Power distribution network big data restoration method based on confrontation game

A data restoration and distribution network technology, applied in the field of electrical engineering, can solve the problems of ignoring the correlation load change law, mode collapse, low reconstruction accuracy of missing data, etc., and achieve the effect of overcoming the problem of unstable training.

Pending Publication Date: 2022-05-06
CHINA PETROLEUM & CHEM CORP +2
View PDF0 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The traditional adversarial game model is prone to the problems of gradient disappearance and mode collapse during the training process, and only considers a single data distribution law, ignoring the correlation between measurement points and collected variables in the power system and the historical load change law. Low reconstruction accuracy for missing data

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Power distribution network big data restoration method based on confrontation game
  • Power distribution network big data restoration method based on confrontation game
  • Power distribution network big data restoration method based on confrontation game

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0039] The present invention provides a method for repairing big data of distribution network based on confrontation game, comprising the steps of,

[0040] S1. Obtain sample category labels: collect electromagnetic data of distribution network and classify and label, and generate corresponding category labels for each sample;

[0041] S2. Build the generator framework: input the data after horizontal concatenation of noise and category labels into the generator, and use the generated samples output by the generator as the input of the discriminator;

[0042] S3. Build a discriminator framework;

[0043] S4, training network;

[0044] S5. Generate missing data of the distribution network according to the network trained in step S4.

[0045] In step S3, the specific method of building the frame of the discriminator is to input the real data or the generated sample and the corresponding category label horizontally into the discriminator, and output it as a judgment value, indi...

Embodiment 2

[0060] The present invention provides a method for repairing big data of distribution network based on confrontation game, comprising the steps of,

[0061] (1) Obtain sample category labels: Collect electromagnetic data of distribution network and classify and label, and generate corresponding category labels for each sample.

[0062] (2) Build a generator framework: input the data after noise and category labels are horizontally concatenated into the generator, and use the generated samples output by the generator as the input of the discriminator.

[0063] (3) Build a discriminator framework: the real data or the generated sample and the corresponding category label are horizontally spliced ​​and input to the discriminator, and the output is a judgment value, indicating the probability that the input sample of the discriminator is a real sample. (4) Training network

[0064] The Wasserstein distance is used to measure the difference between the actual distribution and the ...

Embodiment 3

[0075] see Figure 1 to Figure 6 , figure 1 It is a schematic diagram of a generator and a discriminator based on a conditional confrontational game model. The confrontational game model is a generative model. The sum is zero, and the gain of one party is the loss of the other party), which consists of a generator and a discriminator. The generator captures the mathematical distribution model of real data samples, and generates new data samples from the learned distribution model; the discriminator is a binary classifier, which is used to distinguish whether the input is real data or generated samples. The two continue to learn to improve their generative and discriminative abilities. Such as figure 1 As shown, the model is a conditionally controlled adversarial game network. By adding the same condition Y (such as the label of the data) to the generator and the discriminator, the adversarial game model control condition is realized. Taking the generation of missing data o...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses a power distribution network big data restoration method based on an adversarial game, and relates to the technical field of electrical engineering. Building a generator framework; building a discriminator framework; training the network; and generating missing data of the power distribution network according to the trained network. The method has the beneficial effects that a mutual confrontation game model between the generator and the discriminator formed by the deep neural network is adopted in the scheme; the generator tries to generate samples which are consistent with real data in distribution, the discriminator identifies the samples generated by the generator as much as possible, and the two networks progress together in mutual comparison. By establishing the conditional adversarial game model of gradient penalty optimization, the generation process of multi-class missing samples can be guided, and the problem of unstable training of an original adversarial game model is solved, so that the problem of data missing caused by influences of measurement faults, distributed energy and the like in the power distribution network can be solved.

Description

technical field [0001] The invention relates to the technical field of electrical engineering, in particular to a large data restoration method of a distribution network based on an adversarial game. Background technique [0002] The massive measurement data acquired by the distribution network data acquisition system is of great significance to state estimation, equipment evaluation, and system operation optimization. However, only the research conclusions based on real and reliable collected data have practical application value and can correctly reflect the operating characteristics and objective laws of the power system. In the actual operation of the measurement system, all links of data collection, measurement, transmission, and conversion may fail or be disturbed, resulting in missing and abnormal data. In addition, with the vigorous development of distributed new energy, the monitoring data of large-scale new energy is becoming more and more abundant. It will lead ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
IPC IPC(8): G06F17/15G06V10/764G06V10/82G06K9/62G06N3/04G06N3/08G06Q50/06
CPCG06F17/15G06Q50/06G06N3/084G06N3/045G06F18/241G06F18/2415
Inventor 孙东严川盛庆博李炜范路董伟佳王莉刘聪岳宇张晓菡
Owner CHINA PETROLEUM & CHEM CORP
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap