基于VMD-IMPA-SVM的超短期风电功率预测
Ultra Short Term Wind Power Prediction Based on VMD-IMPA-SVM
刘金朋 1邓嘉明 1高鹏宇 1刘胡诗涵 1孙思源1
作者信息
- 1. 华北电力大学经济与管理学院,北京 102206
- 折叠
摘要
针对风力发电强波动性带来的预测精度不高问题,构建一种基于变模态分解(VMD)、灰狼优化算法(GWO)、海洋捕食者算法(MPA)和支持向量机(SVM)的组合预测模型.采用GWO对VMD的模态数和惩罚因子进行寻优,将原始功率序列分解为子序列进行降噪处理;运用对立学习和柯西变异等方法改进MPA的种群生成与变异方式,得到改进MPA(IMPA)并优化SVM中的核参数与惩罚参数,进而构建VMD-IMPA-SVM组合预测模型,对各子序列进行预测并叠加得到最终预测值.实际算例分析表明,所提组合预测模型具有较高的预测精度,同时具备强鲁棒性.
Abstract
Aiming at the problem of low prediction accuracy caused by the strong volatility of wind power generation,a combined prediction model based on variable modal decomposition(VMD),gray wolf optimization algorithm(GWO),Marine Predator Algorithm(MPA)and support vector machine(SVM)is constructed.Firstly,GWO is used to optimize the modal number and penalty factor of VMD,decompose the original power sequence into subsequence for noise reduction,improve the population generation and variation of MPA by using the methods of opposition learning and Cauchy's variance,and derive the Improved Marine Predator Algorithm(IMPA)to optimize the SVM.The kernel function and penalty parameters are optimized in SVM,and a combined VMD-IMPA-SVM prediction model is constructed,which predicts each subsequence and superimposes to obtain the final prediction value.Practical examples show that the proposed combined prediction model has high prediction accuracy and strong robustness.
关键词
风电功率预测/变模态分解/海洋捕食者算法/支持向量机/灰狼优化算法Key words
wind power prediction/variable modal decomposition/marine predator algorithm/support vector machine/gray wolf optimization algorithm引用本文复制引用
基金项目
国家自然科学基金资助项目(72274060)
出版年
2024
NETL
NSTL
万方数据