Title :
The study of three kinds of wind power prediction methods(60603)
Author :
Jiali Shuai ; Jixian Qu
Author_Institution :
Sch. of Electr. Eng., Wuhan Univ., Wuhan, China
Abstract :
The wind power prediction has great significance on the optimization of grid scheduling, ensuring the power balance and safe operation. First, the ARMA model, the gray prediction model, and the BP neural network model are established by using the sampling time of 15 min wind power time series. The conclusion is that the BP neural network model predicts most accurate. Then, stand-alone wind power and the sum of 58 wind turbines wind power are predicted. The accuracy of 58 wind turbines is higher. It is visible wind power power aggregation can improve the prediction accuracy.
Keywords :
autoregressive moving average processes; backpropagation; neural nets; power engineering computing; time series; wind power plants; wind turbines; ARMA model; BP neural network model; gray prediction model; grid scheduling; power aggregation; power balance; prediction accuracy improvement; safe operation; sampling time; stand-alone wind power; time 15 min; time series; wind power prediction methods; wind turbines; Autoregressive processes; Data models; Educational institutions; Mathematical model; Neural networks; Predictive models; Wind power generation; ARMA; BP neural network; gray prediction; wind power prediction;
Conference_Titel :
Power Engineering and Automation Conference (PEAM), 2012 IEEE
Conference_Location :
Wuhan
Print_ISBN :
978-1-4577-1599-0
DOI :
10.1109/PEAM.2012.6612518
