CN116956042A - Method, device, equipment and medium for establishing fault type detection model - Google Patents
Method, device, equipment and medium for establishing fault type detection model Download PDFInfo
- Publication number
- CN116956042A CN116956042A CN202310944264.1A CN202310944264A CN116956042A CN 116956042 A CN116956042 A CN 116956042A CN 202310944264 A CN202310944264 A CN 202310944264A CN 116956042 A CN116956042 A CN 116956042A
- Authority
- CN
- China
- Prior art keywords
- model
- sample data
- layer
- fault type
- switch
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 197
- 238000000034 method Methods 0.000 title claims abstract description 69
- 238000000605 extraction Methods 0.000 claims abstract description 110
- 238000012549 training Methods 0.000 claims abstract description 86
- 238000011176 pooling Methods 0.000 claims description 73
- 238000004088 simulation Methods 0.000 claims description 22
- 238000004590 computer program Methods 0.000 claims description 17
- 230000010354 integration Effects 0.000 claims description 4
- 238000010276 construction Methods 0.000 claims description 3
- 230000000694 effects Effects 0.000 abstract description 8
- 230000008569 process Effects 0.000 description 11
- 238000010586 diagram Methods 0.000 description 9
- 238000004891 communication Methods 0.000 description 8
- 230000006870 function Effects 0.000 description 8
- 239000011159 matrix material Substances 0.000 description 8
- 238000012935 Averaging Methods 0.000 description 6
- 238000012545 processing Methods 0.000 description 6
- 230000009286 beneficial effect Effects 0.000 description 4
- 238000003062 neural network model Methods 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 239000003990 capacitor Substances 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000004880 explosion Methods 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000007781 pre-processing Methods 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 230000004913 activation Effects 0.000 description 1
- 230000032683 aging Effects 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000001413 cellular effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 239000011521 glass Substances 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 230000000630 rising effect Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000001953 sensory effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0475—Generative networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention discloses a method, a device, equipment and a medium for establishing a fault type detection model, wherein the method for establishing the fault type detection model comprises the following steps: acquiring switch state data in a detection area of the power distribution network; constructing a sample data set according to the switch state data; performing model training on a pre-constructed detection network model by adopting switch sample data until a preset model training ending condition is met; and taking the trained detection network model as a fault type detection model. Through the technical scheme, the content of the characteristic information of the sample data subjected to characteristic extraction by the detection network model can be enriched, and meanwhile, the training efficiency and the training effect of the fault type detection model training are improved.
Description
Technical Field
The present invention relates to the field of power distribution network technologies, and in particular, to a method, an apparatus, a device, and a medium for establishing a fault type detection model.
Background
Electric power, one of the basic demands of modern industry and life, has become an indispensable important energy source in people's daily life. In the power transmission and distribution process, once a fault occurs, not only a power failure or even an accident is caused in a short time, but also little difficulty is brought to maintenance and management of the power system. Especially in rural areas, the power grid faults are more frequent and complex due to factors such as remote regions, severe natural environments, line aging and the like.
At present, a technology for positioning faults of a power distribution network based on a model requires a great deal of priori knowledge and collects a great deal of experimental data, and the faults of the power distribution network cannot be positioned rapidly and accurately.
Disclosure of Invention
The invention provides a method, a device, equipment and a medium for establishing a fault type detection model, so as to improve the precision and efficiency of fault positioning of a power distribution network.
According to an aspect of the present invention, there is provided a method for establishing a fault type detection model, the method including:
acquiring switch state data in a detection area of the power distribution network;
constructing a sample data set according to the switch state data; the sample data set comprises at least one switch sample data with a standard fault class label;
performing model training on a pre-constructed detection network model by adopting the switch sample data until a preset model training ending condition is met;
wherein the detection network model comprises a generator model and a classifier model; the generator model is used for generating simulation sample data; the classifier model is used for carrying out self model training based on the simulation sample data and the switch sample data to obtain output fault categories respectively corresponding to the switch sample data;
And taking the trained detection network model as a fault type detection model.
According to another aspect of the present invention, there is provided a fault type detection method including:
acquiring state data of a switch to be detected in a region to be detected of the power distribution network;
and inputting the switch state data to be detected into a fault type detection model to obtain a fault type detection result.
According to another aspect of the present invention, there is provided an apparatus for building a fault type detection model, including:
the switch state acquisition module is used for acquiring switch state data in a detection area of the power distribution network;
the sample data construction module is used for constructing a sample data set according to the switch state data; the sample data set comprises at least one switch sample data with a standard fault class label;
the model training module is used for carrying out model training on a pre-constructed detection network model by adopting the switch sample data until a preset model training ending condition is met; wherein the detection network model comprises a generator model and a classifier model; the generator model is used for generating simulation sample data; the classifier model is used for carrying out self model training based on the simulation sample data and the switch sample data to obtain output fault categories respectively corresponding to the switch sample data;
And the model determining module is used for taking the trained detection network model as a fault type detection model.
According to another aspect of the present invention, there is provided a fault type detection device including:
the to-be-detected data acquisition module is used for acquiring to-be-detected switch state data in a to-be-detected area of the power distribution network;
the fault type determining module is used for inputting the switch state data to be detected into a fault type detection model to obtain a fault type detection result.
According to another aspect of the present invention, there is provided an electronic apparatus including:
at least one processor; and
a memory communicatively coupled to the at least one processor; wherein,,
the memory stores a computer program executable by the at least one processor, the computer program being executable by the at least one processor to enable the at least one processor to perform the method of building a fault type detection model or the method of detecting a fault type according to any of the embodiments of the present invention.
According to another aspect of the present invention, there is provided a computer readable storage medium storing computer instructions for causing a processor to implement the method for creating a fault type detection model or the method for detecting a fault type according to any of the embodiments of the present invention when executed.
According to the technical scheme, a sample data set is constructed according to the acquired switch state data in the detection area of the power distribution network, simulated sample data is generated through a generator model in the detection network model, and model training is carried out on a classifier model in the detection network model according to the simulated sample data and the switch sample data, so that the detection network model after the training is acquired as a fault type detection model, the characteristic information content of the detection network model for carrying out characteristic extraction on the sample data is enriched by introducing the simulated sample data in the model training, meanwhile, the time spent for collecting the switch sample data is shortened, the training efficiency and the training effect of the model training are improved, and the detection precision of fault detection on the power distribution network by using the fault type detection model is improved.
It should be understood that the description in this section is not intended to identify key or critical features of the embodiments of the invention or to delineate the scope of the invention. Other features of the present invention will become apparent from the description that follows.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly described below, and it is apparent that the drawings in the following description are only some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of a method for building a fault type detection model according to a first embodiment of the present invention;
FIG. 2A is a flowchart of a method for creating a fault type detection model according to a second embodiment of the present invention;
FIG. 2B is a schematic diagram of a generator model according to a second embodiment of the present invention;
fig. 3 is a flowchart of a fault type detection method according to a third embodiment of the present invention;
fig. 4 is a schematic structural diagram of a device for creating a fault type detection model according to a fourth embodiment of the present invention;
fig. 5 is a schematic structural diagram of a fault type detection device according to a fifth embodiment of the present invention;
fig. 6 is a schematic structural diagram of an electronic device implementing a method for establishing a fault type detection model or a method for detecting a fault type according to an embodiment of the present invention.
Detailed Description
In order that those skilled in the art will better understand the present invention, a technical solution in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in which it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the present invention without making any inventive effort, shall fall within the scope of the present invention.
It should be noted that the terms "first," "second," and the like in the description and the claims of the present invention and the above figures are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate such that the embodiments of the invention described herein may be implemented in sequences other than those illustrated or otherwise described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
Example 1
Fig. 1 is a flowchart of a method for establishing a fault type detection model according to an embodiment of the present invention, where the method may be performed by a device for establishing a fault type detection model when a power distribution network fails, and the device for establishing a fault type detection model may be implemented in hardware and/or software, and the device for establishing a fault type detection model may be configured in an electronic device. As shown in fig. 1, the method includes:
S110, acquiring switch state data in a detection area of the power distribution network.
The distribution network detection area may be a preset distribution network area in the distribution network. Alternatively, there may be multiple power distribution network areas in the power distribution network, and at least one intelligent switch may be present in each power distribution network area.
The switch state data may represent the switch state of intelligent switches in the power distribution network detection area, e.g., the switch state may be on or off. Alternatively, the switch state data may be represented in a matrix or vector data form, which is not limited in this embodiment.
Specifically, the intelligent switch state of the power distribution network detection area when faults occur in the history period can be determined to be the switch state data of the power distribution network detection area.
S120, constructing a sample data set according to the switch state data.
Wherein the sample data set may include at least one switch sample data with a standard fault class label. Alternatively, the fault type may refer to a fault type of the power distribution network detection area corresponding to the switch state data when the switch state data is acquired, for example, a fault type such as explosion of a lightning arrester, overload of a distribution voltage device, and a capacitor tube fault. The standard fault class label may refer to the exact fault class to which the switch sample data corresponds.
Specifically, a sample data set may be constructed according to the switch state data obtained in at least one detection area of the power distribution network. Alternatively, the switch sample data in the sample data set may be divided into training sample data and test sample data, for example, seventy percent of the switch sample data in the sample data set may be used as the training sample data and thirty percent of the switch sample data in the sample data set may be used as the test sample data.
Optionally, constructing the sample data set according to the switch state data includes: acquiring topology data of a switch structure corresponding to a detection area of the power distribution network; determining switch sample data according to the switch structure topology data and the switch state data; determining standard fault categories of the switch sample data; a sample dataset is generated that includes at least one switch sample data with a standard fault class label.
The switch structure topology data may represent a topology relationship of intelligent switches in a detection area of the power distribution network. Alternatively, the switch topology data and the switch sample data may be represented in the form of matrix or vector data, which is not limited in this embodiment.
Specifically, the topology data of the switch structure corresponding to the detection area of the power distribution network can be determined according to the topology relation of at least one intelligent switch when the history of the detection area of the power distribution network fails; according to the switch structure topology data corresponding to the power distribution network detection area and the switch state data in the power distribution network detection area, determining the switch structure topology data as switch sample data, determining the corresponding standard fault type of each switch sample data, namely the fault type of the power distribution network detection area, which is in fault history, and generating a sample data set comprising at least one switch sample data with a standard fault type label.
Alternatively, the switch sample data may be determined by the following formula:
P＝G×D；
where P is a matrix representing switch sample data and G is a matrix representing switch state data. D is a matrix representing the topology data of the switch structure.
S130, performing model training on a pre-constructed detection network model by adopting switch sample data until a preset model training ending condition is met.
The detection network model may be a neural network model preset by a related technician, for example, the detection network model may be a generated countermeasure network model (GAN, generative Adversarial Networks), or the like. For example, the generated countermeasure network model may be trained using the switch sample data until a preset model training end condition is satisfied.
The detection network model can also be a network model which is built in advance by relevant technicians according to actual requirements.
By way of example, the detection network model may include a generator model and a classifier model; the generator model may be used to generate simulated sample data; the classifier model can be used for carrying out self model training based on the simulation sample data and the switch sample data to obtain output fault types respectively corresponding to the switch sample data.
Alternatively, the generator model and classifier model may be neural network models; optionally, the simulated sample data is different from the switch sample data, and the simulated sample data is not generated by the switch state data and the switch structure topology data when the history of the detection area of the power distribution network fails, but is generated by the generator model; the two are simultaneously participated in the training process of the classifier model, and the training of the classifier model is completed. By generating the simulated sample data, the characteristic information content of the detection network model for characteristic extraction of the sample data is enriched, a better training effect is achieved on the premise of not increasing the switch sample data, meanwhile, the time spent for collecting the switch sample data is reduced, and the training efficiency of model training is improved.
The preset model training ending condition may refer to that the model training iteration number reaches a preset iteration number threshold; alternatively, the iteration number threshold may be adaptively set by those skilled in the art, for example, the iteration number threshold may be 50.
Specifically, model training may be performed on a pre-constructed detection network model based on the acquired switch sample data until the training process of the detection network model meets a preset model training end condition.
And S140, taking the trained detection network model as a fault type detection model.
Specifically, a detection network model meeting a preset model training ending condition is used as a fault type detection model, so that fault type detection can be performed on a region to be detected of the power distribution network.
According to the technical scheme, a sample data set is constructed according to the acquired switch state data in the detection area of the power distribution network, simulated sample data is generated through a generator model in the detection network model, and model training is carried out on a classifier model in the detection network model according to the simulated sample data and the switch sample data, so that the detection network model after the training is acquired as a fault type detection model, the characteristic information content of the detection network model for carrying out characteristic extraction on the sample data is enriched by introducing the simulated sample data in the model training, meanwhile, the time spent for collecting the switch sample data is shortened, the training efficiency and the training effect of the model training are improved, and the detection precision of fault detection on the power distribution network by using the fault type detection model is improved.
Example two
Fig. 2A is a flowchart of a method for establishing a fault type detection model according to a second embodiment of the present invention, where the method is further refined based on the foregoing embodiment, and specific steps for performing model training on a pre-constructed detection network model by using switch sample data until a preset model training end condition is met are provided. It should be noted that, in the embodiments of the present invention, the details of the description of other embodiments may be referred to, and will not be described herein. As shown in fig. 2A, the method includes:
s210, acquiring switch state data in a detection area of the power distribution network.
S220, constructing a sample data set according to the switch state data.
S230, inputting the pre-constructed random noise data into a generator model in the detection network model to obtain simulation sample data output by the generator model.
The random noise data may be white noise or other noise data, and this embodiment is not limited thereto.
Specifically, the random noise data constructed in advance can be input into a generator model in the detection network model, and the generator model can output simulation sample data according to the input random noise data.
Optionally, in an embodiment of the present invention, a schematic structural diagram of the generator model may be as shown in fig. 2B, and the classifier model may include four deconvolution layers, a first convolution layer, and an attention feature extraction layer; the attention feature extraction layer includes a channel attention feature extraction layer and a spatial attention feature extraction layer in series. Correspondingly, the pre-constructed random noise data is input into a generator model in the detection network model to obtain simulation sample data output by the generator model, and the method comprises the following steps: inputting random noise data into a deconvolution layer of a generator model to perform feature extraction operation, so as to obtain a first extracted feature parameter output by the deconvolution layer; inputting the first extracted feature parameters into a channel attention feature extraction layer for feature extraction operation to obtain channel attention feature parameters, and inputting the channel attention feature parameters into a space attention feature extraction layer for feature extraction operation to obtain space attention feature parameters; and inputting the spatial attention characteristic parameters into a first convolution layer to perform characteristic extraction operation, and obtaining the analog sample data output by the first convolution layer. Alternatively, the analog sample data may be represented in a matrix or vector data format, which is not limited in this embodiment.
Specifically, the pre-constructed random noise data is input to a deconvolution layer in a generator model, and after feature extraction of dimension rising is carried out on the random noise data through four deconvolution layers in sequence, a first extraction parameter feature is output; inputting the output first extracted characteristic parameters to a channel attention characteristic extraction layer for characteristic extraction operation, and outputting channel attention characteristic parameters; inputting the channel attention parameters to a spatial attention feature extraction layer for feature extraction operation, and outputting the spatial attention parameters; and finally, inputting the spatial attention parameter into a first convolution layer for feature extraction of dimension reduction, and outputting analog sample data.
By introducing the attention mechanism of the attention feature extraction layer into the generator model, the generator model can achieve better data generation performance under the condition of not increasing model parameters, and the data generation efficiency of the generator model is improved.
Optionally, the channel attention extraction layer includes a first maximum pooling layer, a first average pooling layer, and a shared network layer; the shared network layer comprises at least one sensor; the spatial attention feature extraction layer includes a second max-pooling layer, a second average pooling layer, and a second convolution layer. Correspondingly, inputting the first extracted feature parameter to the channel attention feature extraction layer for feature extraction operation to obtain the channel attention feature parameter, inputting the channel attention feature parameter to the spatial attention feature extraction layer for feature extraction operation to obtain the spatial attention feature parameter, including: inputting the first extracted characteristic parameters into a first maximum pooling layer for characteristic extraction operation to obtain first maximum pooling characteristic parameters, and inputting the first extracted characteristic parameters into a first average pooling layer for characteristic extraction operation to obtain first average pooling characteristic parameters; the first maximum pooling characteristic parameter and the first average pooling characteristic parameter are respectively input into a shared network layer, and characteristic extraction operation is carried out by each sensor to obtain a channel attention characteristic parameter; inputting the channel attention characteristic parameters to a second maximum pooling layer of the space attention characteristic extraction layer for characteristic extraction operation to obtain second maximum pooling characteristic parameters; inputting the second maximum pooling characteristic parameters into a second averaging pooling layer for characteristic extraction operation to obtain second averaging pooling characteristic parameters; and inputting the second average pooled characteristic parameters into a second convolution layer to perform characteristic extraction operation to obtain the spatial attention characteristic parameters.
Specifically, the first extracted characteristic parameters are input to a first largest pooling layer in the channel attention extraction layer to perform characteristic extraction operation and then output the first largest pooling characteristic parameters, and the first extracted characteristic parameters are input to a first average pooling layer in the channel attention extraction layer to perform characteristic extraction operation and then output the first average pooling largest parameters; the first maximum pooling feature parameter and the first average pooling maximum parameter are respectively input into a shared network layer in the channel attention extraction layer, and feature extraction operation is carried out on the first maximum pooling feature parameter and the first average pooling maximum parameter by a perceptron (MLP, multilayer Perceptron) in the shared network layer, so that the channel attention feature parameter is obtained. The number of the sensors may be at least one, for example, the number of the sensors is two, and the sensors are a first sensor and a second sensor respectively. The first sensor is connected to the output end of the first maximum pooling layer, and the second sensor is connected to the output end of the first average pooling layer.
After obtaining the channel attention parameters, inputting the channel attention characteristic parameters to a space attention characteristic extraction layer, carrying out characteristic extraction operation on the channel attention characteristic parameters by a second maximum pooling layer in the space attention extraction layer, and outputting the second maximum pooling characteristic parameters; inputting the second maximum pooling characteristic parameters into a second averaging pooling layer, and outputting the second averaging pooling characteristic parameters after the characteristic extraction operation; and inputting the second average pooling parameter into a second convolution layer for feature extraction operation so as to obtain the spatial attention feature parameter.
Alternatively, the channel attention profile may be determined by the following formula:
wherein M is c (F) Representing the channel attention characteristic parameter,representing a first averaged pooling characteristic parameter, +.>Representing the first maximally pooled feature parameters, MLP () represents the perceptron feature extraction function and σ represents the sigmoid function.
Alternatively, the spatial attention characteristic parameter may be determined by the following formula:
wherein M is s (F) Representing a spatial attention characteristic parameter, representingRepresenting a second averaged pooling characteristic parameter, +.>Representing a second maximum pooling feature parameter, σ representing a sigmoid function, f m×n The convolution kernel size is m×n, and the convolution kernel size is 7×7.
S240, inputting the simulated sample data and the switch sample data into a classifier model in the detection network model, and obtaining output fault types respectively corresponding to the switch sample data output by the classifier model.
The classifier model comprises a third convolution layer, an attention mechanics learning module, a time sequence pooling layer and a full connection layer.
Specifically, the switch sample data and the simulation sample data generated by the generator model in the detection network model are used as input data, and are input into the classifier model in the detection network model, and the output fault types corresponding to the output switch sample data are obtained after the classifier model is processed.
Optionally, inputting the analog sample data and the switch sample data into a classifier model in the detection network model, to obtain output fault types corresponding to the switch sample data output by the classifier model, including: inputting the analog sample data and the switch sample data into a third convolution layer for feature extraction operation to obtain sample extraction feature parameters; inputting the sample extracted characteristic parameters to an attention mechanics learning module for characteristic sequence information extraction operation to obtain characteristic information sequence parameters; inputting the characteristic information sequence parameters to a time sequence pooling layer for characteristic extraction operation to obtain sequence characteristic parameters; and inputting the sequence characteristic parameters into the full-connection layer to perform characteristic integration operation, and obtaining output fault types respectively corresponding to the switch sample data output by the classifier model.
Specifically, the analog sample data and the switch sample data are input into a third convolution layer in the classifier model, and characteristic parameters are extracted from the output samples after characteristic extraction; taking the sample extracted characteristic parameters output by the third convolution layer as the input of an attention mechanics learning module in the classifier model, carrying out characteristic sequence information extraction operation on the sample extracted characteristic parameters, and outputting characteristic information sequence parameters; inputting the characteristic information sequence parameters to a time sequence pooling layer in the classifier model, and outputting the sequence characteristic parameters after characteristic extraction operation; and finally, inputting the sequence characteristic parameters output by the time sequence pooling layer into a full-connection layer in the classifier model to perform characteristic integration operation, and outputting the output fault types corresponding to the switch sample data.
Alternatively, the sample extraction feature parameter may be determined by the following formula:
F＝MaxPooling(ReLU(Conv2d(x))；
wherein x is analog sample data and switch sample data, F is sample extraction characteristic parameters, conv2d () is a convolution function, reLU () is an activation function, and MaxPooling () is a dimension reduction operation function.
S250, performing model iterative training on the classifier model according to the standard fault type and the output fault type of the switch sample data until a preset model training ending condition is met.
Specifically, model iterative training can be performed on the classifier model according to the standard fault class carried by the switch sample data and the output fault class of the switch sample data output by the classifier model, and the training process of the classifier model is known to meet the preset model training ending condition. Optionally, the standard fault type carried by the switch sample data and the output fault type of the output switch sample data of the classifier model may be used as a judging standard, when the output fault type of the output switch sample data of the classifier model can meet the preset type accuracy, the classifier model is determined to meet the preset model training ending condition, or the preset iteration number is set, when the training process of the classifier model reaches the preset iteration number, that is, the output fault type of the output switch sample data of the classifier model can meet the preset type accuracy. Alternatively, the preset category accuracy may be adaptively set by those skilled in the art.
Optionally, the determining mode of the model accuracy may be that, according to the output sample type and the standard sample type of the switch sample data, whether the output sample type and the standard sample type of any switch sample data are consistent is determined, a first data volume of the switch sample data with consistent types is obtained, and a ratio between the first data volume and the total number of the switch sample data is determined as the model accuracy.
And S260, taking the trained detection network model as a fault type detection model.
According to the technical scheme, the simulation sample data are generated through the generator model in the detection network model, and the classifier model in the detection network model is subjected to model training through the switch sample data, so that the detection network model after training is obtained as a fault type detection model, the characteristic information content of the detection network model for carrying out characteristic extraction on the sample data is enriched through introducing the simulation sample data in model training, meanwhile, the time spent for collecting the switch sample data is shortened, the training efficiency and the training effect of model training are improved, and the detection precision of fault detection on the power distribution network by using the fault type detection model is improved.
Example III
Fig. 3 is a flowchart of a fault type detection method provided in a third embodiment of the present invention, where the present embodiment is applicable to a case of locating a fault when the power distribution network fails, the method may be performed by a fault type detection device, the fault type detection device may be implemented in a form of hardware and/or software, and an apparatus for establishing a fault type detection model may be configured in an electronic device.
S310, acquiring to-be-detected switch state data in a to-be-detected area of the power distribution network.
The area to be detected can be an area with faults in the distribution network; the switch state data to be detected may be a switch state of an intelligent switch in the area to be detected when the area to be detected fails, for example, the switch state may be on or off. Alternatively, the switch state data to be detected may be represented in a matrix or vector data form, which is not limited in this embodiment.
Specifically, the switch state data of the to-be-detected area of the power distribution network is collected, and the collected switch state data is used as the to-be-detected switch state data.
S320, inputting the switch state data to be detected into a fault type detection model to obtain a fault type detection result.
If the data used by the fault type detection model in the training process is the switch state data to be detected, the switch state data to be detected can be directly input into the fault type detection model. If the data used in the training process of the fault type detection model is the data after the data preprocessing of the detection switch state data, the data after the data preprocessing is determined to be the state data of the area to be detected, and the state data of the area to be detected is input into the fault type detection model.
The state data of the area to be detected can be determined by the state data of the switch to be detected and the topology data of the switch structure to be detected. The topology data of the switch structure to be detected may refer to the topology relationship of the intelligent switch in the area to be detected. Alternatively, the topology data of the switch structure to be detected may be represented in a matrix or vector data form, which is not limited in this embodiment.
The fault type detection model may refer to a trained detection network model; by way of example, the detection network model may include a generator model and a classifier model; the generator model may be used to generate simulated sample data; the classifier model can be used for carrying out self model training based on the simulation sample data and the switch sample data to obtain output fault types respectively corresponding to the switch sample data. Alternatively, the generator model and classifier model may be neural network models. Optionally, the simulated sample data is different from the switch sample data, and the simulated sample data is not generated by the switch state data and the switch structure topology data when the history of the detection area of the power distribution network fails, but is generated by the generator model; the two are participated in the training process of the classifier model at the same time, and the training of the classifier model is completed; the switch state data can represent state data of whether an intelligent switch existing in a detection area of the power distribution network is on or not; the switch structure topology data can represent the topology relation of intelligent switches in a detection area of the power distribution network; the distribution network detection area may be a preset distribution network area in the distribution network. Alternatively, there may be multiple power distribution network areas in the power distribution network, and at least one intelligent switch may be present in each power distribution network area.
The fault type detection result may refer to a fault type corresponding to the to-be-detected area state data output by the fault detection model, for example, a fault type such as explosion of a lightning arrester, overload of a distribution voltage device, and a capacitor fault.
Specifically, the to-be-detected switch state data is input into the fault type detection model to detect the to-be-detected switch state data, so that the fault type corresponding to the to-be-detected switch state data output by the fault type detection model can be obtained.
The technical effect of the embodiment of the invention is that the fault type detection result is obtained by inputting the switch state data to be detected into the pre-constructed fault type detection model, the fault location of the fault area in the power distribution network can be rapidly carried out, the fault type information of the fault area to be detected can be timely determined, and the stable operation of the power grid is ensured.
Example IV
Fig. 4 is a schematic structural diagram of a device for establishing a fault type detection model according to a fourth embodiment of the present invention. The device can be realized by hardware and/or software, and the method for establishing the fault type detection model provided by any embodiment of the invention can be executed, and has the corresponding functional modules and beneficial effects of the execution method. As shown in fig. 4, the apparatus includes:
A switch state acquisition module 410, configured to acquire switch state data in a detection area of the power distribution network;
a sample data construction module 420, configured to construct a sample data set according to the switch state data; wherein the sample data set comprises at least one switch sample data with a standard fault class label;
the model training module 430 is configured to perform model training on a pre-constructed detection network model by using the switch sample data until a preset model training end condition is met; the detection network model comprises a generator model and a classifier model; the generator model is used for generating simulation sample data; the classifier model is used for carrying out self model training based on the simulation sample data and the switch sample data to obtain output fault types respectively corresponding to the switch sample data.
The model determining module 440 is configured to take the trained detection network model as a fault type detection model.
According to the technical scheme, a sample data set is constructed according to the acquired switch state data in the detection area of the power distribution network, simulated sample data is generated through a generator model in the detection network model, and model training is carried out on a classifier model in the detection network model according to the simulated sample data and the switch sample data, so that the detection network model after the training is acquired as a fault type detection model, the characteristic information content of the detection network model for carrying out characteristic extraction on the sample data is enriched by introducing the simulated sample data in the model training, meanwhile, the time spent for collecting the switch sample data is shortened, the training efficiency and the training effect of the model training are improved, and the detection precision of fault detection on the power distribution network by using the fault type detection model is improved.
Optionally, the model training module 430 includes:
the simulation sample generation unit is used for inputting the pre-constructed random noise data into a generator model in the detection network model to obtain simulation sample data output by the generator model;
the fault class output unit is used for inputting the simulated sample data and the switch sample data into a classifier model in the detection network model to obtain output fault classes respectively corresponding to the switch sample data output by the classifier model;
and the model training unit is used for carrying out model iterative training on the classifier model according to the standard fault type of the switch sample data and the output fault type until a preset model training ending condition is met.
Optionally, the generator model includes a deconvolution layer, a first convolution layer, and an attention feature extraction layer; the attention feature extraction layer includes a channel attention feature extraction layer and a spatial attention feature extraction layer.
Optionally, the analog sample generating unit includes:
the first characteristic parameter extraction subunit is used for inputting random noise data into the deconvolution layer of the generator model to perform characteristic extraction operation, so as to obtain first characteristic parameter extraction output by the deconvolution layer;
The space attention characteristic subunit is used for inputting the first extracted characteristic parameter into the channel attention characteristic extraction layer for characteristic extraction operation to obtain the channel attention characteristic parameter, and inputting the channel attention characteristic parameter into the space attention characteristic extraction layer for characteristic extraction operation to obtain the space attention characteristic parameter;
and the analog sample subunit is used for inputting the spatial attention characteristic parameters into the first convolution layer to perform characteristic extraction operation, so as to obtain analog sample data output by the first convolution layer.
Optionally, the channel attention extraction layer includes a first maximum pooling layer, a first average pooling layer, and a shared network layer; the shared network layer comprises at least one sensor; the spatial attention feature extraction layer includes a second max-pooling layer, a second average pooling layer, and a second convolution layer.
Optionally, the spatial attention feature subunit is specifically configured to: inputting the first extracted characteristic parameters into a first maximum pooling layer for characteristic extraction operation to obtain first maximum pooling characteristic parameters, and inputting the first extracted characteristic parameters into a first average pooling layer for characteristic extraction operation to obtain first average pooling characteristic parameters; the first maximum pooling characteristic parameter and the first average pooling characteristic parameter are respectively input into a shared network layer, and characteristic extraction operation is carried out by each sensor to obtain a channel attention characteristic parameter; inputting the channel attention characteristic parameters to a second maximum pooling layer of the space attention characteristic extraction layer for characteristic extraction operation to obtain second maximum pooling characteristic parameters; inputting the second maximum pooling characteristic parameters into a second averaging pooling layer for characteristic extraction operation to obtain second averaging pooling characteristic parameters; and inputting the second average pooled characteristic parameters into a second convolution layer to perform characteristic extraction operation to obtain the spatial attention characteristic parameters.
Optionally, the classifier model includes a third convolution layer, an attention learning module, a timing pooling layer, and a full connection layer.
Optionally, the fault class output unit includes:
the sample characteristic subunit is used for inputting the analog sample data and the switch sample data into the third convolution layer to perform characteristic extraction operation to obtain sample extraction characteristic parameters;
the characteristic information sequence subunit is used for inputting the characteristic parameters extracted by the sample into the attention mechanics learning module for characteristic sequence information extraction operation to obtain characteristic information sequence parameters;
the sequence feature subunit is used for inputting the feature information sequence parameters to the time sequence pooling layer for feature extraction operation to obtain sequence feature parameters;
and the fault class output subunit is used for inputting the sequence characteristic parameters to the full-connection layer for characteristic integration operation to obtain output fault classes respectively corresponding to the switch sample data output by the classifier model.
The device for establishing the fault type detection model provided by the embodiment of the invention can execute the method for establishing the fault type detection model provided by any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method.
Example five
Fig. 5 is a schematic structural diagram of a fault type detection device according to a fifth embodiment of the present invention. The device can be realized by hardware and/or software, and the fault type detection method provided by any embodiment of the invention can be executed, and has the corresponding functional modules and beneficial effects of the execution method. As shown in fig. 5, the apparatus includes:
the to-be-detected data acquisition module 510 is configured to acquire to-be-detected switch state data in a to-be-detected area of the power distribution network;
the fault type determining module 520 is configured to input the switch status data to be detected to the fault type detection model, so as to obtain a fault type detection result.
The fault type detection model may be generated by using the method for establishing the fault type detection model in any one of the above embodiments.
The technical effect of the embodiment of the invention is that the fault type detection result is obtained by inputting the switch state data to be detected into the pre-constructed fault type detection model, the fault location of the fault area in the power distribution network can be rapidly carried out, the fault type information of the fault area to be detected can be timely determined, and the stable operation of the power grid is ensured.
The device for establishing the fault type detection model provided by the embodiment of the invention can execute the method for establishing the fault type detection model provided by any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method.
Example six
Fig. 6 shows a schematic diagram of an electronic device 610 that may be used to implement an embodiment of the invention. Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. Electronic equipment may also represent various forms of mobile devices, such as personal digital processing, cellular telephones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be exemplary only, and are not meant to limit implementations of the inventions described and/or claimed herein.
As shown in fig. 6, the electronic device 610 includes at least one processor 611, and a memory, such as a Read Only Memory (ROM) 612, a Random Access Memory (RAM) 613, etc., communicatively coupled to the at least one processor 611, where the memory stores computer programs executable by the at least one processor, and the processor 611 may perform various suitable actions and processes according to the computer programs stored in the Read Only Memory (ROM) 612 or the computer programs loaded from the storage unit 618 into the Random Access Memory (RAM) 613. In the RAM 613, various programs and data required for the operation of the electronic device 610 may also be stored. The processor 611, the ROM 612, and the RAM 613 are connected to each other by a bus 614. An input/output (I/O) interface 615 is also connected to bus 614.
Various components in the electronic device 610 are connected to the I/O interface 615, including: an input unit 616 such as a keyboard, mouse, etc.; an output unit 617 such as various types of displays, speakers, and the like; a storage unit 618, such as a magnetic disk, optical disk, etc.; and a communication unit 619 such as a network card, modem, wireless communication transceiver, etc. The communication unit 619 allows the electronic device 610 to exchange information/data with other devices through computer networks, such as the internet, and/or various telecommunication networks.
Processor 611 may be a variety of general and/or special purpose processing components having processing and computing capabilities. Some examples of processor 611 include, but are not limited to, a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), various specialized Artificial Intelligence (AI) computing chips, various processors running machine learning model algorithms, digital Signal Processors (DSPs), and any suitable processor, controller, microcontroller, etc. The processor 611 performs the respective methods and processes described above, such as a fault type detection model establishment method or a fault type detection method.
In some embodiments, the method of building the fault type detection model or the method of fault type detection may be implemented as a computer program tangibly embodied on a computer-readable storage medium, such as storage unit 618. In some embodiments, some or all of the computer program may be loaded and/or installed onto the electronic device 610 via the ROM 612 and/or the communication unit 619. When the computer program is loaded into the RAM 613 and executed by the processor 11, one or more steps of the above-described method of establishing a fault type detection model or fault type detection method may be performed. Alternatively, in other embodiments, processor 611 may be configured to perform the method of building the fault type detection model or the method of fault type detection in any other suitable manner (e.g., by means of firmware).
Various implementations of the systems and techniques described here above may be implemented in digital electronic circuitry, integrated circuit systems, field Programmable Gate Arrays (FPGAs), application Specific Integrated Circuits (ASICs), application Specific Standard Products (ASSPs), systems On Chip (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs, the one or more computer programs may be executed and/or interpreted on a programmable system including at least one programmable processor, which may be a special purpose or general-purpose programmable processor, that may receive data and instructions from, and transmit data and instructions to, a storage system, at least one input device, and at least one output device.
A computer program for carrying out methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the computer programs, when executed by the processor, cause the functions/acts specified in the flowchart and/or block diagram block or blocks to be implemented. The computer program may execute entirely on the machine, partly on the machine, as a stand-alone software package, partly on the machine and partly on a remote machine or entirely on the remote machine or server.
In the context of the present invention, a computer-readable storage medium may be a tangible medium that can contain, or store a computer program for use by or in connection with an instruction execution system, apparatus, or device. The computer readable storage medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. Alternatively, the computer readable storage medium may be a machine readable signal medium. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
To provide for interaction with a user, the systems and techniques described here can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and a pointing device (e.g., a mouse or a trackball) through which a user can provide input to the electronic device. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic input, speech input, or tactile input.
The systems and techniques described here can be implemented in a computing system that includes a background component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such background, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), wide Area Networks (WANs), blockchain networks, and the internet.
The computing system may include clients and servers. The client and server are typically remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server can be a cloud server, also called a cloud computing server or a cloud host, and is a host product in a cloud computing service system, so that the defects of high management difficulty and weak service expansibility in the traditional physical hosts and VPS service are overcome.
It should be appreciated that various forms of the flows shown above may be used to reorder, add, or delete steps. For example, the steps described in the present invention may be performed in parallel, sequentially, or in a different order, so long as the desired results of the technical solution of the present invention are achieved, and the present invention is not limited herein.
The above embodiments do not limit the scope of the present invention. It will be apparent to those skilled in the art that various modifications, combinations, sub-combinations and alternatives are possible, depending on design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the scope of the present invention.
Claims (10)
1. The method for establishing the fault type detection model is characterized by comprising the following steps of:
acquiring switch state data in a detection area of the power distribution network;
constructing a sample data set according to the switch state data; the sample data set comprises at least one switch sample data with a standard fault class label;
performing model training on a pre-constructed detection network model by adopting the switch sample data until a preset model training ending condition is met;
Wherein the detection network model comprises a generator model and a classifier model; the generator model is used for generating simulation sample data; the classifier model is used for carrying out self model training based on the simulation sample data and the switch sample data to obtain output fault categories respectively corresponding to the switch sample data;
and taking the trained detection network model as a fault type detection model.
2. The method according to claim 1, wherein the model training the pre-constructed detection network model using the switch sample data until a preset model training end condition is satisfied, comprises:
inputting pre-constructed random noise data into a generator model in the detection network model to obtain simulation sample data output by the generator model;
inputting the simulated sample data and the switch sample data into a classifier model in the detection network model to obtain output fault categories respectively corresponding to the switch sample data output by the classifier model;
and carrying out model iterative training on the classifier model according to the standard fault type and the output fault type of the switch sample data until a preset model training ending condition is met.
3. The method of claim 2, wherein the generator model comprises a deconvolution layer, a first convolution layer, and an attention feature extraction layer; the attention feature extraction layer comprises a channel attention feature extraction layer and a space attention feature extraction layer;
correspondingly, the inputting the pre-constructed random noise data into a generator model in the detection network model to obtain the simulation sample data output by the generator model comprises the following steps:
inputting the random noise data to a deconvolution layer of the generator model for feature extraction operation to obtain a first extracted feature parameter output by the deconvolution layer;
inputting the first extracted feature parameters to the channel attention feature extraction layer for feature extraction operation to obtain channel attention feature parameters, and inputting the channel attention feature parameters to the space attention feature extraction layer for feature extraction operation to obtain space attention feature parameters;
and inputting the spatial attention characteristic parameters into the first convolution layer to perform characteristic extraction operation to obtain analog sample data output by the first convolution layer.
4. The method of claim 3, wherein the channel attention extraction layer comprises a first max pooling layer, a first average pooling layer, and a shared network layer; the shared network layer comprises at least one sensor; the spatial attention feature extraction layer comprises a second maximum pooling layer, a second average pooling layer and a second convolution layer;
correspondingly, the step of inputting the first extracted feature parameter to the channel attention feature extraction layer to perform feature extraction operation to obtain a channel attention feature parameter, and inputting the channel attention feature parameter to the spatial attention feature extraction layer to perform feature extraction operation to obtain a spatial attention feature parameter includes:
inputting the first extracted characteristic parameters to the first maximum pooling layer for characteristic extraction operation to obtain first maximum pooling characteristic parameters, and inputting the first extracted characteristic parameters to the first average pooling layer for characteristic extraction operation to obtain first average pooling characteristic parameters;
inputting the first maximum pooling characteristic parameter and the first average pooling characteristic parameter to the shared network layer respectively, and performing characteristic extraction operation by each sensor to obtain a channel attention characteristic parameter;
Inputting the channel attention characteristic parameters to a second maximum pooling layer of the space attention characteristic extraction layer for characteristic extraction operation to obtain second maximum pooling characteristic parameters;
inputting the second maximum pooling characteristic parameters into the second average pooling layer for characteristic extraction operation to obtain second average pooling characteristic parameters;
and inputting the second average pooling characteristic parameters into the second convolution layer to perform characteristic extraction operation to obtain the spatial attention characteristic parameters.
5. The method of claim 2, wherein the classifier model comprises a third convolution layer, an attention learning module, a timing pooling layer, and a fully connected layer;
inputting the simulated sample data and the switch sample data into a classifier model in the detection network model to obtain output fault categories respectively corresponding to the switch sample data output by the classifier model, wherein the method comprises the following steps of:
inputting the analog sample data and the switch sample data to the third convolution layer for feature extraction operation to obtain sample extraction feature parameters;
inputting the sample extracted characteristic parameters to the attention mechanics learning module for characteristic sequence information extraction operation to obtain characteristic information sequence parameters;
Inputting the characteristic information sequence parameters to the time sequence pooling layer for characteristic extraction operation to obtain sequence characteristic parameters;
and inputting the sequence characteristic parameters to the full-connection layer for characteristic integration operation to obtain output fault types respectively corresponding to the switch sample data output by the classifier model.
6. A fault type detection method, comprising:
acquiring state data of a switch to be detected in a region to be detected of the power distribution network;
inputting the switch state data to be detected into a fault type detection model to obtain a fault type detection result; wherein the fault type detection model is generated using the method of any one of claims 1-5.
7. A device for building a fault type detection model, comprising:
the switch state acquisition module is used for acquiring switch state data in a detection area of the power distribution network;
the sample data construction module is used for constructing a sample data set according to the switch state data; wherein the sample data set comprises at least one switch sample data with a standard fault class label;
the model training module is used for carrying out model training on a pre-constructed detection network model by adopting the switch sample data until a preset model training ending condition is met; wherein the detection network model comprises a generator model and a classifier model; the generator model is used for generating simulation sample data; the classifier model is used for carrying out self model training based on the simulation sample data and the switch sample data to obtain output fault categories respectively corresponding to the switch sample data;
And the model determining module is used for taking the trained detection network model as a fault type detection model.
8. A fault type detection device, comprising:
the to-be-detected data acquisition module is used for acquiring to-be-detected switch state data in a to-be-detected area of the power distribution network;
the fault type determining module is used for inputting the switch state data to be detected into the fault type detecting model to obtain a fault type detecting result; wherein the fault type detection model is generated using the method of any one of claims 1-5.
9. An electronic device, the electronic device comprising:
at least one processor; and
a memory communicatively coupled to the at least one processor; wherein,,
the memory stores a computer program executable by the at least one processor to enable the at least one processor to perform the method of building a fault type detection model according to any one of claims 1-5 or the method of fault type detection according to claim 6.
10. A computer readable storage medium storing computer instructions for causing a processor to implement the method of building a fault type detection model according to any one of claims 1-5 or the method of fault type detection according to claim 6 when executed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310944264.1A CN116956042A (en) | 2023-07-28 | 2023-07-28 | Method, device, equipment and medium for establishing fault type detection model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310944264.1A CN116956042A (en) | 2023-07-28 | 2023-07-28 | Method, device, equipment and medium for establishing fault type detection model |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116956042A true CN116956042A (en) | 2023-10-27 |
Family
ID=88461612
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310944264.1A Pending CN116956042A (en) | 2023-07-28 | 2023-07-28 | Method, device, equipment and medium for establishing fault type detection model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116956042A (en) |
-
2023
- 2023-07-28 CN CN202310944264.1A patent/CN116956042A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113657467B (en) | Model pre-training method and device, electronic equipment and storage medium | |
| CN116090605A (en) | Pipe network early warning method and device, electronic equipment and storage medium | |
| CN116975081A (en) | Log diagnosis set updating method, device, equipment and storage medium | |
| CN116956042A (en) | Method, device, equipment and medium for establishing fault type detection model | |
| CN115600607A (en) | Log detection method and device, electronic equipment and medium | |
| CN116127326B (en) | Composite insulator detection method and device, electronic equipment and storage medium | |
| CN115373449B (en) | Data processing method, device, equipment and storage medium | |
| CN116990744B (en) | Electric energy meter detection method, device, equipment and medium | |
| CN115542100B (en) | Insulator fault detection method, device, equipment and medium | |
| CN116468076A (en) | Driving behavior analysis method and device, electronic equipment and storage medium | |
| CN117574146B (en) | Text classification labeling method, device, electronic equipment and storage medium | |
| CN117749676A (en) | Reverse analysis and vulnerability test method, device, equipment and medium of industrial control protocol | |
| CN117667488A (en) | Abnormal positioning method, device, equipment and medium for artificial intelligent computing power cluster | |
| CN116308284A (en) | Operation data detection method, device and equipment of pumped storage equipment | |
| CN117251809A (en) | Power grid time sequence data anomaly detection method, device, equipment and storage medium | |
| CN118074625A (en) | Equipment fault detection method, device, equipment and storage medium | |
| CN118133127A (en) | Substation equipment fault prediction method and device, electronic equipment and readable medium | |
| CN116303071A (en) | Interface testing method and device, electronic equipment and storage medium | |
| CN116795615A (en) | Chip evaluation method, system, electronic equipment and storage medium | |
| CN118097144A (en) | Model training and point cloud segmentation method and device and electronic equipment | |
| CN116307159A (en) | Load prediction method and device, electronic equipment and storage medium | |
| CN116111562A (en) | Method, device, equipment and storage medium for determining reliability of power distribution network | |
| CN117421377A (en) | Data processing method, device, equipment and medium for energy station | |
| CN115223177A (en) | Text recognition method, device, equipment and storage medium | |
| CN117131990A (en) | Power grid infrastructure information management method and device, electronic equipment and storage medium |
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 |