[1]Sun Xiaoyan,Shi Liangzhen,Xu Ruidong,et al.Forecast of wind power generation with uncertainty based on interval sample and echo state network[J].Journal of Zhengzhou University (Engineering Science),2017,38(01):56-.[doi:10.13705/j.issn.1671-6833.2017.01.003]
Copy
Journal of Zhengzhou University (Engineering Science)[ISSN
1671-6833/CN
41-1339/T] Volume:
38
Number of periods:
2017 01
Page number:
56-
Column:
Public date:
2017-02-24
- Title:
-
Forecast of wind power generation with uncertainty based on interval sample and echo state network
- Author(s):
-
Sun Xiaoyan; Shi Liangzhen; Xu Ruidong; Zhang Yong
-
School of Information and Electrical Engineering, China University of Mining and Technology, Xuzhou, Jiangsu, 221116
-
- Keywords:
-
- CLC:
-
-
- DOI:
-
10.13705/j.issn.1671-6833.2017.01.003
- Abstract:
-
The wind power forecasting was essential to the stability control of the grid connected operation,the economical dispatch,and so on.However,due to the variety of nature of wind,wind power had great uncertainties.Effectively expressing the uncertainties in wind power forecasting is crucial for improving the reliability of the forecast.Most existing methods focued on point forecasting,which can hardly quantify the uncertainties.To overcome the weekness,this paper proposed a novel interval-based forecasting model to quantify the uncertainties.A new interval sample selection method was firstly presented to reflect the uncertainties of wind power based on similar days and interval similar metric.Secondly,the echo state network were designed to predict the interval-based wind power in a short time due to its merits in time series predictions.The outstanding stability of the forecasting model was guaranteed by employing the recursive least squares algorithm to adjust the output weights of the echo state network.The prediction interval coverage probability (PICP) and mean prediction interval width (MPIW) were applied to evaluate the performance of our interval forecast on wind power.The experiments empirically demonstrated the feasibility and effectiveness of the proposed algorithm.