| Title |
Short-term wind power forecast using hourly LSTM technique |
| Authors |
최준영(Joonyoung Choi) ; 이송근(Songkeun Lee) |
| DOI |
https://doi.org/10.5370/KIEE.2020.69.6.759 |
| Keywords |
RNN; LSTM; short-term; wind power; forecast |
| Abstract |
Wind power is intermittent and nonlinear. When the amount of power generation exceeding the margin of the power system generation is input from the wind power generator to the grid, disturbance may occur in the grid. Therefore, in order to operate a power system stable, it is necessary to accurately predict wind power generation. For stable system operation, short-term wind power prediction, that is, prediction of the amount of electrical power generated from 1 hour to several hours is required. In this paper, an LSTM model for each forecasted time was created and combined to predict the amount of wind power generation after 1 to 3 hours. The validity of the proposed model was proved by comparing and analyzing the prediction accuracy with the proposed model, LSTM model, and DBN model. |