Short-term active power forecasting for wind farms using artificial neural networks

These days, wind energy plays an increasingly crucial role in the energy sector, posing challenges in its management and operation. Given the current upgrade of Vietnam's 500 kV grid infrastructure, wind farms are concentrated in specific regions. This concentration can lead to significant power influxes into the grid at certain times, causing grid overcurrent. Hence, the National Load Dispatch Center is currently regulating power generation based on forecasted data from generating units. Therefore, short-term power forecasting for wind farms is crucial to mitigate grid overcurrent. This article proposes a short-term forecast of active power in wind farm using a model based on Artificial Neural Network (ANN) on Matlab platform. In the process of building the ANN model, this article considers eliminating the impact of capacity regulation on the power grid. The model was tested using real data from the Ia Pết Đăk Đoa 1 wind farm in Gia Lai province. The time forecast is given in 15-minute intervals for the next 4 hours. The collected results show the superiority of the method in forecasting with low errors and saving calculation time.