Prediction of ultra-short-term wind power based on CEEMDAN-LSTM-TCN

So as to decrease those cacoethic impact of a huge amount of wind energy generation systems associated with the electric power system and improve the utilization rate and the budgetary profits of wind power era, this paper raises a neural network in view of CEEMDAN-LSTM-TCN. Firstly, CEEMDAN is used to break down the wind velocity arrangement to decrease the sway of arbitrariness Furthermore variance about wind velocity. Secondly, the ultra-short-term wind power forecast depend upon LSTM and TCN is built to realize the real-time prediction for wind energy. Finally, the simulation results show that LSTM-TCN can deal with multi time order characteristics and predict ultra-short period wind energy with effect, which is better than LSTM and TCN. It also has a scientific reference for local power dispatching. (c) 2022 The Authors. Published by Elsevier Ltd. Peer-review under responsibility of the scientific committee of the International Conference on Energy Storage Technology and Power Systems, ESPS, 2022.

Automated summary

This paper proposes a neural network model based on CEEMDAN-LSTM-TCN for predicting ultra-short term wind energy. By decomposing wind velocity data and establishing the model, it achieves real-time prediction of wind energy with good forecasting performance.