CN110084106A - Microgrid inverter method for diagnosing faults based on wavelet transformation and probabilistic neural network - Google Patents
Microgrid inverter method for diagnosing faults based on wavelet transformation and probabilistic neural network Download PDFInfo
- Publication number
- CN110084106A CN110084106A CN201910206025.XA CN201910206025A CN110084106A CN 110084106 A CN110084106 A CN 110084106A CN 201910206025 A CN201910206025 A CN 201910206025A CN 110084106 A CN110084106 A CN 110084106A
- Authority
- CN
- China
- Prior art keywords
- fault
- neural network
- probabilistic neural
- phase
- feature vector
- 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
- 238000000034 method Methods 0.000 title claims abstract description 49
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 39
- 230000009466 transformation Effects 0.000 title claims abstract description 27
- 238000012549 training Methods 0.000 claims abstract description 34
- 238000003745 diagnosis Methods 0.000 claims abstract description 24
- 238000004088 simulation Methods 0.000 claims abstract description 24
- 238000009499 grossing Methods 0.000 claims abstract description 20
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 14
- 238000000605 extraction Methods 0.000 claims abstract description 10
- 230000002068 genetic effect Effects 0.000 claims abstract description 10
- 238000004458 analytical method Methods 0.000 claims abstract description 9
- 238000003062 neural network model Methods 0.000 claims description 34
- 210000002569 neuron Anatomy 0.000 claims description 15
- 210000000349 chromosome Anatomy 0.000 claims description 12
- 238000012360 testing method Methods 0.000 claims description 12
- 230000008569 process Effects 0.000 claims description 9
- 108090000623 proteins and genes Proteins 0.000 claims description 9
- 230000035772 mutation Effects 0.000 claims description 6
- 238000010606 normalization Methods 0.000 claims description 6
- 238000011217 control strategy Methods 0.000 claims description 5
- 239000000284 extract Substances 0.000 claims description 4
- 238000012545 processing Methods 0.000 claims description 4
- NAWXUBYGYWOOIX-SFHVURJKSA-N (2s)-2-[[4-[2-(2,4-diaminoquinazolin-6-yl)ethyl]benzoyl]amino]-4-methylidenepentanedioic acid Chemical compound C1=CC2=NC(N)=NC(N)=C2C=C1CCC1=CC=C(C(=O)N[C@@H](CC(=C)C(O)=O)C(O)=O)C=C1 NAWXUBYGYWOOIX-SFHVURJKSA-N 0.000 claims description 3
- 230000005540 biological transmission Effects 0.000 claims description 3
- 230000001186 cumulative effect Effects 0.000 claims description 3
- 238000000354 decomposition reaction Methods 0.000 claims description 3
- 238000013213 extrapolation Methods 0.000 claims description 3
- 230000001960 triggered effect Effects 0.000 claims description 3
- 239000000203 mixture Substances 0.000 claims description 2
- 125000002619 bicyclic group Chemical group 0.000 claims 1
- 238000005457 optimization Methods 0.000 abstract description 4
- 230000009286 beneficial effect Effects 0.000 abstract description 2
- 230000000153 supplemental effect Effects 0.000 abstract description 2
- 238000010276 construction Methods 0.000 abstract 1
- 230000006870 function Effects 0.000 description 10
- 238000010248 power generation Methods 0.000 description 4
- 230000005611 electricity Effects 0.000 description 3
- 238000012706 support-vector machine Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 239000003245 coal Substances 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000004146 energy storage Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 239000003208 petroleum Substances 0.000 description 1
- 238000011160 research Methods 0.000 description 1
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
- G01R31/50—Testing of electric apparatus, lines, cables or components for short-circuits, continuity, leakage current or incorrect line connections
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/047—Probabilistic or stochastic networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/12—Computing arrangements based on biological models using genetic models
- G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/02—Preprocessing
- G06F2218/04—Denoising
- G06F2218/06—Denoising by applying a scale-space analysis, e.g. using wavelet analysis
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biophysics (AREA)
- General Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Physiology (AREA)
- Genetics & Genomics (AREA)
- Probability & Statistics with Applications (AREA)
- Testing Electric Properties And Detecting Electric Faults (AREA)
Abstract
The present invention provides a kind of microgrid inverter method for diagnosing faults based on wavelet transformation and probabilistic neural network, it include: the fault simulation system for building microgrid inverter in simulation software first, then the threephase load electric current under different faults type is acquired as measuring signal, its fault characteristic signals is extracted using wavelet multi_resolution analysis method;In turn by the fault characteristic signals normalized construction feature vector of extraction;The fault feature vector of extraction is finally chosen a part as training data to be trained PNN model, the optimal smoothing of probabilistic neural network is found because of period of the day from 11 p.m. to 1 a.m σ using genetic algorithm simultaneously, when inverter breaks down, current signal is subjected to fault location according to the above method, to realize microgrid inverter fault diagnosis.The beneficial effects of the present invention are: structural model is simple, be easy determination, fast convergence rate and converges on Bayes optimization solution, sample supplemental capabilities are strong, do not need re -training, and precision is high, practicality is strong, and are easy to engineering combination.
Description
Technical field
The present invention relates to power electronics fault diagnosis fields, more particularly to one kind to be based on wavelet transformation and probabilistic neural network
Microgrid inverter method for diagnosing faults.
Background technique
In today that global economy, rapid technological growth and industrially scalable are growing.Demand of the people to the energy
It is growing day by day.Seriously polluted with the continuous failure of the traditional energies such as coal, petroleum and caused by environment, countries in the world start by
Gradually recognize the importance of energy and environment.Therefore the clean energy resourcies such as solar energy, wind energy are greatly developed.Distributed power generation is as clear
The Processes For Effective Conversion of the clean energy, the low-pressure system that micro-capacitance sensor is made of distributed generation resource, energy storage, load etc..Micro-capacitance sensor conduct
The important bridge and tie of bulk power grid and distributed power generation, can effectively reduce distributed power generation can to the impact raising power supply of power grid
By property.And power conversion and power control are mainly undertaken by inverter in micro-capacitance sensor.If inverter breaks down and cannot examine
Disconnected and reparation, will cause the economic loss and security risk that can not be retrieved surely.Therefore inverter is safe and stable, reliability service pair
It is most important in electric system.
Fault diagnosis is one kind of pattern-recognition in fact, and research main contents are to excavate, believe the reason of realization to failure
Breath acquisition, feature extraction and the analysis of system mode and identification.The information collection present invention uses Matlab/Simulink module
In information acquisition module, information characteristics, which extract, mainly has Fourier transformation, wavelet transformation, principle component analysis etc. at present.Failure
Diagnosis positioning mainly has the methods of expert system approach, artificial neural network, support vector machines.Microgrid inverter switching device failure
Type is complicated, and traditional artificial search has been unable to meet the quick universal of extensive micro-capacitance sensor, with artificial intelligence, computer skill
The fast development of art, signal processing technology.Thought based on data-driven reflects operating status using invertor operation in the process
The data such as continually changing current signal carry out data analysis and feature extraction, can orient fault of converter real-time, quickly
Position.
The present invention extracts its feature vector using wavelet transformation, then determines optimal smoothing factor sigma band using genetic algorithm
Enter PNN network model and carries out fault location.Small echo can extract the time domain of signal and the important information of frequency domain simultaneously, and can be with
Local message is amplified and is added certain noise with signal interference present in object simulating, referred to as " mathematics is micro-
Mirror ".And probabilistic neural network is simple since network architecture determines, diagnosis rate is fast, is completed with linear learning algorithm non-thread
The work that inquiry learning algorithm is done.Not only stability is strong, but also converges on Bayes Optimum solution and meet the requirement handled in real time.
Summary of the invention
In order to solve microgrid inverter switching device fault type complexity, traditional artificial lookup has been unable to meet extensive
The quickly universal problem of micro-capacitance sensor, the microgrid inverter based on wavelet transformation and probabilistic neural network that the present invention provides a kind of
Method for diagnosing faults mainly comprises the steps that
S101: according to the operating condition of microgrid inverter to be diagnosed, the fault simulation system of the microgrid inverter is built
System, and threephase load current signal of the fault simulation system under different faults type is acquired as measuring signal;It is described
Fault type is to be numbered according to the preset various faults of failure that may occur;
S102: extracting the fault characteristic signals of the measuring signal using wavelet multi_resolution analysis method, and by extraction
Fault characteristic signals are normalized and are added noise processed, obtain fault sample data;
S103: probabilistic neural network model is built, and using the fault sample data to the probabilistic neural network mould
Type is trained, and obtains trained probabilistic neural network model;Wherein, in the training process of the probabilistic neural network model
In, the optimal smoothing factor of the probabilistic neural network model is found using genetic algorithm;
S104: the physical fault for the microgrid inverter for treating diagnosis using trained probabilistic neural network model is examined
It is disconnected, realize the microgrid inverter fault diagnosis based on wavelet transformation and probabilistic neural network.
Further, it in step S101, is taken on Matlab/Simulink platform using microgrid inverter PQ control strategy
Build fault simulation system, wherein three-phase grid-connected inverter in the fault simulation system uses SVPWM method, and
Fault simulation module is added after pulse-triggered module in the fault simulation system, and then micro- by control trigger pulse simulation
The open-circuit fault of net inverter;It is specific as follows:
S201: the three-phase voltage and three-phase current of grid side are acquired, and then is converted by Clarke and Park by three-phase electricity
Pressure and three-phase current are converted into the current signal under the voltage signal under dq axis and dq axis;
S202: using under dq axis voltage signal and current signal as the input of phaselocked loop, obtained by PHASE-LOCKED LOOP PLL TECHNIQUE
The phase theta of voltage and current;
S203: and then by the voltage and current transmission of phase of acquisition to microgrid central controller;Microgrid central controller root
Power instruction P is issued according to the voltage and current phase receivedrefAnd QrefAnd generate current-order IdrefAnd Iqref, and pass through
Double -loop control structure and Park inverse transformation obtain Uαβ;
S204: the U that will be obtainedαβWith the DC current U of piconet controllerdcAs the input of SVPWM, modulated by SVPWM
Generate driving pulse of 6 pulses as microgrid inverter;A failure trigger module is finally added, by the driving pulse
As the input of failure trigger module, a microgrid inverter fault simulation system is generated;The failure trigger module passes through control
The pulse of switching tube processed carrys out the shutdown of analog switch pipe.
Further, in step S102, the fault signature of the measuring signal is extracted using wavelet multi_resolution analysis method
Signal, and the fault characteristic signals of extraction are normalized and are added noise processed, specific steps include:
S301: it selects 5 layers of decomposition method of db3 small echo to decompose the measuring signal, obtains the institute of the measuring signal
There is frequency component Sph|j;Wherein, ph is phase serial number, value A, B or C;J is the corresponding frequency range serial number of each phase, takes and 0 is
0~125Hz, taking 1 is 126~250Hz;
S302: selecting frequency component Sph|jLow frequency part coefficient carry out energy reconstruct, obtain frequency component Sph|jIt is corresponding
ENERGY Eph|j, shown in reconstruction formula such as formula (1):
In above formula, xph|jkIndicate frequency component Sph|jDiscrete point amplitude, n be frequency component Sph|jDiscrete point it is total
Number;
S303: the vector that the low-frequency range energy of three-phase is formed is as fault feature vector T, as shown in formula (2):
T=[EA|0,EA|1,EB|0,EB|1,EC|0,EC|1] (2)
In above formula, EA|0、EA|1、EB|0、EB|1、EC|0And EC|1The respectively energy of 0~125Hz frequency band of A phase, A phase
The energy of 126~250Hz frequency band, the energy of 0~125Hz frequency band of B phase, B phase 126~250Hz frequency band energy,
The energy of the energy of 0~125Hz frequency band of C phase, 126~250Hz frequency band of C phase;
S304: the normalized as shown in formula (3) is carried out to fault feature vector T:
Ej=EA|j+EB|j+EC|j (3)
Fault feature vector T1 after being normalized, shown in the fault feature vector T1 such as formula (4) after normalization:
S305: the random noise of %5 is added in fault feature vector T1 after normalization, with letters other in object simulating
Number interference, obtain be added random noise after fault feature vector T2;
Under different faults type, multiple measuring signals are acquired respectively, and each measuring signal progress is above-mentioned
After step S301 to S305 process processing, corresponding fault feature vector T2 is obtained, forms fault sample data.
Further, in step S103, the probabilistic neural network model is by input layer, mode layer, summation layer and output
4 layers of composition of layer;
Input layer is used to receive the value of fault feature vector T2 in the fault sample data, by the fault signature to
Amount T2 passes to probabilistic neural network;The dimension of the neuron number of the input layer and the fault feature vector T2 are equal;
Mode layer is used to calculate the matching relationship of each fault type in input fault feature vector T2 and fault sample data,
Shown in calculation formula such as formula (5):
In above formula, WiThe weight connected for input layer to mode layer;X is the corresponding value of fault feature vector T2 of input;δ
For smoothing factor;The number of the mode layer neuron is equal to the sum of the sample number of corresponding fault type, and i is whole greater than 0
Number represents the serial number of corresponding neuron;
Summation layer is used to belong to the Cumulative probability of some fault type, so that the estimated probability for obtaining each fault type is close
Function is spent, probability density function is obtained by Parzen method, as shown in formula (6):
In above formula, XaiFor fault type θAThe corresponding fault feature vector T2 of i-th of training sample;M is fault type
θATotal training sample number;P is probability, is preset value;The value of δ has determined the bell curve centered on sample point
Width;The neuron number of the summation layer is equal with fault type sum;
Output layer is made of thresholding discriminator, is had for selecting one in the estimated probability density of each fault type
The neuron of maximum estimated probability density exports fault type as output.
Further, in step S103, when being trained using the fault sample data to probabilistic neural network model:
The fault sample data are divided into training data and test data two parts;Using training data to probabilistic neural network model
It is trained, to modify the expansion rate of radial basis function in probabilistic neural network model, and utilizes something lost during training
Propagation algorithm finds the optimal smoothing factor of the probabilistic neural network, obtains optimum probability neural network model;Test set is used for
Input probability neural network model, tests it, sees whether prediction output result is consistent with reality output result, is examined
The disconnected indexs such as rate and precision;Continuous circuit training and test, until precision reaches preset value or the number of iterations reaches maximum
When setting value, stop iteration, and using probabilistic neural network model at this time as trained probabilistic neural network model.
Further, in step S103, the optimal smoothing factor of the probabilistic neural network is found using genetic algorithm, is had
Body includes:
Smoothing factor is encoded to chromosome and forms initial population, selected, intersected, made a variation by calculating fitness
With update iteration, to obtain the i.e. optimal smoothing factor of optimal chromosome;
Fitness is calculated by fitness function, then selected, intersected and is made a variation on this basis;Wherein fitness
Shown in function such as formula (7);
In above formula, M is the number of samples of corresponding training data, and i is the classification number of training classification (fault type), and I is class
Not sum;yim^ and yimThe predicted value and actual value of respectively m-th sample, a is a positive minimum, in order to avoid denominator
It is 0.
Selection operation selects roulette method, the i.e. selection strategy based on fitness ratio, the select probability p of each individual ii
Such as formula (8):
In above formula,FiFor the fitness value of individual i, k is coefficient, is preset value;N be population at individual number, i and
The value range of j is identical, is [1, N];
Crossover operation uses real number interior extrapolation method, k-th of Autosome akWith first of chromosome a1J crossover operation sides
Shown in method such as formula (9):
In above formula, akjFor the chromosome generated after crossover operation, random number of the b between [0,1];
Mutation operation chooses j-th of gene a of i-th of individualijIt makes a variation, shown in mutation operation such as formula (10):
In above formula, amaxFor gene aijThe upper bound, aminFor gene aijLower bound;F (g)=r2(1-g/Gmax)2;r2It is one
A random number;G is current the number of iterations;GmaxFor maximum evolution number;Random number of the r between [0,1].
Technical solution provided by the invention has the benefit that common BP neural network classifying quality is initial by it
Weight and model structure are affected, classification results lack transparency and are more slowly easy to fall into part using pace of learning
Optimal solution.And not only structural model is simple, is easy determination, fast convergence rate and converges on for technical solution proposed by the invention
Bayes optimization solution, sample supplemental capabilities are strong, do not need re -training, and precision is high, practicality is strong, and are easy to engineering combination.
Detailed description of the invention
Present invention will be further explained below with reference to the attached drawings and examples, in attached drawing:
Fig. 1 is a kind of microgrid inverter fault diagnosis based on wavelet transformation and probabilistic neural network in the embodiment of the present invention
The flow chart of method;
Fig. 2 is microgrid inverter rotating coordinate system PQ control principle drawing in the embodiment of the present invention;
Grid side three-phase current fault waveform figure when Fig. 3 is partial switch pipe failure in the embodiment of the present invention;
Fig. 4 is 1 failure A electric current wavelet transformation exploded view of switching tube in the embodiment of the present invention;
Fig. 5 is probabilistic neural network basic model figure in the embodiment of the present invention;
Fig. 6 is genetic algorithm optimization probabilistic neural network smoothing factor σ flow chart in the embodiment of the present invention;
Fig. 7 (a)~Fig. 7 (c) is fault diagnosis result figure in part in the embodiment of the present invention.
Specific embodiment
For a clearer understanding of the technical characteristics, objects and effects of the present invention, now control attached drawing is described in detail
A specific embodiment of the invention.
The embodiment provides a kind of microgrid inverter failures based on wavelet transformation and probabilistic neural network to examine
Disconnected method.
Referring to FIG. 1, Fig. 1 is a kind of microgrid inversion based on wavelet transformation and probabilistic neural network in the embodiment of the present invention
The flow chart of device method for diagnosing faults, specifically comprises the following steps:
S101: according to the operating condition of microgrid inverter to be diagnosed, the fault simulation system of the microgrid inverter is built
System, and threephase load current signal of the fault simulation system under different faults type is acquired as measuring signal;It is described
Fault type is to be numbered according to the preset various faults of failure that may occur;
S102: extracting the fault characteristic signals of the measuring signal using wavelet multi_resolution analysis method, and by extraction
Fault characteristic signals are normalized and are added noise processed, obtain fault sample data;
S103: probabilistic neural network model is built, and using the fault sample data to the probabilistic neural network mould
Type is trained, and obtains trained probabilistic neural network model;Wherein, in the training process of the probabilistic neural network model
In, the optimal smoothing factor of the probabilistic neural network model is found using genetic algorithm;
S104: the physical fault for the microgrid inverter for treating diagnosis using trained probabilistic neural network model is examined
It is disconnected, realize the microgrid inverter fault diagnosis based on wavelet transformation and probabilistic neural network.
In step S101, failure mould is built on Matlab/Simulink platform using microgrid inverter PQ control strategy
Quasi- system, wherein the three-phase grid-connected inverter in the fault simulation system uses SVPWM method, and in the failure
Fault simulation module is added after pulse-triggered module in simulation system, and then microgrid inverter is simulated by control trigger pulse
Open-circuit fault;Using fault simulation system analog switch pipe failure generate fault-signal, and be saved into working space into
Row saves;It is illustrated in figure 3 grid side three-phase current fault waveform figure when partial switch pipe failure.
The microgrid inverter PQ control strategy is the common inverter control of the new-energy grid-connecteds such as photovoltaic power generation, wind-power electricity generation
Mode processed, Fig. 2 is microgrid inverter rotating coordinate system PQ control (microgrid inverter PQ control strategy) schematic diagram, according to its topology
Structure builds fault simulation system on Matlab/Simulink platform, specific as follows:
S201: the three-phase voltage (U of grid side is acquireda、UbAnd Uc) and three-phase current (Ia、IbAnd Ic), and then pass through
Clarke and Park converts the voltage signal (U being converted into three-phase voltage and three-phase current under dq axisdAnd Uq) and dq axis under electricity
Flow signal (Id、Iq);
S202: using under dq axis voltage signal and current signal as the input of phaselocked loop, obtained by PHASE-LOCKED LOOP PLL TECHNIQUE
The phase theta of voltage and current;
S203: and then by the voltage and current transmission of phase of acquisition to microgrid central controller;Microgrid central controller root
Power instruction P is issued according to the voltage and current phase receivedrefAnd QrefAnd generate current-order IdrefAnd Iqref, and pass through
Double -loop control structure and Park inverse transformation obtain Uαβ;
S204: the U that will be obtainedαβWith the DC current U of piconet controllerdcAs the input of SVPWM, modulated by SVPWM
Generate driving pulse of 6 pulses as microgrid inverter;A failure trigger module is finally added, by the driving pulse
As the input of failure trigger module, a microgrid inverter fault simulation system is generated;The failure trigger module passes through control
The pulse of switching tube processed carrys out the shutdown of analog switch pipe.
Since switching tube short trouble finally can also be converted into open-circuit fault (route protection), in the present invention only
Consider switching tube open-circuit fault, fault type is (5 major class, 22 groups) as follows:
1) all switching tubes operate normally, switching tube fault-free;
2) single switching tube failure, totally 6 group;
3) with bridge arm switching tube failure, totally 3 group;
4) in upper bridge arm or in lower bridge arm switching tube failure, totally 6 class;
5) switching tube intersects bridge arm failure, totally 6 class.
In step S102, the fault characteristic signals of the measuring signal are extracted using wavelet multi_resolution analysis method, and will
The fault characteristic signals of extraction are normalized and are added noise processed, specific as follows:
S301: it selects 5 layers of decomposition method of db3 small echo to decompose the measuring signal and (is illustrated in figure 4 implementation of the present invention
1 failure A electric current wavelet transformation exploded view of switching tube in example), obtain all frequency component S of the measuring signalph|j;Wherein,
Ph is phase serial number, value A, B or C;J is the corresponding frequency range serial number of each phase, takes 0 as 0~125Hz, take 1 for 126~
250Hz;
S302: since the high frequency section transformation of the different measuring signals is unobvious, so selecting frequency component Sph|j's
Low frequency part coefficient carries out energy reconstruct, obtains frequency component Sph|jCorresponding ENERGY Eph|j, such as reconstruction formula such as formula (1) institute
Show:
In above formula, xph|jkIndicate frequency component Sph|jDiscrete point amplitude, n be frequency component Sph|jDiscrete point it is total
Number;
S303: the vector that the low-frequency range energy of three-phase is formed is as fault feature vector T, as shown in formula (2):
T=[EA|0,EA|1,EB|0,EB|1,EC|0,EC|1] (2)
In above formula, EA|0、EA|1、EB|0、EB|1、EC|0And EC|1The respectively energy of 0~125Hz frequency band of A phase, A phase
The energy of 126~250Hz frequency band, the energy of 0~125Hz frequency band of B phase, B phase 126~250Hz frequency band energy,
The energy of the energy of 0~125Hz frequency band of C phase, 126~250Hz frequency band of C phase;
S304: being normalized fault feature vector T, as shown in formula (3):
Ej=EA|j+EB|j+EC|j (3)
Fault feature vector T1 after being normalized, shown in the fault feature vector T1 such as formula (4) after normalization:
S305: the random noise of %5 is added in fault feature vector T1 after normalization, with letters other in object simulating
Number interference, obtain be added random noise after fault feature vector T2;
Under different faults type, multiple measuring signals are acquired respectively, and each measuring signal progress is above-mentioned
After step S301 to S305 process processing, corresponding fault feature vector T2 is obtained, forms fault sample data.
In step S103, the probabilistic neural network model is made of input layer, mode layer, summation layer and 4 layers of output layer;
It is illustrated in figure 5 probabilistic neural network basic model;
Input layer is used to receive the value of feature vector T2 in the fault sample data, by the fault feature vector T2
Pass to probabilistic neural network;The dimension of the neuron number of the input layer and the fault feature vector T2 are equal;
Mode layer is used to calculate the matching relationship of each fault type in input fault feature vector T2 and fault sample data,
Shown in calculation formula such as formula (5):
In above formula, WiThe weight connected for input layer to mode layer;X is the corresponding value of fault feature vector of input;δ is
Smoothing factor;The number of the mode layer neuron is equal to the sum of the sample number of corresponding fault type, and i is the integer greater than 0,
Represent the serial number of corresponding neuron;
Summation layer is used to belong to the Cumulative probability of some fault type, so that the estimated probability for obtaining each fault type is close
Function is spent, probability density function is obtained by Parzen method, as shown in formula (6):
In above formula, XaiFor fault type θAThe corresponding fault feature vector of i-th of training sample (T2);M is failure classes
Type θATotal training sample number;P is probability, is preset value;The value of δ has determined the bell curve centered on sample point
Width;The neuron number of the summation layer is equal with fault type sum;
Output layer is made of thresholding discriminator, is had for selecting one in the estimated probability density of each fault type
The neuron of maximum estimated probability density exports fault type as output.
In step S103, when being trained using the fault sample data to probabilistic neural network (PNN) model: by institute
It states fault sample data and is divided into training data and (in embodiments of the present invention, the every kind of fault type selection of test data two parts
175 groups of data, wherein 150 groups of data are as training data, and 25 groups of data are as test data);Using training data to PNN
Network model is trained, and to modify the expansion rate of radial basis function in PNN network model, and is utilized during training
Genetic algorithm finds the optimal smoothing factor of the probabilistic neural network, obtains optimal PNN disaggregated model;Test set is for inputting
PNN network model, tests it, see prediction output result it is whether consistent with reality output result, obtain diagnosis rate and
The indexs such as precision;Continuous circuit training and test, until precision (the error percentage of model output and physical fault type label
Than) reach preset value or when the number of iterations reaches maximum set value, stop iteration, and using PNN network model at this time as
Trained probabilistic neural network model.
In step S103, the optimal smoothing factor of the probabilistic neural network is found using genetic algorithm, is specifically included:
Smoothing factor is encoded to chromosome and forms initial population, selected, intersected, made a variation by calculating fitness
With update iteration, to obtain the i.e. optimal smoothing factor of optimal chromosome;Its process is as shown in Figure 5.
Fitness is calculated by fitness function, then selected, intersected and is made a variation on this basis;Wherein fitness
Shown in function such as formula (7);
In above formula, M is the number of samples of corresponding training data, and i is the classification number of training classification (fault type), and I is class
Not sum;yim^ and yimThe predicted value and actual value of respectively m-th sample, a is a positive minimum, in order to avoid denominator
It is 0.
Selection operation selects roulette method, the i.e. selection strategy based on fitness ratio, the select probability p of each individual ii
Such as formula (8):
In above formula,FiFor the fitness value of individual i, k is coefficient, is preset value;N be population at individual number, i and
The value range of j is identical, is [1, N];
Crossover operation uses real number interior extrapolation method, k-th of Autosome akWith first of chromosome a1J crossover operation sides
Shown in method such as formula (9):
In above formula, akjFor the chromosome generated after crossover operation, random number of the b between [0,1];
Mutation operation chooses j-th of gene a of i-th of individualijIt makes a variation, shown in mutation operation such as formula (10):
In above formula, amaxFor gene aijThe upper bound, aminFor gene aijLower bound;F (g)=r2(1-g/Gmax)2;r2It is one
A random number;G is current the number of iterations;GmaxFor maximum evolution number;Random number of the r between [0,1];Specific flow chart is such as
Shown in Fig. 6.
It is as shown in table 1 fault type list of labels used in the embodiment of the present invention:
1 fault type label of table
Fault type | Class label | Fault type | Class label |
1 failure of switching tube | 1 | The simultaneous faults of switching tube 3 and 5 | 35 |
2 failure of switching tube | 2 | The simultaneous faults of switching tube 2 and 4 | 24 |
3 failure of switching tube | 3 | The simultaneous faults of switching tube 2 and 6 | 26 |
4 failure of switching tube | 4 | The simultaneous faults of switching tube 4 and 6 | 46 |
5 failure of switching tube | 5 | The simultaneous faults of switching tube 1 and 4 | 14 |
6 failure of switching tube | 6 | The simultaneous faults of switching tube 1 and 6 | 16 |
The simultaneous faults of switching tube 1 and 2 | 12 | The simultaneous faults of switching tube 3 and 2 | 32 |
The simultaneous faults of switching tube 3 and 4 | 34 | The simultaneous faults of switching tube 3 and 6 | 36 |
The simultaneous faults of switching tube 5 and 6 | 56 | The simultaneous faults of switching tube 5 and 2 | 52 |
The simultaneous faults of switching tube 1 and 3 | 13 | The simultaneous faults of switching tube 5 and 6 | 56 |
The simultaneous faults of switching tube 1 and 5 | 15 | Switching tube fault-free | 0 |
Diagnosis knot to carry out fault diagnosis in the embodiment of the present invention to microgrid inverter using distinct methods as shown in table 2
Fruit table:
2 fault diagnosis result table of table
Method for diagnosing faults | BP neural network | Support vector machines (SVM) | Probabilistic neural network (PNN) |
Fault diagnosis accuracy rate | 91.78% | 96.00% | 97.73% |
Failure diagnosis time | 15-30 seconds | 150-420 seconds | 5-10 seconds |
The partial fault diagnostic result figure being shown such as Fig. 7 (a)~Fig. 7 (c) in the embodiment of the present invention.
The beneficial effects of the present invention are: common BP neural network classifying quality is influenced by its initial weight and model structure
Larger, classification results lack transparency and are more slowly easy to fall into locally optimal solution using pace of learning.And the present invention is mentioned
Not only structural model is simple for technical solution out, be easy determination, fast convergence rate and converges on Bayes optimization solution, and sample is additional
Ability is strong, does not need re -training, and precision is high, practicality is strong, and is easy to engineering combination.
The foregoing is merely presently preferred embodiments of the present invention, is not intended to limit the invention, it is all in spirit of the invention and
Within principle, any modification, equivalent replacement, improvement and so on be should all be included in the protection scope of the present invention.
Claims (6)
1. a kind of microgrid inverter method for diagnosing faults based on wavelet transformation and probabilistic neural network, it is characterised in that: including
Following steps:
S101: according to the operating condition of microgrid inverter to be diagnosed, building the fault simulation system of the microgrid inverter, and
Threephase load current signal of the fault simulation system under different faults type is acquired as measuring signal;The failure classes
Type is to be numbered according to the preset various faults of failure that may occur;
S102: extracting the fault characteristic signals of the measuring signal using wavelet multi_resolution analysis method, and by the failure of extraction
Characteristic signal is normalized and is added noise processed, obtains fault sample data;
S103: building probabilistic neural network model, and using the fault sample data to the probabilistic neural network model into
Row training, obtains trained probabilistic neural network model;Wherein, in the training process of the probabilistic neural network model,
The optimal smoothing factor of the probabilistic neural network model is found using genetic algorithm;
S104: the physical fault for the microgrid inverter for treating diagnosis using trained probabilistic neural network model is diagnosed,
Realize the microgrid inverter fault diagnosis based on wavelet transformation and probabilistic neural network.
2. a kind of microgrid inverter fault diagnosis side based on wavelet transformation and probabilistic neural network as described in claim 1
Method, it is characterised in that: in step S101, built on Matlab/Simulink platform using microgrid inverter PQ control strategy
Fault simulation system, wherein the three-phase grid-connected inverter in the fault simulation system uses SVPWM method, and in institute
Fault simulation module is added after stating the pulse-triggered module in fault simulation system, and then microgrid is simulated by control trigger pulse
The open-circuit fault of inverter;It is specific as follows:
S201: acquiring the three-phase voltage and three-phase current of grid side, so by Clarke and Park transformation by three-phase voltage and
Three-phase current is converted into the current signal under the voltage signal under dq axis and dq axis;
S202: using under dq axis voltage signal and current signal as the input of phaselocked loop, voltage is obtained by PHASE-LOCKED LOOP PLL TECHNIQUE
With the phase theta of electric current;
S203: and then by the voltage and current transmission of phase of acquisition to microgrid central controller;Microgrid central controller is according to connecing
The voltage and current phase received issues power instruction PrefAnd QrefAnd generate current-order IdrefAnd Iqref, and by bicyclic
Control structure and Park inverse transformation obtain Uαβ;
S204: the U that will be obtainedαβWith the DC current U of piconet controllerdcAs the input of SVPWM, 6 are generated by SVPWM modulation
Driving pulse of a pulse as microgrid inverter;Finally add a failure trigger module, using the driving pulse as
The input of failure trigger module generates a microgrid inverter fault simulation system;The failure trigger module is opened by control
The pulse for closing pipe carrys out the shutdown of analog switch pipe.
3. a kind of microgrid inverter fault diagnosis side based on wavelet transformation and probabilistic neural network as described in claim 1
Method, it is characterised in that: in step S102, believed using the fault signature that wavelet multi_resolution analysis method extracts the measuring signal
Number, and the fault characteristic signals of extraction are normalized and are added noise processed, specific steps include:
S301: it selects 5 layers of decomposition method of db3 small echo to decompose the measuring signal, obtains all frequencies of the measuring signal
Rate component Sph|j;Wherein,phFor phase serial number, value A, B or C;J is the corresponding frequency band serial number of each phase, take 0 for 0~
125Hz, taking 1 is 126~250Hz;
S302: selecting frequency component Sph|jLow frequency part coefficient carry out energy reconstruct, obtain frequency component Sph|jCorresponding energy
Eph|j, shown in reconstruction formula such as formula (1):
In above formula, xph|jkIndicate frequency component Sph|jDiscrete point amplitude, n be frequency component Sph|jDiscrete point total number;
S303: the vector that the low-frequency range energy of three-phase is formed is as fault feature vector T, as shown in formula (2):
T=[EA|0,EA|1,EB|0,EB|1,EC|0,EC|1] (2)
In above formula, EA|0、EA|1、EB|0、EB|1、EC|0And EC|1Respectively the energy of 0~125Hz frequency band of A phase, A phase 126~
The energy of 250Hz frequency band, the energy of 0~125Hz frequency band of B phase, the energy of 126~250Hz frequency band of B phase, C phase
The energy of the energy of 0~125Hz frequency band, 126~250Hz frequency band of C phase;
S304: the normalized as shown in formula (3) is carried out to fault feature vector T:
Ej=EA|j+EB|j+EC|j (3)
Fault feature vector T1 after being normalized, shown in the fault feature vector T1 such as formula (4) after normalization:
S305: the random noise of %5 is added in fault feature vector T1 after normalization, with other signals in object simulating
Interference obtains that the fault feature vector T2 after random noise is added;
Under different faults type, multiple measuring signals are acquired respectively, and each measuring signal is subjected to above-mentioned steps
After S301 to S305 process processing, corresponding fault feature vector T2 is obtained, forms fault sample data.
4. a kind of microgrid inverter fault diagnosis side based on wavelet transformation and probabilistic neural network as described in claim 1
Method, it is characterised in that: in step S103, the probabilistic neural network model is by input layer, mode layer, summation layer and output layer 4
Layer composition;
Input layer is used to receive the value of fault feature vector T2 in the fault sample data, by the fault feature vector T2
Pass to probabilistic neural network;The dimension of the neuron number of the input layer and the fault feature vector T2 are equal;
Mode layer is used to calculate the matching relationship of each fault type in input fault feature vector T2 and fault sample data, calculates
Shown in formula such as formula (5):
In above formula, WiThe weight connected for input layer to mode layer;X is the corresponding value of fault feature vector T2 of input;δ is flat
The sliding factor;The number of the mode layer neuron is equal to the sum of the sample number of corresponding fault type, and i is the integer greater than 0, generation
Table corresponds to the serial number of neuron;
Summation layer is used to belong to the Cumulative probability of some fault type, to obtain the estimated probability density letter of each fault type
Number, obtains probability density function by Parzen method, as shown in formula (6):
In above formula, XaiFor fault type θAThe corresponding fault feature vector T2 of i-th of training sample;M is fault type θAIt is total
Training sample number;P is probability, is preset value;The value of δ has determined the width of the bell curve centered on sample point;
The neuron number of the summation layer is equal with fault type sum;
Output layer is made of thresholding discriminator, has maximum for selecting one in the estimated probability density of each fault type
The neuron of estimated probability density exports fault type as output.
5. a kind of microgrid inverter fault diagnosis side based on wavelet transformation and probabilistic neural network as described in claim 1
Method, it is characterised in that: in step S103, when being trained using the fault sample data to probabilistic neural network model: will
The fault sample data are divided into training data and test data two parts;Using training data to probabilistic neural network model into
Row training to modify the expansion rate of radial basis function in probabilistic neural network model, and utilizes heredity during training
Algorithm finds the optimal smoothing factor of the probabilistic neural network, obtains optimum probability neural network model;Test set is for defeated
Enter probabilistic neural network model, it is tested, sees whether prediction output result is consistent with reality output result, is diagnosed
The indexs such as rate and precision;Continuous circuit training and test, until precision reaches preset value or the number of iterations reaches maximum and sets
When definite value, stop iteration, and using probabilistic neural network model at this time as trained probabilistic neural network model.
6. a kind of microgrid inverter fault diagnosis side based on wavelet transformation and probabilistic neural network as described in claim 1
Method, it is characterised in that: in step S103, the optimal smoothing factor of the probabilistic neural network is found using genetic algorithm, specifically
Include:
Smoothing factor is encoded to chromosome and forms initial population, selected, intersected, made a variation and more by calculating fitness
New iteration, to obtain the i.e. optimal smoothing factor of optimal chromosome;
Fitness is calculated by fitness function, then selected, intersected and is made a variation on this basis;Wherein fitness function
As shown in formula (7);
In above formula, M is the number of samples of corresponding training data, and i is the classification number of training classification (fault type), and I is that classification is total
Number;yim^ and yimThe predicted value and actual value of respectively m-th sample, a are a positive minimum, in order to avoid denominator is 0;
Selection operation selects roulette method, the i.e. selection strategy based on fitness ratio, the select probability p of each individual iiSuch as public affairs
Formula (8):
In above formula,FiFor the fitness value of individual i, k is coefficient, is preset value;N is population at individual number, and i and j's takes
It is identical to be worth range, is [1, N];
Crossover operation uses real number interior extrapolation method, k-th of Autosome akWith first of chromosome a1J crossover operation methods such as
Shown in formula (9):
In above formula, akjFor the chromosome generated after crossover operation, random number of the b between [0,1];
Mutation operation chooses j-th of gene a of i-th of individualijIt makes a variation, shown in mutation operation such as formula (10):
In above formula, amaxFor gene aijThe upper bound, aminFor gene aijLower bound;F (g)=r2(1-g/Gmax)2;r2It is random for one
Number;G is current the number of iterations;GmaxFor maximum evolution number;Random number of the r between [0,1].
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910206025.XA CN110084106A (en) | 2019-03-19 | 2019-03-19 | Microgrid inverter method for diagnosing faults based on wavelet transformation and probabilistic neural network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910206025.XA CN110084106A (en) | 2019-03-19 | 2019-03-19 | Microgrid inverter method for diagnosing faults based on wavelet transformation and probabilistic neural network |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110084106A true CN110084106A (en) | 2019-08-02 |
Family
ID=67413258
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910206025.XA Pending CN110084106A (en) | 2019-03-19 | 2019-03-19 | Microgrid inverter method for diagnosing faults based on wavelet transformation and probabilistic neural network |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110084106A (en) |
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111239549A (en) * | 2020-02-18 | 2020-06-05 | 国网信通亿力科技有限责任公司 | Power distribution fault rapid positioning method based on discrete wavelet transform |
CN111366814A (en) * | 2020-03-31 | 2020-07-03 | 上海电力大学 | Power grid fault diagnosis method based on multi-source data and multi-dimensional fault coding space |
CN111679158A (en) * | 2020-08-04 | 2020-09-18 | 国网内蒙古东部电力有限公司呼伦贝尔供电公司 | Power distribution network fault identification method based on synchronous measurement data similarity |
CN111812435A (en) * | 2020-06-28 | 2020-10-23 | 国网新源控股有限公司回龙分公司 | Method for diagnosing fault of static frequency converter based on BP neural network |
CN112068033A (en) * | 2020-09-02 | 2020-12-11 | 河北工业大学 | 1/6 periodic current-based inverter power tube open-circuit fault online identification method |
CN112257335A (en) * | 2020-10-10 | 2021-01-22 | 西南交通大学 | Oil-immersed transformer fault diagnosis method combining PNN and SVM |
CN112255495A (en) * | 2020-09-10 | 2021-01-22 | 西安理工大学 | Micro-grid high-resistance fault detection method |
CN112414446A (en) * | 2020-11-02 | 2021-02-26 | 南昌智能新能源汽车研究院 | Data-driven transmission sensor fault diagnosis method |
CN113075469A (en) * | 2020-01-06 | 2021-07-06 | 株洲中车时代电气股份有限公司 | Inversion overcurrent fault diagnosis method, device and system |
CN113516066A (en) * | 2021-07-05 | 2021-10-19 | 内蒙古工业大学 | Power quality disturbance signal classification method and device, storage medium and electronic equipment |
CN113702862A (en) * | 2021-08-27 | 2021-11-26 | 广东省科学院电子电器研究所 | Power supply state panoramic monitoring method and device based on cloud deployment |
CN116520990A (en) * | 2023-04-28 | 2023-08-01 | 暨南大学 | Sign language identification method and system based on lightweight neural network and glove |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102570392A (en) * | 2012-01-17 | 2012-07-11 | 上海电力学院 | Method for identifying exciting inrush current of transformer based on improved probability neural network |
CN105914736A (en) * | 2016-05-05 | 2016-08-31 | 河海大学 | Inverter power supply modeling method in power distribution network short circuit |
US20170052060A1 (en) * | 2014-04-24 | 2017-02-23 | Alstom Transport Technologies | Method and system for automatically detecting faults in a rotating shaft |
-
2019
- 2019-03-19 CN CN201910206025.XA patent/CN110084106A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102570392A (en) * | 2012-01-17 | 2012-07-11 | 上海电力学院 | Method for identifying exciting inrush current of transformer based on improved probability neural network |
US20170052060A1 (en) * | 2014-04-24 | 2017-02-23 | Alstom Transport Technologies | Method and system for automatically detecting faults in a rotating shaft |
CN105914736A (en) * | 2016-05-05 | 2016-08-31 | 河海大学 | Inverter power supply modeling method in power distribution network short circuit |
Non-Patent Citations (4)
Title |
---|
Z HUANG, Z WANG AND H ZHANG: "A diagnosis algorithm for multiple open-circuited faults of microgrid inverters based on main fault component analysis", 《EEE TRANSACTIONS ON ENERGY CONVERSION》 * |
刘琦: "微网多逆变器并联的故障诊断研究", 《中国优秀硕士学位全文数据库工程科技Ⅱ辑》 * |
周沙: "基于概率神经网络的变压器局部放电模式识别研究", 《中国优秀硕士学位全文数据库工程科技Ⅱ辑》 * |
高杰: "基于优化PNN网络的变压器故障诊断研究", 《中国优秀硕士学位论文全文数据库工程科技Ⅱ辑》 * |
Cited By (16)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113075469A (en) * | 2020-01-06 | 2021-07-06 | 株洲中车时代电气股份有限公司 | Inversion overcurrent fault diagnosis method, device and system |
CN111239549A (en) * | 2020-02-18 | 2020-06-05 | 国网信通亿力科技有限责任公司 | Power distribution fault rapid positioning method based on discrete wavelet transform |
CN111366814A (en) * | 2020-03-31 | 2020-07-03 | 上海电力大学 | Power grid fault diagnosis method based on multi-source data and multi-dimensional fault coding space |
CN111812435A (en) * | 2020-06-28 | 2020-10-23 | 国网新源控股有限公司回龙分公司 | Method for diagnosing fault of static frequency converter based on BP neural network |
CN111679158A (en) * | 2020-08-04 | 2020-09-18 | 国网内蒙古东部电力有限公司呼伦贝尔供电公司 | Power distribution network fault identification method based on synchronous measurement data similarity |
CN112068033A (en) * | 2020-09-02 | 2020-12-11 | 河北工业大学 | 1/6 periodic current-based inverter power tube open-circuit fault online identification method |
CN112068033B (en) * | 2020-09-02 | 2024-03-26 | 河北工业大学 | On-line identification method for open-circuit faults of inverter power tube based on 1/6 period current |
CN112255495A (en) * | 2020-09-10 | 2021-01-22 | 西安理工大学 | Micro-grid high-resistance fault detection method |
CN112255495B (en) * | 2020-09-10 | 2023-10-24 | 西安理工大学 | Micro-grid high-resistance fault detection method |
CN112257335A (en) * | 2020-10-10 | 2021-01-22 | 西南交通大学 | Oil-immersed transformer fault diagnosis method combining PNN and SVM |
CN112414446A (en) * | 2020-11-02 | 2021-02-26 | 南昌智能新能源汽车研究院 | Data-driven transmission sensor fault diagnosis method |
CN113516066A (en) * | 2021-07-05 | 2021-10-19 | 内蒙古工业大学 | Power quality disturbance signal classification method and device, storage medium and electronic equipment |
CN113516066B (en) * | 2021-07-05 | 2023-08-08 | 内蒙古工业大学 | Power quality disturbance signal classification method and device, storage medium and electronic equipment |
CN113702862A (en) * | 2021-08-27 | 2021-11-26 | 广东省科学院电子电器研究所 | Power supply state panoramic monitoring method and device based on cloud deployment |
CN116520990A (en) * | 2023-04-28 | 2023-08-01 | 暨南大学 | Sign language identification method and system based on lightweight neural network and glove |
CN116520990B (en) * | 2023-04-28 | 2023-11-24 | 暨南大学 | Sign language identification method and system based on lightweight neural network and glove |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110084106A (en) | Microgrid inverter method for diagnosing faults based on wavelet transformation and probabilistic neural network | |
US10234495B2 (en) | Decision tree SVM fault diagnosis method of photovoltaic diode-clamped three-level inverter | |
Han et al. | Short-time wavelet entropy integrating improved LSTM for fault diagnosis of modular multilevel converter | |
CN105095566B (en) | A kind of fault of converter diagnostic method based on wavelet analysis and SVM | |
WO2019141040A1 (en) | Short term electrical load predication method | |
CN109933881A (en) | A kind of Fault Diagnosis of Power Electronic Circuits method based on optimization deepness belief network | |
CN108832619A (en) | Transient stability evaluation in power system method based on convolutional neural networks | |
CN112051481B (en) | Alternating current-direct current hybrid power grid fault area diagnosis method and system based on LSTM | |
CN109948833A (en) | A kind of Hydropower Unit degradation trend prediction technique based on shot and long term memory network | |
CN104809722A (en) | Electrical device fault diagnosis method based on infrared thermography | |
CN108732528A (en) | A kind of digitalized electrical energy meter method for diagnosing faults based on depth confidence network | |
CN104616061B (en) | Island detection method based on wavelet packet logarithmic energy entropy and genetic algorithm optimization | |
CN101782625B (en) | Power electronic system fault diagnostic method based on Gradation-boosting algorithm | |
CN104730423A (en) | Island effect detecting method of grid-connected photovoltaic power system | |
CN113659565B (en) | Online prediction method for frequency situation of new energy power system | |
CN103761569A (en) | Fault diagnosis method and device for wind driven generator | |
Wang et al. | State variable technique islanding detection using time-frequency energy analysis for DFIG wind turbine in microgrid system | |
Dai et al. | Fault diagnosis of data-driven photovoltaic power generation system based on deep reinforcement learning | |
Jiang et al. | Application of a hybrid model of big data and BP network on fault diagnosis strategy for microgrid | |
Shi et al. | A fault location method for distribution system based on one-dimensional convolutional neural network | |
CN115754790A (en) | Power supply side fault diagnosis and automatic switching method for low-voltage dual-power system | |
CN112630596A (en) | Comprehensive diagnosis method for open-circuit fault of IGBT device of wind power converter | |
Ruirong et al. | Research on Fault Location Technology of Intelligent Distribution Network based on Neural Network | |
Bhuyan et al. | Fault Classification in a DG Connected Power System using Artificial Neural Network | |
Mehdi et al. | Squaring and lowpass filtering data-driven technique for AC faults in AC/DC lines |
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: 20190802 |
|
RJ01 | Rejection of invention patent application after publication |