基于堆叠稀疏降噪自动编码器的地区风电场群高精度超短期风电功率预测

High Precision Ultra-short-term Wind Power Prediction of Regional Wind Farms Based on Stacked Sparse Noise Reduction Autoencoders

为解决风电功率预测过程中面临的风电数据量大且复杂化以及如何提高预测精度的问题,基于堆叠稀疏降噪自动编码器提出地区风电场群高精度超短期风电功率预测方法。该方法采用自编码器对输入风电功率数据进行降维提取特征,为进一步增强自动编码器的抗干扰性,对其引入稀疏性约束和降噪技术。该方法能够有效降低数据的解析难度和提高特征提取的可靠性。通过实际算例验证,该预测方法可有效提高多风电场功率预测的精度。

In order to solve the problems of large and complex wind power data in the process of wind power prediction and how to improve the prediction accuracy,this paper proposes a high precision ultra short-term wind power prediction method for regional wind farms based on stacked sparse noise reduction autoencoders.This method uses an autoencoder to reduce the dimensionality of the input wind power data to extract features.In order to further enhance the anti-interference performance of the autoencoder,sparsity constraints and noise reduction techniques are introduced to it.This method can effectively reduce the difficulty of data analysis and improve the reliability of feature extraction.Verification by actual calculation examples shows that the prediction method can effectively improve the accuracy of power prediction for multiple wind farms.