基于深度学习的多尺度风功率预测的研究

Study on Multi-Scale Wind Power Prediction Based on Depth Learning

能源结构的改革,新能源行业的高速发展,风电装机容量的逐步增加,使得风电在整个电网体系中占比增多,风电功率预测对风电安全、稳定及高效并入电网具有重要意义,但准确预测风电功率存在一定难度,基于此现状本文展开了相关研究。首先,介绍了风电领域发展背景及国内外研究现状;其次,从数据预处理、特征选择及风功率建模方法 3个方面讲述了风功率预测的整体建模流程;再次,通过实验分析了不同特征选择方法、不同组合算法及不同时间尺度对风功率预测实验结果的影响;最后,结合实验分析,给出了基于深度学习的多尺度风功率预测相关结论。

The reform of the energy structure,the rapid development of the new energy industry,and the gradual increase in wind power installed capacity have led to an increase in the proportion of wind power in the entire power grid system.Wind power prediction is of great significance for the safe,stable,and efficient integration of wind power into the grid.However,accurately predicting wind,power poses certain difficulties.Based on this current situation,this article conducts relevant research.Firstly,the development background and research status in the field of wind power at home and abroad were introduced;Secondly,the overall modeling process of wind power prediction in this article is described from three aspects:data preprocessing,feature selection,and wind power modeling methods.Once again,analyze the impact of different feature selection methods,different combination algorithms,and different time scales on the experimental results of wind power prediction through experiments;Finally,combined with experimental analysis,relevant conclusions on multi-scale wind power prediction based on deep learning were presented.