New energy scene generation method and system

Abstract

The invention discloses a new energy scene generation method and system. The new energy scene generation method comprises the following steps of: constructing a generation network and a discriminationnetwork, carrying out adversarial training of the generation network through employing a random vector of a hidden space, the discrimination network and a training sample set, and generating a plurality of new scenes through employing the generation network after adversarial training; extracting a feature label of each new scene; constructing a regression model, and constructing a characteristicaxis of a random variable by utilizing the slope of the regression model in a high-dimensional space; orthogonalizing the characteristic axis of the random variable; and controlling characteristics onthe characteristic axis of the random variable after orthogonalization, so as to generate different new energy scenes, thereby completing generation of the new energy scenes. According to the new energy scene generation method and the system, the problem of dimensionality disasters in the prior art can be solved, the problem of over-fitting can be avoided, and meanwhile, the feature distributionof the generated anticipated scene can be flexibly controlled, so as to generate specific new energy scenes.

Description

technical field [0001] The invention belongs to the field of new energy consumption and optimized operation of the energy internet, and relates to a method and system for generating new energy scenarios. Background technique [0002] With the continuous development of clean energy such as wind power and photovoltaics and the continuous increase in penetration rate, their intermittent and fluctuating characteristics have brought new challenges to the safe dispatch and economic operation of the Energy Internet. In order to describe the operating situation of the system, it will be difficult to fully characterize the possible operating boundaries and states of the system by simply relying on traditional physical mechanism-based and deterministic methods. [0003] An effective technical method is to express the potential statistical laws of new energy through a certain mathematical model, and to mine the spatio-temporal correlation characteristics and changing laws of the output...

AI Technical Summary

Problems solved by technology

However, since VAE trains the model by minimizing the error between the generated data and the original data, it is easy to cause the problem of model training overfitting, and the diversity of generated new scene samples is insufficient, which makes it difficult to fully describe the needs of random running scenarios;

[0023] (2) When the scene is generated, the model inputs random noise of Gaussian distribution, which cannot flexibly control the distribution of generated samples, and it is difficult to generate scene samples with specified characteristics, such as samples under strong light days, windy days or different seasons

Images