Optical power prediction method based on adaptive classification strategy and hybrid optimization SVR

Abstract

The invention provides an optical power prediction method based on an adaptive classification strategy and hybrid optimization SVR. The method comprises the following steps: S1, performing data preprocessing, i.e., eliminating abnormal samples in historical samples by adopting a dimensionless processing method and a density-based outlier algorithm, and reconstructing or directly deleting missing data by adopting an interpolation algorithm; S2, performing sample screening, i.e., screening out historical samples highly similar to the meteorological data at the prediction moment through an adaptive classification strategy; and S3, performing model training and numerical prediction, i.e., training a prediction model by adopting a quantum-behaved particle swarm algorithm and a grid method hybrid optimization SVR algorithm, and inputting meteorological data at a prediction moment to obtain a power prediction value. According to the method, adverse effects of error data and missing data on prediction precision can be reduced, historical samples highly similar to the meteorological data at the prediction moment can be screened out according to the meteorological data at the prediction moment, hyper-parameters of the SVR algorithm are optimized by adopting a quantum particle swarm and a grid method in a mixed mode, and the generalization ability of the model is improved.

Description

technical field [0001] The invention belongs to the technical field of photovoltaic power prediction, and in particular relates to a photovoltaic power prediction method based on an adaptive classification strategy and a hybrid optimized SVR algorithm. Background technique [0002] The output power of photovoltaic power plants is affected by meteorological characteristics such as solar irradiance, temperature, humidity, wind speed, wind direction, and air pressure. The changing characteristics of these meteorological characteristics make the output power of photovoltaic power plants have randomness, volatility, and intermittent characteristics. The main factors for the limited power operation of photovoltaic power plants. In order to improve the photovoltaic consumption capacity of the grid, the grid-connected photovoltaic power station must be equipped with a high-precision photovoltaic power prediction system. [0003] Photovoltaic power prediction generally includes seve...

AI Technical Summary

Problems solved by technology

[0009] (2) The determination method of similar days is relatively extensive, resulting in the possibility that samples of similar days may be mixed with samples that are significantly different from the meteorological characteristics at the time of prediction, further affecting the accuracy of the prediction model

[0010] (3) The neural network algorithm can fit complex nonlinear relationships, but it is prone to overfitting; the time series algorithm has low prediction accuracy in scenes with drastic changes in weather

[0011] (4) At present, most of the model fusion methods still only stay on the simple linear weighted superposition of multiple models

Images