兵工自动化2026，Vol.45Issue(1)：64-72,9.DOI:10.7690/bgzdh.2026.01.013
基于模态分解融合改进BiGRU的光伏功率预测
PV Power Prediction Based on Modal Decomposition Fusion with Improved BiGRU
摘要
Abstract
In order to meet the high precision requirement of PV power prediction in military energy scheduling,a dung beetle optimization(DBO)algorithm is proposed to integrate attention mechanism to optimize the bidirectional gated recurrent unit(BiGRU)photovoltaic power prediction method.Carry out analysis on influence factors and output characteristic of that photovoltaic power;carrying out multi-scale decomposition on the photovoltaic power data through complete empirical mode decomposition,and effectively separate high-frequency noise and low-frequency trend components;The multi-head self-attention(MHSA)mechanism is introduced to enhance the dynamic focusing ability of the model on key meteorological features,and the dung beetle optimization algorithm is combined to optimize the hyperparameters of the two-way gated recurrent unit network,which significantly improves the generalization performance of the model in complex military environments.The results show that the proposed model is significantly better than traditional comparison model in terms of MAE,RMSE and R²,and has a good prediction effect.关键词
军事能源/功率预测/模态分解/注意力机制Key words
military energy/power prediction/modal decomposition/attention mechanism分类
信息技术与安全科学引用本文复制引用
花国祥,孙炎,李伟伟..基于模态分解融合改进BiGRU的光伏功率预测[J].兵工自动化,2026,45(1):64-72,9.基金项目
江苏省研究生科研与实践创新计划项目基金(SJCX24_0465) （SJCX24_0465）
