Distributed photovoltaic short-term output prediction method

Abstract

The invention relates to a distributed photovoltaic short-term output prediction method. The distributed photovoltaic short-term output prediction method comprises the following steps: collecting photovoltaic distribution transformer data, geographic information and weather condition data; preprocessing data, and constructing adjacent day and similar day characteristics of the prediction day; performing wavelet transform algorithm decomposition on the historical power generation data to obtain a main body sequence and a detail sequence; according to the power generation data and weather data,dividing weather into a sunny day, a shady day and a rainy day; combining weather data, adjacent days, similar days and wavelet decomposition results as input features; modeling in different weather categories by using a random forest regression algorithm respectively to obtain three prediction models of sunny, shady and rainy days; judging the weather category to which the weather data of the next day belongs; and inputting the algorithm model corresponding to the weather category to obtain a power generation data main body sequence and detail sequence prediction data. According to the distributed photovoltaic short-term output prediction method, the friendly grid connection level of distributed power generation is improved; the light abandoning rate is effectively reduced; clean energy consumption is promoted; and safe and stable operation of a power grid is guaranteed; and the distributed photovoltaic short-term output prediction method has extremely high generalizability.

Description

technical field [0001] The invention relates to a method for predicting short-term output of distributed photovoltaics. Background technique [0002] The many advantages of solar power generation make the construction of distributed photovoltaic power generation system a trend, and photovoltaic equipment covers a wide range. Since the photovoltaic output is greatly affected by the environment, and the frequency of the original power generation data collection method is low, its monitoring level and output capacity are limited by the metering device and communication channel, and it has been in a weak management state for a long time, making it difficult to accurately predict the output. The department lacks photovoltaic grid-connected scheduling, market transaction basis, and the increasing amount of abandoned light, which brings obstacles to the consumption of new energy and the stable operation of the power system. Therefore, it is of great theoretical and practical signi...

AI Technical Summary

Problems solved by technology

Since the photovoltaic output is greatly affected by the environment, and the frequency of the original power generation data collection method is low, its monitoring level and output capacity are limited by the metering device and communication channel, and it has been in a weak management state for a long time, making it difficult to accurately predict the output. The department lacks photovoltaic grid-connected scheduling, market transaction basis, and the increasing amount of abandoned light, which brings obstacles to the consumption of new energy and the stable operation of the power system

Images