CN117592513A - Method and system for generating power system operation mode sample based on variation self-encoder and model migration - Google Patents
Method and system for generating power system operation mode sample based on variation self-encoder and model migration Download PDFInfo
- Publication number
- CN117592513A CN117592513A CN202311559870.8A CN202311559870A CN117592513A CN 117592513 A CN117592513 A CN 117592513A CN 202311559870 A CN202311559870 A CN 202311559870A CN 117592513 A CN117592513 A CN 117592513A
- Authority
- CN
- China
- Prior art keywords
- encoder
- power system
- self
- sample
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 41
- 238000013508 migration Methods 0.000 title claims abstract description 17
- 230000005012 migration Effects 0.000 title claims abstract description 17
- 238000012549 training Methods 0.000 claims abstract description 76
- 238000009826 distribution Methods 0.000 claims abstract description 49
- 238000012795 verification Methods 0.000 claims description 15
- 238000011156 evaluation Methods 0.000 claims description 7
- 238000000342 Monte Carlo simulation Methods 0.000 claims description 6
- 238000010606 normalization Methods 0.000 claims description 5
- 238000005070 sampling Methods 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 abstract description 3
- 238000004458 analytical method Methods 0.000 description 9
- 238000012360 testing method Methods 0.000 description 8
- 238000007796 conventional method Methods 0.000 description 3
- 238000005457 optimization Methods 0.000 description 3
- 230000006978 adaptation Effects 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 238000007476 Maximum Likelihood Methods 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 230000035800 maturation Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000000638 solvent extraction Methods 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution 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
- G06N3/045—Combinations of networks
- G06N3/0455—Auto-encoder networks; Encoder-decoder 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
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J2203/00—Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
- H02J2203/10—Power transmission or distribution systems management focussing at grid-level, e.g. load flow analysis, node profile computation, meshed network optimisation, active network management or spinning reserve management
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J2203/00—Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
- H02J2203/20—Simulating, e g planning, reliability check, modelling or computer assisted design [CAD]
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Power Engineering (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention discloses a method for generating a power system operation mode sample based on variation self-encoder and model migration, which comprises the following steps: acquiring operation data of a target power system and constructing a training data set of a variation self-encoder; inputting the corresponding training data set into a variable self-encoder according to different given tasks, and training to obtain the variable self-encoder of the corresponding task; and inputting the hidden characteristic distribution combination of the operation data into a decoder of the variation self-encoder to obtain the operation mode sample data of the power system. According to the method, the basic model is obtained by obtaining the operation data of the target power system through pre-training, only a small amount of samples are needed to train one model in the process of learning and training different distribution combinations according to the sample data of a specific operation mode, the method has the characteristics of low calculation power requirement and less sample requirement, and the problem of poor reliability of the generated samples in the conventional power system operation mode sample generation method is solved.
Description
Technical Field
The invention relates to the technical field of power systems, in particular to a power system operation mode sample generation method and a power system operation mode sample generation system based on variation self-encoder and model migration.
Background
In recent years, with the continuous maturation of power electronics technology, large-scale renewable energy sources have largely replaced traditional synchronous generators in the context of resource limitations and environmental pressures. The distributed and centralized renewable energy sources are connected into the power grid, so that the uncertainty and complexity of the running mode of the power grid are greatly improved. In order to ensure the safety and economy of the power grid, the operation mode analysis of the power system is an indispensable important means, but as the number of the operation systems is continuously increased, the analysis difficulty is also obviously improved, and how to obtain the operation mode test integration required by the analysis is a fundamental difficulty problem. The current method generally extracts according to experience manually, but with the increase of analysis requirements, the traditional method is difficult to meet the requirements, and a sample-driven random analysis method becomes the mainstream, and mainly comprises two main types of random simulation and optimization models, wherein the former generally samples uncertain factors randomly to obtain a test sample, and the latter generally provides constraint based on the analysis requirements and aims to build an optimization model to solve the test sample. However, the electric power system often has a series of problems such as high dimensionality, nonlinearity, uncertainty and the like, so that a random simulation method has a large number of problems, and a typical sample is difficult to obtain in a targeted manner; the optimization method also has a series of problems of high solving difficulty, high time cost and the like.
How to efficiently generate a sample of a typical operation mode of a power system so as to cope with the sample requirement of increasingly complex power system analysis, improve the safety of the power system, and become a problem to be solved in the power system analysis.
In summary, the conventional method for generating the power system operation mode sample has the problem that the generated sample has poor reliability.
Disclosure of Invention
In view of the above, the invention provides a method and a system for generating a power system operation mode sample based on variation self-encoder and model migration, which are used for reconstructing typical sample data of a power system in different operation modes by extracting the power system operation modes of different power system operation modes, thereby solving the problem of poor reliability of the generated sample in the conventional power system operation mode sample generation method.
In order to solve the above problems, the technical scheme of the invention is to adopt a method for generating a power system operation mode sample based on variation self-encoder and model migration, comprising the following steps: acquiring operation data of a target power system and constructing a training data set of a variation self-encoder; inputting the corresponding training data set into a variable self-encoder according to different given tasks, and training to obtain the variable self-encoder of the corresponding task; and inputting the hidden characteristic distribution combination of the operation data into a decoder of the variation self-encoder to obtain the operation mode sample data of the power system.
Optionally, acquiring the operational data of the target power system and constructing the training data set of the variational self-encoder includes: acquiring operation data of a target power system in different operation modes based on a Monte Carlo simulation method, and constructing a sample training set by combining historical data of the target power system; dividing the sample training set based on different evaluation indexes corresponding to a plurality of preset operation modes to obtain sample sets under different operation modes and marking; and carrying out data normalization and denoising on the sample sets under different operation modes to obtain the training data set of the variation self-encoder.
Optionally, according to different given tasks, inputting the corresponding training data set into a variable-score self-encoder, and training to obtain the variable-score self-encoder of the corresponding task, including: dividing the training data set into a training set and a verification set; determining the number of hidden features and the distribution mode of the hidden features according to a given task; configuring the variation self-encoder based on the hidden characteristic distribution mode; and inputting the training set into the variation self-encoder, and obtaining the trained variation self-encoder of the corresponding task after verification based on the verification set.
Optionally, the method for judging that the variation is completed from the training of the encoder comprises the following steps: inputting the training set into the variational self-encoder; the encoder based on the variation self-encoder learns the distribution of hidden features of samples of the typical operation mode of the corresponding power system, and obtains feature samples by sampling the hidden features; inputting the characteristic samples of the hidden characteristics into a decoder of the variation self-encoder for reconstruction to obtain operation mode sample data and reconstruction errors, wherein the operation mode sample data is hidden characteristic distribution of the operation data; and updating the network parameters of the variable self-encoder based on the reconstruction error until the reconstruction error is lower than a preset threshold.
Optionally, inputting the hidden feature distribution combination of the operation data to a decoder of the variation self-encoder to obtain operation mode sample data of the power system, including: and according to the running mode of the electric power system required by a given task, calling the corresponding characteristic sample of the hidden characteristic and inputting the characteristic sample to a decoder of the trained variable self-encoder to obtain the running mode sample data of the electric power system.
Accordingly, the invention provides a system for generating a power system operation mode sample based on variation self-encoder and model migration, comprising: the data acquisition module is used for acquiring the operation data of the target power system and constructing a training data set of the variable self-encoder; the training module inputs the corresponding training data set into the variable self-encoder according to different given tasks, and trains the variable self-encoder to obtain the corresponding tasks; and the sample generation module is used for inputting the hidden characteristic distribution combination of the operation data into a decoder of the variation self-encoder to obtain the operation mode sample data of the power system.
Optionally, the data acquisition module obtains operation data of the target power system under different operation modes based on a Monte Carlo simulation method, and divides the sample training set based on different evaluation indexes corresponding to a plurality of preset operation modes after constructing the sample training set by combining historical data of the target power system, so as to obtain and label the sample set under the different operation modes, and then normalizes and denoises the data of the sample set under the different operation modes to obtain the training data set of the variable self-encoder.
Optionally, the training module determines the number of hidden features and the distribution mode of the hidden features according to a given task, configures the variable self-encoder based on the distribution mode of the hidden features, inputs the training set into the variable self-encoder, and obtains the trained variable self-encoder of the corresponding task after verification based on the verification set.
Optionally, the sample generating module invokes the feature samples of the corresponding hidden features according to the running mode of the power system required by the given task and inputs the feature samples to the decoder of the trained variational self-encoder to obtain the running mode sample data of the power system.
The primary improvement of the invention is that the provided generation method of the power system operation mode sample based on the variation self-encoder and model migration obtains a basic model through pre-training by obtaining the operation data of the target power system, can fully learn the hidden characteristics of the power system typical operation mode sample of different types, is suitable for the generation of the test sample of each type of typical operation mode, and can use the hidden characteristics of different distributions according to actual demands. In addition, only a small amount of samples are needed to train a model according to different distribution combinations of sample data learning of a specific operation mode, the method has the characteristics of low calculation force requirement and less sample requirement, can provide good data support for various typical operation modes of an electric power system, and solves the problem of poor reliability of generated samples in the conventional method for generating the samples of the operation mode of the electric power system.
Drawings
FIG. 1 is a simplified flow diagram of a method of generating a power system run mode sample based on variation self-encoder and model migration of the present invention;
FIG. 2 is a simplified block diagram of a system for generating a power system run mode sample based on variation from encoder and model migration of the present invention.
Detailed Description
In order to make the technical solution of the present invention better understood by those skilled in the art, the present invention will be further described in detail with reference to the accompanying drawings and specific embodiments.
As shown in fig. 1, a method for generating a power system operation mode sample based on variation self-encoder and model migration includes:
s1: acquiring operational data of a target power system and constructing a training data set of a variational self-encoder, comprising: acquiring operation data of a target power system in different operation modes based on a Monte Carlo simulation method, and constructing a sample training set by combining historical data of the target power system; dividing the sample training set based on different evaluation indexes corresponding to a plurality of preset operation modes to obtain sample sets under different operation modes and marking; and carrying out data normalization and denoising on the sample sets under different operation modes to obtain the training data set of the variation self-encoder.
Furthermore, a series of methods such as variable analysis, partitioning, SVD, PCA and the like can be used for performing dimension reduction processing on the operation data of the large-scale power system so as to improve the training effect of the training data set on the variable self-encoder.
Furthermore, although different evaluation indexes corresponding to the preset multiple operation modes may be different due to different research topics, the sample generation step and the model migration training method have universality.
S2: inputting the corresponding training data set into a variable self-encoder according to different given tasks, training to obtain the variable self-encoder of the corresponding tasks, and comprising the following steps: dividing the training data set into a training set and a verification set; determining the number of hidden features and the distribution mode of the hidden features according to a given task; configuring the variation self-encoder based on the hidden characteristic distribution mode; and inputting the training set into the variation self-encoder, and obtaining the trained variation self-encoder of the corresponding task after verification based on the verification set.
Further, the number of hidden features may be the same, but the manner in which the hidden features are distributed may be different, depending on the given task. For example, task one: the test sample of a specific operation mode of the power system is needed, the distribution selection of the hidden features is more random, all the features can obey one distribution (standard normal distribution and even distribution), and each feature can also obey different distributions. Task two: multiple power system modes of operation are required, but targeted generation, selection and task one consistency are not required. Task three: multiple power system modes of operation are required, and targeted generation is required. Such tasks may be split into multiple tasks-i.e., training multiple variations from the encoder for generation of corresponding test samples. It may also be placed in a variable self-encoder, but different distributions, in particular the same hidden feature, need to be set for each class of hidden features for samples, and the hidden feature distribution between each class is as far as possible, otherwise the probability of error in generating test samples is greatly increased. This does not occur for the first approach.
Further, the method for judging the completion of the training of the variation self-encoder comprises the following steps: inputting the training set into the variational self-encoder; the encoder based on the variation self-encoder learns the distribution of hidden features of samples of a typical operation mode of a corresponding power system, constructs a mirror image decoder of the distribution, and obtains feature samples by sampling the hidden features; inputting the characteristic samples of the hidden characteristics into a decoder of the variation self-encoder for reconstruction to obtain operation mode sample data and reconstruction errors, wherein the operation mode sample data is hidden characteristic distribution of the operation data; and updating the network parameters of the variable self-encoder based on the reconstruction error, repeating the operation to train, and stopping training when the reconstruction error and the hidden characteristic distribution learned by the encoder are close to the real hidden characteristic distribution, namely repeating the operation until the reconstruction error is lower than a preset threshold value.
Further, if there is a trained variant self-encoder, the variant self-encoder for a new task can be learned by model migration using a small number of samples as a given task changes.
Further, the variational self-encoder is composed of three parts, the first part is an inference network (encoder) for learning the distribution of hidden features Z; the second part is a sampler, and corresponding sampling is carried out in the distribution of the hidden features Z to obtain hidden features; the third part is a generation network (decoder) that reconstructs data from the hidden feature distribution obtained by the sampler, thereby generating samples that do not exist in the original data.
First, we can describe the training data set input as P (X) = ≡first z P (x|z) P (z) dz, wherein f (z) is replaced by P (x|z), so that X dependence on z can be explicitly expressed by a probability formula, i.e. X can be generated by hiding the feature z.
Specifically, in calculating the KL divergence, q (z|x) is introduced because the posterior distribution P (X|z) is difficult to solve, thenFurther, we can simply write as log P (X) =elbo+kl (q (z|x) |p (z|x)), since the distribution of z is unknown or unknown, we need to do so by making the distribution of the hidden features z of the sample as close as possible to the distribution we set, i.e. minimizing the KL divergence as described above, i.e. q (z|x) as close as possible to P (z|x), gets our first Loss function Loss 1 =kl (q (z|x) |p (z|x)) and then subjected to maximum likelihood estimation L b P (X) can be obtained. At this time, the liquid crystal display device,
then Loss is 2 ＝KL(q(z|x)||P(z))-∫ z q (z|x) log (P (x|z)) dz, i.e. the reconstruction error.
S3: inputting the hidden characteristic distribution combination of the operation data into a decoder of the variation self-encoder to obtain operation mode sample data of the power system, wherein the operation mode sample data comprises the following steps: and according to the running mode of the electric power system required by a given task, calling the corresponding characteristic sample of the hidden characteristic and inputting the characteristic sample to a decoder of the trained variable self-encoder to obtain the running mode sample data of the electric power system.
According to the invention, the basic model is obtained by obtaining the operation data of the target power system through pre-training, so that the hidden characteristics of samples of typical operation modes of different types of power systems can be fully learned, the method is suitable for generating test samples of various types of typical operation modes, and the hidden characteristics with different distributions can be used according to actual requirements. In addition, only a small amount of samples are needed to train a model according to different distribution combinations of sample data learning of a specific operation mode, the method has the characteristics of low calculation force requirement and less sample requirement, can provide good data support for various typical operation modes of an electric power system, and solves the problem of poor reliability of generated samples in the conventional method for generating the samples of the operation mode of the electric power system.
Accordingly, as shown in fig. 2, the present invention provides a system for generating a power system operation mode sample based on variation self-encoder and model migration, including: the data acquisition module is used for acquiring the operation data of the target power system and constructing a training data set of the variable self-encoder; the training module inputs the corresponding training data set into the variable self-encoder according to different given tasks, and trains the variable self-encoder to obtain the corresponding tasks; and the sample generation module is used for inputting the hidden characteristic distribution combination of the operation data into a decoder of the variation self-encoder to obtain the operation mode sample data of the power system.
Further, the data acquisition module obtains operation data of the target power system under different operation modes based on a Monte Carlo simulation method, and divides the sample training set based on different evaluation indexes corresponding to a plurality of preset operation modes after the sample training set is constructed by combining historical data of the target power system, so as to obtain and label the sample set under the different operation modes, and performs data normalization and denoising on the sample set under the different operation modes to obtain the training data set of the variable self-encoder.
Further, the training module determines the number of hidden features and the distribution mode of the hidden features according to a given task, configures the variable self-encoder based on the distribution mode of the hidden features, inputs the training set into the variable self-encoder, and obtains the trained variable self-encoder of the corresponding task after verification based on the verification set.
Further, the sample generation module calls the feature samples of the corresponding hidden features according to the running mode of the power system required by the given task and inputs the feature samples to the decoder of the trained variational self-encoder to obtain the running mode sample data of the power system.
The method and the system for generating the power system operation mode sample based on the variation self-encoder and the model migration provided by the embodiment of the invention are as above. In the description, each embodiment is described in a progressive manner, and each embodiment is mainly described by the differences from other embodiments, so that the same similar parts among the embodiments are mutually referred. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant points refer to the description of the method section. It should be noted that it will be apparent to those skilled in the art that various modifications and adaptations of the invention can be made without departing from the principles of the invention and these modifications and adaptations are intended to be within the scope of the invention as defined in the following claims.
Those of skill would further appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both, and that the various illustrative elements and steps are described above generally in terms of functionality in order to clearly illustrate the interchangeability of hardware and software. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention. The steps of a method or algorithm described in connection with the embodiments disclosed herein may be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. The software modules may be disposed in Random Access Memory (RAM), memory, read Only Memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.
Claims (9)
1. The utility model provides a generation method of a power system operation mode sample based on variation self-encoder and model migration, which is characterized by comprising the following steps:
acquiring operation data of a target power system and constructing a training data set of a variation self-encoder;
inputting the corresponding training data set into a variable self-encoder according to different given tasks, and training to obtain the variable self-encoder of the corresponding task;
and inputting the hidden characteristic distribution combination of the operation data into a decoder of the variation self-encoder to obtain the operation mode sample data of the power system.
2. The method of generating a power system operation mode sample according to claim 1, wherein acquiring the operation data of the target power system and constructing the training data set of the variational self-encoder includes:
acquiring operation data of a target power system in different operation modes based on a Monte Carlo simulation method, and constructing a sample training set by combining historical data of the target power system;
dividing the sample training set based on different evaluation indexes corresponding to a plurality of preset operation modes to obtain sample sets under different operation modes and marking;
and carrying out data normalization and denoising on the sample sets under different operation modes to obtain the training data set of the variation self-encoder.
3. The method for generating a power system operation mode sample according to claim 2, wherein inputting the corresponding training data set into the variable-score self-encoder according to different given tasks, training the variable-score self-encoder to obtain the corresponding tasks, comprises:
dividing the training data set into a training set and a verification set;
determining the number of hidden features and the distribution mode of the hidden features according to a given task;
configuring the variation self-encoder based on the hidden characteristic distribution mode;
and inputting the training set into the variation self-encoder, and obtaining the trained variation self-encoder of the corresponding task after verification based on the verification set.
4. A method of generating a power system operational mode sample according to claim 3, wherein the method of determining that the variation self-encoder training is complete comprises:
inputting the training set into the variational self-encoder;
the encoder based on the variation self-encoder learns the distribution of hidden features of samples of the typical operation mode of the corresponding power system, and obtains feature samples by sampling the hidden features;
inputting the characteristic samples of the hidden characteristics into a decoder of the variation self-encoder for reconstruction to obtain operation mode sample data and reconstruction errors, wherein the operation mode sample data is hidden characteristic distribution of the operation data;
and updating the network parameters of the variable self-encoder based on the reconstruction error until the reconstruction error is lower than a preset threshold.
5. The method for generating a power system operation mode sample according to claim 4, wherein inputting the hidden feature distribution combination of the operation data to the decoder of the variation self-encoder to obtain the power system operation mode sample data comprises:
and according to the running mode of the electric power system required by a given task, calling the corresponding characteristic sample of the hidden characteristic and inputting the characteristic sample to a decoder of the trained variable self-encoder to obtain the running mode sample data of the electric power system.
6. A system for generating a power system run mode sample based on variation self-encoder and model migration, comprising:
the data acquisition module is used for acquiring the operation data of the target power system and constructing a training data set of the variable self-encoder;
the training module inputs the corresponding training data set into the variable self-encoder according to different given tasks, and trains the variable self-encoder to obtain the corresponding tasks;
and the sample generation module is used for inputting the hidden characteristic distribution combination of the operation data into a decoder of the variation self-encoder to obtain the operation mode sample data of the power system.
7. The system for generating the power system operation mode sample according to claim 6, wherein the data acquisition module obtains operation data of the target power system in different operation modes based on a Monte Carlo simulation method, and divides the sample training set based on different evaluation indexes corresponding to a plurality of preset operation modes after the sample training set is built by combining historical data of the target power system, so as to obtain and label the sample set in the different operation modes, and performs data normalization and denoising on the sample set in the different operation modes to obtain the training data set of the variable self-encoder.
8. The system according to claim 6, wherein the training module determines the number of hidden features and the distribution of hidden features according to a given task, configures the variable self-encoder based on the distribution of hidden features, inputs the training set into the variable self-encoder, and obtains the trained variable self-encoder of the corresponding task based on verification of the verification set.
9. The system for generating power system operation mode samples according to claim 6, wherein the sample generation module invokes the feature samples of the corresponding hidden features according to the power system operation mode required by the given task and inputs the feature samples to the decoder of the trained variational self-encoder to obtain the power system operation mode sample data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311559870.8A CN117592513A (en) | 2023-11-21 | 2023-11-21 | Method and system for generating power system operation mode sample based on variation self-encoder and model migration |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311559870.8A CN117592513A (en) | 2023-11-21 | 2023-11-21 | Method and system for generating power system operation mode sample based on variation self-encoder and model migration |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117592513A true CN117592513A (en) | 2024-02-23 |
Family
ID=89912774
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311559870.8A Pending CN117592513A (en) | 2023-11-21 | 2023-11-21 | Method and system for generating power system operation mode sample based on variation self-encoder and model migration |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117592513A (en) |
-
2023
- 2023-11-21 CN CN202311559870.8A patent/CN117592513A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111459700B (en) | Equipment fault diagnosis method, diagnosis device, diagnosis equipment and storage medium | |
| CN111046581B (en) | Power transmission line fault type identification method and system | |
| CN111860982A (en) | Wind power plant short-term wind power prediction method based on VMD-FCM-GRU | |
| CN101105841A (en) | Method for constructing gene controlled subnetwork by large scale gene chip expression profile data | |
| Gai et al. | A Parameter‐Optimized DBN Using GOA and Its Application in Fault Diagnosis of Gearbox | |
| CN110826803A (en) | Electricity price prediction method and device for electric power spot market | |
| CN106897945A (en) | The clustering method and equipment of wind power generating set | |
| CN111275108A (en) | Method for performing sample expansion on partial discharge data based on generation countermeasure network | |
| CN115427968A (en) | Robust artificial intelligence reasoning in edge computing devices | |
| CN115147012B (en) | Neural network model-based carbon emission accounting method and device | |
| EP4050518A1 (en) | Generation of realistic data for training of artificial neural networks | |
| CN115587543A (en) | Federal learning and LSTM-based tool residual life prediction method and system | |
| CN111737249A (en) | Abnormal data detection method and device based on Lasso algorithm | |
| CN112381673A (en) | Park electricity utilization information analysis method and device based on digital twin | |
| CN110674984A (en) | Tri-Training-Lasso-BP network-based static voltage stability margin prediction method | |
| CN112559316A (en) | Software testing method and device, computer storage medium and server | |
| CN117556369B (en) | Power theft detection method and system for dynamically generated residual error graph convolution neural network | |
| Zhang et al. | A web service qos forecasting approach based on multivariate time series | |
| CN113904915A (en) | Intelligent power communication fault analysis method and system based on Internet of things | |
| CN113151842A (en) | Method and device for determining conversion efficiency of wind-solar complementary water electrolysis hydrogen production | |
| CN115859128B (en) | Analysis method and system based on interaction similarity of archive data | |
| CN116722653A (en) | Dynamic state detection method and system for electric power system | |
| CN117592513A (en) | Method and system for generating power system operation mode sample based on variation self-encoder and model migration | |
| CN111061708A (en) | Electric energy prediction and restoration method based on LSTM neural network | |
| Liu et al. | Power system stochastic transient stability assessment based on Monte Carlo simulation |
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 |