摘要
针对电力负荷序列随机性强、不平稳、随时间变化具有非线性等特点,提出一种基于互补式集合经验模态分解(complementary ensemble empirical mode decomposition,CEEMD)和麻雀搜索算法(sparrow search algorithm,SSA)的最小二乘支持向量机(least square support vector machine,LSSVM)的组合模型。首先,该方法利用CEEMD将原始电力负荷序列分解为一系列不同频率尺度的固有分量和剩余分量,减少了不同频率的信号之间的相互影响。然后,针对各个分量不同的特点,采用新型麻雀搜索算法优化核函数相关参数并建立各自对应的预测模型,最后将不同分量预测得到的结果叠加。这种组合方式加入了CEEMD的分解和麻雀搜索算法,能有效地提高预测精度。麻雀搜索算法较之其他组合模型中的优化算法有较强的鲁棒性。与LSSVM、CEEMD-LSSVM和SSA-LSSVM模型进行了对比分析,结果表明所提出的组合模型预测精度高,具有良好的追踪性和泛化性。
In view of the strong randomness,instability and non-linear characteristics of power load series with time variation,a new combined model based on complementary ensemble empirical mode decomposition(CEEMD)and sparrow search algorithm(SSA)least square support vector machine(LSSVM)is proposed.Firstly,the original load sequence is decomposed into a series of intrinsic components and residual components with different frequency scales by CEMMD,which reduces the interaction between signals with different frequencies.Then,according to the different characteristics of each component,the sparrow search algorithm is used to optimize the relevant parameters of the kernel function,and the corresponding prediction models are established.Finally,the prediction results of different components are superimposed.This combination method adds the decomposition of CEEMD and sparrow search algorithm to improve the prediction accuracy effectively.Sparrow search algorithm has stronger robustness than other optimization algorithms in combination model.Compared with LSSVM,CEEMD-LSSVM and SSA-LSSVM models,the results show that the combination model proposed in this paper has high prediction accuracy,good traceability and generalization.
作者
杨海柱
石剑
江昭阳
张鹏
YANG Haizhu;SHI Jian;JIANG Zhaoyang;ZHANG Peng(School of Electrical Engineering and Automation,Henan Polytechnic University,Jiaozuo 454000,China;College of Electrical Engineering and Automation,Tianjin University,Tianjin 300072,China)
出处
《武汉大学学报（工学版）》
CAS
CSCD
北大核心
2022年第6期609-616,共8页
Engineering Journal of Wuhan University
基金
国家自然科学基金项目(编号:61703144)
天津市自然科学基金项目(编号:19JCQNJC06100)。
关键词
电力负荷
最小二乘支持向量机
麻雀搜索算法
组合模型
power load
the least square support vector machine
sparrow search algorithm
combinatorial model
