School of Electrical and Power Engineering, China University of Mining and Technology, Xuzhou 221116, China
In order to improve the interval estimation accuracy of wind power, an interval estimation method based on error classification is proposed combined with the characteristics of prediction error data. Firstly, the K-means clustering algorithm is introduced to classify the overall levels of wind power prediction errors by using the Euclidean distance as the clustering index. Secondly, an error interval estimation model is established, of which the input is the wind power and the historical prediction error, and the output is the prediction error interval. Long short-term memory (LSTM) network is utilized to deeply learn the correlation between the input and output of the model. Finally, the wind power data from Elia website is used for verification. The results show that, compared with other estimation models and traditional error probability distribution methods, the proposed method can capture the characteristics of prediction error data and obtain more accurate interval estimation results of wind power.
This work is supported by National Natural Science Foundation of China (No. 61703404).
|||HAN Li, QIAO Yan, JING Huitian. Interval Estimation of Wind Power Based on Error Classification[J]. Automation of Electric Power Systems,2021,45(1):97-104. DOI:10.7500/AEPS20200227014|