CN114742140A - UHF and CG-BP algorithm-based GIS insulation fault type identification method - Google Patents
UHF and CG-BP algorithm-based GIS insulation fault type identification method Download PDFInfo
- Publication number
- CN114742140A CN114742140A CN202210312079.6A CN202210312079A CN114742140A CN 114742140 A CN114742140 A CN 114742140A CN 202210312079 A CN202210312079 A CN 202210312079A CN 114742140 A CN114742140 A CN 114742140A
- Authority
- CN
- China
- Prior art keywords
- gis
- neural network
- insulation fault
- fault type
- samples
- 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
- 238000009413 insulation Methods 0.000 title claims abstract description 95
- 238000004422 calculation algorithm Methods 0.000 title claims abstract description 55
- 238000000034 method Methods 0.000 title claims abstract description 38
- 238000013528 artificial neural network Methods 0.000 claims abstract description 64
- 238000012549 training Methods 0.000 claims abstract description 47
- 238000003062 neural network model Methods 0.000 claims abstract description 25
- 238000001514 detection method Methods 0.000 claims abstract description 24
- 230000007547 defect Effects 0.000 claims abstract description 9
- 238000004519 manufacturing process Methods 0.000 claims abstract description 6
- 238000002939 conjugate gradient method Methods 0.000 claims description 9
- 239000002184 metal Substances 0.000 claims description 9
- 239000002923 metal particle Substances 0.000 claims description 8
- 230000005284 excitation Effects 0.000 claims description 6
- 238000000605 extraction Methods 0.000 claims description 6
- 238000003860 storage Methods 0.000 claims description 6
- 238000004590 computer program Methods 0.000 claims description 5
- 238000007667 floating Methods 0.000 claims description 5
- 239000000725 suspension Substances 0.000 claims 1
- 230000006870 function Effects 0.000 description 12
- 238000004088 simulation Methods 0.000 description 7
- 238000010586 diagram Methods 0.000 description 6
- 230000008569 process Effects 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 3
- 230000009191 jumping Effects 0.000 description 3
- 239000000463 material Substances 0.000 description 3
- 238000005457 optimization Methods 0.000 description 3
- 230000009467 reduction Effects 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 239000004593 Epoxy Substances 0.000 description 2
- 230000004888 barrier function Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000003745 diagnosis Methods 0.000 description 2
- 238000013095 identification testing Methods 0.000 description 2
- 238000005339 levitation Methods 0.000 description 2
- 210000002569 neuron Anatomy 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000014509 gene expression Effects 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Images
Classifications
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/11—Complex mathematical operations for solving equations, e.g. nonlinear equations, general mathematical optimization problems
- G06F17/13—Differential equations
-
- 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
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
- G06N3/044—Recurrent networks, e.g. Hopfield 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
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Computational Mathematics (AREA)
- Software Systems (AREA)
- Evolutionary Biology (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Operations Research (AREA)
- Algebra (AREA)
- Databases & Information Systems (AREA)
- Testing Relating To Insulation (AREA)
Abstract
The invention relates to a GIS insulation fault type identification method based on UHF and CG-BP algorithms, which comprises the following steps: classifying insulation fault types of the GIS equipment; establishing a CG-BP neural network model, wherein the CG-BP neural network model comprises a CG conjugate gradient algorithm module and a BP neural network, and optimizing the weight and the threshold of the BP neural network by using the CG conjugate gradient algorithm module; manufacturing a plurality of GIS equipment samples with insulation fault defects, acquiring PD partial discharge signals of the GIS equipment samples by using a UHF partial discharge detection sensor, acquiring characteristic parameters of the corresponding GIS equipment samples by using the PD partial discharge signals, and adding insulation fault type labels to the GIS equipment samples to obtain a plurality of training samples; performing iterative training on the CG-BP neural network model by using a training sample, finishing the training after an iteration ending condition is reached, and outputting a GIS insulation fault type identification model; and identifying the fault type of the GIS equipment by using the GIS insulation fault type identification model.
Description
Technical Field
The invention relates to a GIS insulation fault type identification method based on UHF and CG-BP algorithms, and belongs to the technical field of GIS equipment insulation fault type identification.
Background
The closed GIS has high integration, has the characteristics of small occupied area and small electromagnetic hazard, and is widely applied to high-voltage power transformation places. However, due to the current production and manufacturing level, transportation means and installation technology, and uncontrollable factors in the operating environment, the PD partial discharge phenomenon of the GIS device sometimes occurs. Therefore, the GIS equipment insulation fault type diagnosis is beneficial to improving the stable operation capacity of the power grid.
The conventional detection method mainly comprises ultrasonic, chemical materials, optical instruments, pulse current, UHF ultrahigh frequency partial discharge technology and the like. Among them, the UHF sensor has good sensitivity and strong anti-interference capability, and is widely used in practical engineering.
How to accurately identify the insulation fault type of the GIS equipment by using PD data acquired by a UHF sensor is a problem which is urgently needed to be solved at present.
Disclosure of Invention
In order to solve the problems in the prior art, the invention provides a GIS insulation fault type identification method based on UHF and CG-BP algorithms, provides an improved CG-BP neural network model, avoids the problem of non-convergence or local optimization caused by random generation of weight and threshold of the BP neural network, realizes the reduction of the training times of the traditional BP neural network, improves the learning efficiency and accelerates the convergence progress; and a GIS insulation fault type recognition model is trained by using the improved CG-BP neural network model, and compared with the model trained by the traditional BP neural network, the GIS insulation fault type recognition model has higher recognition accuracy.
The technical scheme of the invention is as follows:
on one hand, the invention provides a GIS insulation fault type identification method based on UHF and CG-BP algorithm, comprising the following steps:
classifying insulation fault types of the GIS equipment;
establishing a CG-BP neural network model, wherein the CG-BP neural network model comprises a CG conjugate gradient algorithm module and a BP neural network, and optimizing the weight and the threshold of the BP neural network by using the CG conjugate gradient algorithm module;
manufacturing a plurality of GIS equipment samples with insulation fault defects, acquiring PD partial discharge signals of the GIS equipment samples by using a UHF partial discharge detection sensor, acquiring characteristic parameters of the corresponding GIS equipment samples by using the PD partial discharge signals, and adding insulation fault type labels to the GIS equipment samples to obtain a plurality of training samples consisting of the characteristic parameters of the corresponding GIS equipment samples and the insulation fault type labels;
performing iterative training on the CG-BP neural network model by using a training sample, finishing the training after an iterative finishing condition is reached, and outputting a GIS insulation fault type identification model;
and identifying the fault type of the GIS equipment by using the GIS insulation fault type identification model.
As a preferred embodiment, the method for optimizing the weight and the threshold of the BP neural network by using the CG conjugate gradient algorithm module specifically comprises:
the result objective function for constructing the BP neural network output is as follows:
wherein x isjRepresenting the input of the jth node of the input layer; y isiAn input representing an ith node of the hidden layer; w is akiRepresenting the weight from the Kth node of the output layer to the ith node of the hidden layer; w is aijRepresenting the weight from the ith node of the hidden layer to the jth node of the input layer; thetaiA threshold value representing the ith node of the implicit layer; alpha (alpha) ("alpha")kA threshold value representing the kth node of the output layer; phi represents the excitation function of the hidden layer;
let the single sample quadratic error expectation be:
wherein, TkL is the number of output nodes of the neural network for the desired output;
the total error of the samples is:
wherein P is the total number of samples;
CG conjugate gradient algorithm module adopts conjugate gradient method to output layer weight wkiThe hidden layer weight wijOutput layer threshold value alphakThreshold of hidden layer thetaiAnd correcting to obtain the gradients of all parameters as follows:
wherein eta is the step length of gradient search;
and carrying out gradient descent according to the calculated gradient of each parameter, and iteratively updating the parameters of the BP neural network until a preset iteration termination condition is met.
As a preferred embodiment, the step of correcting any parameter by the conjugate gradient method specifically includes:
initializing, and setting an initial solution vector, a maximum iteration number and a residual threshold;
calculating the residual vector of the iteration according to the solution vector of the iteration;
calculating the direction vector of the iteration according to the residual vector of the iteration;
calculating the step length of the iteration according to the direction vector and the residual vector of the iteration;
updating the solution vector according to the calculated direction vector and step length of the iteration;
setting the termination condition of iteration as reaching the maximum iteration number or the residual vector of the iteration is larger than the residual threshold value;
and after the termination condition of iteration is reached, outputting the current solution vector as the optimal solution vector.
As a preferred embodiment, the acquiring the characteristic parameters of the corresponding GIS device sample by using the PD partial discharge signal includes:
average value of PD partial discharge signalAbsolute mean valueMaximum value xmaxMinimum value xminPeak to peak value xmmMean square valuePeak index IpPulse index CfMargin CeRoot amplitude xarsVariance DXSkewness α, kurtosis β.
As a preferred embodiment, the insulation fault types of GIS devices include metal tip faults, insulated internal air gap or creeping discharge faults, free metal particle faults, and levitation discharge faults.
On the other hand, the invention provides a GIS (geographic information System) absolute cause barrier type identification system based on UHF (ultra high frequency) and CG-BP (CG-BP) algorithms, which comprises the following steps:
the fault type classification module is used for classifying the insulation fault types of the GIS equipment;
the UHF sensor module is used for acquiring PD partial discharge signals of the GIS equipment;
the characteristic parameter extraction module is used for extracting characteristic parameters according to PD partial discharge signals of the GIS equipment;
the system comprises a sample data making module, a characteristic parameter extracting module and a training module, wherein the sample data making module is used for making a plurality of GIS equipment samples with insulation fault defects, a PD partial discharge signal of each GIS equipment sample is obtained by using a UHF partial discharge detection sensor, characteristic parameters of the corresponding GIS equipment sample are obtained by using the characteristic parameter extracting module, an insulation fault type label is added to each GIS equipment sample, and a plurality of training samples consisting of the characteristic parameters of the corresponding GIS equipment sample and the insulation fault type label are obtained;
the CG-BP neural network model comprises a CG conjugate gradient algorithm module and a BP neural network, wherein the CG conjugate gradient algorithm module is used for optimizing the weight and the threshold of the BP neural network; the BP neural network is used for inputting training samples to perform training instead, finishing training after an iteration ending condition is met, and outputting a GIS insulation fault type recognition model; and identifying the fault type of the GIS equipment by using the GIS insulation fault type identification model.
As a preferred embodiment, the method for optimizing the weight and the threshold of the BP neural network by the CG conjugate gradient algorithm module specifically comprises:
the result objective function for constructing the BP neural network output is as follows:
wherein x isjRepresenting the input of the jth node of the input layer; y isiAn input representing an ith node of the hidden layer; w is akiRepresenting the weight from the Kth node of the output layer to the ith node of the hidden layer; w is aijRepresenting the weight from the ith node of the hidden layer to the jth node of the input layer; thetaiA threshold value representing the ith node of the implicit layer; alpha is alphakA threshold value representing the kth node of the output layer; phi represents the excitation function of the hidden layer;
let the single sample quadratic error expectation be:
wherein, TkL is the number of output nodes of the neural network for the desired output;
the total error of the samples is:
wherein P is the total number of samples;
CG conjugate gradient algorithm module adopts conjugate gradient method to output layer weight wkiHidden layer weight wijOutput layer threshold value alphakThreshold of hidden layer thetaiAnd correcting to obtain the gradients of all parameters as follows:
wherein eta is the step length of gradient search;
and performing gradient descent according to the calculated gradient of each parameter, and iteratively updating the parameters of the BP neural network until a preset iteration termination condition is met.
As a preferred embodiment, the feature parameters extracted by the feature parameter extraction module include:
average value of PD partial discharge signalAbsolute mean valueMaximum value xmaxMinimum value xminPeak to peak value xmmMean square valuePeak index IpPulse index CfMargin CeRoot amplitude xarsVariance DXSkewness α, kurtosis β.
As a preferred embodiment, the fault type classification module specifically classifies the absolute fault type of the GIS device as a metal tip fault, an insulated internal air gap or creeping discharge fault, a self-contained metal particle fault, and a levitation discharge fault.
In still another aspect, the present invention further provides a computer-readable storage medium, on which a computer program is stored, which when executed by a processor, implements the GIS insulation fault type identification method according to any one of the embodiments of the present invention.
The invention has the following beneficial effects:
1. the invention relates to a GIS insulation fault type identification method based on UHF and CG-BP algorithm, which classifies insulation fault types of GIS equipment, provides an improved CG-BP neural network model, avoids the problem of non-convergence or local optimization caused by random generation of weight and threshold of a BP neural network, realizes the reduction of training times of the traditional BP neural network, improves the learning efficiency and accelerates the convergence progress; and a GIS insulation fault type recognition model is trained by using the improved CG-BP neural network model, and compared with the model trained by the traditional BP neural network, the GIS insulation fault type recognition model has higher recognition accuracy.
2. The invention relates to a GIS (geographic information System) insulation fault type identification method based on UHF (ultra high frequency) and CG-BP (CG-Back propagation) algorithms, which sets the average value of PD (potential difference) partial discharge signalsAbsolute mean valueMaximum value xmaxMinimum value xminPeak to peak value xmmMean square value ofPeak index IpPulse index CfMargin CeSquare root amplitude xarsVariance DX13 characteristic parameters including skewness alpha and kurtosis beta are adopted, data of multiple dimensions are used as input characteristic parameters of fault type identification, and accuracy and reliability of GIS insulation fault type identification are improved.
Drawings
FIG. 1 is a flowchart of a method according to a first embodiment of the present invention;
FIG. 2 is an exemplary diagram of UHF detection principle of PD signals of GIS equipment in the embodiment of the present invention;
FIG. 3 is a schematic diagram of the location of various types of insulation faults in an embodiment of the present invention;
FIG. 4a is a graph illustrating the variation of training bias of a conventional BP neural network in a simulation experiment according to an embodiment of the present invention;
FIG. 4b is a graph showing the variation of the training bias of the CG-BP neural network model in the simulation experiment according to the embodiment of the present invention;
FIG. 5a is a diagram illustrating a fault identification test of a conventional BP neural network in a simulation experiment according to an embodiment of the present invention;
fig. 5b is a fault identification test chart of the CG-BP neural network model in the simulation experiment according to the embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be described clearly and completely with reference to the accompanying drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
It should be understood that the step numbers used herein are for convenience of description only and are not intended as limitations on the order in which the steps are performed.
It is to be understood that the terminology used in the description of the invention herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. As used in this specification and the appended claims, the singular forms "a", "an", and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.
The terms "comprises" and "comprising" indicate the presence of the described features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.
The term "and/or" refers to any and all possible combinations of one or more of the associated listed items and includes such combinations.
The first embodiment is as follows:
referring to fig. 1, a method for identifying a type of a GIS insulation fault based on UHF and CG-BP algorithms comprises the following steps:
classifying insulation fault types of the GIS equipment;
establishing a CG-BP neural network model, wherein the CG-BP neural network model comprises a CG conjugate gradient algorithm module and a BP neural network, and optimizing the weight and the threshold of the BP neural network by using the CG conjugate gradient algorithm module; compared with the scheme that the initial value is randomly generated in the iteration process of the traditional BP neural network, the output value of the CG conjugate gradient algorithm can search the value to the optimal solution range, so that the learning efficiency of the BP neural network is improved, and the training times are reduced;
manufacturing a plurality of GIS equipment samples with different types of insulation fault defects, acquiring PD partial discharge signals of the GIS equipment samples by using a UHF partial discharge detection sensor, acquiring characteristic parameters of the corresponding GIS equipment samples by using the PD partial discharge signals, and adding insulation fault type labels to the GIS equipment samples to obtain a plurality of training samples consisting of the characteristic parameters of the corresponding GIS equipment samples and the insulation fault type labels;
the method comprises the steps that a built-in UHF partial discharge detection sensor and an external UHF partial discharge detection sensor are adopted in the process of acquiring PD partial discharge signals of samples of the GIS equipment by using the UHF partial discharge detection sensor, the external UHF partial discharge detection sensor mainly detects electromagnetic wave signals transmitted out through epoxy materials to identify whether the PD partial discharge phenomenon occurs in the GIS equipment, and the built-in UHF partial discharge detection sensor is installed in the GIS equipment and can directly identify whether the PD partial discharge phenomenon occurs in the GIS by detecting the electromagnetic wave signals in the GIS equipment; the acquisition precision and reliability of the PD signals of the GIS equipment are greatly improved by the internal and external UHF partial discharge detection sensors, wherein the UHF detection principle of the PD signals of the GIS equipment is shown in figure 2;
performing iterative training on the CG-BP neural network model by using a training sample, finishing the training after an iterative finishing condition is reached, and outputting a GIS insulation fault type identification model;
the GIS insulation fault type identification model is used for identifying the fault type of the GIS equipment to be detected, a PD partial discharge signal of the GIS equipment to be detected is obtained through a UHF partial discharge detection sensor, corresponding characteristic parameters are obtained through the PD partial discharge signal, the characteristic parameters are input into the GIS insulation fault type identification model, and the insulation fault defect type of the current GIS equipment is automatically identified.
As a preferred implementation manner of this embodiment, the method for optimizing the weight and the threshold of the BP neural network by using the CG conjugate gradient algorithm module specifically includes:
the result objective function for constructing the BP neural network output is as follows:
wherein x isjRepresenting the input of the jth node of the input layer; y isiAn input representing an ith node of the hidden layer; w is akiRepresenting the weight from the Kth node of the output layer to the ith node of the hidden layer; w is aijRepresenting the weight between the ith node of the hidden layer and the jth node of the input layer; thetaiA threshold value representing the ith node of the implicit layer; alpha is alphakA threshold value representing the kth node of the output layer; phi represents the excitation function of the hidden layer;
let the single sample quadratic error expectation be:
wherein, TkL is the number of output nodes of the neural network for the desired output;
the total error of the samples is:
wherein P is the total number of samples;
CG conjugate gradient algorithm module adopts conjugate gradient method to output layer weight wkiHidden layer weight wijOutput layer threshold value alphakThreshold of hidden layer thetaiAnd correcting to obtain the gradients of all parameters as follows:
wherein eta is the step length of gradient search;
and performing gradient descent according to the calculated gradient of each parameter, and iteratively updating the parameters of the BP neural network until a preset iteration termination condition is met.
As a preferred implementation manner of this embodiment, the step of correcting any parameter by the conjugate gradient method specifically includes:
step 1: initializing, setting the initial iteration number k to 0, the maximum iteration number kmax to 50, and setting the residual threshold epsilon to 10-3(ii) a m (0) is the initial solution vector, and the solution vector, direction vector and residual vector after k iterations are respectively represented by m (k), d (k) and r (k).
m(0)=0;d(0)=0;r(0)=0;
Step 2: calculating the residual vector of the iteration through the solution vector of the iteration:
r(k)=A×m(k-1)+b
in the formula, A is a weight and b is a relaxation factor;
judging whether the residual vector r (k) of the iteration is smaller than a residual threshold epsilon, if so, jumping to the step 6, otherwise, continuing to execute the step 3;
and step 3: calculating the direction vector of the iteration through the residual vector of the iteration:
and 4, step 4: calculating the step length of the iteration according to the direction vector and the residual vector of the iteration:
and 5: updating a solution vector:
m(k)=m(k-1)+∝(k)×d(k)
k=k+1
and judging whether k is less than 50, if so, jumping to the step 6, otherwise, jumping back to the step 2, and continuing to execute.
Step 6: and ending the program, and outputting the current solution vector as the optimal solution vector.
As a preferred embodiment of this embodiment, acquiring the characteristic parameters of the corresponding GIS device sample by using the PD partial discharge signal includes:
average value of PD partial discharge signalAbsolute mean valueMaximum value xmaxMinimum value xminPeak to peak value xmmMean square valuePeak index IpPulse index CfMargin CeSquare root amplitude xarsVariance DXSkewness α, kurtosis β;
the definition and calculation mode of each parameter are as follows:
mean value ofThe premises and the basis of all other parameters are also static reflections of the signal.
Absolute mean valueIn addition to reflecting the integral magnitude of the waveform signal, it is also a representation of the signal energy.
Maximum value xmaxAnd the minimum value xmin: the maxima and minima reflect the instantaneous signal energy magnitude (one set for each maximum and minimum).
Peak to peak value xmm: the numerical difference between the peak and the peak shows the fluctuation range of the local discharge signal.
xmm=xmax-xmin
The mean square value and the effective value are both used for describing the energy of the vibration signal, are the second moment statistical average of the signal, the effective value Xrms is also called root mean square value and is used for describing the energy of the discharge signal, the stability and the repeatability are good, and the method is an important index for diagnosing the partial discharge fault.
Peak value Xp: the peak value Xp is a single peak value of the vibration waveform, and 10 numbers with the largest absolute value are found out in the total length of one signal sample, and the arithmetic mean of the 10 numbers is used as the peak value Xp.
Pulse index CfAnd peak index IpAre statistical indicators used to detect whether there is an impact in the signal.
Square root amplitude xars: like root mean square, the representation of the overall energy of the signal.
Variance Dx: the variance value represents the degree of fluctuation of the signal.
Skewness α: indicating whether the signal belongs to a symmetric distribution.
Kurtosis β: is a reflection of the steepness of the signal and is sensitive to the pulse value. Therefore, a minute pulse change can be reflected in the detection signal.
As a preferred implementation manner of this embodiment, this embodiment combines the insulation fault types of the GIS device including a metal tip fault, an insulation internal air gap or creeping discharge fault, a free metal particle fault, and a floating discharge fault; the generation positions of the respective types of insulation faults are shown in fig. 3.
Further, the present embodiment provides four kinds of absolute barrier characteristics causing the PD phenomenon of the GIS device and propagation characteristics of the PD signal thereof, as shown in table 1 below:
table 1: GIS insulation defects and characteristics thereof
Furthermore, simulation experiments are performed in the embodiment to verify the effectiveness and the superiority of the method for identifying the type of the GIS insulation fault provided by the invention; the method comprises the following specific steps:
the number of neurons of the BP neural network input layer is set as the number of corresponding characteristic parameters, that is, 13 neurons, the four insulation fault types classified in the above embodiment are used as the output of the neural network (in simulation, the ordinate represents the four insulation faults by 1 to 4 respectively), and 24 layers of hidden layers are set.
The training sample set is characteristic parameter data after dimensionality reduction, wherein 50 groups of PD signals of four fault types are acquired respectively, and 200 groups of PD signal data are obtained in total. In this case, the characteristic parameter data may form a 13 × 200 matrix, and 70% of the characteristic parameter data may be training data, 15% of the characteristic parameter data may be test data, and the rest of the characteristic parameter data may be verification data. Training function and transfer function adopt Translm and Tansig respectively, and the values of college learning times and learning rate are 100 and 0.01. By using MATLAB software for simulation, a training error curve and a training result comparison graph of the model combining the CG conjugate gradient algorithm and the BP neural network and the traditional BP neural network can be obtained, which are respectively shown in fig. 4 and fig. 5.
As can be easily seen from fig. 4 and 5, the model combining the CG conjugate gradient algorithm and the BP neural network proposed in this embodiment is significantly superior to the conventional BP neural network in the GIS insulation fault type identification. The improved CG-BP neural network model optimizes the weight and the threshold of the BP neural network by utilizing a CG conjugate gradient algorithm, further weakens the dependence degree of the BP neural network on the weight and the threshold, improves the convergence speed and reduces the learning and training times.
Fig. 4a is a training deviation variation graph of a conventional BP neural network, in fig. 4a, the 12 th iteration is an inflection point, before that, the errors of three curves are not obvious, after 12 iterations, the error value gradually increases, when the iteration reaches the 24 th iteration, the performance reaches the optimum, and the training is finished.
Fig. 4b is a training deviation variation diagram of the CG-BP neural network model proposed in this embodiment, in fig. 4b, the 6 th iteration is an inflection point, before that, errors of three curves are not obvious, after 6 iterations, an error value is gradually increased, when the iteration reaches the 14 th iteration, the performance reaches the optimum, the classification function is completed, the training is ended, the iteration number is greatly shortened, and the efficiency is high.
FIG. 5a is a diagram of a fault recognition test of a trained conventional BP neural network; FIG. 5b is a diagram of a CG-BP neural network model for testing fault identification according to the present embodiment; table 2 is a fault identification rate table of the conventional BP neural network for each type of insulation fault, and table 3 is a fault identification rate table of the CG-BP neural network model proposed in this embodiment, which is specifically as follows:
table 2: GIS fault type recognition rate based on traditional BP network
Type of insulation failure | Number of samples | Identifying samples | Recognition rate |
Metal tip | 50 | 35 | 70% |
Insulated internal air gap | 50 | 39 | 78% |
Free metal particles | 50 | 45 | 90% |
Discharge of floating potential | 50 | 43 | 86% |
Table 3: GIS fault type identification rate of CG-BP network based on embodiment
Type of insulation failure | Number of samples | Identifying samples | Recognition rate |
Metal tip | 50 | 45 | 90% |
Insulating internal air gap | 50 | 42 | 84% |
Free metal particles | 50 | 49 | 98% |
Discharge of floating potential | 50 | 46 | 92% |
As can be seen from fig. 5, tables 2 and 3, compared with the fault type diagnosis rate of the GIS device of the conventional BP network of 81%, the fault type identification rate of the GIS device of the CG-BP network improved by the present embodiment is as high as 91%. And the identification rate of each single insulation fault type is improved, wherein the metal tip fault type is improved by 18% at the highest rate, and the identification rate of the insulation internal air gap fault type is 98% at the highest rate. This effectively demonstrates that the method for identifying the type of the GIS fault based on the ultrahigh frequency partial discharge technology and the CG-BP algorithm provided by the embodiment has good identification capability.
In the embodiment, the CG conjugate gradient algorithm and the BP neural network are effectively combined, and an improved CG-BP algorithm is provided by utilizing the complementary advantages of the respective algorithms, so that the problem of non-convergence or local optimization caused by random generation of weight and threshold is avoided, the training times of the traditional BP neural network are shortened, the learning efficiency is improved, and the convergence progress is accelerated.
Example two:
the embodiment provides a system for identifying a type of a GIS (geographic information system) insulation fault based on UHF (ultra high frequency) and CG-BP (CG-Back propagation) algorithms, which comprises the following steps:
the fault type classification module is used for classifying the insulation fault types of the GIS equipment;
the UHF sensor module is used for acquiring PD partial discharge signals of the GIS equipment; the UHF sensor module adopts two detection sensors, namely a built-in UHF partial discharge detection sensor and an external UHF partial discharge detection sensor, the external UHF partial discharge detection sensor mainly identifies whether the PD partial discharge phenomenon occurs in the GIS equipment by detecting electromagnetic wave signals transmitted through the epoxy material, and the built-in UHF partial discharge detection sensor is installed in the GIS equipment and can directly identify whether the PD partial discharge phenomenon occurs in the GIS by detecting the electromagnetic wave signals in the GIS equipment; by the aid of the internal and external UHF partial discharge detection sensors, acquisition accuracy and reliability of PD signals of GIS equipment are greatly improved;
the characteristic parameter extraction module is used for extracting characteristic parameters according to the PD partial discharge signals of the GIS equipment acquired by the UHF sensor module;
the system comprises a sample data making module, a characteristic parameter extracting module and a training module, wherein the sample data making module is used for making a plurality of GIS equipment samples with insulation fault defects, a PD partial discharge signal of each GIS equipment sample is obtained by using a UHF partial discharge detection sensor, characteristic parameters of the corresponding GIS equipment sample are obtained by using the characteristic parameter extracting module, an insulation fault type label is added to each GIS equipment sample, and a plurality of training samples consisting of the characteristic parameters of the corresponding GIS equipment sample and the insulation fault type label are obtained;
the CG-BP neural network model comprises a CG conjugate gradient algorithm module and a BP neural network, wherein the CG conjugate gradient algorithm module is used for optimizing the weight and the threshold of the BP neural network; the BP neural network is used for inputting training samples to perform training instead, finishing training after an iteration ending condition is met, and outputting a GIS insulation fault type recognition model; and identifying the fault type of the GIS equipment by using the GIS insulation fault type identification model.
As a preferred embodiment of this embodiment, the method for optimizing the weight and the threshold of the BP neural network by the CG conjugate gradient algorithm module specifically includes:
the result objective function for constructing the BP neural network output is as follows:
wherein x isjRepresenting the input of the jth node of the input layer; y isiAn input representing an ith node of the hidden layer; w is akiRepresenting the weight from the Kth node of the output layer to the ith node of the hidden layer; w is aijRepresenting the weight from the ith node of the hidden layer to the jth node of the input layer; thetaiA threshold value representing the ith node of the implicit layer; alpha is alphakA threshold value representing the kth node of the output layer; phi represents the excitation function of the hidden layer;
let the single sample quadratic error expectation be:
wherein, TkL is the number of output nodes of the neural network for the desired output;
the total error of the samples is:
wherein P is the total number of samples;
CG conjugate gradient algorithm module adopts conjugate gradient method to output layer weight wkiHidden layer weight wijOutput layer threshold value alphakThreshold of hidden layer thetaiCorrecting to obtain the gradient of each parameterRespectively, the following steps:
wherein eta is the step length of gradient search;
and performing gradient descent according to the calculated gradient of each parameter, and iteratively updating the parameters of the BP neural network until a preset iteration termination condition is met.
As a preferred implementation manner of this embodiment, the feature parameters extracted by the feature parameter extraction module include:
average value of PD partial discharge signalAbsolute mean valueMaximum value xmaxMinimum value xminPeak to peak value xmmMean square valuePeak index IpPulse index CfMargin CeSquare root amplitude xarsVariance DXSkewness α, kurtosis β.
As a preferred implementation manner of this embodiment, the fault type classification module specifically classifies the insulation fault type of the GIS device into a metal tip fault, an insulation internal air gap or along surface discharge fault, a free metal particle fault, and a floating discharge fault.
Example three:
the present embodiment proposes a computer-readable storage medium, on which a computer program is stored, which when executed by a processor implements the GIS insulation fault type identification method according to any one of the embodiments of the present invention.
Example four:
this embodiment provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, where the processor executes the computer program to implement the method for identifying a GIS insulation fault type according to any embodiment of the present invention.
In the embodiments of the present application, "at least one" means one or more, "and" a plurality "means two or more. "and/or" describes the association relationship of the associated objects, and means that there may be three relationships, for example, a and/or B, and may mean that a exists alone, a and B exist together, or B exists alone. Wherein A and B can be singular or plural. The character "/" generally indicates that the former and latter associated objects are in an "or" relationship. "at least one of the following" and similar expressions refer to any combination of these items, including any combination of singular or plural items. For example, at least one of a, b, and c may represent: a, b, c, a and b, a and c, b and c or a and b and c, wherein a, b and c can be single or multiple.
Those of ordinary skill in the art will appreciate that the various elements and algorithm steps described in connection with the embodiments disclosed herein can be implemented as electronic hardware, computer software, or combinations of electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the technical solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.
It is clear to those skilled in the art that, for convenience and brevity of description, the specific working processes of the system, the apparatus and the unit described above may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.
In the several embodiments provided in the present application, any function, if implemented in the form of a software functional unit and sold or used as a standalone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application or portions thereof that contribute to the prior art may be embodied in the form of a software product stored in a storage medium and including several instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the methods described in the embodiments of the present application. And the aforementioned storage medium includes: a U disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
The above description is only an embodiment of the present invention, and is not intended to limit the scope of the present invention, and all equivalent structures or equivalent processes performed by the present invention or directly or indirectly applied to other related technical fields are included in the scope of the present invention.
Claims (10)
1. A GIS insulation fault type identification method based on UHF and CG-BP algorithm is characterized by comprising the following steps:
classifying the insulation fault types of the GIS equipment;
establishing a CG-BP neural network model, wherein the CG-BP neural network model comprises a CG conjugate gradient algorithm module and a BP neural network, and optimizing the weight and the threshold of the BP neural network by using the CG conjugate gradient algorithm module;
manufacturing a plurality of GIS equipment samples with insulation fault defects, acquiring PD partial discharge signals of the GIS equipment samples by using a UHF partial discharge detection sensor, acquiring characteristic parameters of the corresponding GIS equipment samples by using the PD partial discharge signals, and adding insulation fault type labels to the GIS equipment samples to obtain a plurality of training samples consisting of the characteristic parameters of the corresponding GIS equipment samples and the insulation fault type labels;
performing iterative training on the CG-BP neural network model by using a training sample, finishing the training after an iteration ending condition is reached, and outputting a GIS insulation fault type identification model;
and identifying the fault type of the GIS equipment by using the GIS insulation fault type identification model.
2. The method for identifying the type of the GIS insulation fault based on the UHF and CG-BP algorithm according to claim 1, wherein the method for optimizing the weight and the threshold of the BP neural network by using the CG conjugate gradient algorithm module specifically comprises the following steps:
the result objective function for constructing the BP neural network output is as follows:
wherein x isjRepresenting the input of the jth node of the input layer; y isiAn input representing an ith node of the hidden layer; w is akiRepresenting the weight from the Kth node of the output layer to the ith node of the hidden layer; w is aijRepresenting the weight from the ith node of the hidden layer to the jth node of the input layer; thetaiA threshold value representing the ith node of the hidden layer; alpha is alphakA threshold value representing the kth node of the output layer; phi denotes the excitation function of the hidden layer;
let the single sample quadratic error expectation be:
wherein, TkL is the number of output nodes of the neural network for the desired output;
the total error of the samples is:
wherein P is the total number of samples;
CG conjugate gradient algorithm module adopts conjugate gradient method to output layer weight wkiHidden layer weight wijOutput layer threshold value alphakThreshold of hidden layer thetaiAnd correcting to obtain the gradients of all parameters as follows:
wherein eta is the step length of gradient search;
and performing gradient descent according to the calculated gradient of each parameter, and iteratively updating the parameters of the BP neural network until a preset iteration termination condition is met.
3. The method for identifying the type of the GIS insulation fault based on the UHF and CG-BP algorithm according to claim 2, wherein the step of correcting any parameter by the conjugate gradient method specifically comprises the following steps:
initializing, and setting an initial solution vector, a maximum iteration number and a residual threshold;
calculating the residual vector of the iteration according to the solution vector of the iteration;
calculating the direction vector of the iteration according to the residual vector of the iteration;
calculating the step length of the iteration according to the direction vector and the residual vector of the iteration;
updating the solution vector according to the calculated direction vector and step length of the iteration;
setting the termination condition of iteration as reaching the maximum iteration number or the residual vector of the iteration is larger than the residual threshold value;
and after the termination condition of iteration is reached, outputting the current solution vector as the optimal solution vector.
4. The GIS insulation fault type identification method based on UHF and CG-BP algorithm according to claim 1, characterized in that the acquisition of the characteristic parameters of the corresponding GIS device sample by using PD partial discharge signal comprises:
5. The method for identifying the type of the GIS insulation fault based on the UHF and CG-BP algorithm according to claim 1, wherein the method comprises the following steps:
the insulation fault types of GIS devices include metal tip faults, insulated internal air gap or creeping discharge faults, free metal particle faults, and floating discharge faults.
6. A GIS insulation fault type identification system based on UHF and CG-BP algorithm is characterized by comprising the following steps:
the fault type classification module is used for classifying the insulation fault types of the GIS equipment;
the UHF sensor module is used for acquiring PD partial discharge signals of the GIS equipment;
the characteristic parameter extraction module is used for extracting characteristic parameters according to PD partial discharge signals of the GIS equipment;
the system comprises a sample data making module, a characteristic parameter extracting module and a training module, wherein the sample data making module is used for making a plurality of GIS equipment samples with insulation fault defects, acquiring PD partial discharge signals of the GIS equipment samples by using a UHF partial discharge detection sensor, acquiring characteristic parameters of the corresponding GIS equipment samples by using the characteristic parameter extracting module, adding insulation fault type labels to the GIS equipment samples and obtaining a plurality of training samples consisting of the characteristic parameters of the corresponding GIS equipment samples and the insulation fault type labels;
the CG-BP neural network model comprises a CG conjugate gradient algorithm module and a BP neural network, wherein the CG conjugate gradient algorithm module is used for optimizing the weight and the threshold of the BP neural network; the BP neural network is used for inputting training samples to perform training instead, finishing training after an iteration ending condition is met, and outputting a GIS insulation fault type recognition model; and identifying the fault type of the GIS equipment by using the GIS insulation fault type identification model.
7. The GIS insulation fault type recognition system based on UHF and CG-BP algorithm according to claim 6, characterized in that the method for optimizing the weight and threshold of the BP neural network by the CG conjugate gradient algorithm module is specifically as follows:
the result objective function for constructing the BP neural network output is as follows:
wherein x isjRepresenting the input of the jth node of the input layer; y isiTo representInput of the ith node of the hidden layer; w is akiRepresenting the weight from the Kth node of the output layer to the ith node of the hidden layer; w is aijRepresenting the weight from the ith node of the hidden layer to the jth node of the input layer; thetaiA threshold value representing the ith node of the hidden layer; alpha is alphakA threshold value representing the kth node of the output layer; phi denotes the excitation function of the hidden layer;
let the single sample quadratic error expectation be:
wherein, TkL is the number of output nodes of the neural network for the desired output;
the total error of the samples is:
wherein P is the total number of samples;
CG conjugate gradient algorithm module adopts conjugate gradient method to output layer weight wkiHidden layer weight wijOutput layer threshold value alphakThreshold of hidden layer thetaiAnd correcting to obtain the gradients of all parameters as follows:
wherein eta is the step length of gradient search;
and performing gradient descent according to the calculated gradient of each parameter, and iteratively updating the parameters of the BP neural network until a preset iteration termination condition is met.
8. The GIS insulation fault type recognition system based on UHF and CG-BP algorithm according to claim 6, characterized in that the characteristic parameters extracted by the characteristic parameter extraction module include:
9. The GIS insulation fault type identification system based on UHF and CG-BP algorithm according to claim 6, characterized in that: the fault type classification module specifically classifies insulation fault types of the GIS equipment into a metal tip fault, an insulation internal air gap or creeping discharge fault, a free metal particle fault and a suspension discharge fault.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, implements the GIS insulation fault type identification method according to any one of claims 1 to 5.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210312079.6A CN114742140B (en) | 2022-03-28 | 2022-03-28 | GIS insulation fault type identification method based on UHF and CG-BP algorithm |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210312079.6A CN114742140B (en) | 2022-03-28 | 2022-03-28 | GIS insulation fault type identification method based on UHF and CG-BP algorithm |
Publications (2)
Publication Number | Publication Date |
---|---|
CN114742140A true CN114742140A (en) | 2022-07-12 |
CN114742140B CN114742140B (en) | 2024-05-31 |
Family
ID=82277327
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210312079.6A Active CN114742140B (en) | 2022-03-28 | 2022-03-28 | GIS insulation fault type identification method based on UHF and CG-BP algorithm |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN114742140B (en) |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102809718A (en) * | 2012-07-25 | 2012-12-05 | 华南理工大学 | Ultra-high-frequency partial discharge signal identification method for gas insulated switchgear (GIS) |
CN108918137A (en) * | 2018-06-08 | 2018-11-30 | 华北水利水电大学 | Fault Diagnosis of Gear Case devices and methods therefor based on improved WPA-BP neural network |
CN110232402A (en) * | 2019-05-14 | 2019-09-13 | 沈阳化工大学 | A kind of adaptive depth confidence network Fault Diagnosis of Roller Bearings of optimization |
CN110263907A (en) * | 2019-07-01 | 2019-09-20 | 上海海事大学 | Based on the ship short trouble diagnostic method for improving GA-PSO-BP |
-
2022
- 2022-03-28 CN CN202210312079.6A patent/CN114742140B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102809718A (en) * | 2012-07-25 | 2012-12-05 | 华南理工大学 | Ultra-high-frequency partial discharge signal identification method for gas insulated switchgear (GIS) |
CN108918137A (en) * | 2018-06-08 | 2018-11-30 | 华北水利水电大学 | Fault Diagnosis of Gear Case devices and methods therefor based on improved WPA-BP neural network |
CN110232402A (en) * | 2019-05-14 | 2019-09-13 | 沈阳化工大学 | A kind of adaptive depth confidence network Fault Diagnosis of Roller Bearings of optimization |
CN110263907A (en) * | 2019-07-01 | 2019-09-20 | 上海海事大学 | Based on the ship short trouble diagnostic method for improving GA-PSO-BP |
Non-Patent Citations (1)
Title |
---|
肖燕;胡浩;郁惟镛;: "基于ARTNN的GIS绝缘故障识别新方法", 高电压技术, no. 12, 15 December 2007 (2007-12-15) * |
Also Published As
Publication number | Publication date |
---|---|
CN114742140B (en) | 2024-05-31 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109188211B (en) | High-voltage equipment insulation fault diagnosis method and system | |
CN111562108A (en) | Rolling bearing intelligent fault diagnosis method based on CNN and FCMC | |
CN112001110B (en) | Structural damage identification monitoring method based on vibration signal space real-time recurrent graph convolutional neural network | |
CN110082136B (en) | Rotary machine fault diagnosis method based on cloud genetic algorithm optimization support vector machine | |
Mas’Ud et al. | Application of an ensemble neural network for classifying partial discharge patterns | |
CN116167010B (en) | Rapid identification method for abnormal events of power system with intelligent transfer learning capability | |
CN106950475A (en) | A kind of local discharge signal extracting method and device based on wavelet transformation | |
CN112596016A (en) | Transformer fault diagnosis method based on integration of multiple one-dimensional convolutional neural networks | |
CN112560981A (en) | Training method, apparatus, device, program and storage medium for generating countermeasure model | |
Jahan et al. | Locating power transformer fault based on sweep frequency response measurement by a novel multistage approach | |
CN111353529A (en) | Mixed attribute data set clustering method for automatically determining clustering center | |
CN114487129A (en) | Flexible material damage identification method based on acoustic emission technology | |
CN114742140B (en) | GIS insulation fault type identification method based on UHF and CG-BP algorithm | |
Wang et al. | Multi‐harmonic sources location based on sparse component analysis and complex independent component analysis | |
CN113987910A (en) | Method and device for identifying load of residents by coupling neural network and dynamic time planning | |
CN112016598A (en) | SDAE-based partial discharge pattern recognition method, computer device and storage medium | |
CN116400168A (en) | Power grid fault diagnosis method and system based on depth feature clustering | |
CN116150687A (en) | Fluid pipeline leakage identification method based on multi-classification G-WLSTSVM model | |
CN115061094A (en) | Radar target identification method based on neural network and SVM | |
CN114818845A (en) | Noise-stable high-resolution range profile feature selection method | |
Wang et al. | Research on rolling bearing fault diagnosis based on Volterra Kernel identification and KPCA | |
CN110032981B (en) | Rotary machine fault identification method based on improved support vector machine | |
CN112183745A (en) | High-voltage cable partial discharge mode identification method based on particle swarm algorithm and DBN | |
Karthikeyan et al. | A heuristic complex probabilistic neural network system for partial discharge pattern classification | |
CN111814108B (en) | Connection type intermittent fault diagnosis method based on self-organizing neural 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 |