CN116008756B - Insulation fault diagnosis method, system, equipment and medium for capacitive voltage transformer - Google Patents
Insulation fault diagnosis method, system, equipment and medium for capacitive voltage transformer Download PDFInfo
- Publication number
- CN116008756B CN116008756B CN202310247981.9A CN202310247981A CN116008756B CN 116008756 B CN116008756 B CN 116008756B CN 202310247981 A CN202310247981 A CN 202310247981A CN 116008756 B CN116008756 B CN 116008756B
- Authority
- CN
- China
- Prior art keywords
- fault
- samples
- sample
- insulation
- voltage
- 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.)
- Active
Links
- 238000009413 insulation Methods 0.000 title claims abstract description 190
- 238000003745 diagnosis Methods 0.000 title claims abstract description 59
- 238000000034 method Methods 0.000 title claims abstract description 56
- 230000009466 transformation Effects 0.000 claims abstract description 30
- 230000002159 abnormal effect Effects 0.000 claims abstract description 26
- 238000012549 training Methods 0.000 claims abstract description 19
- 230000006870 function Effects 0.000 claims description 24
- 239000003990 capacitor Substances 0.000 claims description 20
- 238000011156 evaluation Methods 0.000 claims description 15
- 238000004422 calculation algorithm Methods 0.000 claims description 14
- 239000011159 matrix material Substances 0.000 claims description 10
- 238000010606 normalization Methods 0.000 claims description 7
- 238000011176 pooling Methods 0.000 claims description 7
- 238000004364 calculation method Methods 0.000 claims description 6
- 238000004590 computer program Methods 0.000 claims description 6
- 230000008569 process Effects 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 5
- 238000013507 mapping Methods 0.000 claims description 4
- 238000005070 sampling Methods 0.000 claims description 4
- 238000009421 internal insulation Methods 0.000 description 11
- 230000015556 catabolic process Effects 0.000 description 7
- 210000002569 neuron Anatomy 0.000 description 4
- 238000012544 monitoring process Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 230000005856 abnormality Effects 0.000 description 2
- 238000006731 degradation reaction Methods 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 238000005295 random walk Methods 0.000 description 2
- 230000001131 transforming effect Effects 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000013145 classification model Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000006866 deterioration Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 230000000192 social effect Effects 0.000 description 1
Images
Classifications
-
- 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
- Y04S10/52—Outage or fault management, e.g. fault detection or location
Landscapes
- Testing Of Short-Circuits, Discontinuities, Leakage, Or Incorrect Line Connections (AREA)
- Testing Electric Properties And Detecting Electric Faults (AREA)
Abstract
The invention relates to a method, a system, equipment and a medium for diagnosing insulation faults of a capacitive voltage transformer, wherein the method comprises the following steps: acquiring historical insulation fault data of a plurality of groups of CVTs, and establishing a fault sample data set based on the historical insulation fault data; sample expansion is carried out on insulation fault type fault samples of a few classes in the fault sample data set; performing feature transformation on the fault sample, converting one-dimensional features into three-dimensional feature images, and generating an image sample set; constructing a Resnet-RBF network, and performing iterative training on the Resnet-RBF network by taking an image sample as input and an insulation fault type as output to obtain a trained fault diagnosis model; when the metering state of a group of CVTs is abnormal, extracting a voltage data set of the group of CVTs, performing insulation fault diagnosis on the voltage data set by using a fault diagnosis model, and positioning the CVTs with insulation faults in the group of CVTs based on the contribution index of the voltage data.
Description
Technical Field
The invention relates to a method, a system, equipment and a medium for diagnosing insulation faults of a capacitive voltage transformer, and belongs to the technical field of insulation diagnosis of transformers.
Background
CVT (capacitive voltage transformer) is complex in structure, and the number of internal components is large, and as a primary device for outdoor operation, the insulating state of CVT is deteriorated after long-term operation, and the deterioration of the internal insulating state directly affects the metering accuracy and operation stability of CVT. Similar to measurement error state evaluation, the internal insulation evaluation method widely applied at the present stage is also periodic power failure detection, and the method cannot learn the internal insulation state of the CVT in two adjacent test periods, and because the CVT power failure is required to be arranged in detection, the test work is heavier, so that a large number of CVTs which are not detected for an over period exist in a power grid.
Therefore, there is a need to make intensive studies on how to evaluate the internal insulation degradation state of CVT in real time, and to be able to diagnose the type of abnormality and the degree of abnormality in time when the internal insulation state is abnormally changed.
The invention patent number CN114358092A discloses an on-line diagnosis method for the internal insulation performance of a capacitive voltage transformer, which comprises the following steps: establishing a source domain data set and a target domain data set of a CVT diagnostic model, and determining individual characteristics and common characteristics in CVT data; extracting relevant parameters of the CVT with the internal insulation fault as common characteristics according to a CVT internal insulation performance degradation mechanism; estimating a common characteristic distribution model through a Dirichlet process Gaussian mixture model; constructing an internal insulation fault virtual CVT generation model to generate an internal insulation fault virtual CVT sample; based on the number of normal state CVT samples and virtual CVT samples in the balance target domain of the graph generator, distinguishing nodes on the balance network through a training graph convolution network identifier, and obtaining a fault diagnosis model; and performing internal insulation fault diagnosis on the CVT sample to be tested through a fault diagnosis model. According to the scheme, the sample data requirement can be reduced, the construction of the fault diagnosis model is realized based on the generated virtual sample, and the identification accuracy of the fault diagnosis model can be effectively ensured.
However, the above prior art has the technical problems that 1) a virtual CVT generating model is used to generate a virtual sample, training of a diagnostic model is performed using a training set including the virtual sample, the virtual sample is not real, and training of the model using the not real data causes a decrease in accuracy of the diagnostic model; 2) The characteristic parameters are complex, and the individual characteristics and the common characteristics are used as parameters, so that the complexity of the data characteristics is extremely high, and the individual characteristics of each sample are different, so that the training amount of the diagnosis model is large, and the fitting state is difficult to reach.
Disclosure of Invention
In order to solve the problems in the prior art, the invention provides a method, a system, equipment and a medium for diagnosing insulation faults of a capacitive voltage transformer.
The technical scheme of the invention is as follows:
in one aspect, the invention provides a method for diagnosing insulation faults of a capacitive voltage transformer, which comprises the following steps:
acquiring historical insulation fault data of a plurality of groups of CVTs, acquiring insulation fault types and output voltage data of each group of CVTs based on the historical insulation fault data, calculating positive sequence voltage, negative sequence voltage and zero sequence voltage of each group of CVTs according to the voltage data output by each group of CVTs, putting the positive sequence voltage, the negative sequence voltage and the zero sequence voltage into a fault sample data set, and taking the insulation fault types of each group of CVTs as labels of fault samples;
Oversampling is carried out on a few insulating fault type fault samples in the fault sample data set by adopting an SOMTE algorithm, so that sample expansion is completed;
performing feature transformation on the fault samples in the fault sample data set after sample expansion is completed, converting one-dimensional features into three-dimensional feature images, and generating an image sample set;
constructing a Resnet-RBF network consisting of a Resnet network model and an RBF network model, wherein the pooling layer output of the Resnet network model is used as the input of the RBF network model; taking an image sample as input, taking an insulation fault type as output, and performing iterative training on a Resnet-RBF network to obtain a trained fault diagnosis model;
when the metering state of a group of CVTs is abnormal, extracting a voltage data set of the group of CVTs in an abnormal period, performing insulation fault diagnosis on the voltage data set of the group of CVTs in the abnormal period by using the fault diagnosis model, and positioning the CVTs with insulation faults in the group of CVTs based on the contribution index of the voltage data.
As a preferred embodiment, the insulation fault types include: the capacitor is not broken down, a single high-voltage capacitor is broken down, two high-voltage capacitors are broken down, a single medium-voltage capacitor is broken down, and two medium-voltage capacitors are broken down;
the step of calculating positive sequence voltage, negative sequence voltage and zero sequence voltage of each CVT according to the voltage data output by each CVT comprises the following steps of:
Calculating positive sequence voltage, negative sequence voltage and zero sequence voltage of each CVT according to the collected voltage data output by each CVT:
wherein ,
、/>
、/>
a, B, C three-phase voltages, i=1, 2, …, N, < >, respectively representing i-th group CVT outputs>
、/>
、/>
Respectively representing positive sequence component, negative sequence component and zero sequence component of three-phase voltage of the i-th group of mutual inductor, and operator +.>
，/>
；
Performing park transformation processing on the positive sequence component, the negative sequence component and the zero sequence component of the three-phase voltage obtained through calculation:
wherein the park transformation is performed at an angular velocity in a d-q coordinate system
Rotate counterclockwise (i.e. go up)>
The direction of leading the d-axis to the a-phase axis is taken as positive; />
、/>
、/>
Park transformation results of positive sequence component, negative sequence component and zero sequence component respectively; />
The positive sequence component, the negative sequence component, the zero sequence component and the park transformation result of the three-phase voltage of each group of CVT are taken as fault samples to be put into a sample data set.
As a preferred embodiment, the step of oversampling the insulation fault type fault samples of a minority class in the fault sample data set by adopting the SOMTE algorithm specifically includes:
setting the number of insulation fault type fault samples of a few types after oversampling:
wherein ,
，/>
for the over sampling rate, +.>
、/>
The original number of the insulation fault type fault samples of the majority class and the minority class respectively;
Dividing the insulation fault type fault samples of the minority classes according to the unbalance degree, and dividing the insulation fault type fault samples of each minority class
Calculating its m nearest neighbors from the whole sample dataset Q, wherein the number of insulation failure type failure samples of the majority class is +.>
The number of insulating fault type fault samples of a minority class is expressed by +>
The representation divides the insulation fault type fault samples of a minority class based on the following equation:
where j=1, 2, …,
representing a minority class sample number; eliminating fault samples judged to be noise samples, and eliminating insulation faults divided into minority classes of boundary samples and safety samplesType fault samples are oversampled:
and distributing the number of the oversampling samples to the insulation fault type fault samples of the minority class of the non-noise samples based on the class index R, the distance index L and the density index P of the insulation fault type fault samples of the minority class.
As a preferred embodiment, the method for calculating the category index R, the distance index L, and the density index P specifically includes:
and (3) calculating a category index R:
wherein ,
insulation fault type fault sample set representing a minority class of non-noise>
For sample->
A number of insulation fault type fault samples of a majority class within m neighbors;
Calculating a distance index L:
wherein ,
any sample in the insulation fault type fault sample set representing a minority class of non-noise +.>
Sample clustering center similar to the sample clustering center>
Distance of (2):
wherein ,
is the sample dimension;
calculating a density index P:
for any sample
Calculating K neighbor samples within the insulation fault type fault sample set range of a non-noise minority class as +.>
K=1, 2, … …, K, calculate +.>
And->
The Euclidean distance of (2) is:
computing insulation fault type fault sample sets for a minority class of non-noise
The distance between each data point and its k neighbor is taken as the radius:
calculating a density value:
wherein ,
is expressed as->
The number of non-noise minority class insulation fault type fault samples with a distance less than r; cont () represents a count function.
As a preferred embodiment, in the step of allocating the number of oversampling samples to the insulation fault type fault samples of the minority class of the non-noise samples based on the category index R, the distance index L, and the density index P of the insulation fault type fault samples of the minority class of the non-noise samples, the number of oversampling samples is allocated to the insulation fault type fault samples of the minority class of the non-noise samples based on the following formula:
wherein ,
for allocation to samples->
Is used to sample the sample size.
In a preferred embodiment, the method for transforming the characteristics of the fault samples in the fault sample data set after the sample expansion and converting the one-dimensional characteristics into the three-dimensional characteristic images specifically adopts the gram angle field transformation, which specifically includes the following steps:
normalization processing is carried out, and one-dimensional characteristic data of a fault sample is obtained
Mapping to [0,1 ]]The normalization process formula is as follows:
wherein ,
=1,2,3，……，6；/>
、/>
、/>
positive sequence component, negative sequence component and zero sequence component of three-phase voltage in corresponding fault sample data respectively; />
、/>
、/>
Respectively obtaining park transformation results of positive sequence components, negative sequence components and zero sequence components of three-phase voltages in corresponding fault sample data; />
Normalizing the corresponding characteristic data to obtain a result;
normalized results of feature data
Encoded as the tailpiece angle, re-expressed in polar coordinates as +.>
and />
The following is shown:
wherein W is a constant factor set, u is a constant factor regularizing the polar coordinate system generation space,
is->
Polar angle of point->
Is->
A dot polar diameter;
and respectively carrying out angle summation inner product and difference inner product on the calculated polar coordinates to obtain two types of gram angle field GASF matrixes and GADF matrixes:
And remapping the obtained GASF matrix and GADF matrix with the numerical values distributed in [0, 1] to the pixel value intervals of [0,255], and copying the pixel value intervals to RGB three channels in turn to enable the pixel values to be changed into a standard three-dimensional color image serving as the three-dimensional characteristic image.
As a preferred embodiment, the step of locating the CVT of the insulation fault in the group of CVTs based on the contribution index of the voltage data specifically includes:
acquiring a voltage data set of the CVT in the abnormal period, wherein the time sequence length of the data set is Z;
constructing a contribution rate index:
wherein Y is a contribution index,
contribution rate for one of the three phases, +.>
Voltage of one of three phases, +.>
For time sequence representation +.>
The average value of the amplitude values of the three-phase voltages at the same time sequence is obtained;
contribution rate index in the group of CVTs at abnormal time periods
The largest CVT acts as an insulation failure CVT;
the method also comprises the step of judging the CVT insulation fault under different wiring modes, and specifically comprises the following steps:
determining a primary wiring form of the target set CVT;
when the primary main wiring mode is double-bus wiring, if the CVT on any phase of one bus has insulation fault, forming a new evaluation voltage data set by using the voltage data of the other two phases on the bus and the voltage data of the same phase with the insulation fault CVT on the other bus, and re-executing the CVT insulation fault diagnosis method;
When the primary main wiring is in the form of double bus-section wiring or 3/2 wiring, if the CVT on any phase of one bus has insulation fault, a new evaluation voltage data set is formed by the voltage data of the other two phases on the bus and the voltage data average value of all other CVTs with the high voltage side in phase with the insulation fault CVT, and the CVT insulation fault diagnosis method is re-executed.
On the other hand, the invention also provides a capacitive voltage transformer insulation fault diagnosis system, which comprises:
the data set establishing module is used for acquiring historical insulation fault data of a plurality of groups of CVTs, acquiring insulation fault types and output voltage data of each group of CVTs based on the historical insulation fault data, calculating positive sequence voltage, negative sequence voltage and zero sequence voltage of each group of CVTs according to the voltage data output by each group of CVTs, putting the positive sequence voltage, the negative sequence voltage and the zero sequence voltage into a fault sample data set, and taking the insulation fault types of each group of CVTs as labels of fault samples;
the sample expansion module is used for oversampling a few insulating fault type fault samples in the fault sample data set by adopting an SOMTE algorithm to finish sample expansion;
the feature transformation module is used for carrying out feature transformation on the fault samples in the fault sample data set after the sample expansion is completed, converting one-dimensional features into three-dimensional feature images and generating an image sample set;
The model training module is used for constructing a Resnet-RBF network consisting of a Resnet network model and an RBF network model, wherein the pooling layer output of the Resnet network model is used as the input of the RBF network model; taking an image sample as input, taking an insulation fault type as output, and performing iterative training on a Resnet-RBF network to obtain a trained fault diagnosis model;
and the diagnosis module is used for extracting the voltage data set of the CVT in the abnormal period when the metering state of the CVT in the group is abnormal, performing insulation fault diagnosis on the voltage data set of the CVT in the abnormal period by using the fault diagnosis model, and positioning the CVT with the insulation fault in the CVT based on the contribution index of the voltage data.
In still another aspect, the present invention further provides an electronic device, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor, where the processor implements the insulation fault diagnosis method of the capacitive voltage transformer according to any embodiment of the present invention when executing the program.
In still another aspect, the present invention further provides a computer readable storage medium having stored thereon a computer program, which when executed by a processor, implements a capacitive voltage transformer insulation fault diagnosis method according to any of the embodiments of the present invention.
The invention has the following beneficial effects:
1. according to the insulation fault diagnosis method of the capacitive voltage transformer, the insulation fault type and the voltage data are obtained through the historical insulation fault data of the CVT, the characteristic parameters of the samples are calculated through the voltage data, the characteristic parameters are simple and effective in selection, the SOMTE algorithm is adopted for sample expansion of a few types of samples, the objectivity and the effectiveness of the expanded samples are guaranteed based on the real samples, the training efficiency of a diagnosis model is greatly improved, and the accuracy of CVT insulation fault diagnosis is improved.
2. According to the insulation fault diagnosis method of the capacitive voltage transformer, based on the class index R, the distance index L and the density index P of the minority class samples, the improved SMOTE method is adopted to oversample the minority class fault samples, and the balance of sample data is ensured.
3. The invention discloses an insulation fault diagnosis method of a capacitive voltage transformer, which adopts a Graham angle field transformation to convert an internal insulation one-dimensional characteristic parameter into a three-dimensional characteristic image, thereby effectively enhancing the fault characteristics of one-dimensional original data.
4. According to the insulation fault diagnosis method for the capacitive voltage transformer, provided by the invention, the Resnet network model is optimized, the RBF model is used for replacing the softmax classification model, and the algorithm accuracy is improved.
5. According to the insulation fault diagnosis method for the capacitive voltage transformer, provided by the invention, aiming at the condition that CVT insulation faults occur in different wiring forms, new modes are respectively constructed, so that the judgment of the insulation states of a plurality of CVTs is finished, and the adaptability of the diagnosis method is greatly enhanced.
Drawings
FIG. 1 is a flowchart illustrating a method according to an embodiment of the present invention;
fig. 2 is a diagram illustrating a residual unit in an embodiment of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
It should be understood that the step numbers used herein are for convenience of description only and are not limiting as to the order in which the steps are performed.
It is to be understood that the terminology used in the description of the invention 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 stated 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.
Embodiment one:
referring to fig. 1, the embodiment provides a method for diagnosing insulation faults of a capacitive voltage transformer, which specifically includes the following steps:
s100, acquiring historical insulation fault data of a plurality of groups of CVTs, acquiring insulation fault types and output voltage data of each group of CVTs based on the historical insulation fault data, calculating positive sequence voltage, negative sequence voltage and zero sequence voltage of each group of CVTs according to the voltage data output by each group of CVTs, putting the positive sequence voltage, the negative sequence voltage and the zero sequence voltage into a fault sample data set, and taking the insulation fault types of each group of CVTs as labels of fault samples.
In one of the embodiments, in step S100, the insulation fault type is defined as: a= [ capacitor breakdown, single high voltage capacitor breakdown, two high voltage capacitor breakdown, single medium voltage capacitor breakdown, two medium voltage capacitor breakdown ]; the 5 insulation fault types in set a also constitute a set of labels for the fault samples.
In one embodiment, in step S100, the step of calculating the positive sequence voltage, the negative sequence voltage and the zero sequence voltage of each group of CVT according to the voltage data output by each group of CVT includes:
s101, calculating positive sequence voltage, negative sequence voltage and zero sequence voltage of each group of CVTs (A, B, C phases) according to the collected voltage data output by each group of CVTs, wherein the calculation formula is as follows:
wherein ,
、/>
、/>
a, B, C three-phase voltages, i=1, 2, …, N,
、/>
、/>
respectively representing positive sequence component, negative sequence component and zero sequence component of three-phase voltage of the i-th group of mutual inductor, and operator +.>
，/>
；
S102, performing park transformation processing on the positive sequence component, the negative sequence component and the zero sequence component of the three-phase voltage obtained through calculation:
wherein the park transformation is performed at an angular velocity in a d-q coordinate system
Rotate counterclockwise (i.e. go up)>
The direction of leading the d-axis to the a-phase axis is taken as positive; />
、/>
、/>
Park transformation results of positive sequence component, negative sequence component and zero sequence component respectively.
S103, transforming results by positive sequence component, negative sequence component, zero sequence component and park of three-phase voltages of each group of CVT
、/>
and />
The sample data set is composed as a fault sample:
wherein ,
the i-th sample is represented, i=1, 2, …, N, and the number of samples is represented.
S200, considering that the CVT insulation fault samples have the problem of inter-class unbalance, namely that the insulation fault samples have too many insulation fault types and too few insulation fault types, the embodiment adopts an SOMTE algorithm to oversample the insulation fault type fault samples of a few classes in the fault sample data set, and completes sample expansion.
In one embodiment, the step S200 specifically includes:
s201, setting the number of insulation fault type fault samples of a few types (few types refer to insulation fault types with a small number and a plurality of types are the same) after oversampling:
wherein ,
，/>
for the over sampling rate, +.>
、/>
The original number of insulation fault type fault samples of the majority class and minority class respectively.
S202, dividing insulation fault type fault samples of minority classes according to unbalance degree, and dividing the insulation fault type fault samples of each minority class
Calculating its m nearest neighbors from the whole sample dataset Q, wherein the number of insulation failure type failure samples of the majority class is +.>
The number of insulating fault type fault samples of a minority class is expressed by +>
The representation divides the insulation fault type fault samples of a minority class based on the following equation: / >
Where j=1, 2, …,
representing a minority class sample number; eliminating fault samples judged to be noise samples, and oversampling insulation fault type fault samples divided into a minority of boundary samples and safety samples:
s203, distributing the number of oversampling samples for the insulation fault type fault samples of the minority class of the non-noise samples based on the class index R, the distance index L and the density index P of the insulation fault type fault samples of the minority class.
In one embodiment, in step S203, the calculating method of the category index R, the distance index L, and the density index P specifically includes:
and (3) calculating a category index R:
wherein ,
insulation fault type fault sample set representing a minority class of non-noise>
For sample->
A number of insulation fault type fault samples of a majority class within m neighbors;
calculating a distance index L:
wherein ,
any sample in the insulation fault type fault sample set representing a minority class of non-noise +.>
Sample clustering center similar to the sample clustering center>
Distance of (2):
wherein ,
is the sample dimension;
calculating a density index P:
for any sample
Calculating K neighbor samples within the insulation fault type fault sample set range of a non-noise minority class as +. >
K=1, 2, … …, K, calculate +.>
And->
The Euclidean distance of (2) is:
computing insulation fault type fault sample sets for a minority class of non-noise
The distance between each data point and its k neighbor is taken as the radius:
calculating a density value:
wherein ,
is expressed as->
The number of non-noise minority class insulation fault type fault samples with a distance less than r; cont () represents a count function.
In one embodiment, in step S203, the number of oversampled samples is assigned for insulation fault type fault samples of a minority class of non-noise samples based on the following formula:
wherein ,
for allocation to samples->
Is used to sample the sample size.
S300, performing feature transformation on the fault samples in the fault sample data set after sample expansion, converting one-dimensional features into three-dimensional feature images, and generating an image sample set.
In one embodiment, the method of converting the one-dimensional feature into the three-dimensional feature image in step S300 specifically uses a gram angle field transform, where the gram angle field (Gramian Angular Field, GAF) is a data dimension transform method based on polar-coordinate gram matrix, and the method can convert the one-dimensional sample into the three-dimensional feature image sample, which can effectively enhance the feature of the one-dimensional original data. The specific steps of S300 are as follows:
S301, using a maximum and minimum normalization method to obtain one-dimensional feature data of a fault sample
Mapping to [0, 1]]The normalization process formula is as follows:
wherein ,
=1,2,3，……，6；/>
、/>
、/>
positive sequence component, negative sequence component and zero sequence component of three-phase voltage in corresponding fault sample data respectively; />
、/>
、/>
Respectively obtaining park transformation results of positive sequence components, negative sequence components and zero sequence components of three-phase voltages in corresponding fault sample data; />
And normalizing the corresponding characteristic data.
S302, normalizing the characteristic data to obtain a result
Encoded as the tailpiece angle, re-expressed in polar coordinates as +.>
and />
The following is shown:
wherein W is a normal valueA set of numerical factors, u is a constant factor that regularizes the polar coordinate system generation space,
is->
Polar angle of point->
Is->
Dot pole diameter.
S303, after converting the structured data into polar coordinates, respectively carrying out angle summation inner products and difference inner products on the calculated polar coordinates to obtain two types of gram angle field GASF matrixes and GADF matrixes:
s304, the obtained GASF matrix and GADF matrix with the values distributed in [0, 1] are remapped to pixel value intervals of [0, 255], and are sequentially copied to RGB three channels, so that the RGB three-channel three-dimensional color image is changed into a standard three-dimensional color image (width multiplied by height multiplied by three primary color channel number) to be used as the three-dimensional characteristic image.
S400, constructing a Resnet-RBF network consisting of a Resnet network model and an RBF network model.
Specifically, the Resnet network model is composed of a stack of residual units, where one basic residual unit (as shown in FIG. 2) includes a convolutional layer, a bulk normalization layer (BN), an activation function, and an identity map. The input picture may be based on the output of the last residual unit and then the output is sent to an averaging pooling layer and classification layer. In the model training process, the cross entropy function is used as a loss function to evaluate the fitting condition of the model.
Each residual unit may be represented by the following formula:
wherein ,
indicate->
The outputs of the residual units,/>
Indicate->
The outputs of the residual units,/>
Representing the residual function.
The output of the last residual unit is:
wherein ,
indicate->
The outputs of the residual units.
The classifier behind the Resnet network typically chooses a Softmax function. The traditional Softmax classifier has poor recognition effect on types with small feature difference and large ambiguity, and the RBF model with strong learning ability is used for replacing the Softmax classifier aiming at the defect, and the embodiment adopts the standard constants of the center vector c and the basis function of the RBF by using the F algorithm 
And optimizing to improve the accuracy of model state diagnosis.
The RBF is a three-layer forward neural network, and in this embodiment, the RBF has the following structure:
input layer:
the output of the pooling layer of the resnet network is taken as the input of the RBF network and is defined as B.
Hidden layer:
the RBF hidden layer neuron kernel function is a Gaussian function and performs space mapping transformation on input information. Here, a gaussian kernel function is used as the implicit neuron basis function:
wherein ,
for implicit layer output, < >>
=1, 2, …, q, the number of implicit neurons. B is an input vector of RBF; />
Is the center vector of the Gaussian function; />
Is a normalized constant of the basis function. />
Output layer:
the expression is:
wherein ,
weight information for the q-th hidden neuron, < ->
The classification probability output by the model is the classification result of the maximum sample, p=1, 2, …,5, the corresponding capacitor is not broken down, the single high-voltage capacitor is broken down, the two high-voltage capacitors are broken down, the single medium-voltage capacitor is broken down, and the two medium-voltage capacitors are broken down to obtain 5 insulation fault type classification results.
In this embodiment, the f-algorithm is used for RBFNormalized constant of center vector c and basis function
The optimizing step is as follows:
definition of dayfish
Position information of (a) corresponds to the center vector c in RBF, normalized constant of basis function +. >
；
Setting the speed of male dayfish:
wherein ,
for the t+1st iteration, +.>
Male only, f>
The speed of the dimension; />
For the t-th iteration +.>
Male only, f>
The position of the dimension. />
Is a positive attraction coefficient for the social effect,
representing optimal position of the individual of the falciform->
Representing a global optimum position. />
Is a fixed visibility coefficient. />
Represents the distance of the current location from the optimal location of the individual, < + >>
Representing the distance of the current location from the global optimal location.
Setting the speed of female dayfish:
wherein ,
representing the distance between the female and the male member, and (2)>
For the t-th iteration +.>
Female only, the first one->
Position of dimension->
Is a random walk coefficient, r is a random walk coefficient in the range of [ -1,1]Is a random number of (a) in the memory.
Is->
Fitness of female only dayfish ++>
Is->
Male-only, f-type fitness.
Parameters B1, B2 are defined to adjust the speed function of the dayf, better balancing the local search and local development capabilities:
wherein T is the iteration number, and T is the total iteration number.
Then: the speed optimization of male dayfish is:
the female dayfish's speed is optimized as:
when (when)
When (1):
when (when)
When (1):
based on the optimized f-speed formula, the optimal f-position is found by iterative optimization, and the optimal c is correspondingly found, 
Parameters are input into the RBF network model.
Based on the optimized Resnet-RBF network, taking an image sample as input and taking an insulation fault type (sample label) as output to perform iterative training on the Resnet-RBF network, so as to obtain a trained fault diagnosis model.
In this embodiment, 70% of the image sample set is selected for model training, and the remaining 30% is used for model testing. Defining training image samples
, wherein />
Is->
Three-dimensional color image after GASF matrix transformation of individual image samples, < >>
Is->
Three-dimensional color image after GADF matrix transformation of the individual image samples,>
=1，2，…,/>
，/>
is the total number of image samples.
S500, when the metering state of a group of CVTs is abnormal, extracting a voltage data set of the group of CVTs in an abnormal period, performing insulation fault diagnosis on the voltage data set of the group of CVTs in the abnormal period by using the fault diagnosis model, and positioning the CVTs with insulation faults in the group of CVTs based on the contribution index of the voltage data.
In one embodiment, the step of locating the CVT of the insulation fault in the group of CVTs based on the contribution index of the voltage data specifically includes:
based on the three-phase voltage monitoring data, acquiring a voltage data set of the CVT in an abnormal period, wherein the time sequence length of the data set is Z;
Constructing a contribution rate index:
wherein Y is a contribution index,
contribution rate for one of the three phases, +.>
Voltage for one of the three phases, namely: />
，/>
For the time-series representation,
the average value of the amplitude values of the three-phase voltages at the same time sequence is obtained;
contribution rate index in the group of CVTs at abnormal time periods
The largest CVT acts as an insulation failure CVT.
Based on the embodiment, the invention solves the problem of diagnosing the insulation faults in a single CVT in a group, but the abnormal CVT cannot be replaced immediately due to the fact that the CVT is difficult to power failure, so that the evaluation parameters used by the original evaluation group are changed, and the diagnosis model fails. Based on this technical problem, the present embodiment proposes an adaptive method to enhance the applicability of the evaluation method.
In one embodiment, for the case that a CVT insulation fault occurs in different connection modes, the method of this embodiment further includes a step of determining the CVT insulation fault in different connection modes, specifically:
determining a primary wiring form of the target set CVT; the CVT is widely used for 220: 220 kV and above voltage class power transmission networks, and primary main wiring mainly comprises wiring modes such as double-bus wiring, double-bus sectional wiring, one half-breaker wiring (3/2 wiring) and the like.
When the primary main connection is in the form of a double bus connection, if the CVT on any one phase of one bus has an insulation fault, a new evaluation voltage data set is formed by the voltage data of the other two phases on the bus and the voltage data of the other bus in phase with the insulation fault CVT, for example:
if insulation fault occurs in the primary-secondary A1 phase CVT, a new evaluation voltage data set needs to be constructed to realize continuous monitoring of the B1 and C1 phase CVT: x1= { A2, B1, C1}. Wherein A2 represents voltage data of the ii matrix a-phase CVT. The CVT insulation fault diagnosis method is re-executed based on the new evaluation voltage data set.
When the primary main wiring is in the form of a double bus-section wiring or a 3/2 wiring, if the CVT on any one phase of one bus has an insulation fault, a new evaluation voltage data set is formed by the voltage data of the other two phases on the bus and the voltage data average value of all other CVTs with the high voltage side in phase with the insulation fault CVT, for example:
if insulation faults occur in the I busbar A1 phase CVT, a new evaluation group needs to be constructed in order to realize continuous monitoring of the B1 and C1 phase CVT: x1= { (a2+a3+ … +an)/n-1, B1, C1}, where (a2+a3+ … +an)/n-1 represents the voltage data average value of all other CVT whose high voltage side is in phase with A1. The CVT insulation fault diagnosis method is re-executed based on the new evaluation voltage data set.
Embodiment two:
the invention also provides a system for diagnosing insulation faults of the capacitive voltage transformer, which comprises:
the data set establishing module is used for acquiring historical insulation fault data of a plurality of groups of CVTs, acquiring insulation fault types and output voltage data of each group of CVTs based on the historical insulation fault data, calculating positive sequence voltage, negative sequence voltage and zero sequence voltage of each group of CVTs according to the voltage data output by each group of CVTs, putting the positive sequence voltage, the negative sequence voltage and the zero sequence voltage into a fault sample data set, and taking the insulation fault types of each group of CVTs as labels of fault samples; the module is used for implementing the function of step S100 in the first embodiment, and will not be described here again;
the sample expansion module is used for oversampling a few insulating fault type fault samples in the fault sample data set by adopting an SOMTE algorithm to finish sample expansion; the module is used for implementing the function of step S200 in the first embodiment, and will not be described in detail herein;
the feature transformation module is used for carrying out feature transformation on the fault samples in the fault sample data set after the sample expansion is completed, converting one-dimensional features into three-dimensional feature images and generating an image sample set; the module is used for implementing the function of step S300 in the first embodiment, and will not be described in detail herein;
The model training module is used for constructing a Resnet-RBF network consisting of a Resnet network model and an RBF network model, wherein the pooling layer output of the Resnet network model is used as the input of the RBF network model; taking an image sample as input, taking an insulation fault type as output, and performing iterative training on a Resnet-RBF network to obtain a trained fault diagnosis model; the module is used for realizing the function of step S400 in the first embodiment, and will not be described in detail herein;
the diagnosis module is used for extracting a voltage data set of the CVT in an abnormal period when the metering state of the CVT in the group is abnormal, performing insulation fault diagnosis on the voltage data set of the CVT in the abnormal period by using the fault diagnosis model, and positioning the CVT with the insulation fault in the CVT in the group based on the contribution index of the voltage data; the module is used to implement the function of step S500 in the first embodiment, and will not be described herein.
Embodiment III:
the embodiment provides an electronic device, which comprises a memory, a processor and a computer program stored on the memory and capable of running on the processor, wherein the processor realizes the insulation fault diagnosis method of the capacitive voltage transformer according to any embodiment of the invention when executing the program.
Embodiment four:
the present embodiment proposes a computer-readable storage medium, on which a computer program is stored, which when executed by a processor implements a capacitive voltage transformer insulation fault diagnosis method according to any of the embodiments 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 an association relation of association objects, and indicates that there may be three kinds of relations, for example, a and/or B, and may indicate that a alone exists, a and B together, and B alone exists. Wherein A, B may be singular or plural. The character "/" generally indicates that the context-dependent object is an "or" relationship. "at least one of the following" and the like means any combination of these items, including any combination of single 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 the embodiments disclosed herein can be implemented as a combination of electronic hardware, computer software, and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the 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 will be clear to those skilled in the art that, for convenience and brevity of description, specific working procedures of the above-described systems, apparatuses and units may refer to corresponding procedures in the foregoing method embodiments, and are not repeated herein.
In several embodiments provided herein, any of the functions, if implemented in the form of software functional units and sold or used as a stand-alone product, may be stored in a computer-readable storage medium. Based on such understanding, the technical solution of the present application may be embodied essentially or in a part contributing to the prior art or in a part of the technical solution, in the form of a software product stored in a storage medium, including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in the embodiments of the present application. And the aforementioned storage medium includes: a usb disk, a removable hard disk, a Read-Only Memory (hereinafter referred to as ROM), a random access Memory (Random Access Memory) and various media capable of storing program codes such as a magnetic disk or an optical disk.
The foregoing description is only illustrative of the present invention and is not intended to limit the scope of the invention, and all equivalent structures or equivalent processes or direct or indirect application in other related technical fields are included in the scope of the present invention.
Claims (8)
1. The insulation fault diagnosis method for the capacitive voltage transformer is characterized by comprising the following steps of:
acquiring historical insulation fault data of a plurality of groups of CVTs, acquiring insulation fault types and output voltage data of each group of CVTs based on the historical insulation fault data, calculating positive sequence voltage, negative sequence voltage and zero sequence voltage of each group of CVTs according to the voltage data output by each group of CVTs, putting the positive sequence voltage, the negative sequence voltage and the zero sequence voltage into a fault sample data set, and taking the insulation fault types of each group of CVTs as labels of fault samples;
oversampling is carried out on a few insulating fault type fault samples in the fault sample data set by adopting an SOMTE algorithm, so that sample expansion is completed;
performing feature transformation on the fault samples in the fault sample data set after sample expansion is completed, converting one-dimensional features into three-dimensional feature images, and generating an image sample set;
constructing a Resnet-RBF network consisting of a Resnet network model and an RBF network model, wherein the pooling layer output of the Resnet network model is used as the input of the RBF network model; taking an image sample as input, taking an insulation fault type as output, and performing iterative training on a Resnet-RBF network to obtain a trained fault diagnosis model;
When the metering state of a group of CVTs is abnormal, extracting a voltage data set of the group of CVTs in an abnormal period, performing insulation fault diagnosis on the voltage data set of the group of CVTs in the abnormal period by using the fault diagnosis model, and positioning the CVTs with insulation faults in the group of CVTs based on the contribution index of the voltage data;
the step of oversampling insulation fault type fault samples of a minority class in the fault sample data set by adopting a SOMTE algorithm specifically comprises the following steps:
setting the number of insulation fault type fault samples of a few types after oversampling:
wherein ,
，/>
for the over sampling rate, +.>
、/>
The original number of the insulation fault type fault samples of the majority class and the minority class respectively;
dividing the insulation fault type fault samples of the minority classes according to the unbalance degree, and dividing the insulation fault type fault samples of each minority class
Calculating its m nearest neighbors from the whole sample dataset Q, wherein the number of insulation failure type failure samples of the majority class is +.>
The number of insulating fault type fault samples of a minority class is expressed by +>
The representation divides the insulation fault type fault samples of a minority class based on the following equation:
where j=1, 2, …,
representing a minority class sample number; eliminating fault samples judged to be noise samples, and oversampling insulation fault type fault samples divided into a minority of boundary samples and safety samples:
Based on a class index R, a distance index L and a density index P of the insulation fault type fault samples of the minority class, the number of the oversampling samples is distributed for the insulation fault type fault samples of the minority class of the non-noise samples;
the calculation method of the category index R, the distance index L and the density index P specifically comprises the following steps:
and (3) calculating a category index R:
wherein ,
insulation fault type fault sample set representing a minority class of non-noise>
For sample->
A number of insulation fault type fault samples of a majority class within m neighbors;
calculating a distance index L:
wherein ,
any sample in the insulation fault type fault sample set representing a minority class of non-noise +.>
Sample clustering center similar to the sample clustering center>
Distance of (2):
wherein ,
is the sample dimension;
calculating a density index P:
for any sample
Calculating K neighbor samples within the insulation fault type fault sample set range of a non-noise minority class as +.>
K=1, 2, … …, K, calculate +.>
And->
The Euclidean distance of (2) is: />
Calculating non-noiseInsulation fault type fault sample set for minority classes
The distance between each data point and its k neighbor is taken as the radius:
calculating a density value:
wherein ,
is expressed as- >
The number of non-noise minority class insulation fault type fault samples with a distance less than r; cont () represents a count function.
2. The method for diagnosing an insulation fault of a capacitive voltage transformer according to claim 1, wherein:
the insulation fault types include: the capacitor is not broken down, a single high-voltage capacitor is broken down, two high-voltage capacitors are broken down, a single medium-voltage capacitor is broken down, and two medium-voltage capacitors are broken down;
the step of calculating positive sequence voltage, negative sequence voltage and zero sequence voltage of each CVT according to the voltage data output by each CVT comprises the following steps of:
calculating positive sequence voltage, negative sequence voltage and zero sequence voltage of each CVT according to the collected voltage data output by each CVT:
wherein ,
、/>
、/>
a, B, C three-phase voltages, i=1, 2, …, N,
、/>
、/>
respectively representing positive sequence component, negative sequence component and zero sequence component of three-phase voltage of the i-th group of mutual inductor, and operator +.>
，/>
；
Performing park transformation processing on the positive sequence component, the negative sequence component and the zero sequence component of the three-phase voltage obtained through calculation:
wherein the park transformation is performed at an angular velocity in a d-q coordinate system
Rotate counterclockwise (i.e. go up)>
The direction of leading the d-axis to the a-phase axis is taken as positive; / >
、/>
、/>
Park transformation results of positive sequence component, negative sequence component and zero sequence component respectively;
the positive sequence component, the negative sequence component, the zero sequence component and the park transformation result of the three-phase voltage of each group of CVT are taken as fault samples to be put into a sample data set.
3. The insulation fault diagnosis method of a capacitive voltage transformer according to claim 1, wherein in the step of allocating the number of oversampled samples to the insulation fault type fault samples of the minority class of the non-noise samples based on the category index R, the distance index L, and the density index P of the insulation fault type fault samples of the minority class of the non-noise samples, the number of oversampled samples is allocated to the insulation fault type fault samples of the minority class of the non-noise samples based on the following formula:
wherein ,
for allocation to samples->
Is used to sample the sample size.
4. The insulation fault diagnosis method of a capacitive voltage transformer according to claim 2, wherein the method for converting one-dimensional features into three-dimensional feature images specifically adopts a gram angle field transformation, and specifically comprises the following steps:
normalization processing is carried out, and one-dimensional characteristic data of a fault sample is obtained 
Mapping to [0, 1]]The normalization process formula is as follows:
wherein ,
=1,2,3，……，6；/>
、/>
、/>
positive sequence component, negative sequence component and zero sequence component of three-phase voltage in corresponding fault sample data respectively; />
、/>
、/>
Respectively obtaining park transformation results of positive sequence components, negative sequence components and zero sequence components of three-phase voltages in corresponding fault sample data; />
Normalizing the corresponding characteristic data to obtain a result; />
Normalized results of feature data
Encoded as the tailpiece angle, re-expressed in polar coordinates as +.>
and />
The following is shown:
wherein W is a constant set, u is a constant factor regularizing the polar coordinate system generation space,
is->
Polar angle of point->
Is->
A dot polar diameter;
and respectively carrying out angle summation inner product and difference inner product on the calculated polar coordinates to obtain two types of gram angle field GASF matrixes and GADF matrixes:
and remapping the obtained GASF matrix and GADF matrix with the numerical values distributed in [0, 1] to the pixel value intervals of [0,255], and copying the pixel value intervals to RGB three channels in turn to enable the pixel values to be changed into a standard three-dimensional color image serving as the three-dimensional characteristic image.
5. The method for diagnosing insulation faults of a capacitive voltage transformer according to claim 1, wherein the step of locating the CVT of the group of CVT based on the contribution index of the voltage data is specifically:
Acquiring a voltage data set of the CVT in the abnormal period, wherein the time sequence length of the data set is Z;
constructing a contribution rate index:
wherein Y is a contribution index,
contribution rate for one of the three phases, +.>
Voltage of one of three phases, +.>
For time sequence representation +.>
The average value of the amplitude values of the three-phase voltages at the same time sequence is obtained;
contribution rate index in the group of CVTs at abnormal time periods
The largest CVT acts as an insulation failure CVT;
the method also comprises the step of judging the CVT insulation fault under different wiring modes, and specifically comprises the following steps:
determining a primary wiring form of the target set CVT;
when the primary main wiring mode is double-bus wiring, if the CVT on any phase of one bus has insulation fault, forming a new evaluation voltage data set by using the voltage data of the other two phases on the bus and the voltage data of the same phase with the insulation fault CVT on the other bus, and re-executing the CVT insulation fault diagnosis method;
when the primary main wiring is in the form of double bus-section wiring or 3/2 wiring, if the CVT on any phase of one bus has insulation fault, a new evaluation voltage data set is formed by the voltage data of the other two phases on the bus and the voltage data average value of all other CVTs with the high voltage side in phase with the insulation fault CVT, and the CVT insulation fault diagnosis method is re-executed.
6. A capacitive voltage transformer insulation fault diagnostic system, comprising:
the data set establishing module is used for acquiring historical insulation fault data of a plurality of groups of CVTs, acquiring insulation fault types and output voltage data of each group of CVTs based on the historical insulation fault data, calculating positive sequence voltage, negative sequence voltage and zero sequence voltage of each group of CVTs according to the voltage data output by each group of CVTs, putting the positive sequence voltage, the negative sequence voltage and the zero sequence voltage into a fault sample data set, and taking the insulation fault types of each group of CVTs as labels of fault samples;
the sample expansion module is used for oversampling a few insulating fault type fault samples in the fault sample data set by adopting an SOMTE algorithm to finish sample expansion;
the feature transformation module is used for carrying out feature transformation on the fault samples in the fault sample data set after the sample expansion is completed, converting one-dimensional features into three-dimensional feature images and generating an image sample set;
the model training module is used for constructing a Resnet-RBF network consisting of a Resnet network model and an RBF network model, wherein the pooling layer output of the Resnet network model is used as the input of the RBF network model; taking an image sample as input, taking an insulation fault type as output, and performing iterative training on a Resnet-RBF network to obtain a trained fault diagnosis model;
The diagnosis module is used for extracting a voltage data set of the CVT in an abnormal period when the metering state of the CVT in the group is abnormal, performing insulation fault diagnosis on the voltage data set of the CVT in the abnormal period by using the fault diagnosis model, and positioning the CVT with the insulation fault in the CVT in the group based on the contribution index of the voltage data;
the step of oversampling insulation fault type fault samples of a minority class in the fault sample data set by adopting a SOMTE algorithm specifically comprises the following steps:
setting the number of insulation fault type fault samples of a few types after oversampling:
wherein ,
，/>
for the over sampling rate, +.>
、/>
The original number of the insulation fault type fault samples of the majority class and the minority class respectively;
dividing the insulation fault type fault samples of the minority classes according to the unbalance degree, and dividing the insulation fault type fault samples of each minority class
Calculating its m nearest neighbors from the whole sample dataset Q, wherein the number of insulation failure type failure samples of the majority class is +.>
The number of insulating fault type fault samples of a minority class is expressed by +>
The representation divides the insulation fault type fault samples of a minority class based on the following equation:
where j=1, 2, …,
representing a minority class sample number; eliminating fault samples judged to be noise samples, and oversampling insulation fault type fault samples divided into a minority of boundary samples and safety samples:
Based on a class index R, a distance index L and a density index P of the insulation fault type fault samples of the minority class, the number of the oversampling samples is distributed for the insulation fault type fault samples of the minority class of the non-noise samples;
the calculation method of the category index R, the distance index L and the density index P specifically comprises the following steps:
and (3) calculating a category index R:
wherein ,
insulation fault type fault sample set representing a minority class of non-noise>
For sample->
A number of insulation fault type fault samples of a majority class within m neighbors;
calculating a distance index L:
wherein ,
insulation faults representing a minority class of non-noiseAny sample in a sample set of type faults +.>
Sample clustering center similar to the sample clustering center>
Distance of (2):
wherein ,
is the sample dimension; />
Calculating a density index P:
for any sample
Calculating K neighbor samples within the insulation fault type fault sample set range of a non-noise minority class as +.>
K=1, 2, … …, K, calculate +.>
And->
The Euclidean distance of (2) is:
computing insulation fault type fault sample sets for a minority class of non-noise
The distance between each data point and its k neighbor is taken as the radius:
calculating a density value:
wherein ,
is expressed as- >
The number of non-noise minority class insulation fault type fault samples with a distance less than r; cont () represents a count function.
7. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the method for diagnosing insulation faults of a capacitive voltage transformer according to any of claims 1 to 5 when executing the program.
8. A computer-readable storage medium, on which a computer program is stored, characterized in that the program, when executed by a processor, implements the capacitive voltage transformer insulation fault diagnosis method according to any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310247981.9A CN116008756B (en) | 2023-03-15 | 2023-03-15 | Insulation fault diagnosis method, system, equipment and medium for capacitive voltage transformer |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310247981.9A CN116008756B (en) | 2023-03-15 | 2023-03-15 | Insulation fault diagnosis method, system, equipment and medium for capacitive voltage transformer |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116008756A CN116008756A (en) | 2023-04-25 |
| CN116008756B true CN116008756B (en) | 2023-06-09 |
Family
ID=86033804
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310247981.9A Active CN116008756B (en) | 2023-03-15 | 2023-03-15 | Insulation fault diagnosis method, system, equipment and medium for capacitive voltage transformer |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116008756B (en) |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114048769A (en) * | 2021-11-08 | 2022-02-15 | 太原科技大学 | Multi-source multi-domain information entropy fusion and model self-optimization method for bearing fault diagnosis |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111275204B (en) * | 2020-02-25 | 2023-04-07 | 西安工程大学 | Transformer state identification method based on hybrid sampling and ensemble learning |
| CN112910859B (en) * | 2021-01-19 | 2022-06-14 | 山西警察学院 | Internet of things equipment monitoring and early warning method based on C5.0 decision tree and time sequence analysis |
| CN113033837A (en) * | 2021-03-05 | 2021-06-25 | 国网电力科学研究院武汉南瑞有限责任公司 | Artificial intelligence fault identification system and method based on transient waveform of power transmission line |
| CN113505655B (en) * | 2021-06-17 | 2023-10-13 | 电子科技大学 | Intelligent bearing fault diagnosis method for digital twin system |
| CN113780412B (en) * | 2021-09-10 | 2024-01-30 | 齐齐哈尔大学 | Fault diagnosis model training method and system and fault diagnosis method and system |
| CN114595730A (en) * | 2022-03-31 | 2022-06-07 | 南京工业大学 | Rolling bearing fault diagnosis method based on GAF-DRSN |
| CN115641283A (en) * | 2022-11-07 | 2023-01-24 | 广东电网有限责任公司 | Transformer fault diagnosis method and system based on multi-sensor information fusion |
-
2023
- 2023-03-15 CN CN202310247981.9A patent/CN116008756B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114048769A (en) * | 2021-11-08 | 2022-02-15 | 太原科技大学 | Multi-source multi-domain information entropy fusion and model self-optimization method for bearing fault diagnosis |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116008756A (en) | 2023-04-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11875500B2 (en) | Failure diagnosis method for power transformer winding based on GSMallat-NIN-CNN network | |
| CN110532859B (en) | Remote sensing image target detection method based on deep evolution pruning convolution net | |
| CN109829402B (en) | GS-SVM-based bearing damage degree diagnosis method under different working conditions | |
| CN109215034B (en) | Weak supervision image semantic segmentation method based on spatial pyramid covering pooling | |
| Gao et al. | Rolling bearing fault diagnosis based on intelligent optimized self-adaptive deep belief network | |
| CN114048769A (en) | Multi-source multi-domain information entropy fusion and model self-optimization method for bearing fault diagnosis | |
| JP6892606B2 (en) | Positioning device, position identification method and computer program | |
| CN116503399B (en) | Insulator pollution flashover detection method based on YOLO-AFPS | |
| CN114355240A (en) | Power distribution network ground fault diagnosis method and device | |
| CN113191429A (en) | Power transformer bushing fault diagnosis method and device | |
| CN115546558A (en) | Electrical equipment insulation fault state classification method and device and storage medium | |
| CN116612098B (en) | Insulator RTV spraying quality evaluation method and device based on image processing | |
| Somawirata et al. | Road detection based on the color space and cluster connecting | |
| CN113705396A (en) | Motor fault diagnosis method, system and equipment | |
| CN111179270A (en) | Image co-segmentation method and device based on attention mechanism | |
| CN115496144A (en) | Power distribution network operation scene determining method and device, computer equipment and storage medium | |
| Wang et al. | A novel semi-supervised generative adversarial network based on the actor-critic algorithm for compound fault recognition | |
| Zheng et al. | Benchmarking unsupervised anomaly detection and localization | |
| CN110135428A (en) | Image segmentation processing method and device | |
| CN114187261A (en) | Non-reference stereo image quality evaluation method based on multi-dimensional attention mechanism | |
| CN116008756B (en) | Insulation fault diagnosis method, system, equipment and medium for capacitive voltage transformer | |
| Fomin et al. | Construction of the diagnostic model based on combining spectral characteristics of nonlinear dynamic objects | |
| CN111950635A (en) | Robust feature learning method based on hierarchical feature alignment | |
| JP6950647B2 (en) | Data determination device, method, and program | |
| Shi et al. | Hyperspectral bands reduction based on rough sets and fuzzy c-means clustering |
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 |