CN115587309A - Method, device and equipment for extracting key features of short-circuit resistance of transformer - Google Patents
Method, device and equipment for extracting key features of short-circuit resistance of transformer Download PDFInfo
- Publication number
- CN115587309A CN115587309A CN202211307860.0A CN202211307860A CN115587309A CN 115587309 A CN115587309 A CN 115587309A CN 202211307860 A CN202211307860 A CN 202211307860A CN 115587309 A CN115587309 A CN 115587309A
- Authority
- CN
- China
- Prior art keywords
- singular value
- input
- transformer
- modal
- input sample
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Complex Calculations (AREA)
Abstract
The invention relates to the technical field of transformer data analysis, in particular to a method, a device and equipment for extracting key features of short-circuit resistance of a transformer, wherein the method comprises the following steps: acquiring real-time monitoring data of the transformer as an input sample; dividing an input sample into a training set and a test set, and performing data preprocessing on the training set and the test set; obtaining at least one modal component according to the preprocessed input sample, and calculating a singular value entropy based on variational modal decomposition according to the modal component; and (4) inputting the singular value entropy as a feature vector to the SOM neural network for training, and outputting a fault diagnosis result. The method can freely determine the number of modal components, and better separate the noise in the original data; the singular value entropy can screen out the characteristic data which can represent the most relevant signal information, and the accuracy of the clustering data is improved; the influence of the data noise problem caused by the severe working condition environment of the transformer on the extraction of the key features of the short-circuit resistance of the transformer and the diagnosis of the short-circuit fault is reduced.
Description
Technical Field
The invention relates to the technical field of transformer data analysis, in particular to a method, a device and equipment for extracting key features of short-circuit resistance of a transformer.
Background
The transformer is the main equipment in the operation of the power grid, undertakes the transformation task in the transmission process, and the operation condition of the transformer is closely related to the whole power grid. With the development of digitization and the deep fusion of intelligent technology and traditional operation and maintenance, massive transformer state information can be shared on a unified platform through a sensor.
Short-circuit faults are one of typical problems in transformer operation, and in more and more transformer information, transformer short-circuit fault related data are clustered, so that key features influencing the short-circuit resistance of the transformer are extracted, and the method is important content for evaluating the short-circuit resistance of the transformer. The technology for extracting the key characteristics of the short-circuit resistance of the transformer can master the running condition of the transformer in advance and reflect the change trend of the short-circuit fault indexes of the transformer, so that preventive measures are taken in advance to avoid major power accidents.
At present, a power transformer is influenced by environmental working condition factors, performance of a sensor is reduced, monitoring equipment is failed, and signal interruption in data transmission frequently occurs, so that the operation monitoring data of the transformer is full of noise, and extraction of key features of short-circuit resistance of the transformer and diagnosis of short-circuit faults are seriously influenced.
Disclosure of Invention
In view of this, the present invention provides a method, an apparatus, and a device for extracting a key feature of a transformer with short-circuit resistance, so as to solve the problem in the prior art that data noise caused by a severe working condition environment of the transformer seriously affects extraction of the key feature of the transformer with short-circuit resistance and diagnosis of a short-circuit fault.
According to a first aspect of the embodiments of the present invention, a method for extracting key features of a transformer in short-circuit resistance capability is provided, including:
acquiring real-time monitoring data of the transformer as an input sample;
dividing the input sample into a training set and a test set, and performing data preprocessing on the training set and the test set;
obtaining at least one modal component according to the preprocessed input sample, and calculating a singular value entropy based on variational modal decomposition according to the modal component;
and inputting the singular value entropy as a feature vector to an SOM neural network for training, and outputting a fault diagnosis result.
Preferably, the obtaining at least one modal component according to the preprocessed input sample, and calculating a singular value entropy based on a variational modal decomposition according to the modal component includes:
decomposing the preprocessed input sample into a plurality of modal components by using a variational modal decomposition method;
extracting a characteristic matrix according to the modal component;
and calculating to obtain the singular value entropy of the input sample based on the variational modal decomposition according to the feature matrix.
Preferably, the decomposing the preprocessed input sample into a plurality of modal components by using a variational modal decomposition method includes:
constructing a constraint variation problem model;
converting the constraint variation problem model into an unconstrained variation problem model;
using an alternative multiplier method to solve saddle points in the unconstrained variational problem model, obtaining a corresponding variable updating formula, updating the unconstrained variational problem model, and performing iterative computation;
and after a preset iteration stop condition is reached, carrying out variation modal decomposition on the input sample to obtain a modal component of the input sample.
Preferably, the calculating, according to the feature matrix, the singular value entropy of the input sample based on variational modal decomposition includes:
and combining the modal components into a feature matrix, and performing singular value decomposition on the feature matrix to construct a singular value entropy of the input sample.
Preferably, the step of inputting the singular value entropy as a feature vector to an SOM neural network for training and outputting a fault diagnosis result includes:
combining the singular value entropies, and selecting an entropy combination containing the most input sample information to form an input vector;
initializing an SOM neural network, setting a connection weight and a learning rate between neurons of an input layer and neurons of a competition layer, calculating Euclidean distances between the neurons and input vectors, and marking the neuron with the largest distance from the input vector in the competition layer as a winning neuron;
updating a neuron connection weight between the input layer and the competition layer through a preset formula, and performing iterative computation;
and outputting a fault diagnosis result when the iterative calculation reaches a convergence condition or a cycle upper limit.
Preferably, before the dividing the input samples into the training set and the test set, the method further includes:
using the formula x i ＝(x' i -x min )/(x max -x min ) Normalizing different input characteristics of the input sample;
wherein, x' i An original value representing the input feature; x is the number of i Representing a normalized value of the input feature; x is the number of max And x min Respectively representing the maximum and minimum values of the set of input features.
Preferably, the converting the constrained variation problem model into an unconstrained variation problem model includes:
introducing a Lagrange multiplier and a punishment factor, solving the constraint variation problem model, and obtaining an augmented Lagrange equation;
and taking the augmented Lagrange equation as a non-constrained variation problem model.
Preferably, the performing singular value decomposition on the feature matrix to construct singular value entropy of the input sample includes:
performing singular value decomposition on the feature matrix;
arranging the decomposed singular values into a singular value spectrum according to the order of the singular values;
calculating the proportion of each singular value in the sum of all singular values;
and constructing singular value entropy according to the specific gravity.
According to a second aspect of the embodiments of the present invention, there is provided a transformer short-circuit resistance key feature extraction apparatus, including:
the data acquisition module is used for acquiring real-time monitoring data of the transformer as an input sample; the system is also used for dividing the input samples into a training set and a testing set and carrying out data preprocessing on the training set and the testing set;
the calculation module is used for obtaining at least one modal component according to the preprocessed input sample and calculating a singular value entropy based on variational modal decomposition according to the modal component;
and the neural network module is used for inputting the singular value entropy as a feature vector to the SOM neural network for training and outputting a fault diagnosis result.
According to a third aspect of the embodiments of the present invention, there is provided a transformer short-circuit resistance capability key feature extraction device, including:
the main controller and the memory connected with the main controller;
the memory having stored therein program instructions;
the master is configured to execute program instructions stored in a memory to perform the method of any of the above.
The technical scheme provided by the embodiment of the invention can have the following beneficial effects:
it can be understood that the invention takes the real-time monitoring data of the transformer as the input sample; dividing an input sample into a training set and a test set, and performing data preprocessing on the training set and the test set; obtaining at least one modal component according to the preprocessed input sample, and calculating a singular value entropy based on variational modal decomposition according to the modal component; and (4) inputting the singular value entropy as a feature vector to the SOM neural network for training, and outputting a fault diagnosis result. It can be understood that the number of modal components can be freely determined, and noise in the original data can be better separated; the singular value entropy can screen out the characteristic data which can represent the related signal information most, and the accuracy of the clustering data is improved. Finally, the problem of data noise caused by severe working condition environment of the transformer is solved, the key characteristic which can represent the short-circuit resistance of the transformer is screened out, and guidance is provided for improving the short-circuit resistance of the transformer and early warning of short-circuit faults.
It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention, as claimed.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and together with the description, serve to explain the principles of the invention.
FIG. 1 is a schematic diagram illustrating steps of a transformer short-circuit immunity key feature extraction method according to an exemplary embodiment;
FIG. 2 is a flow diagram illustrating a method for extracting key features of a transformer for short-circuit immunity, according to an example embodiment;
FIG. 3 is a detailed flow diagram illustrating a transformer short circuit immunity key feature extraction method in accordance with an exemplary embodiment;
fig. 4 is a schematic block diagram illustrating a transformer short-circuit immunity key feature extraction apparatus according to an exemplary embodiment.
Detailed Description
Reference will now be made in detail to the exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, like numbers in different drawings represent the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the invention, as detailed in the appended claims.
Example one
Fig. 1 is a schematic diagram illustrating steps of a transformer short-circuit capability resisting key feature extraction method according to an exemplary embodiment, and referring to fig. 1, a transformer short-circuit capability resisting key feature extraction method is provided, which includes:
s11, acquiring real-time monitoring data of the transformer as an input sample;
wherein the transformer monitoring data includes: and the current amplitude, the voltage amplitude, the active value, the passive value, the winding temperature and the like of the high-voltage side, the medium-voltage side and the low-voltage side of the transformer.
S12, dividing the input sample into a training set and a test set, and performing data preprocessing on the training set and the test set;
s13, obtaining at least one modal component according to the preprocessed input sample, and calculating a singular value entropy based on variational modal decomposition according to the modal component;
and S14, inputting the singular value entropy serving as a feature vector to an SOM neural network for training, and outputting a fault diagnosis result.
It can be understood that, in the present embodiment, real-time monitoring data of the transformer is obtained as an input sample; dividing an input sample into a training set and a test set, and performing data preprocessing on the training set and the test set; obtaining at least one modal component according to the preprocessed input sample, and calculating a singular value entropy based on variational modal decomposition according to the modal component; and (4) inputting the singular value entropy as a feature vector to the SOM neural network for training, and outputting a fault diagnosis result. It can be understood that the number of modal components can be freely determined, and noise in the original data can be better separated; the singular value entropy can screen out the characteristic data which can represent the related signal information most, and the accuracy of the clustering data is improved. Finally, the problem of data noise caused by severe working condition environment of the transformer is solved, the key characteristics which can represent the short-circuit resistance of the transformer are screened out, and guidance is provided for improvement of the short-circuit resistance of the transformer and early warning of short-circuit faults.
It should be noted that, in step S13, the obtaining at least one modal component according to the preprocessed input sample, and calculating a singular value entropy based on a variational modal decomposition according to the modal component includes:
decomposing the preprocessed input sample into a plurality of modal components by using a variational modal decomposition method;
extracting a characteristic matrix according to the modal component;
and calculating to obtain the singular value entropy of the input sample based on the variational modal decomposition according to the feature matrix.
In specific practice, the collected transformer monitoring data is subjected to variation modal decomposition to obtain a Variation Modal Decomposition (VMD) decomposition sequence of the transformer monitoring data, namely a plurality of modal components, and then a characteristic matrix is extracted to obtain a singular value entropy of the characteristic matrix. As shown in fig. 2, a flowchart of the method for extracting key features of short-circuit resistance of a transformer is that monitoring data of the transformer is input first, then the input data is preprocessed, and after the preprocessing, the input data is subjected to variation modal decomposition to generate a plurality of modal components (IMF) 1 ～IMF k ) And then, calculating a singular spectrum entropy value, and inputting the result into the SOM neural network to obtain a clustering result output by the neural network.
It should be noted that, referring to fig. 3, the decomposing the preprocessed input sample into several modal components by using the variational modal decomposition method includes:
s21, constructing a constraint variational problem model;
s22, converting the constraint variation problem model into an unconstrained variation problem model;
s23, solving saddle points in the unconstrained variational problem model by using an alternative multiplier method to obtain a corresponding variable updating formula, updating the unconstrained variational problem model and performing iterative computation;
and S24, after a preset iteration stop condition is reached, performing variation modal decomposition on the input sample to obtain a modal component of the input sample.
In specific practice, in step S21, a constraint variational problem model is constructed, and the corresponding mathematical formula is:
wherein k represents the current mode number; n represents the total number of modes;
representing a gradient operator; j denotes a demodulation operator; t represents time; { u k Represents a collection of modalities; { omega [ [ omega ] ] k Represents the collection of center frequencies of each mode; δ (t) represents the dirichlet function.
In step S22, it should be noted that the converting the constraint variational problem model into an unconstrained variational problem model includes:
introducing a Lagrange multiplier and a penalty factor, solving the constraint variation problem model, and obtaining an augmented Lagrange equation;
and taking the augmented Lagrange equation as a non-constrained variation problem model.
The non-constraint variational problem model is characterized in that a corresponding mathematical formula is as follows:
in step S23, the saddle point in the unconstrained variational problem model is solved by using an alternative multiplier method, and the corresponding variable update formula is obtained as follows:
can use the alternative multiplier method pair
And
updating, wherein in the formula, omega represents the center frequency of the current mode; n represents the current number of iterations. τ denotes the update parameter of the lagrange multiplier.
In step S24, the mathematical formula corresponding to the iteration stop condition is:
in the formula, ε represents convergence accuracy, and ε >0.
Firstly, the number of the decomposition modal components is specified, and the frequency of the components is initialized to obtain
And
respectively iterating and updating according to corresponding variable updating formulasAnd (4) newly corresponding component signals, the center frequency and the Lari-Langi multiplier until a mathematical formula corresponding to an iteration stop condition is met, and then outputting k IMF components after VMD decomposition.
After step S24, the calculating, according to the feature matrix, singular value entropy of the input sample based on the variational modal decomposition includes:
and S25, combining the modal components into a feature matrix, and performing singular value decomposition on the feature matrix to construct a singular value entropy of the input sample.
In a specific practice, in the step S25, the singular value entropy is calculated according to the following mathematical formula:
e i ＝p i log(p i )
in the three formulae, q i Singular values of the feature matrix, p i The proportion of each singular value in the sum of all singular values is shown.
It should be noted that, the performing singular value decomposition on the feature matrix to construct singular value entropy of the input sample includes:
performing singular value decomposition on the feature matrix;
arranging the decomposed singular values into a singular value spectrum according to the order of the singular values;
calculating the proportion of each singular value in the sum of all singular values;
and constructing singular value entropy according to the specific gravity.
In specific practice, the obtained k IMF components are combined into a feature matrix, singular value decomposition is carried out on the feature matrix, and then singular value spectrums are arranged according to the order of magnitude of singular values, wherein the singular value spectrums are described in the singular value spectrumsThe outlier spectrum is shown below: q = { Q = i I =1,2. According to the formula
Calculating the proportion of each singular value in the sum of all singular values, and calculating the singular spectrum entropy according to a formula
Is shown in the specification, wherein e i ＝p i log(p i )。
It should be noted that the step of inputting the singular value entropy as a feature vector to the SOM neural network for training and outputting a fault diagnosis result includes:
s31, combining the singular value entropies, selecting an entropy combination containing the most input sample information, and forming an input vector;
s32, initializing the SOM neural network, setting a connection weight and a learning rate between neurons of an input layer and neurons of a competition layer, calculating Euclidean distances between the neurons and input vectors, and marking the neuron with the largest distance from the input vector in the competition layer as a winning neuron;
s33, updating the neuron connection weight between the input layer and the competition layer through a preset formula, and performing iterative computation;
and step S34, outputting a fault diagnosis result when the iterative computation reaches a convergence condition or a cycle upper limit.
In specific practice, step S31 can select an entropy combination containing most information of the original input from the singular value entropies to form an input vector X;
step S32 can initialize the SOM network, the neuron connection weight value W of the input layer and the competition layer is set to be a random value in a [0,1] closed interval, the initial value eta (0) of the learning rate eta (t) is set, and the eta is greater than 0 and smaller than 1; calculating Euclidean distance between the ith neuron and an input vector X, and marking the neuron with the largest distance from the input vector in a competition layer as a winning neuron i (X); wherein, the Euclidean distance between the neuron and the input vector is calculated, and the corresponding mathematical formula is as follows:
in the formula (d) j Representing the Euclidean distance between the input vector and the jth neuron of the competition layer; w is a ij Representing the connection weight between the ith neuron of the input layer and the jth neuron of the competition layer; w j Representing the weight vector of the jth neuron of the competition layer, X is the input vector, X i Is the input vector of the ith neuron of the input layer.
Step S33 can adjust the neuron connection weight between the input layer and the competition layer by a formula, and the specific corresponding mathematical formula is:
w ij (t+1)＝w ij (t)+ηT j,i(x) (t)(x i -w ij )
in the formula, T j,i(x) Representing the domain function around the winning neuron i (x),
represents the distance between node j and the winning neuron i (x), and t represents the number of iterations.
After updating the neuron connection weight between the input layer and the competition layer, the steps S32 and S33 need to be repeated to perform iterative computation until a convergence condition or a loop upper limit is reached.
Finally, the failure diagnosis result is output through step S34.
It should be noted that, before the dividing the input samples into the training set and the test set, the method further includes:
using the formula x i ＝(x' i -x min )/(x max -x min )，Performing normalization operation on different input characteristics of the input sample;
wherein, x' i An original value representing the input feature; x is the number of i Representing a normalized value of the input feature; x is the number of max And x min Respectively representing the maximum and minimum values of the set of input features.
In specific practice, in order to avoid the influence of dimensions of different input features on a clustering result and maintain the independence between the features, normalization operation needs to be performed on input feature vectors.
Example two
Fig. 4 is a schematic block diagram illustrating a transformer short-circuit resistance capability key feature extraction apparatus according to an exemplary embodiment, and referring to fig. 4, there is provided a transformer short-circuit resistance capability key feature extraction apparatus including:
the data acquisition module 101 is used for acquiring real-time monitoring data of the transformer as an input sample; the system is also used for dividing the input samples into a training set and a testing set and carrying out data preprocessing on the training set and the testing set;
the calculation module 102 is configured to obtain at least one modal component according to the preprocessed input sample, and calculate a singular value entropy based on a variational modal decomposition according to the modal component;
and the neural network module 103 is used for inputting the singular value entropy as a feature vector to the SOM neural network for training and outputting a fault diagnosis result.
It can be understood that the real-time monitoring data of the transformer is acquired as an input sample by the data acquisition module 101; dividing an input sample into a training set and a test set, and performing data preprocessing on the training set and the test set; obtaining at least one modal component according to the preprocessed input sample through the calculation module 102, and calculating a singular value entropy based on variational modal decomposition according to the modal component; the singular value entropy is used as a feature vector by the neural network module 103, and is input to the SOM neural network for training, and a fault diagnosis result is output. It can be understood that the number of modal components can be freely determined, and noise in the original data can be better separated; the singular value entropy can screen out the characteristic data which can represent the related signal information most, and the accuracy of the clustering data is improved. Finally, the problem of data noise caused by severe working condition environment of the transformer is solved, the key characteristic which can represent the short-circuit resistance of the transformer is screened out, and guidance is provided for improving the short-circuit resistance of the transformer and early warning of short-circuit faults.
EXAMPLE III
Provided is a transformer short circuit resistance capability key feature extraction device, including:
the main controller and the memory connected with the main controller;
the memory having stored therein program instructions;
the master is configured to execute program instructions stored in the memory to perform any of the methods described above.
It is understood that the same or similar parts in the above embodiments may be mutually referred to, and the same or similar parts in other embodiments may be referred to for the content which is not described in detail in some embodiments.
It should be noted that the terms "first," "second," and the like in the description of the present invention are used for descriptive purposes only and are not to be construed as indicating or implying relative importance. Further, in the description of the present invention, the meaning of "a plurality" means at least two unless otherwise specified.
Any process or method descriptions in flow charts or otherwise described herein may be understood as representing modules, segments, or portions of code which include one or more executable instructions for implementing specific logical functions or steps of the process, and alternate implementations are included within the scope of the preferred embodiment of the present invention in which functions may be executed out of order from that shown or discussed, including substantially concurrently or in reverse order, depending on the functionality involved, as would be understood by those reasonably skilled in the art of the present invention.
It should be understood that portions of the present invention may be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, various steps or methods may be implemented in software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, any one or combination of the following techniques, which are known in the art, may be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application specific integrated circuit having an appropriate combinational logic gate circuit, a Programmable Gate Array (PGA), a Field Programmable Gate Array (FPGA), or the like.
It will be understood by those skilled in the art that all or part of the steps carried by the method for implementing the above embodiments may be implemented by hardware related to instructions of a program, which may be stored in a computer readable storage medium, and when the program is executed, the program includes one or a combination of the steps of the method embodiments.
In addition, functional units in the embodiments of the present invention may be integrated into one processing module, or each unit may exist alone physically, or two or more units are integrated into one module. The integrated module can be realized in a hardware mode, and can also be realized in a software functional module mode. The integrated module, if implemented in the form of a software functional module and sold or used as a separate product, may also be stored in a computer-readable storage medium.
The storage medium mentioned above may be a read-only memory, a magnetic or optical disk, etc.
In the description herein, references to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention, and that variations, modifications, substitutions and alterations can be made to the above embodiments by those of ordinary skill in the art within the scope of the present invention.
Claims (10)
1. A method for extracting key features of short-circuit resistance of a transformer is characterized by comprising the following steps:
acquiring real-time monitoring data of the transformer as an input sample;
dividing the input sample into a training set and a testing set, and carrying out data preprocessing on the training set and the testing set;
obtaining at least one modal component according to the preprocessed input sample, and calculating a singular value entropy based on variational modal decomposition according to the modal component;
and inputting the singular value entropy as a feature vector to an SOM neural network for training, and outputting a fault diagnosis result.
2. The method according to claim 1, wherein the deriving at least one modal component from the preprocessed input samples, and the computing of the singular value entropy based on the variational modal decomposition from the modal component comprises:
decomposing the preprocessed input sample into a plurality of modal components by using a variational modal decomposition method;
extracting a characteristic matrix according to the modal component;
and calculating to obtain the singular value entropy of the input sample based on the variational modal decomposition according to the feature matrix.
3. The method of claim 2, wherein decomposing the preprocessed input samples into modal components using a variational modal decomposition method comprises:
constructing a constraint variation problem model;
converting the constraint variation problem model into an unconstrained variation problem model;
using an alternative multiplier method to solve saddle points in the unconstrained variational problem model, obtaining a corresponding variable updating formula, updating the unconstrained variational problem model, and performing iterative computation;
and after a preset iteration stop condition is reached, carrying out variation modal decomposition on the input sample to obtain a modal component of the input sample.
4. The method according to claim 3, wherein the calculating the singular value entropy of the input sample based on the variational modal decomposition according to the feature matrix comprises:
and combining the modal components into a feature matrix, and performing singular value decomposition on the feature matrix to construct a singular value entropy of the input sample.
5. The method according to claim 1, wherein the inputting the singular value entropy as a feature vector to an SOM neural network for training and outputting a fault diagnosis result comprises:
combining the singular value entropies, and selecting an entropy combination containing the most input sample information to form an input vector;
initializing an SOM neural network, setting a connection weight and a learning rate between neurons of an input layer and a competition layer, calculating Euclidean distances between the neurons and input vectors, and marking the neuron with the largest distance from the input vector in the competition layer as a winning neuron;
updating a neuron connection weight between the input layer and the competition layer through a preset formula, and performing iterative computation;
and outputting a fault diagnosis result when the iterative calculation reaches a convergence condition or a cycle upper limit.
6. The method of claim 1, further comprising, prior to said partitioning said input samples into a training set and a test set:
using the formula x i ＝(x′ i -x min )/(x max -x min ) For differences in the input samplesInputting the characteristics to perform normalization operation;
wherein, x' i An original value representing the input feature; x is the number of i Representing a normalized value of the input feature; x is the number of max And x min Respectively representing the maximum and minimum values of the set of input features.
7. The method of claim 3, wherein transforming the constrained variational problem model to an unconstrained variational problem model comprises:
introducing a Lagrange multiplier and a penalty factor, solving the constraint variation problem model, and obtaining an augmented Lagrange equation;
and taking the augmented Lagrange equation as a non-constrained variation problem model.
8. The method of claim 4, wherein the performing the singular value decomposition on the feature matrix to construct the singular value entropy of the input sample comprises:
performing singular value decomposition on the feature matrix;
arranging the decomposed singular values into a singular value spectrum according to the order of the singular values;
calculating the proportion of each singular value in the sum of all singular values;
and constructing a singular value entropy according to the specific gravity.
9. The utility model provides an anti short circuit ability key feature extraction element of transformer which characterized in that includes:
the data acquisition module is used for acquiring real-time monitoring data of the transformer as an input sample; the system is also used for dividing the input samples into a training set and a testing set and carrying out data preprocessing on the training set and the testing set;
the calculation module is used for obtaining at least one modal component according to the preprocessed input sample and calculating a singular value entropy based on variational modal decomposition according to the modal component;
and the neural network module is used for inputting the singular value entropy as a feature vector to the SOM neural network for training and outputting a fault diagnosis result.
10. A transformer anti short circuit ability key feature extraction equipment which characterized in that includes:
the main controller and the memory connected with the main controller;
the memory having stored therein program instructions;
the master is adapted to execute program instructions stored in a memory to perform the method of any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211307860.0A CN115587309A (en) | 2022-10-25 | 2022-10-25 | Method, device and equipment for extracting key features of short-circuit resistance of transformer |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211307860.0A CN115587309A (en) | 2022-10-25 | 2022-10-25 | Method, device and equipment for extracting key features of short-circuit resistance of transformer |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115587309A true CN115587309A (en) | 2023-01-10 |
Family
ID=84781866
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211307860.0A Pending CN115587309A (en) | 2022-10-25 | 2022-10-25 | Method, device and equipment for extracting key features of short-circuit resistance of transformer |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115587309A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116307829A (en) * | 2023-01-17 | 2023-06-23 | 福建实达集团股份有限公司 | Method and device for evaluating influence of infectious diseases on social bearing capacity based on information entropy |
| CN116992248A (en) * | 2023-09-27 | 2023-11-03 | 国网江苏省电力有限公司电力科学研究院 | Short-circuit resistance evaluation method and device for coiled iron core transformer based on short-circuit test |
| CN117196123A (en) * | 2023-11-06 | 2023-12-08 | 深圳市粤能电气有限公司 | Data control method, device and equipment for digital production of power distribution equipment |
-
2022
- 2022-10-25 CN CN202211307860.0A patent/CN115587309A/en active Pending
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116307829A (en) * | 2023-01-17 | 2023-06-23 | 福建实达集团股份有限公司 | Method and device for evaluating influence of infectious diseases on social bearing capacity based on information entropy |
| CN116307829B (en) * | 2023-01-17 | 2024-03-29 | 福建实达集团股份有限公司 | Method and device for evaluating influence of infectious diseases on social bearing capacity based on information entropy |
| CN116992248A (en) * | 2023-09-27 | 2023-11-03 | 国网江苏省电力有限公司电力科学研究院 | Short-circuit resistance evaluation method and device for coiled iron core transformer based on short-circuit test |
| CN116992248B (en) * | 2023-09-27 | 2023-12-05 | 国网江苏省电力有限公司电力科学研究院 | Short-circuit resistance evaluation method and device for coiled iron core transformer based on short-circuit test |
| CN117196123A (en) * | 2023-11-06 | 2023-12-08 | 深圳市粤能电气有限公司 | Data control method, device and equipment for digital production of power distribution equipment |
| CN117196123B (en) * | 2023-11-06 | 2024-03-19 | 深圳市粤能电气有限公司 | Data control method, device and equipment for digital production of power distribution equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115587309A (en) | Method, device and equipment for extracting key features of short-circuit resistance of transformer | |
| CN111325095B (en) | Intelligent detection method and system for equipment health state based on acoustic wave signals | |
| CN107066759B (en) | Steam turbine rotor vibration fault diagnosis method and device | |
| CN108809948B (en) | Abnormal network connection detection method based on deep learning | |
| CN112307950B (en) | Detail feature quantity extraction and intelligent analysis method for GIS vibration defect identification | |
| CN108398268A (en) | A kind of bearing performance degradation assessment method based on stacking denoising self-encoding encoder and Self-organizing Maps | |
| CN108647707B (en) | Probabilistic neural network creation method, failure diagnosis method and apparatus, and storage medium | |
| CN108051660A (en) | A kind of transformer fault combined diagnosis method for establishing model and diagnostic method | |
| CN109784692B (en) | Rapid safety constraint economic dispatching method based on deep learning | |
| CN108960423A (en) | motor monitoring system based on machine learning | |
| CN111680875A (en) | Unmanned aerial vehicle state risk fuzzy comprehensive evaluation method based on probability baseline model | |
| CN111178585A (en) | Fault reporting amount prediction method based on multi-algorithm model fusion | |
| Silva et al. | Damage detection in a benchmark structure using AR-ARX models and statistical pattern recognition | |
| CN105137354A (en) | Motor fault detection method based on nerve network | |
| CN113225346A (en) | Network operation and maintenance situation assessment method based on machine learning | |
| CN117371207A (en) | Extra-high voltage converter valve state evaluation method, medium and system | |
| CN110991471A (en) | Fault diagnosis method for high-speed train traction system | |
| CN116451514A (en) | Bridge structure state evaluation method and device and electronic equipment | |
| CN115331696A (en) | Multi-channel voiceprint signal blind source separation method for transformer abnormity diagnosis | |
| CN112488171A (en) | Variable mode decomposition SRU fault prediction method | |
| CN115600088A (en) | Distribution transformer fault diagnosis method based on vibration signals | |
| Tan et al. | Fault diagnosis method for rolling bearing based on VMD and improved SVM optimized by METLBO | |
| CN104504265B (en) | A kind of method of servicing bridges monitoring information security evaluation | |
| CN112362292B (en) | Method for anomaly detection of wind tunnel test data | |
| CN117609836A (en) | Electromagnetic sensitivity prediction and health management method for integrated module |
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 |