Short-term prediction of the power of a new wind turbine based on IAO-LSTM

Abstract

Short-term wind power forecasting is of great significance to the real-time dispatching of power systems, but the short-term forecasting accuracy of wind power is not high. To this end, this paper proposes a hybrid prediction model that combines the Isolated Forest algorithm, the Synchronous Squeeze Wavelet Transform (SWT) method, the Aquila Optimizer (AO) and the Long Short-term Memory network (LSTM). Firstly, the Isolated Forest algorithm is used to detect abnormal data. Secondly, the SWT method is used to denoise the original power signal of the new wind turbine. Then, the wind power prediction model is established through the long short-term memory network algorithm. The OA is used to optimize the LSTM structure parameters to solve the influence of random parameters on the prediction accuracy. Finally, perform example verification. The results show that the proposed model is effective in power prediction of new wind turbine.

Keywords

• Wind power prediction
• New wind turbine
• Aquila Optimizer
• LSTM network