Photovoltaic power generation power short-term rolling prediction method based on improved MKPLS

Abstract

A photovoltaic power generation power short-term rolling prediction method based on improved MKPLS comprises the steps that firstly, a multi-source factor historical data set influencing power generation power is acquired to serve as an explanatory variable data set, a corresponding power generation power historical data set serves as a dependent variable data set, threshold filtering of a historical data set is realized through a method of combining complete empirical mode decomposition based on adaptive noise and permutation entropy, and effective removal of high-frequency noise and retention of useful information are realized; secondly, a correlation coefficient matrix of the variable data set and the dependent variable data set is explained through threshold filtering, and feature extraction is achieved; then, a multi-scale Gaussian kernel function is introduced to map an input space to a high-dimensional feature space, and a nonlinear prediction model of dependent variables about explanatory variables is established; and finally, the multi-scale Gaussian kernel function of the prediction model is updated in real time by introducing a sliding window strategy, and photovoltaic power generation power short-term rolling prediction is finally realized. The method is easy to implement, small in calculation amount and high in adaptive capacity.

Description

technical field [0001] The invention is applied to the field of prediction of photovoltaic power generation, and relates to a short-term rolling prediction method of photovoltaic power generation based on improved multi-scale kernel function partial least squares (MKPLS). Background technique [0002] In recent years, new power systems with new energy as the main body have been developed and applied on a large scale, such as photovoltaic power generation systems. Through the integration of photovoltaic power generation into the grid and gradually increasing the proportion of photovoltaic power generation, the existing energy structure will be effectively optimized, the country's energy conservation and emission reduction will be promoted, and the goal of "carbon peaking and carbon neutrality" will be achieved. However, the power of photovoltaic power generation system has the characteristics of intermittency, volatility and randomness, which will seriously affect the safe an...

AI Technical Summary

Problems solved by technology

[0005] In order to solve the technical defects of the existing short-term prediction method of photovoltaic power generation, such as poor nonlinear fitting effect, large historical data samples and incapability of rolling prediction, the present invention proposes a method based on improved multi-scale kernel function partial minimum for small samples Short-term Rolling Forecast Method of Photovoltaic Power Generation Based on Multiplication of Two (MKPLS)

Images