Solar irradiation intensity minute-scale predication method

Abstract

The invention relates to a solar irradiation intensity minute-scale predication method and belongs to the field of new energy predication. The solar irradiation intensity minute-scale predication method aims to improve energy conservation efficiency by performing minute-scale predication on solar irradiation intensity, providing more accurate input parameters for solar predication and optimizing system operation and control. The method includes the predication steps: selecting input learning samples, performing normalization processing, determining structure parameters of a network, creating a transfer function, adopting a Levenberg-Marquardt training method, setting network parameters, performing simulation and predication and computing error. The solar irradiation intensity for the next 24 hours is subjected to minute-scale predication, prediction time scale is ultrashort, and predication time interval is five minutes. The method is of great significance to output power of a photovoltaic power generation system, scientific bases and method support are provided for grid-connected operation of photovoltaic power generation, and a foundation is laid for solar resource assessment technique research.

Description

technical field [0001] The invention belongs to the field of new energy forecasting. Background technique [0002] Solar energy is a renewable clean energy. Its total amount is the largest energy that can be developed in the world today. It can be directly developed and utilized without mining and transportation. Its reserves are unlimited, universal, and economical, and it will be an important continuous energy source and strategic energy source in the future. [0003] Large-scale photovoltaic power generation is a way to effectively utilize solar energy, but the photovoltaic power generation system is affected by factors such as radiation intensity and weather type, ambient temperature, sunshine hours, wind direction and wind level, air density, relative humidity, atmospheric pressure, etc. These Factors are unstable, random, and intermittent, so they have a great impact on the power after grid connection. The disturbance of the output power of the photovoltaic system wil...

AI Technical Summary

Problems solved by technology

[0006] The defined weather types are limited. There are four basic types: sunny, cloudy, cloudy, and rainy. It does not take into account the sudden change of weather types such as sunny to showers, and considers very few factors that affect the intensity of solar radiation. It is limited to weather types , temperature, climate, this type of BP neural network prediction method has a large and cumbersome workload, and there are few factors that affect the intensity of solar radiation, making the prediction results inaccurate, and the prediction time interval is long, basically 1 Hours

Images