A photovoltaic ultra-short-term forecasting method and system

Abstract

This disclosure proposes a photovoltaic ultra-short-term prediction method, including: performing pixel standardization processing on satellite visible light cloud images, and performing bottom-out operations on the standardized cloud images to obtain images containing only cloud information; The cloud image of the target area at any time is predicted; the predicted cloud image is input into the convolutional neural network to obtain the cloud shading factor; the cloud shading factor and the influencing factors related to the photovoltaic power or irradiance are combined to establish a correlation with the photovoltaic power or irradiance Mapping relationship, use this relationship to predict the quantity to be predicted. Satellite visible light cloud image data acquisition methods are relatively simple and extensive, and at the same time, the coverage and field of view are large, which can reflect the movement and thickness of clouds. The convolutional neural network is especially suitable for the extraction and learning of image features. Combining the convolutional neural network with satellite cloud images can better capture the influence of clouds on radiation rays.

Description

technical field [0001] The disclosure belongs to the field of power grid new energy forecasting, and in particular relates to a photovoltaic ultra-short-term forecasting method and system. Background technique [0002] The statements in this section merely provide background information related to the present disclosure and do not necessarily constitute prior art. [0003] Global warming and energy crisis promote the use and development of sustainable and clean energy. Solar energy is considered to be one of the most promising renewable energy sources, and the development of new energy sources represented by photovoltaics has become the consensus of all countries in the world. However, as an intermittent energy source, photovoltaic power generation brings huge fluctuations to the smart grid, and poses severe challenges to system stability, power balance, reactive power compensation, and frequency response. To ensure the safe and economical integration of photovoltaics into...

AI Technical Summary

Problems solved by technology

The model of the above method is simple and the calculation speed is fast, but it is impossible to deeply dig out the factors that affect the change of photovoltaic power or irradiance, and cannot guarantee high accuracy

[0006] In recent years, artificial intelligence methods have begun to be used in the field of new energy forecasting, such as artificial neural networks, random forests and other methods are gradually integrated into photovoltaic forecasting technology, but traditional artificial intelligence methods rely on a large amount of historical data, image feature extraction and depth Not good at learning

Images