基于变分模态分解和LSSVM的风电场短期风速预测

SHORT-TERM PREDICTION OF WIND SPEED FOR WIND FARM BASED ON VARIATIONAL MODE DECOMPOSITION AND LSSVM MODEL

针对风速序列的随机性和非平稳性带来的难以准确预测的问题,提出一种基于变分模态分解和最小二乘支持向量机的风速预测模型。首先利用变分模态分解将风速序列分解为一系列限带内禀模态函数以降低其非平稳性对预测性能的影响,然后对各模态分别建立最小二乘支持向量机预测模型,并利用改进的差分进化算法对其关键核参数寻优,最后将各子序列预测结果叠加组合得到最终风速预测值。实例研究表明,所提出的预测模型在短期风速预测上有较高的预测精度。

In view of difficulty in accurate forecasting wind speed sequence influenced by randomness and nonstationarity, a wind speed prediction model based on variational mode decomposition （VMD） and least squares support vector machine （LSSVM） is presented. Firstly, the mean hourly wind speed data is decomposed into a series of band-limited intrinsic mode functions （BIMF） using VMD method to decrease the instability of wind speed series. Then the LSSVM forecasting models are established respectively for each mode and the key kernel parameters are optimized using improved differential evolution algorithm. Finally, the wind speed forecasting model is obtained by superposing each predicting subsequence. Case study shows that the proposed forecasting model has relatively high predicting accuracy on short-term wind speed prediction.