现代电子技术2025，Vol.48Issue(10)：57-62,6.DOI:10.16652/j.issn.1004-373x.2025.10.010
基于ICEEMDAN-PE-GDBO-LSSVM的风电功率预测
Wind power prediction based on ICEEMDAN-PE-GDBO-LSSVM
摘要
Abstract
With the high penetration of renewable energy,especially wind power,the power grid is facing unprecedented challenges of uncertainty and volatility.In order to accurately predict wind power,a combined model based on improved complete ensemble empirical mode decomposition with adaptive noise(ICEEMDAN)-permutation entropy(PE)-improved dung beetle optimization algorithm(GDBO)-least support squares vector machine(LSSVM)is proposed.ICEEMDAN is used to decompose the wind power data to reduce the complexity.The components obtained after decomposition are aggregated according to PE,and then the key parameters of LSSVM are optimized by means of GDBO algorithm to obtain the best prediction model.The optimization model is used to predict and superimpose the aggregation components to obtain the total prediction result.The experimental verification is conducted based on the domestic wind farm dataset.Tthe results show that the proposed method has high prediction accuracy,and the root mean square error is 61.39%lower than that of the single LSSVM model,which has a broader application prospect in engineering practice.关键词
风电功率预测/自适应噪声完全集合经验模态分解/改进的蜣螂优化算法/排列熵/改进的完全集合经验模态分解/最小支持二乘向量机/分量聚合Key words
wind power prediction/ICEEMDAN/GDBO/PE/improved complete ensemble empirical mode decomposition/LSSVM/component polymerization分类
信息技术与安全科学引用本文复制引用
汪繁荣,张旭东..基于ICEEMDAN-PE-GDBO-LSSVM的风电功率预测[J].现代电子技术,2025,48(10):57-62,6.基金项目
国家自然科学基金资助项目(52307239) （52307239）
