Research on Power Prediction of Offshore Wind Power Based on BiSTM

• Published in: 2024 6th International Conference on Energy Systems and Electrical Power (ICESEP) pp. 161 - 164
• Main Authors: Zhao, Qianhao, Zhang, Jiazheng, Meng, Yudong
• Format: Conference Proceeding
• Language: English
• Published: IEEE 21-06-2024
• Subjects: Accuracy; bidirectional long short-term memory network; Correlation; medium and long-term power forecast; ocean current factors; Oceans; offshore wind power; power prediction; Power system stability; Predictive models; Wind power generation; Wind speed
• DOI: 10.1109/ICESEP62218.2024.10651659

The prediction of wind power is of great significance for maintaining the stability of the power system frequency. In addressing the issue of offshore wind power prediction, this paper conducts a correlation analysis between factors such as ocean currents, wind speed in the sea area near Zhejiang, and wind power, and proposes a method for predicting offshore wind power that takes into account ocean currents. Experimental verification is then carried out. The experiment uses wind speed and ocean current data as input features and compares them with traditional meteorological data input models. Subsequently, a Bidirectional Long Short Term Memory (BiSTM) neural network is constructed for model training with input of meteorological and ocean current data, and wind power prediction simulations are performed for both medium to long-term and extreme weather conditions. The simulation results show that the wind power prediction model considering ocean currents significantly improves the accuracy of both medium to long-term and extreme weather predictions compared to traditional models. The method proposed in this study is more effective and practical for offshore wind power.