摘要
针对风能的波动性、非平稳性导致风电功率预测精度不高的问题,研究并提出一种基于可变模式分解(VMD)技术和改进灰狼算法(DIGWO)优化核极限学习机(KELM)的短期风电功率预测模型。将功率信号进行分解得到若干个不同带宽的模式分量,对各个模式分量建立核极限学习机预测模型。为提高核极限学习机的寻优能力,采用改进的灰狼算法对核极限学习机的参数进行优化,得到各个模式分量的预测值,将分量预测值进行叠加后得到风电功率最终预测。采用实际风电功率数据进行实验仿真,实验结果表明,该模型的RMSE和MAE分别是1.5%和1.16%,相比其他模型提高了风电功率预测精度。
The volatility and nonstationarity of wind energy lead to the low accuracy of wind power prediction.This paper studies and proposes a short-term wind power prediction model based on variable mode decomposition(VMD)and improved grey wolf algorithm(DIGWO)so as to optimize the kernel extreme learning machine(KELM).The power signal was decomposed to obtain several mode components with different bandwidth.KELM prediction model was constructed for each mode component.In order to improve the ability of optimization,we adopted the improved grey wolf algorithm to optimize the parameters of KELM,and obtained the predictive values of each mode component.The final prediction of wind power was obtained after the component predictive values were superimposed.The actual wind power data was used for experimental simulation.The results show that the RMSE and MAE of this model are 1.5%and 1.16%respectively,which improves the accuracy of wind power prediction compared with other models.
作者
朱昶胜
赵奎鹏
Zhu Changsheng;Zhao Kuipeng(Lanzhou University of Technology,Lanzhou 730050,Gansu,China)
出处
《计算机应用与软件》
北大核心
2022年第5期291-298,共8页
Computer Applications and Software
基金
国家自然科学基金项目(11364024,51661020)。
关键词
风电功率预测
可变模态分解
灰狼算法
核极限学习机
预测精度
Wind power prediction
Variable mode decomposition
Grey wolf algorithm
Kernel extreme learning machine
Prediction accuracy
