Optical power prediction method and system based on two-stage feature extraction and BiLSTM improvement

Abstract

The invention discloses an optical power prediction method and system based on two-stage feature extraction and improved BiLSTM. The method comprises the following steps: S1, performing shallow feature extraction on optical power data by using a partial autocorrelation function, and then performing normalization processing; s2, constructing a CNN (Convolutional Neural Network), and sending the processed data into the CNN for deep feature extraction; s3, a BiLSTM model is constructed; s4, introducing Lorenz mapping to improve the initial population of the whale algorithm, and adopting an improved whale optimization algorithm to optimize the parameters of the BiLSTM; and S5, sending the data after CNN depth feature extraction to the improved BiLSTM for prediction. According to the method, through two-stage feature extraction of shallow feature extraction and depth feature extraction, the mutual relevance between the features can be further mined, noise and unstable components of optical power data are filtered out, the processed data are sent to the improved BiLSTM model for optical power prediction, and the prediction precision can be effectively improved.

Description

technical field [0001] The invention belongs to the technical field of optical power prediction, and in particular relates to an optical power prediction method and system based on two-stage feature extraction and improved BiLSTM. Background technique [0002] With the reduction of non-renewable energy reserves, the development of clean and clean renewable energy has become a research hotspot in recent years. As one of the renewable energy sources, optical power energy has been widely concerned and researched in the field of power generation. However, due to the instability of optical power energy, it hinders the use of optical power energy for large-scale photovoltaic power generation in the grid. Accurate prediction of optical power can better guide the power grid for power generation and dispatching, and prevent factors that pose a greater threat to the safe operation of the power grid. Therefore, accurate and reliable optical power prediction is very necessary. [000...

AI Technical Summary

Problems solved by technology

Statistical models require data to be stationary, while traditional machine learning models have insufficient modeling and processing capabilities for multi-dimensional data input. The above shortcomings limit their prediction accuracy in optical power

Images