%0 Journal Article
%T 基于LS-SVM和小波分解的风电功率预测模型研究
Research on Wind Power Forecasting Model Based on LS-SVM and Wavelet Decomposition
%A 王阳
%J Advances in Energy and Power Engineering
%P 215-222
%@ 2328-0506
%D 2024
%I Hans Publishing
%R 10.12677/aepe.2024.126024
%X 随着风能在电力市场的高度渗透，开发高效的风电预测模型成为迫切需求。本文利用历史数据和数值天气预报，应用多种混合预测方法进行风电功率预测，特别对复杂地形风电场的发电量进行了对比研究。研究评估了带有小波分解(WD)的最小二乘支持向量机(LS-SVM)在不同时间范围内的性能，并与其他混合预测方法进行了比较。结果表明，基于LS-SVM和WD的混合方法在大多数情况下优于其他预测方法。通过对均方根误差的分解，深入分析了预测值与实际测量值之间的差异，并比较了不同模型的准确性。此外，研究还进行了敏感性分析，探讨了各输入变量对LS-SVM模型训练过程的影响，并对WD技术下LS-SVM模型的分解成分进行了灵敏度分析。
With the increasing penetration of wind energy in the power market, the development of accurate and efficient wind power forecasting models has become a pressing requirement. This paper leverages historical data and numerical weather prediction to apply various hybrid forecasting methods for wind power prediction, with a particular emphasis on a comparative study of power generation in wind farms situated in complex terrains. The performance of the Least Squares Support Vector Machine (LS-SVM) integrated with Wavelet Decomposition (WD) is evaluated over different forecasting horizons, and the results are compared with those of other hybrid forecasting methods. The findings indicate that the LS-SVM and WD-based hybrid approach outperforms most alternative forecasting techniques in most cases. A detailed analysis of the discrepancies between predicted and actual measurements is conducted through the decomposition of root mean square error, and the accuracy of various models is further compared. Additionally, a sensitivity analysis is performed to examine the influence of different input variables on the training process of the LS-SVM model, and a sensitivity analysis of the decomposition components of the LS-SVM model under the WD technique is also presented.
%K 风电功率预测，
%K 最小二乘支持向量机(LS-SVM)，
%K 小波分解
Wind Power Prediction
%K Least Squares Support Vector Machine
%K Wavelet Decomposition
%U http://www.hanspub.org/journal/PaperInformation.aspx?PaperID=104104