可再生能源2025，Vol.43Issue(11)：1458-1468,11.
基于特征优选的DBO-CNN-BiLSTM-AM短期光伏发电预测
DBO-CNN-BiLSTM-AM short-term photovoltaic power generation prediction based on feature optimization
摘要
Abstract
With the gradual increase of photovoltaic installed capacity,the stable and safe operation of power system is facing challenges.By improving the prediction accuracy of photovoltaic power generation,the operation stability and safety of power system can be improved.Therefore,this paper proposes a DBO-CNN-BiLSTM-AM short-term photovoltaic power generation prediction method based on feature optimization.Firstly,Variance Inflation Factor(VIF)and Extremely Randomized Trees(ERT)are used to calculate and select the features with low VIF value and high importance score as input.Secondly,a CNN-BiLSTM-AM prediction model is established by using Convolutional Neural Network(CNN),Bi-directional Long Short Term Memory(BiLSTM)and Attention Mechanism(AM)with dynamic adjustment ability of feature weights.Finally,the Dung Beetle Optimization algorithm(DBO)is used to perform global optimization of hyperparameters to maximize the prediction accuracy.The experimental results show that the prediction accuracy of the short-term photovoltaic power prediction method proposed in this paper is significantly higher than that of the traditional method,which provides a good basis for optimizing power dis-patching and stable operation of power system.关键词
光伏发电/双向长短期神经网络/注意力机制/蜣螂优化算法Key words
photovoltaic power generation/BiLSTM/attention mechanism/dung beetle optimization algorithm分类
能源与动力引用本文复制引用
杜立,李振华,李振兴,魏伟,徐艳春..基于特征优选的DBO-CNN-BiLSTM-AM短期光伏发电预测[J].可再生能源,2025,43(11):1458-1468,11.基金项目
国家自然科学基金(52277012) （52277012）
武汉强磁场学科交叉基金资助(WHMFC202202). （WHMFC202202）
