.png){kind=icon}
返回
Decomposition based deep projection-encoding echo state network for multi-scale and multi-step wind speed prediction
.png){kind=icon}
PRE
AI
DOI：10.1016/j.eswa.2024.126074.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
Accurate wind speed forecasting is essential to improve the scheduling and the utilization ratio of wind power. However, it is challenging to accurately forecast the wind speed, especially for multi-step prediction horizons, due to the complex features of wind speed data and the existence of accumulative error. The existence of accumulative error makes traditional multi-step models unable to obtain expected prediction results. Therefore, this paper proposes a novel efficient wind speed multi-step forecasting model that can effectively decrease the accumulated error, by learning the temporal dependency among future time steps. Specifically, the variational mode decomposition (VMD) is firstly adopted as the processing technique to decompose original wind series into several relatively simple subsequences with different features. Subsequently, a novel deep projection- encoding echo state network (DEESN) is designed for multi-step wind speed prediction. The most distinctive characteristic of DEESN is that it can learn the temporal dependency among future time steps through stacked echo state networks and concatenated extreme learning machine encoders. Finally, to validate the performance of VMD-DEESN, three wind speed datasets at two timescales are collected from Tanghe city, Henan Province, China. Three prediction evaluation metrics and computation time are adopted to evaluate the performance of the proposed model for multi-step wind speed forecasting. Simulation results demonstrate that the proposed VMD-DEESN could obtain better performance compared with six baseline models, with normalized root mean square error (NRMSE) 4.253%, 5.894%, 5.631%, 12.461%, 12.511% and 11.500% for multi-step per ten minutes and hourly wind speed prediction, respectively.
Keyword:
Wind speed prediction
Variational mode decomposition
Echo state network
Encoder
期刊
IF:
7.5
论文数:
3.0W
被引数:
10.2W
机构
暂无机构信息
引用论文
Explainable temporal dependence in multi-step wind power forecast via decomposition based chain echo state networks.png)
ENERGY
ENERGYIF9.4
PREAI
分享收藏A compound framework incorporating improved outlier detection and correction, VMD, weight-based stacked generalization with enhanced DESMA for multi-step short-term wind speed forecastingAPPLIED ENERGY
APPLIED ENERGYIF11
PREAI
分享收藏Short-term wind power prediction using a novel model based on butterfly optimization algorithm-variational mode decomposition-long short-term memoryAPPLIED ENERGY
APPLIED ENERGYIF11
PREAI
分享收藏Weather image-based short-term dense wind speed forecast with a ConvLSTM-LSTM deep learning model
PREAI
分享收藏Short-term wind speed forecasting based on recurrent neural networks and Levy crystal structure algorithmENERGY
ENERGYIF9.4
PREAI
分享收藏
查看更多内容