To address the key problem of insufficient measurement locations of wind speed for the individual offshore wind power structure,a transient expansion method of the wind farm ahead of offshore wind turbines based on limited measurement data for wind speed is proposed.The prior farm data is decomposed into feature information for both the temporal coefficients and the spatial modes based on proper orthogonal decomposition(POD).A nonlinear mapping relationship from insufficient measurement locations of wind speed to the global wind farm is established by recurrent neural networks(RNN)to construct the offshore transient wind farm in real time.The results show that the proposed POD-RNN method can accurately reconstruct the wind farm ahead of the offshore wind turbine using limited measurement data for wind speed,where the root mean square error(RMSE)of the transient reconstruction of wind farm is within 1.8136 m/s.
维普
万方数据