Neural network photovoltaic power generation output prediction method based on grey correlation analysis

Abstract

The present invention relates to a neural network photovoltaic power generation output prediction method based on grey correlation analysis, which comprises grey correlation degree analysis, neural network training and output result prediction analysis. In the grey correlation degree analysis, an optimal sample is obtained by computing and sorting grey correlation degrees of hour period samples comprising factors which influence the photovoltaic output; according to the neural network training, the optical sample is utilized to train a BP neural network which is optimized by a genetic algorithm, so that the trained neural network is obtained; and according to the output result prediction analysis, weather parameter information in each hour period on prediction day is used as an input condition and is combined with the trained neural network to carry out prediction on an output in each time period by using hour as a step length and a mean absolute percentage error is adopted to carry out evaluation on prediction capacity of a system. According to the prediction method disclosed by the present invention, not only is prediction accuracy when weather is suddenly changed improved, but also the defect, that a prediction result is easy to fall into the local optimization, is avoided.

Description

technical field [0001] The invention relates to the field of photovoltaic power generation, in particular to a neural network photovoltaic power generation output prediction method based on gray relational analysis. Background technique [0002] With the deepening of the global energy crisis and the increasingly prominent climate issues, the development and utilization of renewable energy has received extensive attention. Solar photovoltaic power generation has become one of the most popular green energy sources due to its clean, safe and non-polluting features. However, solar photovoltaic power generation has the characteristics of volatility and intermittency, and large-scale photovoltaic grid connection will bring security, stability, power quality, reliability and other issues to the power grid. Therefore, the output prediction of photovoltaic power generation is of great significance to power grid planning and operation. [0003] At present, the models for photovoltaic...

AI Technical Summary

Problems solved by technology

[0004] However, the BP neural network has disadvantages such as low learning efficiency, slow convergence speed, and easy to fall into local optimum, which is not conducive to the improvement of prediction accuracy; the method of using the traditional similar day standard as the basis for training sample selection cannot meet the prediction accuracy requirements under sudden weather conditions. Require

Images