基于RIME-IAOA的混合模型短期光伏功率预测  

Short-Term Photovoltaic Power Prediction Using a Hybrid Model Based on RIME and IAOA

作　　者：王仁明[1] 魏逸明 席磊 WANG Renming;WEI Yiming;XI Lei(College of Electrical Engineering&New Energy,China Three Gorges University,Yichang 443002,China)

机构地区：[1]三峡大学电气与新能源学院,湖北宜昌443002

出　　处：《三峡大学学报（自然科学版）》2025年第1期81-88,共8页Journal of China Three Gorges University：Natural Sciences

基　　金：国家自然科学基金项目(52277108)。

摘　　要：光伏发电在如今的新能源发展中逐渐成为重点,其中光伏功率预测成为研究的主要方向.为了提升光伏功率预测的精度和效率,提出了RIME-VMD-IAOA-LSTM模型.该模型通过霜冰优化算法(RIME)优化变分模态分解(VMD)的参数来提升分解效率;引入余弦控制因子的动态边界策略来控制算数优化算法(AOA)数值的增长速率从而提升算法的精度和稳定性;利用自适应T分布变异策略来改进AOA的局部搜索能力和全局开发能力,更好地避免局部最优解.两种智能优化算法的加入使得整体模型的预测效率和速度都有很大提升,实验结果表明组合模型RIMEVMD-IAOA-LSTM相比于其他预测模型有较高的光伏功率预测精度.Photovoltaic power generation is becoming a focal point in the development of new energy sources today.Especially,the prediction of photovoltaic power is becoming a major research direction.To improve the accuracy and efficiency of photovoltaic power prediction,the model of RIME-VMD-IAOA-LSTM is proposed.The model enhances the decomposition efficiency by optimizing the parameters of variational mode decomposition(VMD)while the frost ice optimization algorithm(IAOA)is adopted.A dynamic boundary strategy with a cosine control factor is utilized to regulate the growth rate of the AOA values,thereby,the precision and stability of the algorithm is improved.An adaptive T-distribution mutation strategy is adopted to enhance the local search capability and global exploration ability of AOA in order to effectively avoid the local optima.The integration of two intelligent optimization algorithms significantly improves the prediction efficiency and the speed of the overall model.The results demonstrate that the combined model RIME-VMDIAOA-LSTM achieves higher accuracy of photovoltaic power prediction compared with that of other prediction models.

关 键 词：霜冰优化算法 变分模态分解 算术优化算法 余弦控制因子策略 自适应T分布策略 短期光伏功率预测 

分 类 号：TM715[电气工程—电力系统及自动化]