CN108241298A - A kind of aerogenerator method for diagnosing faults based on FWA-RNN models - Google Patents
A kind of aerogenerator method for diagnosing faults based on FWA-RNN models Download PDFInfo
- Publication number
- CN108241298A CN108241298A CN201810017397.3A CN201810017397A CN108241298A CN 108241298 A CN108241298 A CN 108241298A CN 201810017397 A CN201810017397 A CN 201810017397A CN 108241298 A CN108241298 A CN 108241298A
- Authority
- CN
- China
- Prior art keywords
- aerogenerator
- fwa
- rnn
- fault
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
- G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
- G05B13/04—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators
- G05B13/042—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators in which a parameter or coefficient is automatically adjusted to optimise the performance
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention relates to a kind of aerogenerator method for diagnosing faults based on FWA RNN models, belong to generator state monitoring and fault diagnosis field.This method key step is as follows:(1) failure mode analysis (FMA) and test signal are chosen;(2) it determines aerogenerator measured signal and passes through sensor to acquire mass data under various working conditions, be transferred to computer later and carry out data storage;(3) mass data of acquisition is done into FFT transform, and be normalized.The data of acquisition are divided into training sample and test sample;(4) binding test sample data optimizes all initial parameters of RNN using FWA, establishes neural network model, and failure modes are carried out using training sample as input;(5) detection model accuracy.The present invention has good data adaptive ability and robustness, can effectively improve aerogenerator failure modes detection efficiency and accuracy rate.
Description
Technical field
The present invention relates to a kind of aerogenerator failures for being based on FWA-RNN (fireworks algorithm-recurrent neural network) model
Diagnostic method belongs to generator state monitoring and fault diagnosis field.
Background technology
Generator is the core of electric system, once breaking down, not only threatens the stable operation of electric system, it is also possible to
Damage gen-set.With the continuous increase of single-machine capacity, the operational reliability of generator is particularly important and protrudes.It is special
It is not in aviation electric system, brushless AC generator has conclusive shadow to the vitality of aircraft whether normal operation
It rings.Therefore, aerogenerator equipment fault diagnosis is researched and analysed, proposes more effective fault diagnosis system, more
The operational safety of aircraft is ensured well, is had great significance to the development of aircraft industry.
In aerogenerator the Study on Fault field, educational circles is generally classified as mechanical breakdown and is carried out with two major class of electric fault
It inquires into, wherein electric fault is broadly divided into stator winding faults, machines under rotor winding faults, three kinds of faults in rotating rectifiers again, machinery
Failure is mainly two kinds of bearing fault, shaft failure.Currently used method for diagnosing faults mainly uses manual analysis or signal
The method of processing, these methods come with some shortcomings, such as analysis efficiency is low, the degree of automation is low.
Recurrent neural network is a kind of neural network for possessing fixed weights and threshold value, because its internal structure is steady
It is fixed, it is made to be more easy to realize, has the characteristics that principle is simple, the speed of service is fast, local search ability is strong, but the setting of initial parameter
It is improper that its convergence rate may be caused slow or be absorbed in Local Minimum.The Inspiration Sources of fireworks algorithm are the fireworks of explosion, are a kind of
Intelligent algorithm in novel group.It possesses preferable part and ability of searching optimum, is suitable for a variety of optimization problems.
Invention content
The present invention proposes a kind of aerogenerator method for diagnosing faults based on FWA-RNN models, excellent using FWA algorithms
Change the initial parameter (all weights and threshold value) of RNN, neural network convergence rate can be accelerated, it is made to be not easy to be absorbed in local pole
It is small.This method has good data adaptive ability and robustness, can effectively improve aerogenerator failure modes detection
Efficiency and accuracy rate.
The present invention is adopted the following technical scheme that solve its technical problem:
A kind of aerogenerator method for diagnosing faults based on FWA-RNN models, mainly includes the following steps that:
(1) failure mode analysis (FMA) and test signal are chosen:The fault mode type of aerogenerator is analyzed, determines failure classes
Type number;
(2) signal acquisition:It determines aerogenerator measured signal and is acquired under various working conditions greatly using sensor
Data are measured, computer is transferred to later and carries out data storage;
(3) data prediction:The mass data of acquisition is done into Fast Fourier Transform (FFT), and be normalized;It will obtain
The data obtained are divided into training sample and test sample;
(4) network operation:Binding test sample data optimizes all initial parameters of RNN using FWA, establishes neural network mould
Training sample is carried out failure modes by type;
(5) performance test:Detection model accuracy.
Detailed process is as follows in the step (4):
First, several groups of RNN initial parameters that will be randomly provided, i.e., all weights and threshold value as initial explosion point coordinates,
Input test sample is counted respectively using based on the Hadoop Distributed Computing Platforms under GFS, MapReduce, Bigtable frame
Calculate the fitness value of each demolition point;
Secondly, initial explosion point coordinates is optimized using FWA algorithms to obtain optimal explosion point coordinates, as
Recurrent neural network parameter.
Finally, input training sample is classified.
The present invention has the beneficial effect that:
The present invention proposes a kind of aerogenerator method for diagnosing faults based on FWA-RNN models, can be quick and precisely
Ground diagnoses generator failure type.This method is applied in aerogenerator condition monitoring and fault diagnosis field, tool
There are preferable part and ability of searching optimum, speed and the accuracy of aerogenerator fault diagnosis can be effectively improved.
Description of the drawings
Fig. 1 FWA-RNN Troubleshooting Flowcharts.
Fig. 2 fireworks algorithm (FWA) structure chart.
Fig. 3 MapReduce fundamental diagrams.
Fig. 4 recurrent neural networks (RNN) illustraton of model.
Specific embodiment
The invention is described in further details below in conjunction with the accompanying drawings.
The present invention proposes a kind of aerogenerator method for diagnosing faults based on FWA-RNN models, and this method is mainly wrapped
Include fault mode and test signal analysis, signal acquisition, data prediction, with fireworks algorithm optimization neural network initial parameter,
Fault detect, several parts of performance detection are carried out using recurrent neural network.The flow of use is as shown in Figure 1, concrete operations include
Following steps:
First, the major failure type and number of aerogenerator are determined by consulting literatures and experiment simulation.Aviation is sent out
The main typical fault pattern of motor includes the internal short circuit fault of stator winding, the insulation fault of stator winding, rotor windings
Shorted-turn fault, rotating rectifier rectifying tube single tube open circuit fault, two-tube open circuit fault, bearing fault, shaft failure etc., institute
The signal that need to be acquired is generally the machinery of main generator excitation current signal, main generator three-phase output voltage signal and generator
Vibration signal.
Then, current sensor, voltage sensor, vibrating sensor and data acquisition are connected on more aerogenerators
The equipment such as card, and it is made to operate under normal and various fault modes.Acquire their main generator excitation current signal, main hair
Motor three-phase output voltage signal and vibration signal, and import in computer and stored, form original sampled data.
The collected three kinds of signals of institute are subjected to Fast Fourier Transform (FFT) respectively, time-domain signal is converted into frequency-region signal,
Then it is normalized to eliminate dimension.Finally, the data of acquisition are divided into two class of training sample and test sample.Normalizing
Changing mathematic(al) representation is:
In formula, L is number of samples,For the sample set after normalization,G for sample set X
A sample, xgFor g-th of sample before normalization, xmeanRepresent the mean value of X, xstdRepresent the standard deviation of X.
Finally, FWA-RNN programs are run:
It is 3 to design recurrent neural networks model output layer according to seven kinds of failure modes, and collected 5 groups of signals are done
Input layer is used as after FFT (fast Fourier) transformation.Maximum iteration Iterator is setmax。
Use fireworks algorithm optimization RNN parameters, such as Fig. 2:It is assumed that number of parameters is m in recurrent neural network, one is established
A m dimension coordinates system.Light n pieces of fireworks at random.
Since the calculating of each point to best demolition point is independent from each other, it can utilize and map stipulations model
(MapReduce) such as Fig. 3, parallel computation frame excavate its degree of parallelism, and the time of best demolition point optimization is greatly shortened with this,
Mapping tasks 1 are using the coordinate of demolition point as key assignments, using the point coordinates that explodes as the RNN models fitness value of parameter as defeated
Go out, stipulations task 1 finds out the position of optimal demolition point using the output of mapping tasks 1 as input.Wherein using training sample as
It is RNN model fitness values to input obtained output error value.
The best demolition point for retaining last iteration is calculated as next-generation fireworks demolition point position, and according to formula (1) and (2)
Go out fireworks burst radius AmSpark number S is generated with explosionm, in order to the good fireworks position of restraining error value will not generate it is excessive quick-fried
Fried spark, while the fireworks position of error value difference will not generate very few spark particle, and spark number is limited with formula (3).
Mapping tasks 2 are using the point coordinates that explodes as key assignments, with Am、SmAs output, stipulations task 2 is using the output of mapping tasks 2 as defeated
Enter, explosive spark position is obtained according to formula (4).
In formula:Xm be the m pieces fireworks coordinate vector, ymin=min (f (xm)), (m=1,2 ..., N) it is current fireworks kind
Fitness minimum value in group, ymax=max (f (xm)), (m=1,2 ..., N) it is fitness maximum value in current fireworks population.
It is a constant, for adjusting burst radius size, a, b, M is constant, and for adjusting the explosive spark number size of generation, ε is
One machine minimum, for avoiding except Z-operation, U (1, -1) is uniformly distributed to be obeyed in (1, -1) section.
Some fireworks are randomly choosed in fireworks population, then randomly choosing a certain number of dimensions to the fireworks carries out height
This mutation operation.For fireworks xmThe obtained dimension k of some selection to perform Gaussian mutation operated (5)
In formula:E~N (1,1), N (1,1) represent mean value be 1, variance be 1 Gaussian Profile, xmkIt is tieed up for the m pieces fireworks in k
The value of degree,For result of the m pieces fireworks after K dimension values carry out Gaussian mutation.
During explosion operator and mutation operator generate explosive spark and Gaussian mutation spark respectively, it is possible to create fire
Flower fair exceeds the bounds of feasible zone Ω.As spark xmBeyond boundary on dimension k, will be reflected by the mapping ruler of formula (6)
It is mapped to a new position.
In formula:xUB,kFor feasible zone coboundary, xLB, kFor lower boundary.
System selects a certain number of individuals as next-generation fireworks in fireworks, explosive spark and Gaussian mutation spark.
Assuming that set of candidates is K, fireworks Population Size is N.It is selected to a being determined property of cognition of error minimum in set of candidates
To the next generation as fireworks, and the selection of remaining N-1 fireworks is carried out using the method for roulette in set of candidates
Selection.For candidate xm, the calculation formula for being chosen probability is (7)
Wherein:xnCoordinate for n-th piece of candidate fireworks.
(7) formula top half removes x for current individual to set of candidates KmThe sum of the distance between all individuals.It is waiting
In the person's of choosing set, if population density is higher, i.e., when having a lot of other candidate individuals around the individual, the individual is selected
Probability can reduce.
To obtained spark xm, perform mutation operator formula (8), crossover operator formula (9), selection opertor formula (10) and obtain later
New demolition point position if it is nondominated solution, is then regarded as new demolition point.
vm=xm+γ(xm2-xm3) (8)
γ is in formula>0 coefficient, crFor crossover probability 0<cr<1,rmIt is 0 to the integer between k, xm2, xm3To select at random
The two explosion point coordinates taken, vmFor demolition point xmThe new explosion point coordinates obtained after mutation operator is performed, For demolition point
vm xmThe value of j-th of element,For the value of j-th of element after execution crossover operator, umCrossover operator after-explosion point coordinates is performed,
It is obtained what n demolition point correspondence inputted using stipulations model (MapReduce) is mapped according to updated demolition point
Error chooses the point of error minimum, its stability is analyzed by Lyapunov methods.Design a model energy function E (11), according to
Formula (12) calculates the RNN model stabilities by the use of the point coordinates as parameter, if qualification is just as optimal solution.
E=F (wji,xi,Ii), j=1,2 ..., n;I=1,2 ..., n (11)
W in formulaji,xi,IiRespectively corresponding neuron weights, input, threshold value.
If the iterations Iterator having reached the maximummax, then using result as RNN initial parameters, otherwise update
Demolition point repeats the operation of the demolition point after being optimized of setting off fireworks.
Using test sample as input run RNN models, such as Fig. 4, hidden layer kinetics equation be (13), each neuron
State expression formula is (14).Using pretreated sensor signal as input, exported.
zj(k+1)=Wxj(k)+Ij, j=1,2 ... n;K=1,2 ... N (13)
W is weight matrix in formula,For neuron activation functions, sigmoid functions are usedIjFor
The initial threshold of j-th of neuron, xj(k) output for j-th of neuron, zj(k+1) finally entering for j-th neuron.
Claims (5)
1. a kind of aerogenerator method for diagnosing faults based on FWA-RNN models, which is characterized in that mainly including following step
Suddenly:
(1) failure mode analysis (FMA) and test signal are chosen:The fault mode type of aerogenerator is analyzed, determines fault type
Number;
(2) signal acquisition:It determines aerogenerator measured signal and a large amount of numbers is acquired under various working conditions using sensor
According to, be transferred to later computer carry out data storage;
(3) data prediction:The mass data of acquisition is done into Fast Fourier Transform (FFT), and be normalized;By acquisition
Data are divided into training sample and test sample;
(4) network operation:Binding test sample data optimizes all initial parameters of RNN using FWA, establishes neural network model,
Failure modes are carried out using training sample as input;
(5) performance test:Detection model accuracy.
2. according to a kind of aerogenerator method for diagnosing faults based on FWA-RNN models described in claim 1, feature
It is:Detailed process is as follows in the step (4):
First, several groups of RNN initial parameters that will be randomly provided, i.e., all weights are with threshold value as initial explosion point coordinates, input
Test sample is calculated respectively using based on the Hadoop Distributed Computing Platforms under GFS, MapReduce, Bigtable frame
The fitness value of each demolition point;
Secondly, initial explosion point coordinates is optimized using FWA algorithms to obtain optimal explosion point coordinates, as recurrence
Neural network parameter;
Finally, input training sample is classified.
3. according to a kind of aerogenerator method for diagnosing faults based on FWA-RNN models described in claim 1, feature
It is:Step(1)Described in aerogenerator fault mode type be the internal short circuit fault of stator winding, stator winding
Insulation fault, rotor interturn short-circuit failure.
4. according to a kind of aerogenerator method for diagnosing faults based on FWA-RNN models described in claim 1, feature
It is:Step(1)Described in the fault mode type of aerogenerator be rotating rectifier rectifying tube single tube open circuit fault, two-tube
Open circuit fault, bearing fault, shaft failure.
5. according to a kind of aerogenerator method for diagnosing faults based on FWA-RNN models described in claim 1, feature
It is:Step(2)Described in sensor for current sensor, voltage sensor, vibrating sensor.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810017397.3A CN108241298A (en) | 2018-01-09 | 2018-01-09 | A kind of aerogenerator method for diagnosing faults based on FWA-RNN models |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810017397.3A CN108241298A (en) | 2018-01-09 | 2018-01-09 | A kind of aerogenerator method for diagnosing faults based on FWA-RNN models |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108241298A true CN108241298A (en) | 2018-07-03 |
Family
ID=62698503
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810017397.3A Pending CN108241298A (en) | 2018-01-09 | 2018-01-09 | A kind of aerogenerator method for diagnosing faults based on FWA-RNN models |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108241298A (en) |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109308522A (en) * | 2018-09-03 | 2019-02-05 | 河海大学常州校区 | A kind of GIS failure prediction method based on Recognition with Recurrent Neural Network |
CN109521404A (en) * | 2018-10-12 | 2019-03-26 | 上海交通大学 | The evaluation of accuracy and system of vibration measurement based on fmcw radar |
CN109946080A (en) * | 2019-04-08 | 2019-06-28 | 西安交通大学 | A kind of mechanical equipment health status recognition methods based on embedded recirculating network |
CN110853010A (en) * | 2019-11-11 | 2020-02-28 | 江苏君英天达人工智能研究院有限公司 | High-speed railway cable detection method based on FWA and SM |
CN111753751A (en) * | 2020-06-28 | 2020-10-09 | 辽宁大学 | Fan fault intelligent diagnosis method for improving firework algorithm |
CN112327208A (en) * | 2020-11-02 | 2021-02-05 | 国网江苏省电力有限公司电力科学研究院 | Fault diagnosis method and device for turn-to-turn short circuit of phase modulator rotor winding |
CN112491009A (en) * | 2020-11-15 | 2021-03-12 | 陕西航空电气有限责任公司 | Excitation positive short-circuit fault protection method for generator controller of aviation power system |
CN112486147A (en) * | 2020-12-14 | 2021-03-12 | 江西洪都航空工业集团有限责任公司 | Airplane fault detection and maintenance management system and management method thereof |
CN113255546A (en) * | 2021-06-03 | 2021-08-13 | 成都卡莱博尔信息技术股份有限公司 | Diagnosis method for aircraft system sensor fault |
CN115236507A (en) * | 2022-06-27 | 2022-10-25 | 陕西航空电气有限责任公司 | Neural network-based aviation generator electrical fault diagnosis method and device |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2002057858A3 (en) * | 2000-12-01 | 2003-12-11 | Active Control Experts Inc | Method and device for motion control |
CN101907681A (en) * | 2010-07-15 | 2010-12-08 | 南京航空航天大学 | Analog circuit dynamic online failure diagnosing method based on GSD-SVDD |
CN105607575A (en) * | 2016-01-27 | 2016-05-25 | 电子科技大学 | FA-LSSVM based numerical control machine tool main shaft heat drift modeling method |
CN105652952A (en) * | 2016-04-18 | 2016-06-08 | 中国矿业大学 | Maximum power point tracking method for photovoltaic power generation system based on fireworks algorithm |
CN106443447A (en) * | 2016-09-26 | 2017-02-22 | 南京航空航天大学 | An aero-generator fault feature extraction method based on iSDAE |
CN107169565A (en) * | 2017-04-27 | 2017-09-15 | 西安工程大学 | Yarn quality prediction method based on fireworks algorithm improvement BP neural network |
CN107247231A (en) * | 2017-07-28 | 2017-10-13 | 南京航空航天大学 | A kind of aerogenerator fault signature extracting method based on OBLGWO DBN models |
-
2018
- 2018-01-09 CN CN201810017397.3A patent/CN108241298A/en active Pending
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2002057858A3 (en) * | 2000-12-01 | 2003-12-11 | Active Control Experts Inc | Method and device for motion control |
CN101907681A (en) * | 2010-07-15 | 2010-12-08 | 南京航空航天大学 | Analog circuit dynamic online failure diagnosing method based on GSD-SVDD |
CN105607575A (en) * | 2016-01-27 | 2016-05-25 | 电子科技大学 | FA-LSSVM based numerical control machine tool main shaft heat drift modeling method |
CN105652952A (en) * | 2016-04-18 | 2016-06-08 | 中国矿业大学 | Maximum power point tracking method for photovoltaic power generation system based on fireworks algorithm |
CN106443447A (en) * | 2016-09-26 | 2017-02-22 | 南京航空航天大学 | An aero-generator fault feature extraction method based on iSDAE |
CN107169565A (en) * | 2017-04-27 | 2017-09-15 | 西安工程大学 | Yarn quality prediction method based on fireworks algorithm improvement BP neural network |
CN107247231A (en) * | 2017-07-28 | 2017-10-13 | 南京航空航天大学 | A kind of aerogenerator fault signature extracting method based on OBLGWO DBN models |
Non-Patent Citations (1)
Title |
---|
崔江 等: "一种无刷发电机旋转整流器自适应故障特征提取方法", 《中国电机工程学报》 * |
Cited By (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109308522B (en) * | 2018-09-03 | 2022-02-22 | 河海大学常州校区 | GIS fault prediction method based on recurrent neural network |
CN109308522A (en) * | 2018-09-03 | 2019-02-05 | 河海大学常州校区 | A kind of GIS failure prediction method based on Recognition with Recurrent Neural Network |
CN109521404A (en) * | 2018-10-12 | 2019-03-26 | 上海交通大学 | The evaluation of accuracy and system of vibration measurement based on fmcw radar |
CN109521404B (en) * | 2018-10-12 | 2020-09-04 | 上海交通大学 | FMCW radar-based vibration measurement accuracy evaluation method and system |
CN109946080A (en) * | 2019-04-08 | 2019-06-28 | 西安交通大学 | A kind of mechanical equipment health status recognition methods based on embedded recirculating network |
CN109946080B (en) * | 2019-04-08 | 2020-06-16 | 西安交通大学 | Mechanical equipment health state identification method based on embedded circulation network |
CN110853010A (en) * | 2019-11-11 | 2020-02-28 | 江苏君英天达人工智能研究院有限公司 | High-speed railway cable detection method based on FWA and SM |
CN110853010B (en) * | 2019-11-11 | 2022-05-13 | 江苏君英天达人工智能研究院有限公司 | High-speed railway cable detection method based on FWA and SM |
CN111753751A (en) * | 2020-06-28 | 2020-10-09 | 辽宁大学 | Fan fault intelligent diagnosis method for improving firework algorithm |
CN112327208A (en) * | 2020-11-02 | 2021-02-05 | 国网江苏省电力有限公司电力科学研究院 | Fault diagnosis method and device for turn-to-turn short circuit of phase modulator rotor winding |
CN112491009A (en) * | 2020-11-15 | 2021-03-12 | 陕西航空电气有限责任公司 | Excitation positive short-circuit fault protection method for generator controller of aviation power system |
CN112491009B (en) * | 2020-11-15 | 2023-02-10 | 陕西航空电气有限责任公司 | Excitation positive short-circuit fault protection method for generator controller of aviation power system |
CN112486147A (en) * | 2020-12-14 | 2021-03-12 | 江西洪都航空工业集团有限责任公司 | Airplane fault detection and maintenance management system and management method thereof |
CN113255546A (en) * | 2021-06-03 | 2021-08-13 | 成都卡莱博尔信息技术股份有限公司 | Diagnosis method for aircraft system sensor fault |
CN115236507A (en) * | 2022-06-27 | 2022-10-25 | 陕西航空电气有限责任公司 | Neural network-based aviation generator electrical fault diagnosis method and device |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108241298A (en) | A kind of aerogenerator method for diagnosing faults based on FWA-RNN models | |
Hang et al. | Application of multi-class fuzzy support vector machine classifier for fault diagnosis of wind turbine | |
Amraee et al. | Transient instability prediction using decision tree technique | |
CN109376801B (en) | Wind driven generator blade icing diagnosis method based on integrated deep neural network | |
CN102074955A (en) | Method based on knowledge discovery technology for stability assessment and control of electric system | |
Mazhari et al. | A hybrid fault cluster and thévenin equivalent based framework for rotor angle stability prediction | |
CN107247231A (en) | A kind of aerogenerator fault signature extracting method based on OBLGWO DBN models | |
Lu et al. | Fault diagnosis of rolling bearing based on improved VMD and KNN | |
CN108921347B (en) | Wind power converter fault prediction method | |
CN111652479A (en) | Data driving method for dynamic security assessment of power system | |
CN116169675B (en) | Power system dynamic stability online evaluation method considering operation mode change | |
Colby et al. | Optimizing ballast design of wave energy converters using evolutionary algorithms | |
CN114595858A (en) | Short-term wind speed prediction method and system based on rolling time series and support vector machine | |
CN110635474B (en) | Power grid dynamic trajectory trend prediction method based on long-term and short-term memory network | |
Radhi et al. | Faults diagnosis in stator windings of high speed solid rotor induction motors using fuzzy neural network | |
CN110969214A (en) | Transient state security domain online construction method based on support vector machine comprehensive model | |
Chu et al. | Rolling bearing fault diagnosis by a novel fruit fly optimization algorithm optimized support vector machine | |
Darabi et al. | Employing adaptive particle swarm optimization algorithm for parameter estimation of an exciter machine | |
Medoued et al. | Classification of induction machine faults using time frequency representation and particle swarm optimization | |
CN116305683A (en) | Power system transient stability evaluation method and system based on sample equalization | |
Narasimhulu et al. | Detection and classification of high impedance fault in power distribution system using hybrid technique | |
CN110261773A (en) | A kind of aerogenerator failure symptom extracting method and system | |
CN112651628A (en) | Power system transient stability evaluation method based on capsule neural network | |
CN109901030B (en) | Reactor turn-to-turn insulation state monitoring method, system and application | |
CN107860972A (en) | Method for detecting harmonic wave and m-Acetyl chlorophosphonazo |
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: 20180703 |
|
RJ01 | Rejection of invention patent application after publication |