is monitored point three classes signal
Analysis, including levels of current classification, voltage level classification and vibrational state assessment, finally integrate three classes situation to the state of breaker
It is diagnosed.
Current status assessment, voltage status assessment and vibrational state assessment account for 0.3,0.3,0.4 weight respectively, comprehensive
Divide and be equal to D=0.3 × a+0.3 × b+0.4 × c, a is current status scoring, and b is voltage status scoring, and c comments for vibrational state
Point.Work as D=0, circuit-breaker status is 1 grade;When D ∈ (0,30], circuit-breaker status be 2 grades;When D ∈ (30,60], circuit-breaker status
It is 3 grades;When D ∈ (60,80], circuit-breaker status be 4 grades;When D ∈ (80,100], circuit-breaker status be 5 grades.
Grade 1 represents insignificant failure, and grade 2 represents slight failure, and grade 3 represents moderate failure, etc.
Grade 4 represents serious failure, and class 5 represents catastrophic failure.After measuring after fault level, system will be pushed away by information
Send, remind on-call maintenance, when grade reaches 4 grades, delay 20 minutes will automatic short circuit power supply, stop breaker work, when
When grade reaches 5 grades, directly stopping breaker work.
Specific embodiment described herein is only an example for the spirit of the invention.The neck of technology belonging to the present invention
The technical staff in domain can make various modifications or additions to the described embodiments or replace by a similar method
In generation, however, it does not deviate from the spirit of the invention or beyond the scope of the appended claims.
Claims (9)
1. a kind of monitoring and diagnosis method of breaker Internet of Things information system, it is characterized in that: the following steps are included:
S100: current status assessment, voltage status assessment and vibrational state assessment are carried out respectively to breaker, obtains corresponding electricity
Stream mode scoring, voltage status scoring and vibrational state scoring;
S200: each condition grading obtains comprehensive score multiplied by summing after respective weight;
S300: circuit-breaker status grade is obtained according to comprehensive score;
S400: breaker is performed corresponding processing according to circuit-breaker status grade.
2. the monitoring and diagnosis method of breaker Internet of Things information system according to claim 1, it is characterized in that: the vibration
Status assessment the following steps are included:
S101: Circuit breaker vibration signal is obtained;
S102: intercepted samples segment;
S103: albefaction, standardization and naturalization processing are successively carried out to sample segment;
S104: will treated sample segment input-bound Boltzmann machine;
S105: sample characteristics are obtained;
S106: step S102-S105 is repeated several times, intercepts different sample segments, obtains different sample characteristics, obtain sample
Feature set;
S107: the median of each sample characteristics is taken to obtain vibration signal characteristics;
S108: vibration signal characteristics are inputted into Softmax classifier;
S109:Softmax classifier is to assume the probability numbers of each state of function representation breaker;
S110: taking maximum probability numerical value corresponding states is current breaker vibrational state.
3. the monitoring and diagnosis method of breaker Internet of Things information system according to claim 2, it is characterized in that: described is limited
Boltzmann machine and Softmax classifier need to be trained before use, comprising the following steps:
S201: the vibration signal set of each state of breaker is obtained;
S202: intercepted samples segment constitutes sample data set in vibration signal;
S203: albefaction, standardization and naturalization processing are successively carried out to sample data set, obtains sample training data set;
S204: pass through sample training data set and the limited Boltzmann machine of contrast divergence algorithm training;
S205: the multiple segment input-bound Boltzmann machines for intercepting same state vibration signal obtain sample training feature set;
S206: take the median of each sample training feature as the state vibration signal characteristics;
S207: vibration signal characteristics are inputted into Softmax classifier;
S208: the tag value of the maximum probability for the hypothesis function for taking Softmax classifier to export is as the state vibration signal
Corresponding equipment state.
4. the monitoring and diagnosis method of breaker Internet of Things information system according to claim 1, it is characterized in that: the electric current
Circuit breaker electric stream mode is divided into safety according to the value of breaker actual current and rated current by status assessment, small amount transfinites, greatly
Volume transfinites and is shorted.
5. the monitoring and diagnosis method of breaker Internet of Things information system according to claim 1, it is characterized in that: the voltage
Circuit breaker electric stream mode is divided into normal, low pressure, ultralow pressure according to the value of breaker virtual voltage and voltage rating by status assessment
And high pressure.
6. the monitoring and diagnosis method of breaker Internet of Things information system according to claim 3, it is characterized in that: the hypothesis
Function expression is as follows:
Wherein, θ=[θ1,θ2,…,θK]TIt is the parameter of classifier Softmax,It is entraining agent, χiIndicate sample characteristics
Or sample training feature, βiIndicate all kinds of state, that is, tag values of breaker.
7. the monitoring and diagnosis method of breaker Internet of Things information system according to claim 1, it is characterized in that: the vibration
Device state includes normal, control power supply air switch initial failure, control power supply air switch later period failure, coil early stage event
Barrier, coil later period failure, tank circuit initial failure, tank circuit later period failure, resultant fault.
8. a kind of monitoring and diagnosis system of breaker Internet of Things information system, it is characterized in that: including cloud server (5), database
(4), current sensor (1), voltage sensor (2) and the vibrating sensor (3) of processor (6) and setting on-board the circuit breaker, institute
It states current sensor, voltage sensor and vibrating sensor and breaker data is transferred to by cloud server, cloud by network
Server by utilizing data database storing, processor pass through network call database data.
9. the monitoring and diagnosis system of breaker Internet of Things information system according to claim 8, it is characterized in that: the processing
Device (6) is PC machine.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810758937.3A CN109085495A (en) | 2018-07-11 | 2018-07-11 | A kind of monitoring and diagnosis method and system of breaker Internet of Things information system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810758937.3A CN109085495A (en) | 2018-07-11 | 2018-07-11 | A kind of monitoring and diagnosis method and system of breaker Internet of Things information system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109085495A true CN109085495A (en) | 2018-12-25 |
Family
ID=64837486
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810758937.3A Pending CN109085495A (en) | 2018-07-11 | 2018-07-11 | A kind of monitoring and diagnosis method and system of breaker Internet of Things information system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109085495A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111106958A (en) * | 2019-12-19 | 2020-05-05 | 国网冀北电力有限公司信息通信分公司 | Internet of things data processing method and device for power switch cabinet |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN202929174U (en) * | 2012-10-24 | 2013-05-08 | 云南电网公司保山供电局 | High voltage circuit breaker on-line monitoring system |
CN103575525A (en) * | 2013-11-18 | 2014-02-12 | 东南大学 | Intelligent diagnosis method for mechanical fault of circuit breaker |
CN104267337A (en) * | 2014-08-01 | 2015-01-07 | 云南电力试验研究院(集团)有限公司电力研究院 | All-in-one detecting platform and state recognition system based on circuit breaker operating mechanism |
CN106769048A (en) * | 2017-01-17 | 2017-05-31 | 苏州大学 | Self adaptation depth confidence network Method for Bearing Fault Diagnosis based on Nesterov momentum methods |
CN107451760A (en) * | 2017-09-04 | 2017-12-08 | 西安交通大学 | Based on when the limited Boltzmann machine of window sliding Fault Diagnosis of Roller Bearings |
-
2018
- 2018-07-11 CN CN201810758937.3A patent/CN109085495A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN202929174U (en) * | 2012-10-24 | 2013-05-08 | 云南电网公司保山供电局 | High voltage circuit breaker on-line monitoring system |
CN103575525A (en) * | 2013-11-18 | 2014-02-12 | 东南大学 | Intelligent diagnosis method for mechanical fault of circuit breaker |
CN104267337A (en) * | 2014-08-01 | 2015-01-07 | 云南电力试验研究院(集团)有限公司电力研究院 | All-in-one detecting platform and state recognition system based on circuit breaker operating mechanism |
CN106769048A (en) * | 2017-01-17 | 2017-05-31 | 苏州大学 | Self adaptation depth confidence network Method for Bearing Fault Diagnosis based on Nesterov momentum methods |
CN107451760A (en) * | 2017-09-04 | 2017-12-08 | 西安交通大学 | Based on when the limited Boltzmann machine of window sliding Fault Diagnosis of Roller Bearings |
Non-Patent Citations (4)
Title |
---|
孙鹏等: "基于贝叶斯概率赋权的SF6高压断路器综合状态评估", 《云南电力技术》 * |
王有元等: "基于层次分析结构的变压器健康状态与综合寿命评估模型", 《电网技术》 * |
程国建等: "深度信念网络应用于图像分类的可行性研究", 《电脑知识与技术》 * |
颜丹: "基于深度自动编码器的特征提取算法研究", 《中国优秀硕士学位论文数据库 信息科技辑》 * |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111106958A (en) * | 2019-12-19 | 2020-05-05 | 国网冀北电力有限公司信息通信分公司 | Internet of things data processing method and device for power switch cabinet |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Zhang et al. | Data-based line trip fault prediction in power systems using LSTM networks and SVM | |
CN106202886B (en) | Track circuit red band Fault Locating Method based on fuzzy coarse central and decision tree | |
CN109187025A (en) | A kind of rolling bearing remaining life prediction technique of integrated KELM | |
CN109635928A (en) | A kind of voltage sag reason recognition methods based on deep learning Model Fusion | |
WO2023044978A1 (en) | Adversarial-flow-model-based unsupervised fault diagnosis method for mechanical device | |
CN107766879A (en) | The MLP electric network fault cause diagnosis methods of feature based information extraction | |
CN110376522A (en) | A kind of Method of Motor Fault Diagnosis of the deep learning network of data fusion | |
CN107368854A (en) | A kind of circuit breaker failure diagnostic method based on improvement evidence theory | |
CN103901882A (en) | Online monitoring fault diagnosis system and method of train power system | |
CN109948808A (en) | The banking process in substation equipment fault case library, fault diagnosis method and system | |
CN107316057A (en) | Based on the nuclear power unit method for diagnosing faults being locally linear embedding into K nearest neighbor classifiers | |
CN108710756A (en) | The method for diagnosing faults of lower multicharacteristic information Weighted Fusion is analyzed based on spectral clustering | |
CN105606914A (en) | IWO-ELM-based Aviation power converter fault diagnosis method | |
CN107340766A (en) | Power scheduling alarm signal text based on similarity sorts out and method for diagnosing faults | |
CN108257365A (en) | A kind of industrial alarm designs method based on global nonspecific evidence dynamic fusion | |
CN112200694A (en) | Dominant instability mode identification model construction and application method based on graph neural network | |
CN110991472A (en) | Micro fault diagnosis method for high-speed train traction system | |
CN108919104A (en) | A kind of circuit breaker failure diagnostic method based on Fisher identification and classification method | |
CN113255458A (en) | Bearing fault diagnosis method based on multi-view associated feature learning | |
CN108280424A (en) | A kind of rolling bearing method for predicting residual useful life based on sparse coding | |
CN107491058A (en) | A kind of industrial control system sequence attack detection method and equipment | |
CN109085495A (en) | A kind of monitoring and diagnosis method and system of breaker Internet of Things information system | |
CN114091549A (en) | Equipment fault diagnosis method based on deep residual error network | |
CN109580260A (en) | A kind of inferior health diagnostic method of track vehicle door system | |
CN108121796A (en) | Electric energy metering device failure analysis methods and device based on confidence level |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181225 |