| 刘娜,段钰锴.基于改进蜣螂优化算法的短期光伏发电功率预测研究[J].高技术通讯(中文),2025,35(9):1024~1036 |
| 基于改进蜣螂优化算法的短期光伏发电功率预测研究 |
| Study on short-term photovoltaic power prediction based on improved dung beetle optimization algorithm |
| |
| DOI：10. 3772 / j. issn. 1002-0470. 2025. 09. 011 |
| 中文关键词: 长短期记忆网络； 蜣螂优化算法； 多策略改进； 混沌映射； 螺旋搜索策略； Levy飞行； t分布变异策略； 变分模态分解 |
| 英文关键词: long short-term memory, dung beetle optimization, multi-strategy improvement, chaotic mapping, spiral search strategy, Levy flight, t-distribution variational strategy, variational modal decomposition |
| 基金项目: |
| 作者 | 单位 | | 刘娜 | (上海电机学院商学院上海 201306) | | 段钰锴 | |
|
| 摘要点击次数: 16 |
| 全文下载次数: 15 |
| 中文摘要: |
| 随着“双碳”战略的提出，提高光伏并网稳定性的短期光伏发电功率预测研究成为热点。本文提出了一种基于改进蜣螂优化算法的短期光伏发电功率预测组合模型，该模型以蜣螂优化算法为基础，针对整体搜索能力不足、算法易陷入局部极值等问题进行了多策略改进，其策略为混沌映射、螺旋搜索策略、Levy飞行、t分布变异策略，有效提升了原始蜣螂优化算法的性能；而长短期记忆网络(long short-term memory,LSTM)与变分模态分解(variational mode decomposition,VMD)、改进蜣螂优化算法（dung beetle optimizer,DDBO）组合后的VMD-DDBO-LSTM能有效避免分解过程中出现模态混叠现象，在收敛性、鲁棒性和精度方面远胜单一优化模型，实验证明了改进算法与组合模型优势互补的可行性和优越性。 |
| 英文摘要: |
| With the proposal of ‘double carbon’ strategy, short-term photovoltaic power prediction research to improve the stability of photovoltaic grid-connected power generation has also become a major hotspot. In this paper, a short-term photovoltaic power prediction combined model based on improved dung beetle optimization algorithm(DDBO) is proposed, which is based on the dung beetle optimization algorithm (DBO), and improves multi-strategy to address the problems of insufficient overall search capability and easiness to fall into the local extremes of the algorithm. The strategies of the dung beetle optimization algorithm include chaotic mapping, helical search, Levy flight, and t-distribution variation, which can improve the performance of the original dung beetle optimization algorithm. The combination of long short-term memory （LSTM） with variational modal decomposition (VMD) and DDBO, VMD-DDBO-LSTM can effectively avoid the phenomenon of modal aliasing during the decomposition process, and is much better than the single optimization model in terms of convergence, robustness and accuracy, and the feasibility and superiority of the improved algorithm complementing the advantages of the combined model are clarified with experiments. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|
