CN110503004A - A kind of online method of discrimination of Switching Power Supply operating status - Google Patents
A kind of online method of discrimination of Switching Power Supply operating status Download PDFInfo
- Publication number
- CN110503004A CN110503004A CN201910689142.6A CN201910689142A CN110503004A CN 110503004 A CN110503004 A CN 110503004A CN 201910689142 A CN201910689142 A CN 201910689142A CN 110503004 A CN110503004 A CN 110503004A
- Authority
- CN
- China
- Prior art keywords
- class
- power supply
- input
- sample
- frequency
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 30
- 238000012360 testing method Methods 0.000 claims abstract description 21
- 238000012549 training Methods 0.000 claims abstract description 19
- 238000012216 screening Methods 0.000 claims abstract description 10
- 239000000284 extract Substances 0.000 claims abstract description 7
- 230000004069 differentiation Effects 0.000 claims abstract description 5
- 239000013598 vector Substances 0.000 claims description 13
- 238000004458 analytical method Methods 0.000 claims description 11
- 238000012423 maintenance Methods 0.000 claims description 7
- 238000012544 monitoring process Methods 0.000 claims description 7
- 238000013507 mapping Methods 0.000 claims description 5
- 230000007257 malfunction Effects 0.000 claims description 4
- 230000003321 amplification Effects 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims description 3
- 238000002790 cross-validation Methods 0.000 claims description 3
- 238000000354 decomposition reaction Methods 0.000 claims description 3
- 230000005611 electricity Effects 0.000 claims description 3
- 230000005484 gravity Effects 0.000 claims description 3
- 238000003199 nucleic acid amplification method Methods 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims description 3
- 230000009466 transformation Effects 0.000 claims description 3
- 238000007619 statistical method Methods 0.000 claims description 2
- 238000003909 pattern recognition Methods 0.000 abstract description 9
- 239000011159 matrix material Substances 0.000 abstract description 2
- 230000000007 visual effect Effects 0.000 abstract description 2
- 238000012706 support-vector machine Methods 0.000 description 6
- 230000003044 adaptive effect Effects 0.000 description 4
- 238000013461 design Methods 0.000 description 4
- 238000000605 extraction Methods 0.000 description 4
- 230000005856 abnormality Effects 0.000 description 3
- 239000003990 capacitor Substances 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 238000004088 simulation Methods 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 241001269238 Data Species 0.000 description 1
- 230000002411 adverse Effects 0.000 description 1
- 238000004378 air conditioning Methods 0.000 description 1
- 230000004888 barrier function Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 239000012141 concentrate Substances 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000009472 formulation Methods 0.000 description 1
- 230000001939 inductive effect Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 230000004083 survival effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Economics (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The present invention relates to a kind of online method of discrimination of Switching Power Supply operating status, are first input with collected voltage, electric current and temperature signal, carry out multi-field multi-class fault signature to single sample signal respectively and extract;Secondly, carrying out a point dimension visual analyzing to the fault signature matrix of generation, dimension-reduction treatment is carried out to initial huge feature set using a kind of self-adaptive features screening technique based on average sample class spacing, to improve the precision of follow-up mode identification;Finally, after obtained low-dimensional sensitive features collection is normalized and completes the division of training set and test set, construct failure modes and identification model based on SVM, the Fault Pattern Recognition and disaggregated model that model training generation corresponds to existing fault database are carried out to SVM, then test data set is input in the Fault Pattern Recognition and disaggregated model of training completion, the class label of each test sample is obtained, to complete the online differentiation of intelligence to switch power supply equipment operating status.
Description
Technical field
The present invention relates to a kind of on-line intelligence method of discrimination of novel Switching Power Supply operating status, belong to Switching Power Supply shape
State identification technology field.
Background technique
Switching Power Supply with the advantages such as its high efficiency, wide voltage stabilized range, small size, safe and reliable be widely used in including
In the power supply system of each class of electronic devices such as computer, communication equipment, air-conditioning system.However, by high pressure, high current, high power frequency
Etc. bad working environments influence, the failure rate of Switching Power Supply remains high always.In addition structure is complicated, component for switching power circuit
It forms many kinds of, virtually exacerbates the performance degradation of Switching Power Supply, and increase the difficulty of maintenance maintenance.Therefore, split
The research that the real-time online of powered-down source operating status differentiates becomes of crucial importance.The identification of existing Switching Power Supply operating status is more
It is to trigger by the alarm of power supply itself to realize, on the one hand, lead to subsequent event due to lacking detailed running state information
Barrier positioning becomes extremely difficult with diagnosis;On the other hand, the precision of identification places one's entire reliance upon the performance design parameter of Switching Power Supply,
Erroneous judgement easily occurs or fails to judge.In recent years, related fields researcher is constantly by fault tree, BP neural network, Wavelet Neural Network
The operating status that the methods of network is introduced into Switching Power Supply achieves certain achievement in differentiating.It is noted that switching at present
Power supply operating status method of discrimination is studied either on the etection theory of the characteristic parameter of characterization power failure, or in reality
The identification of existing power supply status mode is theoretically not yet mature.Therefore, there is an urgent need to carry out the differentiation of Switching Power Supply operating status
The online intelligent recognition of Switching Power Supply operating status is realized in the research of correlation theory and method.
Summary of the invention
The purpose of the present invention is seeking a kind of online method of discrimination of Switching Power Supply operating status, this method can utilize input
End electric current, voltage signal differentiate the operating status at Switching Power Supply current time online, to instruct follow-up maintenance plan
Formulation slightly.
The purpose of foregoing invention is realized by the following technical solution:
A kind of online method of discrimination of Switching Power Supply operating status, includes the following steps:
(1) Switching Power Supply input terminal is selected as monitoring point, electric current, the voltage signal of power input is acquired, as opening
Powered-down source monitoring running state data source;
(2) time domain, frequency domain statistical analysis are carried out to collected current signal, extract its time-domain and frequency-domain statistical nature,
Including 10 kinds of Time-domain Statistics features: variance, mean-square value, root-mean-square value, the degree of bias, kurtosis, wave index, the nargin of time domain waveform refer to
Mark, pulse index, peak index and kurtosis index;With 5 kinds of frequency domain statistical natures: mean frequency value, gravity frequency, frequency are square
Root, frequency variance and frequency kurtosis;
(3) further to extract non-linear, the non-stationary property that contain in current signal, current signal is based on
The Time-Frequency Analysis of experience wavelet transformation and empirical mode decomposition, and extract the energy feature of each component signal obtained after decomposition
And it is based on Lempel-Ziv complexity characteristics;
(4) in addition, it is contemplated that input voltage signal is bulk power grid voltage, Opposed Current signal is not vulnerable to load end equipment shape
The influence of morphotype formula is further input with electric current and voltage signal, extracts the dedicated of five kinds of Switching Power Supply operating statuses differentiations
Characteristic parameter, it may be assumed that voltage and the amplitude amplification coefficient of current signal, the related coefficient of voltage and electric current amplitude-versus-frequency curve, electricity
Pressure and the phase difference of current signal, the related coefficient and power factor of voltage and current signal phase-frequency characteristic curve;
(5) it is input with the Switching Power Supply state feature set constructed in step (2)~step (4), devises one kind and be based on
The self-adaptive features screening technique of average sample class spacing carries out dimension-reduction treatment to initial huge feature set, to improve follow-up mode
The precision of identification.The Switching Power Supply state feature set constructed in the step (2)~step (4) is i.e.: when 10 kinds in step (2)
The energy feature for each component signal that domain statistical nature and 5 kinds of frequency domain statistical natures, step (3) are extracted and it is based on Lempel-
5 kinds of specific features vectors of Ziv complexity characteristics and step (4).The self-adaptive features sieve based on average sample class spacing
Choosing method is: divergence S in the class of featurew, class scatter SbAnd in class-class between comprehensive divergence SS calculation expression it is as follows:
C in formula --- the number of malfunction classification;
The total number of N --- sample, wherein NiIndicate the number for belonging to the i-th class sample,
--- the characteristic value of j-th of sample in the i-th class;
--- the mean value of characteristic value in the i-th class;
--- the mean value of characteristic value in all categories.
(6) based on the electric current of historical data concentration different conditions classification, voltage signal, with what is obtained in step (5)
Low-dimensional sensitiveness feature set is input, carries out classification based training to support vector machines (SVM) model, wherein model parameter --- it punishes
Penalty factor C and nuclear parameter σ seeks optimum combination by way of cross validation, and then obtains the support vector machines point that training is completed
Class model, for follow-up test link calling.Switching Power Supply operation state mode principle of classification based on supporting vector machine model
It is: according to structural risk minimization principle in former input space Rd(linear classification) or the former input space are obtained through Nonlinear Mapping
Higher-dimension Hilbert space H (linearly inseparable) in find one and meet classificating requirement and guarantee class interval maximumlly most
Excellent Optimal Separating Hyperplane.
Wherein, history data set refers to that current time pervious input terminal current and voltage signals, these signals are all corresponding
There are the fault condition mode in such circumstances or normal mode of respective Switching Power Supply.
The sample set of a given two class linearly inseparable problemsWherein, xi∈RdIndicate d dimension input to
Amount, yi∈ [± 1] indicates the corresponding category label of sample, and n is sample number.Based on nonlinear mapping functionTo input sample
Data carry out space reflection, and the optimal separating hyper plane constructed is expressed as follows:
ω in formula --- weight vectors;
B --- classification thresholds;
εi——εi>=0 is slack variable.
ω and b decides the position of Optimal Separating Hyperplane.The determination of optimal separating hyper plane can be following convex secondary by solving
Plan optimization problem is realized.
C in formula --- indicate penalty factor, generalization ability and misclassification rate for balanced sort device.
(7) it is input with the power input electric current at current time, voltage signal, successively executes step (2)~(4) and obtain
More analysis domain polymorphic type state feature sets of current status data execute the low-dimensional that step (5) obtain current status data again
Each of the feature set that intrinsic characteristics collection, i.e. step (2)~(4) obtain feature can be in the formula of step (5)The self-adaptive features screening of step (5) is for single feature, so being one by one to feature set during screening
In each feature carry out the calculating of formula (1)~(3), then therefrom select the preferable Partial Feature parameter of Clustering Effect as after
Candidate feature in continuous low-dimensional feature set.And the low-dimensional intrinsic characteristics collection is input to the branch that training is completed in step (6)
It holds in vector machine disaggregated model, test sample can determine the sample class belonging to it according to the position of optimal separating hyper plane,
Obtain the status categories conclusion of current switch power supply.If normal condition, then monitoring process is continued to execute;Otherwise, then according to
Corresponding maintenance maintenance measure is formulated according to the power failure type indicated in output result.
The invention has the following advantages:
1. the present invention is from different angles described the operating status of Switching Power Supply, comprehensive multiple analysis domains are a variety of
Extraction and subsequent stateful pattern recognition and classification of the characteristic condition parameter of type for feature, are conducive to comprehensively utilize
Various aspects information achievees the purpose that ensure algorithm accuracy of identification on the basis of;
2. the self-adaptive features screening technique based on average sample class spacing that the present invention designs, can excavate original height
Dimensional feature concentrates the low-dimensional extrinsic information that contains, be conducive to realize non-sensitive characteristic parameter it is superseded while reduce feature set
Dimension, avoid the generation for inducing " dimension disaster ", and then be conducive to the promotion of succeeding state accuracy of identification;
3. the present invention realizes that the on-line intelligence of Switching Power Supply operating status differentiates using supporting vector machine model, can be with few
It measures sample data and completes the training of disaggregated model, and then effectively determine the current status mode of power supply, shape with higher
State accuracy of identification and efficiency meet the requirement of engineer application.
Detailed description of the invention
Fig. 1 is implementation process block diagram of the invention;
Fig. 2 is the circuit structure block diagram of Switching Power Supply fault simulation experiment porch in the present invention;
Fig. 3 is the time domain waveform of the sample data in the present invention under five kinds of state models of Switching Power Supply;
Fig. 4 is the amplitude-frequency spectrogram of the sample data in the present invention under five kinds of state models of Switching Power Supply;
Fig. 5 is the phase spectrogram of the sample data in the present invention under five kinds of state models of Switching Power Supply;
Fig. 6 is the classification results figure of the test data set after Feature Selection in the present invention.
Specific embodiment
The present invention provides a kind of online method of discrimination of Switching Power Supply operating status, and proposed by the invention to verify
The validity of method, builds that Switching Power Supply as shown in Figure 2 is fault simulation test bed, which simulates respectively including complete
Normally, overload, filter capacitor failure, overtemperature and unnamed abnormality totally five kinds of Switching Power Supply operating statuses, and realize
The acquisition and upload of input terminal electric current, voltage signal under different conditions.As long as can be realized the Switching Power Supply failure of above-mentioned function
Simulator stand may be incorporated for verifying method of the invention.It different will be run according to Fig. 1 and in conjunction with what testing stand provided below
Test data under state is further described technical solution of the present invention and its specific embodiment, but is not limited to
This, all modifying or equivalently replacing the technical solution of the present invention, without departing from the spirit of the technical scheme of the invention and model
It encloses, should all cover within the protection scope of the present invention.
Present embodiments provide a kind of switch combined based on more analysis domain polymorphic type features with supporting vector machine model
Power supply operating status on-line intelligence method of discrimination, specific implementation process are as follows:
(1) acquisition and preliminary analysis of state simulation and corresponding states data.In attached drawing 1 " equipment state signal
Acquisition " corresponds to part realization.
Consider from the feasibility and safety perspective of test, drafted the fault condition mode in such circumstances of four kinds of Switching Power Supplies, wraps
It includes: overload, filter capacitor failure, overtemperature and another unnamed abnormality, and give reality on experiment porch respectively
It is existing, for generating the required data source of following model training.Enable Switching Power Supply respectively with four kinds of fault modes drafting and just
Normal state model operation selectes Switching Power Supply input terminal as monitoring point, acquires Switching Power Supply input current, input voltage in real time
And the parameters such as temperature in chassis and it is stored in host computer designated position.In view of Switching Power Supply is not in data acquisition
The status data of the band model jump process generated with the switching of state model needs pair the adverse effect of following model training
Collected status data carries out preliminary waveform analysis, to select the reset condition data text of really reflection different faults mode
This, final choice goes out electric current, voltage signal under five kinds of states of Switching Power Supply, time domain waveform and spectrogram respectively as Fig. 3~
Shown in Fig. 5.
(2) extraction of state feature and adaptive selection." it is more that multiple domain is carried out to single sample signal respectively in attached drawing 1
Classification fault signature extraction ", " the self-adaptive features screening based on classification distance ", " normalized of fault signature data set
With division " correspond to part realization.
For further from input current under above-mentioned five kinds of state models for collecting Switching Power Supply including normal condition,
The state feature set for containing deeper reflection Switching Power Supply state model, design are extracted in the data sets such as input voltage
More than the 40 kinds of different types of characteristic parameters of different analysis domains, including 10 kinds of Time-domain Statistics features: the variance of time domain waveform, square
Value, root-mean-square value, the degree of bias, kurtosis, wave index, margin index, pulse index, peak index and kurtosis index, 5 kinds of frequency domain systems
Feature: mean frequency value, gravity frequency, frequency root mean square, frequency variance and frequency kurtosis is counted, 2* (n-1) kind is based on empirical mode
Decompose the time-frequency domain energy and Lempel-Ziv complexity characteristics of (EMD).Here n refers to the eigen mode point obtained by EMD
The number of amount, since EMD is a kind of adaptive signal Time-Frequency Analysis Method, n is unknown in advance, and 6 are based on wavelet transformation
(WT) time-frequency domain energy and Lempel-Ziv complexity characteristics, in addition, in order to reflect between current signal and voltage signal
Corresponding relationship, five kinds of additional designs are dedicated for differentiating the characteristic parameter of Switching Power Supply state model, it may be assumed that voltage and electric current
The amplitude amplification coefficient of signal, the related coefficient of voltage and electric current amplitude-versus-frequency curve, voltage and current signal phase difference, electricity
The related coefficient and power factor of pressure and current signal phase-frequency characteristic curve.The corresponding data collection in the case where completing different conditions mode
Feature extraction work after, be every kind of characteristic parameter of entry evaluation to the validity of consequent malfunction pattern recognition and classification, to life
At fault signature matrix carry out point a dimension visual analyzing, by analysis it is found that several characteristic quantities are for different conditions pattern-recognition
Effect it is poor, in line with the thought of the survival of the fittest, devise a kind of self-adaptive features screening side based on average sample class spacing
Method carries out dimension-reduction treatment to initial huge feature set, to improve the precision of follow-up mode identification.It is described to be based between average sample class
Away from self-adaptive features screening technique are as follows:
Divergence S in the class of featurew, class scatter SbAnd in class-class between comprehensive divergence SS calculation expression it is as follows:
C in formula --- the number of malfunction classification;
The total number of N --- sample, wherein NiIndicate the number for belonging to the i-th class sample,
--- the characteristic value of j-th of sample in the i-th class;
--- the mean value of characteristic value in the i-th class;
--- the mean value of characteristic value in all categories.
(3) training and test of Fault Pattern Recognition and disaggregated model.
Firstly, after the low-dimensional sensitive features collection after feature adaptive selection is normalized, with different conditions
Preceding 30 groups of samples in 50 groups of sample datas under mode are training data, and rear 20 samples are test data, and building is based on SVM
Failure modes and identification model training dataset and test data set.Wherein, the training dataset of the present embodiment totally 150
Sample, test data set totally 100 samples.
Secondly, being that input carries out model training to SVM with training sample set, to generate 5 kinds for corresponding to aforementioned analog
The Switching Power Supply Fault Pattern Recognition and disaggregated model of state model, parameter preset C and σ in SVM model pass through cross validation
Mode seek optimal solution.The method for seeking optimal solution is as follows:
The sample set of a given two class linearly inseparable problemsWherein, xi∈RdIndicate d dimension input to
Amount, yi∈ [± 1] indicates the corresponding category label of sample, and n is sample number.Based on nonlinear mapping functionTo input sample
Data carry out space reflection, and the optimal separating hyper plane constructed is expressed as follows:
ω in formula --- weight vectors;
B --- classification thresholds;
εi——εi>=0 is slack variable.
ω and b decides the position of Optimal Separating Hyperplane.The determination of optimal separating hyper plane can be following convex secondary by solving
Plan optimization problem is realized.
C in formula --- indicate penalty factor, generalization ability and misclassification rate for balanced sort device.
Finally, test sample can determine the sample class belonging to it according to the position of optimal separating hyper plane, that is, obtain
The status categories conclusion of current switch power supply.Test data set is input to the Switching Power Supply Fault Pattern Recognition of training completion
In disaggregated model, the class label of each test sample is obtained, it is above-mentioned that acquisition is compared by corresponding physical tags
The classification results of the measuring accuracy of Switching Power Supply Fault Pattern Recognition and disaggregated model, test set are as shown in Figure 6.In Fig. 6, mark
Remember that the classification output result of lap representative model is identical as the concrete class label of test sample, that is, is representing identification completely just
Really, non-overlap part represents identification mistake.It will be appreciated from fig. 6 that after feature adaptive selection, the state recognition precision of test set
It greatly improves, improves to 73%, wherein overload identifies correctly completely, and filter capacitor failure only goes out with unnamed abnormality
Existing one group of misrecognition sample.Thus validity of the method for the present invention in identification switch power supply operating status is demonstrated.
Claims (1)
1. a kind of online recognition method of Switching Power Supply operating status, which is characterized in that this method comprises the following steps:
(1) Switching Power Supply input terminal is selected as monitoring point, acquires electric current, the voltage signal of power input, as switch electricity
Source monitoring running state data source;
(2) time domain, frequency domain statistical analysis are carried out to collected current signal, extracts its time-domain and frequency-domain statistical nature, including
10 kinds of Time-domain Statistics features: the variance of time domain waveform, mean-square value, root-mean-square value, the degree of bias, kurtosis, wave index, margin index,
Pulse index, peak index and kurtosis index;With 5 kinds of frequency domain statistical natures: mean frequency value, gravity frequency, frequency root mean square, frequency
Rate variance and frequency kurtosis;
(3) further to extract non-linear, the non-stationary property that contain in current signal, current signal is carried out based on experience
The Time-Frequency Analysis of wavelet transformation and empirical mode decomposition, and extract decompose after the energy feature of each component signal that obtains and
Based on Lempel-Ziv complexity characteristics;
(4) in addition, it is contemplated that input voltage signal is bulk power grid voltage, Opposed Current signal is not vulnerable to load end equipment state mould
The influence of formula is further input with electric current and voltage signal, extracts the specific features of five kinds of Switching Power Supply operating statuses differentiation
Parameter, it may be assumed that voltage and the amplitude amplification coefficient of current signal, the related coefficient of voltage and electric current amplitude-versus-frequency curve, voltage with
The related coefficient and power factor of the phase difference of current signal, voltage and current signal phase-frequency characteristic curve;
It (5) is input with the Switching Power Supply state feature set constructed in step (2)~step (4), using based on average sample class
The self-adaptive features screening technique of spacing carries out dimension-reduction treatment to initial huge feature set, to improve the essence of follow-up mode identification
Degree;The self-adaptive features screening technique based on average sample class spacing are as follows: divergence S in the class of featurew, class scatter SbAnd
In class-class between comprehensive divergence SS calculation expression it is as follows:
C in formula --- the number of malfunction classification;
The total number of N --- sample, wherein NiIndicate the number for belonging to the i-th class sample,
--- the characteristic value of j-th of sample in the i-th class;
--- the mean value of characteristic value in the i-th class;
--- the mean value of characteristic value in all categories.
(6) based on the electric current of historical data concentration different conditions classification, voltage signal, with the low-dimensional obtained in step (5)
Sensitiveness feature set is input, carries out classification based training to supporting vector machine model, wherein model parameter --- penalty factor with
Nuclear parameter σ seeks optimum combination by way of cross validation, and then obtains the support vector cassification model that training is completed, with
It is called for follow-up test link.Switching Power Supply operation state mode principle of classification based on supporting vector machine model is: according to knot
Structure risk minimization principle is in former input space RdThe higher-dimension that (linear classification) or the former input space are obtained through Nonlinear Mapping
One is found in the space Hilbert H (linearly inseparable) to meet classificating requirement and guarantee the maximized optimal classification in class interval
Hyperplane.
The sample set of a given two class linearly inseparable problemsWherein, xi∈RdIndicate d dimensional input vector, yi∈
[± 1] the corresponding category label of sample is indicated, n is sample number.Based on nonlinear mapping functionTo input sample data into
Row space reflection, the optimal separating hyper plane constructed are expressed as follows:
ω in formula --- weight vectors;
B --- classification thresholds;
εi——εi>=0 is slack variable.
ω and b decides the position of Optimal Separating Hyperplane.The determination of optimal separating hyper plane can be by solving following convex quadratic programming
Optimization problem is realized.
C in formula --- indicate penalty factor, generalization ability and misclassification rate for balanced sort device.
(7) it is input with the power input electric current at current time, voltage signal, successively executes step (2)~(4) and obtain currently
More analysis domain polymorphic type state feature sets of status data, the low-dimensional for executing step (5) acquisition current status data again are intrinsic
Feature set, and be input in the support vector cassification model that training is completed in step (6), it is final to obtain to current switch
The differentiation result of power supply operating status.If normal condition, then monitoring process is continued to execute;Otherwise, then according in output result
The power failure type of instruction formulates corresponding maintenance maintenance measure.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910689142.6A CN110503004B (en) | 2019-07-29 | 2019-07-29 | On-line judging method for operating state of switching power supply |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910689142.6A CN110503004B (en) | 2019-07-29 | 2019-07-29 | On-line judging method for operating state of switching power supply |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110503004A true CN110503004A (en) | 2019-11-26 |
CN110503004B CN110503004B (en) | 2022-03-22 |
Family
ID=68587668
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910689142.6A Active CN110503004B (en) | 2019-07-29 | 2019-07-29 | On-line judging method for operating state of switching power supply |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110503004B (en) |
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111308185A (en) * | 2020-02-28 | 2020-06-19 | 宁波三星医疗电气股份有限公司 | Non-invasive load identification method |
CN111539374A (en) * | 2020-05-07 | 2020-08-14 | 上海工程技术大学 | Rail train bearing fault diagnosis system and method based on multidimensional data space |
CN111856209A (en) * | 2020-07-23 | 2020-10-30 | 广东电网有限责任公司清远供电局 | Power transmission line fault classification method and device |
CN111859815A (en) * | 2020-07-31 | 2020-10-30 | 中国汽车工程研究院股份有限公司 | Pattern clustering method and accident feature identification technology for battery alarm feature data |
CN112486096A (en) * | 2020-12-09 | 2021-03-12 | 中国兵器装备集团自动化研究所 | Machine tool operation state monitoring method |
CN113255771A (en) * | 2021-05-26 | 2021-08-13 | 中南大学 | Fault diagnosis method and system based on multi-dimensional heterogeneous difference analysis |
CN113255795A (en) * | 2021-06-02 | 2021-08-13 | 杭州安脉盛智能技术有限公司 | Equipment state monitoring method based on multi-index cluster analysis |
CN116381542A (en) * | 2023-06-05 | 2023-07-04 | 深圳和润达科技有限公司 | Health diagnosis method and device of power supply equipment based on artificial intelligence |
CN116433226A (en) * | 2023-06-14 | 2023-07-14 | 浙江亿视电子技术有限公司 | Alternating current power supply equipment management method and system based on electrical parameter AI analysis |
CN116505738A (en) * | 2023-06-26 | 2023-07-28 | 易充新能源(深圳)有限公司 | Control method and system for energy-saving consumption-reducing power supply |
CN116824512A (en) * | 2023-08-28 | 2023-09-29 | 西华大学 | 27.5kV visual grounding disconnecting link state identification method and device |
CN115169405B (en) * | 2022-07-14 | 2024-02-02 | 北京威控科技股份有限公司 | Hotel guest room equipment fault diagnosis method and system based on support vector machine |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101783578A (en) * | 2010-02-03 | 2010-07-21 | 北京奥福瑞科技有限公司 | Intelligent online detection optimizing management control method of high-frequency switch power supply and device thereof |
CN101833056A (en) * | 2010-03-26 | 2010-09-15 | 中国电力科学研究院 | Method for diagnosing deformation of transformer winding based on frequency response characteristics |
CN102608545A (en) * | 2012-03-01 | 2012-07-25 | 西安电子科技大学 | Non-contact switch power failure diagnosis system |
CN103235188A (en) * | 2013-05-02 | 2013-08-07 | 合肥工业大学 | Method for measuring and predicting capacitor ESR (Equivalent Series Resistance) values of switching power supplies on line |
CN103234767A (en) * | 2013-04-21 | 2013-08-07 | 蒋全胜 | Nonlinear fault detection method based on semi-supervised manifold learning |
CN105095675A (en) * | 2015-09-07 | 2015-11-25 | 浙江群力电气有限公司 | Switch cabinet fault feature selection method and apparatus |
CN107086545A (en) * | 2017-06-14 | 2017-08-22 | 扬州万泰电子科技有限公司 | A kind of alternating-current charging pile intelligent electric energy meter Switching Power Supply and its method of work |
CN107590506A (en) * | 2017-08-17 | 2018-01-16 | 北京航空航天大学 | A kind of complex device method for diagnosing faults of feature based processing |
CN108764265A (en) * | 2018-03-26 | 2018-11-06 | 海南电网有限责任公司电力科学研究院 | A kind of method for diagnosing faults based on algorithm of support vector machine |
-
2019
- 2019-07-29 CN CN201910689142.6A patent/CN110503004B/en active Active
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101783578A (en) * | 2010-02-03 | 2010-07-21 | 北京奥福瑞科技有限公司 | Intelligent online detection optimizing management control method of high-frequency switch power supply and device thereof |
CN101833056A (en) * | 2010-03-26 | 2010-09-15 | 中国电力科学研究院 | Method for diagnosing deformation of transformer winding based on frequency response characteristics |
CN102608545A (en) * | 2012-03-01 | 2012-07-25 | 西安电子科技大学 | Non-contact switch power failure diagnosis system |
CN103234767A (en) * | 2013-04-21 | 2013-08-07 | 蒋全胜 | Nonlinear fault detection method based on semi-supervised manifold learning |
CN103235188A (en) * | 2013-05-02 | 2013-08-07 | 合肥工业大学 | Method for measuring and predicting capacitor ESR (Equivalent Series Resistance) values of switching power supplies on line |
CN105095675A (en) * | 2015-09-07 | 2015-11-25 | 浙江群力电气有限公司 | Switch cabinet fault feature selection method and apparatus |
CN107086545A (en) * | 2017-06-14 | 2017-08-22 | 扬州万泰电子科技有限公司 | A kind of alternating-current charging pile intelligent electric energy meter Switching Power Supply and its method of work |
CN107590506A (en) * | 2017-08-17 | 2018-01-16 | 北京航空航天大学 | A kind of complex device method for diagnosing faults of feature based processing |
CN108764265A (en) * | 2018-03-26 | 2018-11-06 | 海南电网有限责任公司电力科学研究院 | A kind of method for diagnosing faults based on algorithm of support vector machine |
Non-Patent Citations (2)
Title |
---|
张金格: ""开关电源故障诊断方法研究"", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
章叔昌: ""基于LabVIEW 的变压器绕组变形测试仪"", 《江西电力》 * |
Cited By (16)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111308185A (en) * | 2020-02-28 | 2020-06-19 | 宁波三星医疗电气股份有限公司 | Non-invasive load identification method |
CN111539374A (en) * | 2020-05-07 | 2020-08-14 | 上海工程技术大学 | Rail train bearing fault diagnosis system and method based on multidimensional data space |
CN111856209A (en) * | 2020-07-23 | 2020-10-30 | 广东电网有限责任公司清远供电局 | Power transmission line fault classification method and device |
CN111859815A (en) * | 2020-07-31 | 2020-10-30 | 中国汽车工程研究院股份有限公司 | Pattern clustering method and accident feature identification technology for battery alarm feature data |
CN111859815B (en) * | 2020-07-31 | 2023-05-23 | 中国汽车工程研究院股份有限公司 | Mode clustering method of battery alarm feature data and accident feature recognition technology |
CN112486096A (en) * | 2020-12-09 | 2021-03-12 | 中国兵器装备集团自动化研究所 | Machine tool operation state monitoring method |
CN113255771A (en) * | 2021-05-26 | 2021-08-13 | 中南大学 | Fault diagnosis method and system based on multi-dimensional heterogeneous difference analysis |
CN113255795A (en) * | 2021-06-02 | 2021-08-13 | 杭州安脉盛智能技术有限公司 | Equipment state monitoring method based on multi-index cluster analysis |
CN115169405B (en) * | 2022-07-14 | 2024-02-02 | 北京威控科技股份有限公司 | Hotel guest room equipment fault diagnosis method and system based on support vector machine |
CN116381542A (en) * | 2023-06-05 | 2023-07-04 | 深圳和润达科技有限公司 | Health diagnosis method and device of power supply equipment based on artificial intelligence |
CN116381542B (en) * | 2023-06-05 | 2023-07-25 | 深圳和润达科技有限公司 | Health diagnosis method and device of power supply equipment based on artificial intelligence |
CN116433226A (en) * | 2023-06-14 | 2023-07-14 | 浙江亿视电子技术有限公司 | Alternating current power supply equipment management method and system based on electrical parameter AI analysis |
CN116505738A (en) * | 2023-06-26 | 2023-07-28 | 易充新能源(深圳)有限公司 | Control method and system for energy-saving consumption-reducing power supply |
CN116505738B (en) * | 2023-06-26 | 2024-01-16 | 易充新能源(深圳)有限公司 | Control method and system for energy-saving consumption-reducing power supply |
CN116824512A (en) * | 2023-08-28 | 2023-09-29 | 西华大学 | 27.5kV visual grounding disconnecting link state identification method and device |
CN116824512B (en) * | 2023-08-28 | 2023-11-07 | 西华大学 | 27.5kV visual grounding disconnecting link state identification method and device |
Also Published As
Publication number | Publication date |
---|---|
CN110503004B (en) | 2022-03-22 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110503004A (en) | A kind of online method of discrimination of Switching Power Supply operating status | |
US20210117770A1 (en) | Power electronic circuit troubleshoot method based on beetle antennae optimized deep belief network algorithm | |
US11549985B2 (en) | Power electronic circuit fault diagnosis method based on extremely randomized trees and stacked sparse auto-encoder algorithm | |
CN106682303B (en) | A kind of three-level inverter method for diagnosing faults based on empirical mode decomposition and decision tree RVM | |
CN103076547B (en) | Method for identifying GIS (Gas Insulated Switchgear) local discharge fault type mode based on support vector machines | |
CN109974782B (en) | Equipment fault early warning method and system based on big data sensitive characteristic optimization selection | |
CN102707256B (en) | Fault diagnosis method based on BP-Ada Boost nerve network for electric energy meter | |
CN109444656A (en) | A kind of inline diagnosis method of deformation of transformer winding position | |
CN103033362A (en) | Gear fault diagnosis method based on improving multivariable predictive models | |
CN104849633A (en) | Switchgear partial discharge mode recognition method | |
CN112383052A (en) | Power grid fault repairing method and device based on power internet of things | |
CN109165604A (en) | The recognition methods of non-intrusion type load and its test macro based on coorinated training | |
CN107817404A (en) | A kind of Portable metering automatization terminal trouble-shooter and its diagnostic method | |
CN109284672A (en) | A kind of Mechanical Failure of HV Circuit Breaker diagnostic method based on PSO-Kmeans algorithm | |
CN103968937A (en) | Method for diagnosing mechanical states of distribution switch on basis of EMD sample entropy and FCM | |
CN116683648B (en) | Intelligent power distribution cabinet and control system thereof | |
CN104966161A (en) | Electric energy quality recording data calculating analysis method based on Gaussian mixture model | |
CN101738998A (en) | System and method for monitoring industrial process based on local discriminatory analysis | |
CN106344004A (en) | Electrocardiosignal feature point detecting method and device | |
CN103018537B (en) | The Classification of Transient Power Quality Disturbances recognition methods of kurtosis is composed based on CWD | |
CN110118928A (en) | A kind of circuit breaker failure diagnostic method based on Back Propagation Algorithm | |
CN111476299A (en) | Improved convolutional neural network and power grid intelligent alarm system based on same | |
CN109829627A (en) | A kind of safe confidence appraisal procedure of Electrical Power System Dynamic based on integrated study scheme | |
CN109039280A (en) | Diagnosing failure of photovoltaic array method based on non-primary component data characteristics | |
CN105137354A (en) | Motor fault detection method based on nerve network |
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 |